Recent Advances in Adhesive Science and Technology - Wojciech Gutowski PDF

o f l
H n o r itta
o M
i n ash
r. K
D
R E C E N T A D VA N C E S I N
Adhesion Science
and Technology
EDITED BY
Wojciech (Voytek) Gutowski
Hanna Dodiuk
o f l
no o r itta
i n H ash M
r. K
D
Adhesion Science
and Technology
o f l
no o r itta
i n H ash M
r. K
D
Adhesion Science
and Technology
EDITED BY
Wojciech (Voytek) Gutowski

Hanna Dodiuk
Boca Raton London New York
CRC Press is an imprint of the

Taylor & Francis Group, an informa business
Cover figure courtesy of Matthew Gutowski (www.morfecs.com).
CRC Press
Taylor & Francis Group
6000 Broken Sound Parkway NW, Suite 300
Boca Raton, FL 33487-2742
© 2014 by Taylor & Francis Group, LLC
CRC Press is an imprint of Taylor & Francis Group, an Informa business
No claim to original U.S. Government works

Version Date: 20131118
International Standard Book Number-13: 978-90-04-20174-3 (eBook - PDF)
This book contains information obtained from authentic and highly regarded sources. Reasonable
efforts have been made to publish reliable data and information, but the author and publisher cannot
assume responsibility for the validity of all materials or the consequences of their use. The authors and
publishers have attempted to trace the copyright holders of all material reproduced in this publication
and apologize to copyright holders if permission to publish in this form has not been obtained. If any
copyright material has not been acknowledged please write and let us know so we may rectify in any
future reprint.
Except as permitted under U.S. Copyright Law, no part of this book may be reprinted, reproduced,
transmitted, or utilized in any form by any electronic, mechanical, or other means, now known or
hereafter invented, including photocopying, microfilming, and recording, or in any information stor-
age or retrieval system, without written permission from the publishers.
For permission to photocopy or use material electronically from this work, please access www.copy-
right.com (http://www.copyright.com/) or contact the Copyright Clearance Center, Inc. (CCC), 222
Rosewood Drive, Danvers, MA 01923, 978-750-8400. CCC is a not-for-profit organization that pro-
vides licenses and registration for a variety of users. For organizations that have been granted a pho-
tocopy license by the CCC, a separate system of payment has been arranged.
Trademark Notice: Product or corporate names may be trademarks or registered trademarks, and are
used only for identification and explanation without intent to infringe.
Visit the Taylor & Francis Web site at
http://www.taylorandfrancis.com
and the CRC Press Web site at
http://www.crcpress.com
Contents
Dedication..................................................................................................................ix
Preface....................................................................................................................xvii
Acknowledgments....................................................................................................xix
Editors......................................................................................................................xxi
Contributors.......................................................................................................... xxiii
Section I Interfaces, Wettability and Adhesion
Chapter 1 Study of Metal–Polymer Interfaces: Relevance to Adhesion................ 3

Jörg F. Friedrich
Chapter 2 Lateral Adhesion Force Measurements at the Drop–Surface

Interface............................................................................................... 31
Aisha Leh, Hartmann E. N’guessan, Prashant Bahadur,
Rajiv Jaini, and Rafael Tadmor
Chapter 3 Particle Adhesion in the Pharmaceutical Sciences............................. 41

Frank M. Etzler
Chapter 4 The Effects of Nanostructure and Composition of Solid

Surfaces on Ice Adhesion.................................................................... 67
Hanna Dodiuk, S. Kenig, and A. Dotan
Chapter 5 A Review on the Fabrication of Nanostructured

Superhydrophobic Aluminum Surfaces.............................................. 83
Dilip K. Sarkar and N. M. Saleema
Section II Surface Modification of Polymers
Chapter 6 Surface Modification of ABS for Electroless Copper Plating

without Chromic Acid Etching......................................................... 127
Norihiro Inagaki
v
vi Contents
Chapter 7 Achieving Nanoscale Surface Structure on Polyester Film by

Atmospheric Pressure Plasma Treatment.......................................... 141
W. Y. I. Tsoi, C. W. Kan, C. W. M. Yuen, T. M. Choi, and T. B. Tang
Chapter 8 Aerosol-Based DBD: A Simple Way to Provide Polymers with

Functional Groups for Adhesion Promotion..................................... 155
Renate Mix, H. Yin, and Jörg F. Friedrich
Section III Adhesion Aspects of Bio-Based

Materials and Bio-Adhesion
Chapter 9 Wood Joints Adhesion and Performance in Mechanical Friction

Welding of Wood without Adhesives................................................ 173
A. Pizzi
Chapter 10 Effect of Pretreatment of Rice Straw Used as a Bio-Filler in

Reinforced Polypropylene Composite............................................... 181
A. Karim, M. Masudul Hassan, and Mubarak A. Khan
Chapter 11 “Green” Composites Based on Bacterial Cellulose Produced

Using Novel Low-Cost Carbon Source and Soy Protein Resin........ 193
Kaiyan Qiu and Anil N. Netravali
Chapter 12 Preparing Medium Density Fiberboard Surfaces for Powder

Coating: Effects of Atmospheric Plasma Treatment, Infrared
Irradiation, and Climatization...........................................................209
Christoph Jocham, Judith Sinic, Thomas W. Schmidt,
Uwe Müller, and Andreas Kandelbauer
Chapter 13 Adhesion Aspects in Dentistry: Bonding to Zirconia....................... 227

Jukka P. Matinlinna and Moustafa N. Aboushelib
Chapter 14 Escherichia coli O157:H7 Interactions and Transport in

Unsaturated Porous Media................................................................ 239
Yongjun Mao and Gang Chen
Contents vii
Section IV Adhesives and Their Testing
Chapter 15 Improved Approaches to Analyze the Nonlinear Behavior of

Adhesives in Bonded Assemblies..................................................... 255
Jean Yves Cognard, Romain Créac’hcadec, Julien Maurice,
Laurent Sohier, Bernard Gineste, and Peter Davies
Chapter 16 Soy Proteins as Wood Adhesives...................................................... 277

Charles R. Frihart, Christopher G. Hunt, and Michael J. Birkeland
Section V Nanomaterials and Nanocomposites
Chapter 17 Carbon Nanotube-Based Interphase Sensor for Structural

Health Monitoring in Continuous Glass Fiber-Reinforced
Polypropylene.................................................................................... 295
Julius Rausch and Edith Mäder
Chapter 18 Reaction of Multiwalled Carbon Nanotubes with Gaseous

Oxygen and Chlorine Atoms............................................................. 319
L. Oliveira, T. Debies, and G. A. Takacs
Chapter 19 Nanoparticle Deposition on Packaging Materials by

Liquid Flame Spray: Generation of Superhydrophilic and
Superhydrophobic Coatings.............................................................. 331
Hannu Teisala, Mikko Tuominen, Mikko Aromaa,
Milena Stepien, Jyrki M. Mäkelä, Jarkko J. Saarinen,
Martti Toivakka, and Jurkka Kuusipalo
Dedication
DR. KASHMIRI LAL MITTAL
From Humble Beginnings to Stellar Career …
A famous Latin phrase per aspera ad astra, “through hardships to the stars,” best epito-
mizes Dr. Mittal’s life.
The early years of Kash, as he is known affectionately, in a very small Indian vil-
lage of Kilrodh in the state of Punjab (now known as Haryana), where he was born
on the 15th of October 1945, gave him a cocktail of character-shaping exposures,
paving a path toward a successful future.
The ancient culture and rich history of India, centuries of mind-boggling exploits of
inhabitants, and their deeply ingrained art of survival against adversities and adversar-
ies were inadvertently blended with Dr. Mittal’s own exposure to true hardships, coping
with which undoubtedly turned him into a resourceful, battle-hardened and yet gentle
individual who knew from his early years that dreams, talent, and skills alone were not
enough to achieve success.
Traits such as determination, hard work, self-drive, fighting spirit, resolve, and
intelligence aided by a bit of kismet (good luck) in which he believes deeply were
immediately put to test and received fine-tuning from the earliest stages of his life
due to the remoteness of his native village, absence of a local school, and the lack of
ix
x Dedication
electricity. From his tender years, he was required to work hard and impose tough
self-discipline and a sense of mission: daily walks to a remote school in all weather
conditions, severe cold in winter and scorching summer heat, and reading by candle-
light or natural moonlight. To make things worse, Kash lost sight in his right eye at
the age of 13, which evoked his stoic, humorously positive comment: “If I had both
eyes, I could read and produce double.”
The initial stages of Dr. Mittal’s education, all crowned with the highest marks,
were accomplished in India. After completing his high school education, he enrolled in
Nehru College Jhajjar (named after the first prime minister of free India, Pt Jawaharlal
Nehru) and earned his BSc degree from Panjab University in 1964 ranked 4th among
approximately 40,000 students. This was followed by an MSc (first position) in chemis-
try in 1966 from the Indian Institute of Technology, New Delhi.
In 1966, he migrated to the United States to study at the University of Southern
California, Los Angeles, and earned his PhD degree in colloid chemistry in 1970
after submitting his thesis on the topic of “Factors influencing ultracentrifugal stabil-
ity of oil-in-water emulsions.”
After a brief stint (1970–1971) as a postdoc at The Pennsylvania State University,
he joined the Electrochemistry Laboratory of the University of Pennsylvania as a
postdoc and worked on passivation of metals and heterogeneous electron transfer
mechanisms. Some of this work resulted in a publication in Nature in 1971.
In 1972, Dr. Mittal accepted a postdoctoral position at IBM, San Jose, California,
where he made his first foray into the wonderful world of adhesion science and tech-
nology, and in his own humorous words, “he got stuck” in this field since then. In
1974, he moved to IBM, Poughkeepsie, New York, as a regular employee and then
worked in East Fishkill and Thornwood, New York, until January 1994.
The focus of Dr. Mittal’s scientific work at IBM was on surface chemical aspects
of adhesion and cleaning targeting thin-layered structures for electronic subsystems
under development for computing applications. This included acid–base interac-
tions, surface modification, silanes, tailoring interfaces in thin-film multimaterial
structures, and characterization of surface and interface chemistry by the then most
advanced and just evolving analytical tools such as x-ray photoelectron spectroscopy
(XPS) and inelastic electron tunneling spectroscopy (IETS). In this work, he devel-
oped an avid passion for interfaces, and in his own words, “an interface is the seat of
action, or where the rubber meets the road.”
Dr. Mittal is an ardent believer in teaching and education and is a gifted lecturer,
able to succinctly and lucidly present and explain complex phenomena from their sci-
entific fundamentals to practical aspects. With this premise, he developed and subse-
quently taught courses in surface chemistry and adhesion for IBM employees. At the
same time, he developed a widely offered, full-fledged course on “Adhesion Science
and Technology,” which has been taken by about 10,000 individuals around the globe.
Because of his persistent passion for teaching and sharing scientific knowledge, he says
in his own words, “my destiny is to spread the ‘gospel’ of adhesion science.”
Dr. Mittal always believed that further education of industrial scientists and aca-
demics combined with rapid dissemination of the most recent scientific develop-
ments through publication of fast-response scientific journals were prerequisites for
the steep progress of science and technology. In line with this belief, in 1986, he
Dedication xi
became a founding editor of a very successful scientific journal in this field, Journal
of Adhesion Science and Technology (JAST), for which he was the editor-in-chief
until April 2012. He is also the founding editor of the new journal, Reviews of
Adhesion and Adhesives, which made its debut in February 2013, as well as of
the new book series entitled Adhesion and Adhesives: Fundamental and Applied
Aspects.
Due to his highly successful initiative as a publisher and an indefatigable lec-
turer, Dr. Mittal became a renowned and sought-after speaker disseminating adhe-
sion science and technology through numerous invited presentations, lectures, and
courses covering multifaceted aspects in this field of science for scientific and indus-
trial organizations and societies in many countries all over the globe, and has been
around the world three times to spread the message of adhesion.
Dr. Mittal’s contributions to advancement in the fields of surface and colloid sci-
ence and in particular in adhesion have resulted in global, peer-recognized acclaim
through the following honors and awards (listed in reverse chronological order):
• Title of Doctor honoris causa awarded by the Maria Curie-Sklodowska

University, Lublin/Poland (2003)
• Establishment of the biennial Kash Mittal Award by the worldwide surface
and colloid science community (awarded to peer-assessed scientists active
in this field) in recognition of his large contributions to the field of colloid
and interface chemistry (2002)
• Adhesives Age Award (1997)
• Adhesives Award of ASTM International Committee D-14 (1997)
• John A. Wagnon Technical Achievement Award of the International
Microelectronics and Packaging Society (IMAPS) (1977)
• Thomas D. Callinan Award of the Dielectric Science and Technology
Division of the Electrochemical Society (1995)
• Robert L. Patrick Fellow title of the Adhesion Society (1990)
• Charles B. Dudley Award of ASTM International (1990)
• “Recognition Plaque for Continued Leadership and Distinguished Profes
sional Service” presented by the international surface and colloid science
community comprising prominent scientists from 51 countries at the 6th
International Symposium on Surfactants in Solution (SIS), Delhi (1986)
• Invitation by the International Advisory Panel and Chinese Review
Commission (under the auspices of the World Bank) as a project specialist
to visit Shanghai Jiao Tong University, Shanghai, China (1985)
Dr. Mittal’s outstanding contributions to the global adhesion community and

his achievements in disseminating the science of adhesion have received accolades
through the following global events and associated publications (listed in reverse
chronological order):
• Special issue of Colloids and Surfaces A: Physicochemical and Engineering

Aspects (volume 391, issues 1–3, 2011), dedicated to him to honor the pub-
lication of his 100th edited book.
xii Dedication
• Special Symposium on “Surfactant Science and Technology: Retrospects

and Prospects” in Melbourne, Australia (November 2010), organized in his
honor to commemorate the publication of his 100th edited book, and docu-
mented in a Festschrift volume dedicated to him.
• Special Symposium on “Recent Advances in Adhesion Science and
Technology,” organized in honor of the publication of his 100th edited book at
the American Chemical Society (ACS) meeting in Boston, August 2010. The
current volume represents the Festschrift from this event dedicated to him.
• Special issue of Particulate Science and Technology—An International Journal
(volume 25, no. 1, Jan./Feb. 2007), dedicated to him for his 60th birthday.
• Advances in Colloid and Interface Science (volumes 123–126, 2006), dedi-
cated to him in honor of his 60th birthday.
• The 1st International Congress on Adhesion Science & Technology held
in his honor for his 50th birthday in Amsterdam, the Netherlands, October
1995 (235 papers from 38 countries were presented).
• Special issue (volume 13, nos. 3 and 4, July–Dec. 1995) of Particulate
Science and Technology—An International Journal, dedicated to him on
the occasion of his 50th birthday.
• Adhesives Age (September 1995)—An interview with him recognizing
his contributions marked by the Amsterdam 1st International Congress on
Adhesion Science & Technology.
He is listed in numerous biographical references: American Men and Women

of Science, International Who’s Who of Contemporary Achievement, Men of
Achievement, Who’s Who in Frontier Science and Technology, International Who’s
Who of Intellectuals, Who’s Who in the East (USA), Who’s Who in Technology
Today, and others. He has also served on the advisory boards and committees of
numerous international conferences.
Dr. Mittal has been, or is, a member of the editorial boards of the following
journals:
• Adhesives Age
• Advances in Colloid and Interface Science
• Journal of Adhesion
• Journal of Coatings Technology
• Journal of Polymer Materials
• Journal of Surface Science and Technology
• Particulate Science and Technology: An International Journal
• Precision Cleaning
• Progress in Organic Coatings
• Solid State Technology
• Southern Brazilian Journal of Chemistry
• Surface Innovations
He is also a member of the editorial advisory board of the Encyclopedia of Surface

& Colloid Science.
Dedication xiii
During his scientific research career, Dr. Mittal published approximately 75 papers
in the areas of surface and colloid chemistry, adhesion, polymers, surface cleaning, etc.
He has organized and chaired a large number of very successful international symposia
and remarkably is the editor of 112 published books (see the list on the following page)
and counting!
As the editors of this Festschrift documenting papers presented at the Special
Symposium on “Recent Advances in Adhesion Science and Technology” held at
the American Chemical Society meeting in honor of Dr. Mittal’s publication of his
100th edited book and coinciding with his 65th birthday (100 at 65!), on behalf of
the international coterie of adhesionists attending this event, we collectively salute
him, extending our sincere wishes for good health and good luck to Kash in the years
to come.
W. (Voytek) Gutowski
CSIRO Materials Science and Engineering
Highett-Melbourne, Australia
Hanna Dodiuk
Shenkar College of Engineering and Design
Ramat-Gan, Israel
xiv Dedication
LIST OF BOOKS EDITED BY DR. KASH MITTAL

Title Year
1. Adsorption at Interfaces 1975
2. Colloidal Dispersions and Micellar Behavior 1975
3. Micellization, Solubilization, and Microemulsions, Vol. 1 1977
4. Micellization, Solubilization, and Microemulsions, Vol. 2 1977
5. Adhesion Measurement of Thin Films, Thick Films and Bulk Coatings 1978
6. Surface Contamination: Genesis, Detection and Control, Vol. 1 1979
7. Surface Contamination: Genesis, Detection and Control, Vol. 2 1979
8. Solution Chemistry of Surfactants, Vol. 1 1979
9. Solution Chemistry of Surfactants, Vol. 2 1979
10. Solution Behavior of Surfactants: Theoretical and Applied Aspects, Vol. 1 1982
11. Solution Behavior of Surfactants: Theoretical and Applied Aspects, Vol. 2 1982
12. Physicochemical Aspects of Polymer Surfaces, Vol. 1 1983
13. Physicochemical Aspects of Polymer Surfaces, Vol. 2 1983
14. Adhesion Aspects of Polymeric Coatings 1983
15. Surfactants in Solution, Vol. 1 1984
18. Adhesive Joints: Formation, Characteristics and Testing 1984
19. Polyimides: Synthesis, Characterization and Applications, Vol. 1 1984
20. Polyimides: Synthesis, Characterization and Applications, Vol. 2 1984
24. Treatise on Clean Surface Technology, Vol. 1 1987
25. Surface and Colloid Science in Computer Technology 1987
26. Particles on Surfaces 1: Detection, Adhesion and Removal 1988
27. Opportunities and Research Needs in Adhesion Science and Technology 1988
28. Particles in Gases and Liquids 1: Detection, Characterization and Control 1989
34. Metallized Plastics 1: Fundamental and Applied Aspects 1989
35. Polymers in Information Storage Technology 1989
37. Acid–Base Interactions: Relevance to Adhesion Science and Technology 1991
41. Silanes and Other Coupling Agents 1992
44. Contact Angle, Wettability and Adhesion 1993
45. Handbook of Adhesive Technology 1994
46. Plasma Surface Modification of Polymers: Relevance to Adhesion 1994
Dedication xv
47. Particles on Surfaces: Detection, Adhesion and Removal 1995

48. Fundamentals of Adhesion and Interfaces 1995
49. Adhesion Measurement of Films and Coatings 1995
50. Polymer Surface Modification: Relevance to Adhesion 1996
51. Polyimides: Fundamentals and Applications 1996
52. Surfactants in Solution 1996
53. Polymer Surfaces and Interfaces: Characterization, Modification and Application 1997
54. Metallized Plastics: Fundamentals and Applications 1998
55. Metallized Plastics 5 & 6: Fundamental and Applied Aspects 1998
56. Adhesion Promotion Techniques: Technological Applications 1999
57. Handbook of Microemulsion Science and Technology 1999
58. Particles on Surfaces 5 & 6: Detection, Adhesion and Removal 1999
59. Emulsions, Foams and Thin Films 2000
60. Apparent and Microscopic Contact Angles 2000
61. Acid–Base Interactions: Relevance to Adhesion Science and Technology, Vol. 2 2000
62. Silanes and Other Coupling Agents, Vol. 2 2000
63. Polymer Surface Modification: Relevance to Adhesion, Vol. 2 2000
64. Adhesion Measurement of Films and Coatings, Vol. 2 2001
65. Adhesion Aspects of Thins Films, Vol. 1 2001
66. Polyimides and Other High Temperature Polymers: Synthesis, Characterization
and Applications, Vol. 1 2001
68. Adhesive Joints: Formation, Characteristics and Testing, Vol. 2 2002
70. Contact Angle, Wettability and Adhesion, Vol. 2 2002
71. Adsorption and Aggregation of Surfactants in Solution 2003
72. Adhesion Aspects of Polymeric Coatings, Vol. 2 2003
73. Handbook of Adhesive Technology, 2nd edition 2003
74. Surface Contamination and Cleaning, Vol. 1 2003
81. Adhesion, Aspects of Thin Films, Vol. 2 2005
82. Atomic Force Microscopy in Adhesion Studies 2005
88. Adhesion Aspects of Thin Films, Vol. 3 2007
89. Developments in Surface Contamination and Cleaning: Fundamentals and Applied
Aspects 2008
90. Surfactants in Tribology 2008
xvi Dedication

92. Electrically Conductive Adhesives 2008
96. Superhydrophobic Surfaces 2009
98. Adhesion Aspects in Dentistry 2009
99. Handbook of Sealant Technology 2009
100. Developments in Surface Contamination and Cleaning, Vol. 2 2010
101. Surfactants in Tribology, Vol. 2 2011
103. Wood Adhesives 2011
104. Surface and Interfacial Aspects of Cell Adhesion 2011
105. Adhesion Aspects in MEMS/NEMS 2011
107. Surfactants in Tribology, Vol. 3 2013
109. Advances in Contact Angle, Wettability and Adhesion, Vol. 1 2013
110. Atmospheric Pressure Plasma Treatment of Polymers: Relevance to Adhesion 2013
112. Advances in Modeling and Design of Adhesively Bonded Systems 2013
Preface
This Festschrift volume dedicated to Dr. Kashmiri Lal Mittal (affectionately known
as Kash by his peers and friends) contains selected papers presented at the Special
Symposium on Recent Advances in Adhesion Science and Technology held in honor
of Kash to commemorate the publication of his 100th edited book (a remarkable and
rare achievement!) at the American Chemical Society meeting in Boston, August
2010. Our idea of organizing the Symposium as a tribute to Kash immediately
received enthusiastic response from the global adhesion community. This, in itself,
is a testimonial to the high esteem in which Kash is held. The final program for this
symposium spanned over nine sessions and contained 71 papers covering many fac-
ets of adhesion science and adhesive technology and presented by scientists hailing
from many corners of the globe.
It is difficult to imagine any contemporary product used at home, in industry, land
transportation, aerospace, in the medical field, or anywhere else that does not use, in
some manner, an adhesive, a sealant, or a decorative coating. Just about every article
or object we encounter has a surface that has been cleaned or treated in some way
during its manufacture or decorative finishing. The surface is what we first recognize
when we see or touch an article or object. The surface treatment is responsible for the
decoration, surface feel, and protection of surfaces, including corrosion protection.
New developments, for example, nanotechnology and encapsulation techniques,
have opened up new opportunities for incorporation of new surface functions.
In addition to adhesive bonding and coatings, the phenomena of adhesion is per-
vasive in many other areas (e.g., microelectronics, multilayer devices, metallized
plastics, thin film technology, optics, biomedical, etc.). The subject of adhesion is
truly interdisciplinary, multidisciplinary, and transdisciplinary, and for a holis-
tic approach to adhesion, a synergistic interaction is required among the various
disciplines involved, viz., surface chemistry, rheology, and fracture mechanics. Of
course, surface chemistry constitutes the “marrow of adhesion science.”
In spite of the tremendous importance of adhesion, not many universities (of course,
there are some) offer formal courses and/or degrees in adhesion science and adhesive
technology. Consequently, many of us who currently work with adhesion or adhesives
quite possibly became “experts” in adhesion science and technology through self-
learning and on-the-job training. Accordingly, right now, we have a new breed of experts
known as adhesion scientists (or adhesionists or more humorously as stickologists).
In our quest for finding answers to the challenging questions related to coat-
ing adhesion or adhesive bonding, most likely our librarian pointed to the books
published by K.L. Mittal. This availability of publications by Kash helped in our
romance with adhesion science.
This Festschrift volume, containing 19 articles, is divided into five sections as follows:
Section I: Interfaces, Wettability and Adhesion

Section II: Surface Modification of Polymers
xvii
xviii Preface
Section III: Adhesion Aspects of Bio-Based Materials and Bio-Adhesion

Section IV: Adhesives and Their Testing
Section V: Nanomaterials and Nanocomposites
The topics covered include metal–polymer interfaces and ways to improve adhe-
sion; lateral force at liquid–solid interface; particle adhesion in pharmaceutical
sciences; effect of solid surface characteristics on ice adhesion; superhydropho-
bic aluminum surfaces; surface modification of polymers by various techniques
to improve their adhesion property; wood joints formed without use of adhesives;
reinforced polymer composites using different fillers; “green” composites; medium
density fiber board (MDF) surfaces for powder coating; adhesion aspects in den-
tistry; E. coli interactions in porous media; analysis of adhesive behavior in bonded
assemblies; soy proteins as wood adhesives; carbon nanotube based interphase sen-
sor; reaction of multiwalled carbon nanotubes with gaseous atoms; and generation of
superhydrophilic and superhydrophobic coatings by deposition of particles on pack-
aging materials.
It should be noted that although there has been a delay in publishing this book, the
contents are very current as the authors were asked not only to revise but also update
their articles. Consequently, contributions in this book offer an insight into recent
developments and perspectives on many significant ramifications of adhesion sci-
ence and adhesive technology. Hence, it should be of interest to individuals engaged
and/or interested in the world of adhesion in academia, government research labora-
tories, a variety of industries, and other research institutes. The articles are written
by world-renowned researchers.
Prof. Dr.-Ing. W. (Voytek) Gutowski

Chief Research Scientist and CSIRO Fellow
E-mail: Voytek.Gutowski@csiro.au
Prof. Hanna Dodiuk

Head of Plastics Engineering and Polymers Department
The Pernick Faculty of Engineering
Shenkar College
Ramat-Gan, Israel
E-mail: Hannad@shenkar.ac.il
Acknowledgments
Now it is our pleasure to acknowledge those who helped in many ways to materialize
this Festschrift volume, which we are sure Kash will cherish.
First, we are thankful to the appropriate officials of the Division of Colloid and
Surface Chemistry of the American Chemical Society for enthusiastically accommo-
dating such a large symposium in their overall technical program. Also, thanks are
due to Brill (publisher of the Journal of Adhesion Science and Technology [JAST]
at that time) for providing financial support for this symposium and for arranging
and hosting the banquet in honor of Kash; we are particularly grateful to Michiel
Thijssen from Brill.
Our most sincere thanks are extended to the authors for their contribution, enthusi-
asm, patience, and cooperation, without which this special tribute to Kash would not
have been possible. We acknowledge with thanks the time and effort of the reviewers
in providing valuable comments. Our appreciation is also extended to Barbara Glunn
(CRC Press) for her steadfast interest and unwavering help in bringing this book to
fruition. This book could not have been completed without the coaching we received
with respect to editing from our “Teacher-Superior” himself—Kash Mittal.
In closing, the editors of this book would like to extend very special thanks to
their closest families without whose help the accomplishment of this task would have
been very difficult. In particular, Hanna would like to offer her deepest thanks to her
husband Sam Kenig, daughters Lani and Roni, and grandchildren Jonathan, Coral,
and Daria, while Voytek expresses his deepest gratitude to his wife Carolyn for her
most unselfish attitude and support for this commitment despite her own ongoing
battle with debilitating illness.
Prof. Dr.-Ing. W. (Voytek) Gutowski

Chief Research Scientist and CSIRO Fellow
E-mail: Voytek.Gutowski@csiro.au
Prof. Hanna Dodiuk

Head of Plastics Engineering and Polymers Department
The Pernick Faculty of Engineering
Shenkar College
Ramat-Gan, Israel
E-mail: Hannad@shenkar.ac.il
xix
Editors
Voytek (Wojciech) Gutowski earned his PhD in mate-
rials science from the Technical University of Szczecin
(Stettin) in Poland in 1975 where he worked until 1978
investigating adhesion between organic and inorganic
binders and refractory materials. His developments
received numerous national awards including Minister
of Science Awards (1973 and 1976) and the Medal of
Nicolaus Copernicus (1978).
In 1981, he migrated to Australia where he joined
the Commonwealth Scientific and Industrial Research
Organisation (CSIRO), in 1983. He works in the Division of Materials Science and
Engineering as a team leader of the “functional interphases & coatings” team. His
research covers adhesion science and technology and intelligent multifunctional
interphases for polymers and composites. His achievements in science and its trans-
lation to industrial technologies have brought him the highest Australian and inter-
national acclaim through numerous prestigious scientific and industrial awards,
including the Montell Global Innovation Award (1997), International Plueddemann
Prize (2000), two Society of Plastics Engineers (SPE) Awards in 2008, and the
Victoria Prize in 2010. He was appointed to CSIRO Fellow position in 2012.
Dr. Gutowski is or has been a member of the editorial boards of five interna-
tional journals: Composite Interfaces, Journal of Adhesion Science and Technology,
Journal of Natural Fibres, Reviews of Adhesion and Adhesives, and International
Journal of Automotive Composites. His publications include 78 generic patent fami-
lies and patent applications, 1 book, 22 book chapters, and 120 scientific papers.
Professor Hanna Dodiuk earned her PhD with honors

in chemistry in 1977 from Tel-Aviv University, Israel.
She joined the Israeli Armament Development Authority
(ADA) in 1979. From September 1991 to June 1997,
she held the position of director of the Materials and
Processes Department in ADA. She has spent sabbati-
cals as a visiting professor at the Polytechnic University
of New York, U.S.A.; at the HPM Company in Los
Angeles, California, U.S.A.; at Bayer AG, Leverkusen,
Germany; at the New Jersey Institute of Technology,
Newark, New Jersey, U.S.A.; and at the Center for High-Rate Nanomanufacturing,
University of Massachusetts Lowell, Lowell, Massachusetts, U.S.A.
Dr. Dodiuk was the president of the Israeli Polymers & Plastics Society in 1998–
2000 and was nominated as an honorary fellow in 2013. She serves as a representative
and expert evaluator on the European Community’s R&D Grants within Frameworks
xxi
xxii Editors
4–7. In 1997, she joined Shenkar College of Engineering and Design, where she was
appointed the head of the Plastics Engineering and Polymers Department in 2006.
Prof. Dodiuk specializes in the areas of adhesion science and adhesives tech-
nology, characterization and formulation of polymer adhesives, special coatings,
surface and interfaces analysis, nanotechnology, and aging of polymeric materials.
She is a member of the editorial boards of the Journal of Adhesion Science and
Technology and Reviews of Adhesion and Adhesives and is the author or a coauthor
of 130 scientific publications and 15 patents and the editor of 2 books. She has deliv-
ered numerous (160) lectures in conferences.
Contributors
Moustafa N. Aboushelib Jean Yves Cognard
Dental Biomaterials Department Brest Laboratory of Mechanics and
Faculty of Dentistry Systems
The University of Alexandria ENSTA Bretagne
Alexandria, Egypt 29806 Brest, France
Mikko Aromaa Romain Créac’hcadec

Aerosol Physics Laboratory Brest Laboratory of Mechanics and
Department of Physics Systems
Tampere University of Technology ENSTA Bretagne
P.O. Box 692, FI-33101 29806 Brest, France
Tampere, Finland
Peter Davies
Prashant Bahadur Materials and Structures Group
Dan F. Smith Department of IFREMER
Chemical Engineering Brest Centre
Lamar University 29280 Plouzané, France
Beaumont, Texas, 77710, USA
T. Debies
Michael J. Birkeland Xerox Analytical Services
Heartland Resource Technologies Xerox Corporation
3037 Artesian Lane Webster, New York 14580, USA
Madison, Wisconsin, 53713, USA
Hanna Dodiuk
Gang Chen Department of Plastics Engineering
Department of Civil and Shenkar College of Engineering and
Environmental Engineering Design
FAMU-FSU College of Engineering 12 Anna Frank Street
2525 Pottsdamer Street Ramat-Gan 52526, Israel
Tallahassee, Florida, 32310, USA
A. Dotan
T. M. Choi Department of Plastics Engineering
Institute of Textiles and Clothing Shenkar College of Engineering and
The Hong Kong Polytechnic University Design
Hung Hom Kowloon, Hong Kong 12 Anna Frank Street
Ramat-Gan 52526, Israel
xxiii
xxiv Contributors
Frank M. Etzler Christoph Jocham

School of Pharmacy WOOD Carinthian Competence Center
Lake Erie College of Kompetenzzentrum Holz GmbH
Osteopathic Medicine Klagenfurterstrasse 87-89
1858 West Grandview Blvd. A-9300 St. Veit an der Glan, Austria
Erie, Pennsylvania 16509, USA
C. W. Kan
JÖrg F. Friedrich Institute of Textiles and Clothing
Federal Institute for Materials Research The Hong Kong Polytechnic University
and Testing (BAM) Hung Hom Kowloon, Hong Kong
Unter den Eichen 87
12205 Berlin, Germany Andreas Kandelbauer
School of Applied Chemistry
Charles R. Frihart Reutlingen University
Forest Products Laboratory Alteburgstrasse 150, D-72762
One Gifford Pinchot Drive Reutlingen, Germany
Madison, Wisconsin 53726, USA
A. Karim
Bernard Gineste Department of Chemistry
Brest Laboratory of Mechanics and M C College
Systems National University
UBO, 29285 Brest, France Sylhet-3100, Bangladesh
M. Masudul Hassan S. Kenig

Department of Chemistry Department of Plastics Engineering
M C College Shenkar College of Engineering and
National University Design
Sylhet-3100, Bangladesh 12 Anna Frank Street
Ramat-Gan 52526, Israel
Christopher G. Hunt
Forest Products Laboratory Mubarak A. Khan
One Gifford Pinchot Drive Institute of Radiation and
Madison, Wisconsin 53726, USA Polymer Technology
Bangladesh Atomic Energy
Norihiro Inagaki Commission
Laboratory of Polymer Chemistry Dhaka-1000, Bangladesh
Faculty of Engineering
Shizuoka University Jurkka Kuusipalo
3-5-1 Johoku Paper Converting and Packaging
Hamamatsu, 432-8561 Japan Technology
Department of Energy and Process
Rajiv Jaini Engineering
Dan F. Smith Department of Tampere University of Technology
Chemical Engineering P.O. Box 589, FI-33101
Lamar University Tampere, Finland
Beaumont, Texas 77710, USA
Contributors xxv
Aisha Leh Uwe Müller

Dan F. Smith Department of WOOD Carinthian Competence Center
Chemical Engineering Kompetenzzentrum Holz GmbH
Lamar University Klagenfurstrasse 87-89, A9300
Beaumont, Texas 77710, USA St. Veit an der Glan, Austria
Edith Mäder Anil N. Netravali

Department of Composites Department of Fiber Science and
Leibniz Institute of Polymer Apparel Design
Research Dresden Cornell University
Hohe Strasse 6 Ithaca, New York 14853-4401, USA
01069 Dresden, Germany
Hartmann E. N’guessan
Jyrki M. Mäkelä Dan F. Smith Department of
Aerosol Physics Laboratory Chemical Engineering
Department of Physics Lamar University
Tampere University of Technology Beaumont, Texas 77710, USA
P.O. Box 692, FI-33101
Tampere, Finland L. Oliveira
The School of Chemistry and
Yongjun Mao Materials Science
Medical College Hospital of Rochester Institute of Technology
Qingdao University Rochester, New York 14623, USA
16 Jiangsu Road
Qingdao 266003 A. Pizzi
People’s Republic of China LERMAB-ENSTIB
University of Lorraine
Jukka P. Matinlinna 27 rue Philippe Seguin
Dental Materials Science 88000 Epinal, France
Faculty of Dentistry
The University of Hong Kong Kaiyan Qiu
Hong Kong SAR, People’s Republic of Department of Fiber Science and
China Apparel Design
Cornell University
Julien Maurice Ithaca, New York 14853-4401, USA
Brest Laboratory of Mechanics and
Systems Julius Rausch
ENSTA Bretagne Department of Composites
29806 Brest, France Leibniz Institute of Polymer
Research Dresden
Renate Mix Hohe Strasse 6
Federal Institute for Materials Research 01069 Dresden, Germany
and Testing (BAM)
Unter den Eichen 87
12205 Berlin, Germany
xxvi Contributors
Jarkko J. Saarinen Rafael Tadmor

Laboratory of Paper Coating and Dan F. Smith Department of
Converting Chemical Engineering
Center for Functional Materials Lamar University
Abo Akademi University Beaumont, Texas 77710, USA
Porthansgatan 3, FI-20500
Abo/Turku, Finland G. A. Takacs
The School of Chemistry and
N. M. Saleema Materials Science
National Research Council of Canada Rochester Institute of Technology
(NRC) Rochester, New York 14623, USA
501 Boulevard University East
Saguenay, Québec, G7H 8C3, Canada T. B. Tang
Department of Physics
Dilip K. Sarkar Hong Kong Baptist University
Centre Universitaire de Recherche sur 224 Waterloo Road
L’aluminium (CURAL) Kowloon Tong, Hong Kong
Université du Québec à Chicoutimi
(UQAC) Hannu Teisala
555 Blvd. Université, Chicoutimi Paper Converting and Packaging
Saguenay, Québec G7H 2B1, Canada Technology
Department of Energy and Process
Thomas W. Schmidt Engineering
TIGER Coatings GmbH & Co. KG Tampere University of Technology
Negrellistraβe 36 P.O. Box 589, FI-33101
A-400 Wels, Austria Tampere, Finland
Judith Sinic Martti Toivakka

WOOD Carinthian Competence Center Laboratory of Paper Coating and
Kompetenzzentrum Holz GmbH Converting
Klagenfurterstrasse 87-89 Center for Functional Materials
A-9300 St. Veit an der Glan, Austria Abo Akademi University
Laurent Sohier Abo/Turku, Finland
Brest Laboratory of Mechanics and
Systems W. Y. I. Tsoi
UBO, 29285 Brest, France Institute of Textiles and Clothing
The Hong Kong Polytechnic University
Milena Stepien Hung Hom Kowloon, Hong Kong
Laboratory of Paper Coating and
Converting
Center for Functional Materials
Abo Akademi University
Abo/Turku, Finland
Contributors xxvii
Mikko Tuominen C. W. M. Yuen

Paper Converting and Packaging Institute of Textiles and Clothing
Technology The Hong Kong Polytechnic University
Department of Energy and Process Hung Hom Kowloon, Hong Kong
Engineering
Tampere University of Technology
P.O. Box 589, FI-33101
Tampere, Finland
H. Yin
Federal Institute for Materials Research
and Testing (BAM)
Unter den Eichen 87
12205 Berlin, Germany
Section I
Interfaces, Wettability
and Adhesion
1 Study of Metal–
Polymer Interfaces*
Relevance to Adhesion
Jörg F. Friedrich†
CONTENTS
1.1 Introduction....................................................................................................... 4
1.1.1 First Attempts for Introduction of Chemical Bonds at Metal–
Polymer Interface...................................................................................4
1.1.2 Interactions at Metal–Polymer Interfaces..............................................5
1.2 Factors Influencing Adhesion between Metals and Polymers........................... 6
1.2.1 Role of Surface Impurities..................................................................... 6
1.2.2 Improved Wetting by Polymer Surface Oxidation................................7
1.2.3 Improved Adhesion by Polymer Surface Oxidation..............................8
1.2.4 Dynamics at Plasma-Treated Polymer Surfaces before Metal
Deposition............................................................................................ 10
1.2.5 Mobility of Metal Atoms on Polymer Surfaces................................... 12
1.2.6 Redox Reactions at Metal–Polymer Interfaces................................... 12
1.2.7 Degradation of Double Bonds by Transition Metals........................... 13
1.2.8 Influence of Metal–Polymer Interactions on Near-Interface
Polymer Interphases............................................................................ 14
1.3 New Procedures for Enhancing Adhesion and Its Durability......................... 17
1.3.1 Polymer Surface Modification with Monosort Functional Groups..... 17
1.3.2 Covalent Bonds for Maximal Adhesion.............................................. 18
1.3.3 Architecture of Ideal Interface............................................................ 19
1.3.4 Introduction of Flexibilizing Spacers or Adhesion-Promoting Layers......21
1.3.5 Adhesion of Spacer-Equipped Metal–Polyolefin Laminates............... 23
1.3.6 The “Ideal” Nonpeelable and Humidity-Resistant Al–PP Laminate....... 24
1.4 Outlook............................................................................................................25
1.5 Summary.........................................................................................................26
Acknowledgments.....................................................................................................26
References.................................................................................................................26
* Feature report dedicated to K. L. Mittal on his 65th birthday.

† Prof. Jörg F. Friedrich can be contacted at joerg.friedrich@bam.de.
3
4 Recent Adv. in Adhesion Sci. & Technol. in Honor of Dr. Kash Mittal
To enhance earlier concepts, a new architecture of interfaces between metals

and polymers is presented. It comprises the substitution of all as far as pos-
sible weak (physical) interactions by stronger chemical bonds, the insertion
of chemically linked flexible long-chain spacer molecules, and use of water-
repellent and barrier-like elements.
Chemical bonds have bonding energies much higher than the usual physi-
cal interactions along the metal–polymer interface. The introduction of flexible
spacer molecules compensates for the considerably different thermal expansions
of polymers and metals and thus the resulting mechanical stress. The covalent
bonding of spacer molecules needs the introduction of monosort functional
groups in sufficient concentration on the polymer surface. However, such selec-
tive polymer pretreatment did not exist and had to be developed. Low-pressure
plasma surface bromination and copolymer deposition were successfully used
as well as atmospheric processes such as electrospray polymer deposition.
Additional introduction of hydrophobic spacer elements protects the metal–
polymer bonds from hydrolysis, rearrangement within redox processes across the
interface, and reactions of π-electron containing polymer bonds with d orbitals of
transition metals (dπ → pπ interactions) followed by carbide formation. Such inter-
face design produces highly adherent and humidity-resistant multilayer systems.
1.1 INTRODUCTION
1.1.1 First Attempts for Introduction of Chemical Bonds
at Metal–Polymer Interface
In the 1960s and 1970s, the substitution of weak physical interactions at metal–polymer
or inorganic–polymer composite interfaces by strong chemical bonds was the main
goal. Inspired by the work of Hollahan [1], Friedrich also used the ammonia plasma
for introducing amino groups onto the surface of poly(ethylene terephthalate) (PET)
foils and carbon fiber (CF) surfaces at the beginning of the 1970s [2–4]. PET was
subsequently electrolessly coated with copper (copper sulfate and formaldehyde) and
then sputtered with cobalt for production of magnetic tapes. The plasma-introduced
amino groups increased the adhesion of palladium nuclei needed for copper deposi-
tion. In this way, the peel strength of copper was increased, perhaps by formation
of Pd– and Cu–amino complexes. By introducing a new process, the Pd deposition
could be eliminated by deposition of thin Ni layers using the decomposition of nickel
tetracarbonyl on exposure to plasma:
Ni(CO)4 + plasma → Ni↓ + 4 CO↑
Amino groups introduced onto CF surfaces by exposing them to the ammonia

plasma were linked with epoxy resin. Thus, covalent bonds between a CF surface
and epoxy resin matrix were established [2]:
CF−NH2 + epoxy−R → CF−NH−CH2−CHOH−R

Study of Metal–Polymer Interfaces 5
Since SiOx inorganic filler surfaces did not form interactions or covalent bonds
with polystyrene matrix, the filler particles were coated with a cross-linked plasma
polystyrene, thus transferring mechanical forces to the inorganic material [2,5].
Using vinylsilane instead of styrene monomer for plasma polymer coating of SiOx
filler particles and then initiating the grafting of remaining vinyl groups to the
polystyrene matrix by peroxide, a covalent linking of plasma polymer and polymer
matrix was achieved.
1.1.2 Interactions at Metal–Polymer Interfaces

So far, only physical interactions at metal–polymer interfaces were considered, and
chemical bonds were an exception. Various types of interactions and adhesion mod-
els at metal–polymer interfaces were proposed: electrostatic [6]; mechanical hook-
ing (roughness); interpenetration/diffusion (metal aggregation and diffusion) [7];
weak boundary layer (WBL) (contaminants, weak oxide layers, degraded polymers,
redox reaction modified interfaces) [8]; thermodynamics/adsorption (wetting, adhe-
sion; Dupré) [9]; chemical bonds (Me-O-C or MeC); interfacial redox reactions
(polymer reduction by electronegative metals; Friedrich) [4]; weak interactions
(acid–base, van der Waals, hydrogen bond); and polymer–polymer interpenetra-
tion [10]. Metal clustering at polymer surfaces is due to the high cohesive energy
of the metal and the low metal–polymer interaction energy. Low deposition rates
of metals, however, hinder clustering and enhance metal diffusion into the polymer
matrix [7].
Weak physical interactions such as van der Waals forces (Keesom, Debye, and
Heitler–London), hydrogen bonds, acid–base, and several types of complex forma-
tion [11] characterize metal–polymer interactions [10]. Moreover, electrostatic forces
and forces in colloids (hard sphere, electrostatic, entropic, van der Waals forces) were
also discussed [12]. Covalent bonds between polymer and metal were assumed to
produce the highest adhesion strength [13,14]. A precondition for establishing cova-
lent bonds is to have high concentrations of reactive monosort functional groups at
the polymer surface. However, polyolefins and several other polymers do not possess
in their native state such monosort functional groups at the surface. These monotype
groups can be produced in varying density by the deposition of adhesion-promoting
thin layers produced by chemical copolymerization by varying the comonomer ratio
[15]. Another way is to deposit thin adhesion-promoting layers of functional groups
carrying polymers or copolymers by atmospheric electrospray ionization (ESI) deposi-
tion or using an atmospheric dielectric barrier discharge (DBD) (atmospheric-pressure
chemical ionization [APCI]) [16]. Low-pressure bromine or bromoform plasma
exposure of polymer surfaces was the most successful in introducing the highest
concentrations of monosort bromine groups. This process was highly selective in
C-Br group production and yielded maximal bromination of 120% Br/C for polyeth-
ylene (PE) and 100% Br/C for polypropylene (PP) [15]. It should be considered that
initiation of reactions or interactions between such produced monosort functional
groups at polymer surfaces and evaporated metal atoms needs a maximal distance of
0.1–0.2 nm. Thus, a layer of contaminants hinders chemical reactions or interactions.
1.2 FACTORS INFLUENCING ADHESION

BETWEEN METALS AND POLYMERS
1.2.1 Role of Surface Impurities
Additives, impurities, fingerprints, adsorption layers, oxidized surface layers, and so
forth inhibit physical and chemical interactions between metal and polymer. Their
removal by mechanical elimination, dissolution, sonication, etching, or plasma
stripping often allows the formation of interactions and thus improves the adhesion
of laminates. The following example shows the dependence of (lap) tensile shear
strength of polyurethane–steel laminates on the surface purity of steel (Figure 1.1).
This cleanliness was measured in terms of the C1s/Fe2p intensity ratio measured
using x-ray photoelectron spectroscopy (XPS) [17]. The results obtained from these
experiments show that a combination of rinsing with acetone, grinding with glass
sand (also called glass powder), use of a resorcinol primer, and use of low-pressure
oxygen or hydrogen plasma produces the highest lap shear strength. It was assumed
that the high purity of the cleaned steel surface and the plasma surface modification
of the residual oxide layers are the reasons for high lap shear strength values.
Polymer surfaces are also often contaminated with additives, catalysts, fingerprints,
dust, oil, and water from ambient air. Such a contamination layer blocks the interac-
tions between metal and polymer and must, therefore, be avoided, or if present, it must
be removed by solvent washing, sonication, or plasma etching. The contamination on
an industrial polyethylene foil was analyzed and then removed by solvent extraction
because metallization prior to cleaning did not produce significant adhesion. The rea-
son for poor adhesion was the presence of a slip agent (unsaturated fatty acid, here,
erucic acid amide = H3C−(CH2)7−CH=CH−(CH2)11−CO−NH2) (Figure 1.2).
35
Hydrogen low-pressure plasma
Lap shear strength of PUR-steel (MPa)
30
Oxygen
25 low-pressure
plasma
20
Grinding
with glass sand
15
Polishing
10
5 Rinsing in acetone
Treatment in an ultrasonic (25°C, 10 min)
bath of acetone (25°C)
0
0 5 10 15 20
C1s/Fe2p intensity ratio (degree of cleanliness) (arbitrary unit [a.u.])
FIGURE 1.1 Dependence of tensile shear strength of polyurethane (PUR)–steel laminates

on surface purity of steel achieved using different kinds of surface treatments.
140
120
Transmittance (%) 100
80
60
40
20
Extraction with ether (10 min)
0 As received
4000 3000 2000 1000

Wavenumber (cm–1)
FIGURE 1.2 Contamination of polyethylene foils with erucic acid amide measured by
FTIR before and after ether extraction.
1.2.2 Improved Wetting by Polymer Surface Oxidation

Polyolefins are characterized by the absence of any functional groups and, therefore,
by low surface energy with low or zero polar contribution. Oxidation of the surface
increases the surface energy linearly to 20 O/100 C and then reaches a plateau with
constant surface energy (Figure 1.3).
The same type of dependence exists also for the polar contribution of OH groups
on the surface of plasma-produced ethylene–allyl alcohol copolymers of varying
composition, as shown in Figure 1.4.
70
60
Low-pressure O2 plasma
Surface energy (mJ/m2)
50
40
Dielectric barrier discharge
30 (DBD) in air
20
10
0
0 5 10 15 20 25
Oxygen concentration (O/100 C)
FIGURE 1.3 Surface energy of polyethylene as a function of oxygen concentration for low-
pressure and atmospheric-pressure plasma treatments.
35
30
Polar contribution (mJ/m2)
25
20
15
10
0
0 2 4 6 8 10
OH group concentration (OH/100 C)
FIGURE 1.4 Calculated polar contribution of surface energy (σtotal = σpolar + σdispersive) in
dependence on concentration of OH groups of ethylene–allyl alcohol copolymers deposited
under low-pressure plasma exposure conditions.
These dependencies of surface energy σ (σ[σpolar] = f[O or OH]) do not consider

the existence of low-molecular-weight oxidized material (LMWOM) at the poly-
mer surface [18]. Such layers may be dissolved in the test liquids used to measure
contact angles, thus influencing the results. Moreover, Bikerman [8] has found that
such degraded molecules at the surface form a WBL, which decreases the adhesion
to metals because of the weakened bonding of these molecular fragments to the
polymer bulk. The leveling off at about 30 OH/100 C atoms is due to the maximum
possible concentration of OH groups in these copolymers given by the stoichiometry
of pure poly(allyl alcohol), which possesses, theoretically, 33 OH/100 C.
1.2.3 Improved Adhesion by Polymer Surface Oxidation

The surface energies of metals and inorganics are higher by about two orders of mag-
nitude than those of polymers. This strong difference results in weak interactions
between metals and polymers, such as aluminum and polyethylene. It is a common
approach to increase the surface energy of polyolefins by oxidation for improved
interactions between metal and polymers along their interface and for strongly
adherent laminates. However, after polyolefin oxidation and therefore increasing the
surface energy by about 50%–100% (cf. Figure 1.3), the large difference in sur-
face energies of metals and polymers remains, nevertheless, nearly unchanged. This
remaining difference is the driving force for further (redox) reactions. Therefore, the
interactions between metals and oxidized polyolefin surfaces have to be character-
ized as metastable. The chemical bonds (covalent bonds with high polar contribu-
tion) formed between metal and polymer are not the final reaction product; they
are only an intermediate (metastable) step, which undergoes further reactions when
activation energy is present, maybe hydrolysis. The conclusion is that the desired
(most efficient) chemical bonds are unstable over long periods. Thus, these (meta-
stable) bonds must be stabilized to preserve them for highly adherent laminates.
In contrast to the behavior of test liquids for contact angle measurements on poly-
mer surfaces, the metal layers are solid phases, which cannot dissolve and distribute
low-molecular-weight products (LMWOM) [18]. The low-molecular-weight products
remain at the metal–polymer interface as a weak boundary interlayer and thus may
produce a different behavior. For this comparison, the measured peel strength of
Al–polyolefin and Al–PET laminates [19] as well as tensile shear strengths of poly-
propylene–polyurethane laminates [20] were also referenced to oxygen percentage at
the surface of oxygen plasma–exposed polymers (Figure 1.5).
Deviation from the theoretical elemental composition of PET (C10O4 = 40 O/100 C)
produced low peel strength, which may be due to the destruction of original PET
structure and formation of a WBL at the interface.
Extensive oxygen plasma exposure (>20 O/100 C) of polyolefins may also exten-
sively produce LMWOM, which is not linked to such polymer bulk, and there-
fore, metal deposits are not linked to the polymer substrate, and the measured peel
strength became lower for >20 O/100 C. However, this LMWOM can be removed
before depositing metal using solvent washing. It is interesting to note that the sur-
face energy of PE and the Al peel strength of Al–PE laminates show a different
dependence on oxygen concentration (cf. Figures 1.4 and 1.5).
Distribution of LMWOM on the polypropylene surface produced by oxygen
plasma pretreatment within the viscous polyurethane adhesive by dissolution and dif-
fusion was expected for polyurethane–polyolefin laminates, which show a different
Theoretical
stoichiometry
500 500
Al-PE
400 400
Peel strength (N/m)
300 300
Al-PP
Al-PET
200 200
100 100
0 0
0 10 20 30 40 50 0 10 20 30 40 50
O concentration (O/100 C)
FIGURE 1.5 Peel strength of Al–polymer laminates in relation to XPS-measured O con-

centration introduced by low-pressure oxygen plasma exposure (radio-frequency [rf] plasma,
6 Pa, 100 W).
Low-pressure oxygen plasma
Tensile shear strength (MPa)

4
Atmospheric plasma jet

3
1
Atmospheric dielectric barrier discharge (DBD)
0
0 5 10 15 20 25 30
FIGURE 1.6 Tensile shear strengths of PU-PP specimen as a function of oxygen percentage
at polypropylene surfaces introduced by atmospheric and low-pressure plasmas.
type of adhesion dependence on oxygen percentage (Figure 1.6). Until 20–25 O/100
C at the polypropylene surface, a decrease in lap shear strength was not observed.
The peel strength increases rapidly with O concentration and achieves a maximal
plateau from 5 to 25 O/100 C. The DBD treatment produced the lowest tensile shear
strength. It is assumed that much degradation to LMWOM occurs using this type of
atmospheric plasma, thus forming a pronounced WBL as demonstrated by Strobel
et al. [18].
1.2.4 Dynamics at Plasma-Treated Polymer

Surfaces before Metal Deposition
Some of the oxidative or plasma-introduced functional groups are not permanently
fixed at the polymer surface. They can slowly diffuse from the uppermost layer to
the bulk to eliminate any concentration gradient. Rotation of functional groups on
a C-C bond axis and diffusion of complete polymer segments equipped with func-
tional groups are possible. This mobility of functional groups and macromolecule
segments is described by the terms “surface dynamics” or “hydrophobic recovery”
[12,21–23]. The diffusion of polymer sequences carrying functional groups is ther-
modynamically driven by the need for equilibration of the concentration gradient
between the surface and polymer bulk. Physical, hydrogen, or chemical bonding of
functional groups to water, adhesives, or metals at the surface prevents such diffu-
sion (Figure 1.7) [18].
The formation of such interactions between polar groups and counterparts over-
compensates the energy gains by equilibration of the concentration gradient [24].
The situation is more complicated if functional groups become protonated or find
counterions, are rearranged, or even grafted; then, their mobility is strongly influ-
enced [25].
X X
Surface Surface
CH2 - CH - CH2 CH2 - CH2 - CH2 CH2
Water drop
Bulk Bulk H 2O
OH
HO OH
Air Surface
Surface Surface
CH2 - CH - CH2 CH2 - CH2-
CH2 CH2
X OH
Bulk Bulk X
OH
Rotation around C-C bonds Segmental movements

Bulk
“Hydrophobic recovery”
FIGURE 1.7 Schematics of surface dynamics and “hydrophobic recovery” caused by dif-
ferent environments.
In Figure 1.8, the changes in surface energy and measured lap shear strength of
polypropylene–polyurethane joints are presented in relation to the time of exposure
to the ambient air subsequent to plasma exposure of polypropylene before bonding it
with polyurethane for joint formation. It is noticeable that even short-time storage of
treated surfaces in air strongly influences the adhesion. This effect was attributed to
the hydrophobic recovery (and auto-oxidation).
50
40
Polar contribution
30
20 Dispersion contribution
10
0
0 100 200 300 400 500 600 700 800 900 1000
Lap shear strength (MPa)
4
3
2 Interfacial failure (%)
1
Cohesive failure (%)
0
0 100 200 300 400 500 600 700 800 900 1000
Time of exposure to ambient air (h)
FIGURE 1.8 Relation between surface energy and adhesion during storage of plasma-
modified polypropylene before bonding with polyurethane.
1.2.5 Mobility of Metal Atoms on Polymer Surfaces

Surface dynamics of polymers is dependent on the mobility of metal atoms at a poly-
mer surface, that is, on the type and strength of metal–polymer interaction, the type
of layer growth, and clustering and metal cluster diffusion into the polymer matrix.
Metal cluster formation is determined by the cohesive energy of clusters, their immo-
bilization tendency, and the type of growths (Volmer-Weber, Stranski-Krastanoff, van
der Merve) (cf. Ref. [7]). The diffusion of metal clusters into the polymer increases
the contact area with the polymer. The interaction between metal and polymer matrix
influences the diffusion ability (or immobilization). As mentioned before, the surface
energies of metals and polymers are strongly different. It is well known that differ-
ences in energy tend to be balanced. This is only possible by chemical redox reactions
between metal and polymer to a thermodynamic equilibrium state. Another reaction
pathway is the hydrolysis of the polar covalent Me–O–C bonds:
Me−O−C + H2O → Me−OH + C−OH.
1.2.6 Redox Reactions at Metal–Polymer Interfaces

Evaporating metal atoms in vacuum and depositing them onto a polymer surface pro-
duces direct contact between metal atoms and functional groups or building blocks
of the polymer. Strongly electronegative metals, such as aluminum, magnesium, or
alkali metals, are very reactive toward functional groups and tend to form Me–O–C
or Me-Hal bonds or oxides (MexOy) or halides (MexHaly). The last ones are products
of typical redox reactions. In this way, functional groups and building blocks are
fragmented [26–28]. Redox reactions are often accompanied by scission of covalent
bonds of the polymer backbone. The example of a redox reaction between alkaline
metals and poly(tetrafluoroethylene) (PTFE) elucidates the situation
−[CF2−CF2]n − + 4n K → 2n C + 4n KF,
where n is number of segments and atoms, as evidenced by the C1s signal of XPS
(Figure 1.9) [28].
Aluminum also picks up oxygen from the polymer and forms Al2O3 [27]. In con-
trast with this, noble metals with positive redox potential have the tendency to con-
vert their oxide or sulfide layers at the surface into the metallic state. Thus, CuO in
contact with polyethylene is reduced to Cu and vice versa PE oxidized at the surface
[29–31]:
2 CuO + polyethylene → Cu2O + polyethylene−O or 2 Cu + polyethylene−O2.
Typical oxidation products are carboxylic groups appearing as copper carboxyl-

ates [30]. A CuO oxidation layer on the surface of copper contacts from printed circuit
boards (PCBs) has been shown to oxidize polyimide [32] and poly(phenylquinoxaline)
[26] at the interface and, surprisingly, also aromatic rings of these polymers.
C1s peak PTFE as received PTFE

after K deposition
Intensity (a.u.)
Shift after
K deposition
(thermal evaporation)
CF2 C
300 295 290 285 280

Binding energy (eV)
FIGURE 1.9 C1s signals before and after thermal evaporation of potassium onto PTFE [19].
1.2.7 Degradation of Double Bonds by Transition Metals

Aromatic rings and other π-electron–containing bond systems in polymers can
be attacked oxidatively [26] or by transition metals and with formation of dπ −
pπ-interactions [28]. It could be shown that transition metals, such as chromium,
form complexes. PET forms such complexes, which are converted to σ-bonded Cr to
the ring and finally converted to chromium carbide (Figure 1.10) [28].
Cr
Cr O Cr Cr Cr
O O O
C C C C
O O CH2 CH2 O O CH2 CH2
Cr
Cr
Cr
Cr Cr
Cr O Cr
O O O
C C C C
O O CH2 CH2 O O CH2 CH2
Cr
Cr Cr
Cr
O O Cr
Cr4C3 + Cr2O3
C C
O O CH2 CH2
FIGURE 1.10 Proposed reactions between Cr and aromatic rings of poly(ethylene

terephthalate).
7
CK-edge, 55° π*-resonance
Normalized photo electron yield (PEY) (a.u.) Fit
Multilayer Virgin Phenyl
0.4 nm Cr 6
53% Phenyl-Cr-phenyl
5
Monolayer
0.1 nm Cr
4
78% Monolayer
0.1 nm Cr
Cr
3 Multilayer
A 0.4 nm Cr
2
100% 1 Phenyl
A'
B Virgin
C1s σ* Phenyl-Cr-phenyl
0
280 290 300 310 320 330 284 285 286 287
Photon energy (eV)
FIGURE 1.11 NEXAFS spectra of C1s → π* transition of virgin poly(styrene), polystyrene

evaporated with a Cr monolayer or evaporated with a Cr multilayer of Cr on surfaces of poly-
styrene (PEY = partial electron yield).
Using polystyrene, the formation of sandwich complexes with aromatic phenyl

rings was evidenced using commercial Cr complexes as reference and near-edge
x-ray absorption fine structure (NEXAFS) spectroscopy (Figure 1.11) [33,34].
The absorption spectra in Figure 1.11 (left-hand side) show the CK edge region
with the sharp and intense π* transitions. The right side shows details of this π*
transition characteristic for the aromatic rings in poly(styrene).
Moreover, the conversion of the Cr complex (dπ −pπ interactions) to chromium
bonded to aromatic rings at two positions with formation of a two-ring system was
shown schematically in Figure 1.10 [28]. This was evidenced by blocking these two
positions using poly(2,5dimethyl-1,4-phenylene oxide) (PDMPO; cf. Figure 1.12).
The methyl side groups at the aromatic ring hinder the attachment of Cr under for-
mation of a bicyclic system.
The appearance of a new C1s peak of Cr-evaporated polystyrene at the binding
energy of 283.1 eV as well as peaks at 283.5 eV for bisphenol A polycarbonate and
283.4 eV for PET (not shown) were assigned to the formation of chromium carbide,
the end product of all reactions of Cr with the aromatic systems [35].
Moreover, additional redox reactions of Cr with O functional groups of the poly-
mer occur. It was seen by the formation of Cr(III) oxide through pickup of oxygen
from the polymer (Figure 1.12).
1.2.8 Influence of Metal–Polymer Interactions on

Near-Interface Polymer Interphases
Polymer surfaces in contact with metals often show a new supermolecular structure.
A prominent example is that of the “transcrystalline structure” of polymers along the
2000
8000
CH3 C1s Virgin O1s Virgin
O 1500
6000
CH3 1000 C-O-C

4000 n
CHx 500
2000
Shake-up
C-O
Intensity (cts.)
0 0
294 292 290 288 286 284 282 280 540 538 536 534 532 530 528 526
2000
Cr Cr 4 mono-
8000 C1s O1s 4 monolayers Cr
layers Cr 1500
6000
1000
4000
C-O-C
CHx
2000 500
Cr2O3
Shake-up
C-O
0 0
294 292 290 288 286 284 282 280 540 538 536 534 532 530 528 526
Binding energy (eV)
FIGURE 1.12 C1s- and O1s signals (XPS spectra) of poly(dimethylphenylene oxide)
(PDMPO) with and without Cr evaporation.
metal–polymer interface, as proposed by Schonhorn and coworkers [36,37]. Such

special surface orientation in micrometer dimensions was also found using “etch
gravimetry” showing more etch resistance toward oxygen plasma than for the poly-
mer bulk [38].
However, more dramatic changes in molecular orientation occur caused by
interfacial reactions in the nanometer scale. Thus, it could be shown by means of
NEXAFS spectroscopy that biaxially stretched PET loses the macromolecular ori-
entation in an approximately 3-nm-thick contact zone to chromium or potassium (cf.
Figures 1.13 and 1.14) [28].
The difference spectrum of 90–20° characterizes the molecular orientation of
macromolecules within the sampling depth of NEXAFS. If the difference is zero or
near zero, the polymer is isotropic; if the difference is significant, then the polymer
is anisotropic. It should be considered that the information depths are about 3 nm for
the CK edge and 5 nm for the OK edge. It is interesting to note that the loss in orienta-
tion near the interface to metals is similar to that of oxygen plasma exposure used for
increasing the adhesion property of the polymer surface. Both processes produce fast
and complete loss in polymer orientation (Figure 1.14).
The changes in structure and orientation become maximal when the oxygen
plasma–pretreated polymer surface is in contact with aluminum atoms during ther-
mal evaporation under low-pressure conditions [39].
Using self-assembled monolayers (SAMs) made from octadecyltrichlorosilane
(OTS) as models of paraffin or polyolefin surfaces, strong disorientation is caused by
short-time exposure to low-pressure oxygen plasma (Figure 1.15).
10 A B 10 10
D Virgin 1 monolayer Cr 4 monolayer Cr
8 CK 20° 8 20° 8 20°
CEF G
6 55° 6 6 55°
55°
4 4 4
90° 90° 90°
2 2 2
90‒20° 90‒20° 90‒20°
Normalized partial electron yield (a.u.)
0 A=π*ring; B=π*C=O; 0 0
C-G=σ* 52% 12%
–2 100% C-C, C=O, C-O –2 –2
280 300 320 280 300 320 280 300 320
H K Virgin 1 monolayer Cr 4 monolayer Cr
6 OK IJ 6 20° 6
L 20° 20°
4 55° 4 55° 4 55°
90° 90° 90°
2 2 2
90‒20° 90‒20° 90‒20°
0 0 0
H, I =π*C=O; J, K, L=σ*C=O
53%
–2 100% –2 73% –2
520 540 560 520 540 560 520 540 560
6 Virgin 6 1 monolayer K 6 4 monolayer K

OK 20° 20° 20°
4 55° 4 55° 4 55°
2 90° 2 90° 2 90°

90‒20° 90‒20° 90‒20°
0 0 0
33%
–2 100% –2 69% –2
520 540 560 520 540 560 520 540 560
Photon energy (eV)
FIGURE 1.13 NEXAFS CK and OK edge spectra in relation to interaction with monolayer
and multilayer of Cr and K.
Number of evaporated metal monolayers

0 1 2 3 4 5
CK, OK edges 90‒20° difference spectra
100 Biaxially stretched PET foil, as received

related order parameter (%)
80
OK ‒ Cr
60
OK ‒ K
40
OK ‒ O2 plasma CK ‒ Cr
20
CK ‒ O2 plasma
0
0 20 40 60
Exposure to low-pressure O2 d.c. plasma (s)
FIGURE 1.14 NEXAFS-determined order parameter of biaxially stretched PET in relation

to metal thickness and oxygen plasma exposure.
100
30
Orientation-related parameter (0 s = 100 %)

80 25
Orientation of self-assembled OTS layer
after exposure to O2 plasma
Orientation of self-assembled OTS layer
20
60 after exposure to O2 plasma and
deposition with 1 monolayer Cr
Oxygen introduction into self-assembled 15
40 OTS layer
10
20
5
0 0
0 2 4 6 8 10 12 14 16
Exposure to oxygen plasma (s)
FIGURE 1.15 Loss of orientation and oxygen introduction in self-assembled monolayers of

octadecyltrichlorosilane exposed to oxygen plasma and Cr deposition.
Much more enhanced disorientation of the OTS-SAM is observed after additional

evaporation of Cr, thus causing redox reactions with the oxygen plasma-introduced
O functional groups. This was demonstrated by evaluation of angle-dependent
NEXAFS CK-edge spectra (cf. Figure 1.13) and creating an orientation-related
parameter (Figure 1.15) [40].
The NEXAFS spectra show a complete loss of orientation of OTS molecules after
2 s oxygen plasma exposure + one monolayer Cr coating. Thus, extraordinarily strong
redox reactions were initiated between plasma-introduced O-containing groups and
metallic chromium with formation of Cr2O3 and reduction of plasma-introduced O
functional groups (cf. Figure 1.15) [41].
1.3 NEW PROCEDURES FOR ENHANCING

ADHESION AND ITS DURABILITY
1.3.1 Polymer Surface Modification with Monosort Functional Groups
A precondition for establishing metal–polymer covalent bonds or spacer bonding
to surfaces is a high-density monosort functionalization of the polyolefin surface.
Oxidations are not well suited because of the simultaneous formation of several types
of O-containing groups. Only by postplasma wet-chemical reduction with LiAlH4
[42,43] or B2H6 [44–46] could the dominance of monosort OH-group formation be
achieved. The yield in OH groups was 11 or more per 100 C, and the total O concen-
tration was about 20 O/100 C; thus, the selectivity in OH formation (%OH=[OH]/
[Ototal]) was 55%. Pulsed plasma polymerization of allyl alcohol yielded many more
OH groups (30 OH/100 C); however, it was, in principle, limited by its stoichiometry
(33 OH/100 C). The OH selectivity achieved was 90%.
140
Unwashed
Concentration of Br and O (per 100 C)

120 10 s PE (Br)
PP (Br)
100
Loss in loosely
bonded surface
80 layer by solvent
extraction in
tetrahydrofurane
60 (THF) for
15 min, RT
Washed
40 PE (Br)
PP (Br)
20
PE (O)
PP (O)
0
0 50 100 150 200 250 300
Time of exposure to the CHBr3 plasma (s)
FIGURE 1.16 Time dependence of bromine introduction onto PE and PP surfaces using rf
plasma before washing and after washing.
Atmospheric pressure electro spray ionization (ESI) deposition of polymers as

adhesion-promoting ultrathin layers does not change the composition of polymers.
Therefore, stoichiometry, chemical structure, and concentration of functional groups
in the deposited polymer layer are the same as in the original polymer, for exam-
ple, for poly(vinyl alcohol) with 50 OH/100 C and with an OH selectivity of 100%.
Additional use of corona or DBDs during ESI deposition called APCI may increase
the adhesion to the polymer substrate because of plasma activation of both deposit
and substrate material. However, significant amounts of functional groups are con-
verted, fragmented, or split off.
Underwater plasma (UWP) with or without H2O2 addition is a further alternative
to produce preferentially OH groups, however, only in concentrations of about 10
OH/100 C. The most efficient process was found to be the polyolefin surface func-
tionalization in bromine-containing low-pressure plasmas (bromoform, bromine).
Plasma bromination using bromine or bromoform as precursors produces bromi-
nation densities of 100 Br/100 C for polypropylene, 120 Br/100 C for polyethylene,
and 160 Br/100 C using combination of CHBr3/Br2 [47]. About 10 s was needed to
produce these high Br densities with very low postplasma oxidation (1–3 O/100 C)
(Figure 1.16).
After removing the LMWOM by solvent extraction, 20–40 Br/100 C remains.
This amount was sufficient for all following nucleophilic substitutions, click reac-
tions, and atom transfer radical polymerization (ATRP) [48,49].
1.3.2 Covalent Bonds for Maximal Adhesion

Copolymers of different compositions made from poly(vinylchloride) and maleic
anhydride show a characteristic dependence of peel strength of evaporated aluminum
700 Failure in polypropylene
600
COOH
500
Peel strength (N/m)
400
300
OH
200
100
NH2
Mechanically
0 unstable
0 5 10 15 20 25 30
OH, NH2 and COOH concentration (per 100 C)
FIGURE 1.17 Dependence of Al peel strength on concentration of functional groups at PP

surfaces produced by depositing 50-nm-thick layers of plasma polymerized allylamine, allyl
alcohol, and acrylic acid.
topcoats on concentration of carboxylic groups [10]. However, such bulk copolymers

of different compositions possess varying bulk properties. In contrast, polymer sur-
face functionalization or deposition of ultrathin adhesion-promoting layers does not
change the bulk properties.
Depositing plasma polymers from acrylic acid (COOH), allyl alcohol (OH), and
allylamine (NH2) for adhesion promotion of 150-nm-thick Al evaporated layers, dif-
ferent efficiencies in adhesion promotion can be observed (Figure 1.17) [50–53].
Carboxylic groups show the highest adhesion, OH groups show moderate enhance-
ment in adhesion, and amino groups show very low improvement. It is assumed that
metallic aluminum forms covalent Al-O-C bonds with OH groups and Al-O-CO
groups with carboxylic groups bonded as monodentate or bidentate ligands [54–56].
1.3.3 Architecture of Ideal Interface

The formation of various polar groups at the polyolefin surface by oxidation reactions
and the thus produced weak physical interactions to metal deposits were overcome
by introduction of unique chemical (covalent) metal–polymer bonds as described
before. Metal–polymer interfaces are characterized by an abrupt transition of metal
properties compared to those of the polymer. All stress introduced by mechanical
load applied to the metal–polymer laminate or by thermal heating, considering the
difference of two orders of magnitude in thermal expansion coefficients for metals
and polymers, is focused on the chemical bonds bridging the interface and on neigh-
boring polymer layers.
Direct linking of metal to polymer by covalent bonds makes the interface inelas-
tic, rigid, and therefore, brittle (cf. Figure 1.18a).
20
1. Covalent bonds 2. Introduction of 3. Introduction of 4. Water-repellent 5. Water-repellent and

instead of physical flexibilizing spacer spacers with different hydrophobic elements barrier-forming elements
interactions molecules chain lengths in spacers with spacers (blocking ion
diffusion, redox reactions)
Aluminium
Al Al Aluminium
Al Al
Al
Aluminium Aluminium O O O
Al
Al Al Al Al O O Si O Si
OO OO
Al Al Al Al Si Si
Aluminium O O O O
Al Al O O O O CH CH
Al O
Al O
O O O O O
O CH N CH N
O O O O O O O
O N N
Polyolefin Polyolefin Polyolefin Polyolefin Polyolefin
FIGURE 1.18 Schematics of proposed metal–polyolefin interface design.

Recent Adv. in Adhesion Sci. & Technol. in Honor of Dr. Kash Mittal
All mechanical stress is transferred to near-interface polymer molecules. In addi-

tion, these near-interface molecules are often degraded by pretreatment or redox
reactions. Therefore, dissipation of this mechanical stress is necessary by inserting
long and flexible spacer molecules instead of direct chemical bonding (Figure 1.18b).
With increasing length of spacer molecules, the adhesion was increased, as demon-
strated later. Simultaneous attachment of spacers with two different chain lengths
may produce a two-step fracture under mechanical load, thus hindering abrupt fail-
ure (Figure 1.18c).
Introduction of additional water-repellent elements into the flexible spacer mol-
ecules hinders the hydrolysis of metal–polymer bonds, which is essential for long-
term durability of metal–polymer laminates (Figure 1.18d). These water-repellent
units should preferably consist of siloxane units equipped with methyl, isopropyl,
or tert-butyl groups. Fluorine-containing segments may undergo redox reactions
with metal atoms. Since redox reactions are associated with ion transportation
across the interface, an intermolecularly linked siloxane-containing barrier layer
was introduced to hinder such ion diffusion across the interface. This architecture
was achieved by coupling aminosilane onto aldehyde-modified polymer surfaces.
It is assumed that silanol groups formed after hydrolysis are responsible for strong
interactions to the metal (may be Al-O-Si bonds) and for barrier formation by partial
condensation (Si-O-Si formation) (Figure 1.18e).
Different types of spacer molecules were grafted to the polyolefin surface. Plasma-
introduced monosort functional groups, such as NH2 or Br groups, were used as
anchoring points for spacer molecules. Thus, different types of ethylene glycols,
diols and alcohols, diamines and amines with different end groups, chain lengths,
and branching degrees were introduced. The number of CH2− or CH2−CH2−O−
units in ethylene glycols was varied from 1 to about 100, those of diols and diamines
from 2 to 12 [57]. Different spacer end groups were used (NH2, OH, COOH, CHO,
SH, phenol, methoxy), and the spacer structure was varied from rigid to flexible or
branched (cf. Figure 1.19).
Additionally, siloxane elements within the spacer molecule enhance long-term
durability by their water repellence.
It has to be added that C-Br groups could also be converted wet-chemically
into other functional groups, such as OH, NH 2, CN, NO2, and so forth. Azide
groups (CN3) introduced by reaction of C-Br with NaN3 have served as a start-
ing point for click chemistry [48,58]. Moreover, atomic transfer radical polym-
erization (ATRP) was achieved at the polymer surface [57] (cf. Figure 1.19), as
was grafting of organic–inorganic polyhedral oligomer of silsesquioxane (POSS)
molecules onto C-Br groups (nucleophilic substitution) in noticeable quantities
(cf. Figure 1.19).
1.3.4 Introduction of Flexibilizing Spacers or

Adhesion-Promoting Layers
Spacer molecules at polyolefin surfaces increase the adhesion to metals, polymers,
and paints [57,59–61]. The concentration (graft density) of wet-chemically grafted
organic spacer molecules on brominated polypropylene or polyethylene surfaces
OH
22
OH OH OH Si O Si
NH2 NH2 OH
OH OH O O
NH2 NH2 OH OH Si O Si
OH O NH NH
OH O NH NH
O OH OH O O NH NH
NH NH
Aminosilane
OH OH OH OH OH OH
OH OH OH OH OH OH COOH CO OH
CH2 CH2 NH2
O O O O O O CH CH
HOOC NH HOOC NH
O O O O O O NH2
Si
O O O NH2
OH OH
H2N Si
OH O Si OSi
OH O O Si OSiO
i
SO O O
OH OH OH OH O O COOH CO OH NH2 NH2 H 2N Si NH2
O N N
OH OH OH OH O COOH COOH
NH2 N NH2 N H 2N POSS
O O O O O O O N N
O
O NH NH
O O O O O O O NH
Variation of spacer Variation of chain Variation of end groups Branched - dendrimer Grafting of inorganic coats
concentration length structure
ARGET-ATRP O O O
Click-chemistry O
CH2 N
O O O O O O
CF3
N H H ARGET-ATRP H
CH2 CH N Br H Br Br Br H Br Br
Br N3 n n n
Possible substitution after plasma “Grafting from” with glycidyl methacrylate

bromination
FIGURE 1.19 Architectures produced by grafting of spacer molecules onto plasma-brominated polypropylene surfaces (or allylamine plasma
polymer–coated surface).
OH OH OH
OH OH OH
O O O
O O O
22
20
Spacer concentration (spacer/100 C)
18
16
14
12
NH2
10
NH2
8 Graph
ene Po
lyo
Si
O O O NH2
6 lef H 2N Si
O OSi
O O Si OSiO
Si
in OOO
Si
NH2
4 H2N
Si
H2N POSS
2
NH
0 ESI deposit
1 2 5 10 20 polyolefin
Length of spacer (number CH2-units or C-atoms)
FIGURE 1.20 Maximum achievable spacer densities (concentrations) at polyolefin and

graphene surfaces by nucleophilic substitution or electrophilic addition.
shows a logarithmic dependence of spacer concentration/density on spacer length, as

depicted in Figure 1.20.
All spacer molecules were introduced by nucleophilic substitutions onto poly-
olefin surfaces or by electrophilic addition onto (fully substituted) aromatic double
bonds of graphene (no H).
It can be seen that the maximum achievable spacer density relates to each fifth car-
bon atom of a polyolefin surface (22 spacers per 100 C atoms) being substituted with a
short spacer, here, ethylenediamine. The largest molecules linked to the surface were
poly(ethylene glycol) 5000, octaaminophenylene-T8-POSS and (poly(amido amine))
(PAMAM) amino group–carrying dendrimers of the third, fourth, and fifth generation.
1.3.5 Adhesion of Spacer-Equipped Metal–Polyolefin Laminates

The peel strength of aluminum to spacer-equipped polypropylene surfaces was max
imal for COOH end groups. About two C10 spacers with COOH end groups per
100 C on the polypropylene surface produced high peel strength and cohesive fail-
ure, when aluminum layers were peeled off. Linear, branched, and dendrimer struc-
tures show the same results if equipped with COOH end groups (Figure 1.21). OH
groups containing spacers also promote adhesion. Maximal adhesion promotion was
found with OH-terminated long-chain spacers.
Moreover, the spacer molecules were changed from flexible aliphatic chains to
rigid chains, thus showing the advantage in measured peel strength for using flexible
400
COOH COOH
COOH
Peel strength increment (N/m)
300
(referenced to 1 group)
OH
200 NH2
H2N NH2
H2N NH2
COOH
NH
100
OH
NH2
FIGURE 1.21 Contribution of a single spacer molecule, PAMAM dendrimers, and octaami-
nophenylene-T8-POSS, with shown architecture and end groups to the measured peel strength.
molecules as spacers in Al–polyolefin laminates [57]. The concentration of grafted

POSS and PAMAM-dendrimer molecules as well as PEG oligomers (500–5000)
amounts to one to two molecules per 100 C or 0.4 to 0.8 molecules per square nano-
meter. Incorporation of SiOx structures by grafting POSS had enhanced the durabil-
ity of Al–PP laminates. POSS molecules covalently bonded to the polyolefin surface
are water repellent, in particular when the SiOx cage is substituted by isopropyl or
tert-butyl groups, as shown by enhanced durability toward moisture.
1.3.6 The “Ideal” Nonpeelable and Humidity-Resistant Al–PP Laminate

The most efficient adhesion-promoting variant was the modification of the polypro-
pylene surface with a thin layer (50 nm) of plasma-deposited poly(allylamine). Then,
glutaraldehyde was grafted onto the primary amino groups of the plasma polymer.
In the following step, aminosilane (3-aminopropyltriethoxysilane) was attached onto
the introduced aldehyde groups. The silane was hydrolyzed, partially linked by con-
densation; the residual formed silanol groups reacted with evaporated aluminum
atoms and formed probably covalent bonds [62]:
▮░-NH2 + OHC-R-OHC→▮░-N=CH-R-CHO + H2N-R-(SiOR)3→▮░-N=CH-

R-CH=N-R-Si(OH)3 →▮░-N=CH-R-CH=N-R-Si(OH)2-O-Al
where ▮ = polypropylene substrate and ░ = layer of plasma-deposited poly(allylamine).

Alternatively, the plasma bromination of polypropylene is simpler. After bromi-
nation, the C-Br groups can be directly reacted by aminosilane:
▮-Br + H2N-R-(SiOR)3 → ▮-NH-R-Si(OH)3 → ▮-NH-R-Si(OH)2-O-Al.

For adhesion promotion to aluminum, silanol groups may be responsible [62].

Remaining silanol groups can react intermolecularly with formation of Si-O-Si
bonds. The thus modified Al–PP laminates could not be peeled because of the strong
adhesion. Moreover, the adhesion was durable for 6 months on exposure to ambient
air with 90% humidity and 40 cent.
1.4 OUTLOOK
Metallized films are widely used for decorative purposes and food packaging as well
as for special applications such as insulation and electronics. Metal and polymer films
are often laminated or bonded by using an adhesive or joined by mechanical fasten-
ing or by welding [63]. Thin evaporated metal films need the development of new
techniques of polymer pretreatment, which is a key subject in the industry. Adhesive
bonding has proven to be a very effective method for joining dissimilar materials. The
primary function of adhesive-bonded joints is the transfer of load by shear [64]. The
most used industrial adhesives are the cyanoacrylates, silicon-based ones, and polyure-
thanes [65]. The main advantages of this joining method in comparison to mechani-
cal fastening and welding are the following [66,67]: bonding of dissimilar materials,
low stress concentration, improvement of fatigue resistance, sealing, weight reduction,
good surface finishing, assembly of thin or flexible substrates, no holes required, and
ease of repairs. Disadvantages are difficulty of disassembling, good surface prepara-
tion required, low engineering design confidence compared with mechanical fasten-
ers or welding, resistance only to shear loading, difficulty in predicting bond failure,
temperature sensitivity, high purchase and disposal costs, assembly rate limitations,
special handling requirements (hazardous chemicals and solvents), emission control,
and adhesive susceptibility to thermal and environmental degradation [68].
Metallization of polymer foils is performed using physical vapor deposition
(PVD). Aluminum is the most common metal used for deposition, but other met-
als such as nickel or chromium are also used. The metal is thermally evaporated
or sputtered under vacuum. The metal coating thickness varies from a few tens of
nanometers to about 500 nm. The coating will not fade or discolor over time. Typical
applications are production of capacitor foils; decorative films; barrier layers; or use
in packaging, automotive, and transportation [69].
In particular, stretched polymer foils were metallized such as polypropylene and
PET but also poly(amide) and polyethylene. In electronics, polyimide is often used as
a component of PCBs, which is the platform on which electronic components such as
integrated circuit chips and capacitors are mounted [69]. However, the poor adhesion
properties of polyimide must first be overcome by wet-chemical or plasma-chemical
processing [70,71].
Another way to metallize polymers is the process of electroless metallization. It
requires polymer surface modification; surface activation (in the case of electroless
processes, one needs to adsorb metallic nuclei, for example, palladium, that can initi-
ate metal deposition); and (electroless) metallization of polymers such as PET [3,5],
polycarbonate [72], or polyimide [70]. To metallize polyimide, say with copper,
often, a metal interlayer between polyimide and metal consisting of well-adherent
metals (Cr, Ti, Al) is deposited [73].
1.5 SUMMARY
The introduction of a new molecular architecture at the interface of metal–poly-
olefin laminates was successful. Adhesion strength and laminate durability were
significantly improved. Weak interactions, hydrolysis sensitivity, concentration of
mechanical stress at the interface, and formation of WBLs in the polymer zone con-
tacting the metal were widely avoided. The new concept is based on covalent metal–
polymer bonds, flexibilization by long-chain spacer molecules, and insertion of
water-repellent groups and barrier-forming structures. Such modified polymer sur-
faces produced highly adherent and durable metal–polymer laminates, which could
not be peeled off, or only cohesive failure occurred, indicating excellent adhesion. It
should be noted that such modified polyolefin surfaces have many other applications
such as in medicine, for nanostructures, friction reduction, or sensors, and so forth.
ACKNOWLEDGMENTS
I would like to thank for assistance and discussions Mrs. Gundula Hidde, Mrs. Dr.
Renate Mix, Mr. Dr. W. Unger, Dr. S. Wettmarshausen, Dr. G. Kühn, Dr. A. Meyer-
Plath, and many other coworkers.
REFERENCES
1. J. R. Hollahan, B. B. Stafford, R. D. Falb, and S. T. Payne, Attachment of amino groups
to polymer surfaces by radio-frequency plasmas, J. Appl. Polym. Sci., Polym. Symp., 13,
807–814 (1969).
2. J. Gähde, H. Wittrich, E. Schlosser, J. Friedrich, and G. Kaiser, Verfahren zur Herstellung
von verstärkten Polymerformstoffen, DD-patent 106 052 (1972).
3. J. Friedrich, I. Loeschcke, J. Gähde, H. Frommelt, and H. Wittrich, Verfahren zur
Herstellung von metallisierten Polyäthylenterephthalat-Formstoffen, DD-patent 115
708 (1974).
4. J. Friedrich, J. Gähde, H. Frommelt, and H. Wittrich, Modifizierung von
Feststoffoberflächen in einer HF-Entladung. III. Plasmachemisches Aufbringen
funktioneller Gruppen und selektiver Abbau teilkristalliner Polymere, Faserforsch.
Textiltechn./Z. Polymerenforsch., 27, 604–608 (1976).
5. J. Friedrich, J. Gähde, H. Frommelt, and H. Wittrich, Modifizierung von Feststoffoberflächen
in einer HF-Entladung. II. Schichtuntersuchungen, Reaktionsmechanismus und
Verbundergebnisse, Faserforsch. Textiltechn./Z. Polymerenforsch., 27, 599–603 (1976).
6. B. V. Derjaguin, V. M. Müller, and Y. Toporov, Effect of contact deformations on the
adhesion of particles, J. Colloid Interface Sci., 53, 314–320 (1975).
7. T. Strunskus, M. Kiene, R. Willecke, A. Thran, C. V. Bechtolsheim, and F. Faupel,
Chemistry, diffusion and cluster formation at metal-polymer interfaces, Mater. Corros.
49(3), 180–188 (1998).
8. J. J. Bikermann, The Science of Adhesive Joints. Academic Press, New York (1968).
9. S. Wu, Polymer Interface and Adhesion. Marcel Dekker, New York (1982).
10. S. S. Voyutski, Autohesion and Adhesion of High Polymers. Interscience Publ., New
York (1963).
11. C. Bischof, and W. Possart (Eds.), Adhäsion. Akademie-Verlag, Berlin (1982).
12. F. Garbassi, M. Morra, and E. Occhiello, Polymer Surfaces. John Wiley Sons, Chichester,
U.K. (1998).
13. J. Gähde, Rolle chemischer Bindungen, in: Adhäsion. C. Bischof, and W. Possart (Eds.),
p. 173, Akademie-Verlag, Berlin (1982).
14. B. M. de Koven, and P. L. Hagans, XPS studies of metal/polymer interfaces—Thin films
of Al on polyacrylic acid and polyethylene, Appl. Surf. Sci., 27, 199–212 (1986).
15. J. F. Friedrich, R. Mix, R.-D. Schulze, A. Meyer-Plath, R. Joshi, and S. Wettmarshausen,
New plasma techniques for polymer surface modification with monotype functional
group, Plasma Process. Polym., 5, 407–423 (2008).
16. J. Friedrich, R. Mix, R.-D. Schulze, and A. Rau, Ultra-thin polymer layer deposition
by aerosol-barrier-discharge and electrospray at atmospheric pressure, J. Adhesion Sci.
Technol., 24, 1329–1350 (2010).
17. J. Gähde, J. Friedrich, I. Loeschcke, R. Gehrke, and J. Sachse, Adhesion of polyurethane
to surface-modified steel, J. Adhesion Sci. Technol., 6, 569–580 (1992).
18. M. Strobel, C. Dunatov, J. M. Strobel, C. S. Lyons, S. J. Perron, and M. C. Morgen, Low-
molecular-weight materials on corona-treated polypropylene, J. Adhes. Sci. Technol.
3(1), 321–335 (1989).
19. J. F. Friedrich, W. E. S. Unger, A. Lippitz, R. Giebler, I. Koprinarov, S. Weidner, and G.
Kühn, Significance of interfacial redox reactions and formation of metal-organic com-
plexes for the adhesion of metals on pristine and plasma-treated polymers, in: Polymer
Surface Modification: Relevance to Adhesion, Vol. 2. K. L. Mittal (Ed.), pp. 137–172,
VSP, Utrecht (2000).
20. J. Friedrich, W. Unger, A. Lippitz, T. Gross, P. Rohrer, W. Saur, J. Erdmann, and H.-V.
Gorsler, The improvement in adhesion of polyurethane-polypropylene composites
by short-time exposure of polypropylene to low and atmospheric pressure plasmas,
J. Adhesion Sci. Technol., 9, 575–598 (1995).
21. T. Yasuda, T. Okuno, K. Yoshida, and H. Yasuda, A study of surface dynamics of poly-
mers. II. Investigation by plasma surface implantation of fluorine–containing moieties,
J. Polym. Sci., B: Polym. Phys., 26, 1781–1790 (1988).
22. J. Behnisch, A. Holländer, and H. Zimmermann, Factors influencing the hydrophobic
recovery of oxygen-plasma-treated polyethylene, Surf. Coat. Technol., 59, 356–361
(1993).
23. F. Truica-Marasescu, S. Guimond, P. Jedrzejowski, and M. R. Wertheimer, Hydrophobic
recovery of VUV/NH3 modified polyolefin surfaces: Comparison with plasma treat-
ments in nitrogen, Nucl. Instr. Meth. Phys. Res. B, 236, 117–122 (2005).
24. A. Chilkoti, and B. D. Ratner, Chemical derivatization methods for enhancing the ana-
lytical capabilities of X-ray photoelectron spectroscopy and static secondary ion mass
spectrometry, in: Surface Characterization of Advanced Polymers. L. Sabbatini, and
P. G. Zambonin (Eds.), p. 221, VCH Verlag, Weinheim (1993).
25. D. S. Everhart, and C. H. Reilley, Polymer functional group mobility. 2—Partition of
ion pairs between hydrophobic and hydrophilic phases of plasma oxidized polyethylene,
Surf. Interface Anal., 3, 258–270 (1981).
26. J. Friedrich, B. Falk, I. Loeschcke, B. Rutsch, K. Richter, H.-D. Reiner, U. Throl, and
H. Raubach, Untersuchungen zur Haftung zwischen plasmachemisch oxidiertem Kupfer
und Polyphenylchinoxalin, Acta Polymerica, 36, 310–320 (1985).
27. J. Friedrich, I. Loeschcke, and J. Gähde, Zur Adhäsion von Aluminium auf Polypropylen,
Acta Polymerica, 37, 687–695 (1986).
28. J. F. Friedrich, I. Koprinarov, R. Giebler, A. Lippitz, and W. E. S. Unger, Reactions and
intermediates at the metal-polymer interface—Detected by XPS and NEXAFS spectro-
scopy, J. Adhesion, 71, 297–310 (1999).
29. M. G. Chan, and D. L. Allara, Infrared reflection studies of the mechanism of oxidation
at a copper—Polyethylene interface, J. Colloid Interface Sci., 47, 697–704 (1974).
30. J. R. G. Evans, and D. E. Packham, Adhesion of polyethylene to copper: Reactions bet-
ween copper oxides and the polymer, J. Adhesion, 9, 267–277 (1978).
31. J. R. G. Evans, and D. E. Packham, Adhesion of polyethylene to copper: Importance of

substrate topography, J. Adhesion, 10, 39–47 (1979).
32. M. C. Burrell, P. J. Codella, J. A. Fontana, J. J. Chera, and M. D. McConnell, Interfacial
reactions at copper surfaces coated with polyimide films prepared from poly(amide–
acid) precursors, J. Vac. Sci. Technol., A7, 55–58 (1989).
33. I. Koprinarov, A. Lippitz, J. F. Friedrich, W. E. S. Unger, and C. Wöll, Oxygen plasma
induced degradation of the surface of poly(styrene), poly(bisphenol-A-carbonate)
and poly(ethylene terephthalate) as observed by soft X-ray absorption spectroscopy
(NEXAFS), Polymer, 39, 3001–3009 (1998).
34. N. van Order, W. E. Geiger, T. E. Bitterwolf, and A. L. Rheingolds, Mixed-valent cations
of dinuclear chromium arene complexes: Electrochemical, spectroscopic, and structural
considerations, J. Am. Chem. Soc. 109, 5680–5690 (1987).
35. J. Friedrich, W. Unger, A. Lippitz, I. Koprinarov, G. Kühn, S. Weidner, and L. Vogel,
Chemical reactions at polymer surfaces interacting with a gas plasma or with Cr
atoms—Their relevance to adhesion, Surf. Coat. Technol., 116–119, 772–782 (1999).
36. R. H. Hansen, and H. Schonhorn, A new technique for preparing low surface energy
polymers for adhesive bonding, J. Polymer Sci., B4, 203–209 (1966).
37. T. K. Kwei, H. Schonhorn, and K. L. Frisch, Dynamic mechanical properties of the
transcrystalline regions in two polyolefins, J. Appl. Phys., 38, 2512 (1967).
38. W. Possart, and J. Friedrich, Plaste und Kautschuk, 33, 273–279 (1986).
39. J. Friedrich, S. Geng, W. Unger, and A. Lippitz, Modelling the plasma-induced reactions
on polymer surfaces by using aliphatic self-assembling and LB layers as substrates,
Surf. Coatings Technol., 98, 1132–1141 (1998).
40. W. E. S. Unger, A. Lippitz, J. F. Friedrich, C. Wöll, and L. Nick, The use of octadecyl-
trichlorosilane self-assembled layers as a model for the assessment of plasma treatment
and metallization effects on polyolefines, Langmuir, 15, 1161–1166 (1999).
41. W. E. S. Unger, J. F. Friedrich, A. Lippitz, I. Koprinarov, K. Weiss, and C. Wöll, The
application of X-ray absorption spectroscopy (NEXAFS) and X-ray photoelectron
spectroscopy (ESCA) in surface analysis of chromium-evaporated polymers and orga-
nic films, in: Metallized Plastics 5&6: Fundamental and Applied Aspects. K. L. Mittal
(Ed.), pp. 147–168, VSP, Utrecht (1998).
42. G. Kühn, S. Weidner, R. Decker, A. Ghode, and J. Friedrich, Selective surface functiona-
lization of polyolefins by plasma treatment followed by chemical reduction, Surf. Coat.
Technol., 116–119, 796–801 (1999).
43. G. Kühn, A. Ghode, S. Weidner, I. Retzko, W. E. S. Unger, and J. F. Friedrich, Chemically
well-defined polymer layers by grafting or pulsed plasma polymerization, in: Polymer
Surface Modification: Relevance to Adhesion, Vol. 2. K. L. Mittal (Ed.), pp. 45–64, VSP,
Utrecht (2000).
44. H. C. Brown, H. I. Schlesinger, and A. B. Burg, Hydrides of boron. XI. The reaction of
diborane with organic compounds containing a carbonyl group, J. Amer. Chem. Soc., 61,
673–680 (1939).
45. H. C. Brown, and B. C. Subba Rao, Hydroboration. III. The reduction of organic com-
pounds by diborane, an acid-type reducing agent, J. Amer. Chem. Soc., 82, 681–686
(1960)
46. R. G. Nuzzo, and G. Smolinsky, Preparation and characterization of functionalized
polyethylene surfaces, Macromolecules, 17, 1013–1019 (1985).
47. J. Friedrich, S. Wettmarshausen, and M. Hennecke, Haloform plasmas applied to poly-
mers, Surf. Coat. Technol., 203, 3647–3655 (2009).
48. S. Wettmarshausen, R. Mix, A. Meyer-Plath, H.-U. Mittmann, and J. Friedrich,
Plasmabromination—The selective way to monotype functionalized polymer surfaces,
in: Polymer Surface Modification, Vol. 5. K. L. Mittal (Ed.), pp. 3–18, Brill, Leiden
(2009).
49. J. F. Friedrich, The Plasma Chemistry of Polymer Surfaces: Advanced Techniques for
Surface Design. p. 272, Wiley-VCH, Weinheim (2012).
50. J. Friedrich, R. Mix, R.-D. Schulze, and G. Kühn, Contribution of chemical interactions
to the adhesion between metals and functional groups of different (mono-) type and
density at polymer surfaces, in: Adhesion, W. Possart (Ed.), pp. 265–288, Wiley-VCH,
Weinheim (2005).
51. J. F. Friedrich, R. Mix, and G. Kühn, Adhesion of metals to plasma-induced functional
groups at polymer surfaces, Surf. Coat. Technol., 200, 565–568 (2005).
52. R. Mix, G. Kühn, and J. Friedrich, Contribution of chemical interactions between Al
atoms and different types of functional groups to the adhesion of Al-polymer systems,
in: Adhesion Aspects of Thin Films, Vol. 2. pp. 123–144, K. L. Mittal (Ed.), VSP, Utrecht
(2005).
53. R. Mix, J. F. Friedrich, and G. Kühn, Adhesion between metal atoms and polymer sur-
faces fitted with different types and variable density of functional groups, in: Plasma
Polymers and Related Materials. M. Mutlu, G. Dinescu, R. Förch, J. M. Martin-Martinez,
and J. Vyskocil (Eds.), pp. 107–114, Hacettepe University Press, Ankara (2005).
54. T. P. Nguyen, H. Ettaik, S. Lefrant, G. Leising, and F. Stelzer, Studies of the polyparaphe-
nylene/aluminium interface, Synth. Met., 38, 69–76 (1990).
55. P. S. Ho, P. O. Hahn, J. W. Bartha, G. W. Rubloff, F. K. LeGoues, and B. D. Silvermann,
Chemical bonding and reaction at metal/polymer interfaces, J. Vac. Sci. Technol., A3,
739–745 (1985).
56. M. R. Alexander, G. Beamson, C. J. Blomfield, G. Leggett, and T. M. Duc, Interaction of
carboxylic acids with the oxyhydroxide surface of aluminium: Poly(acrylic acid), acetic
acid and propionic acid on pseudoboehmite, J. Electr. Spectr. Rel. Phenom., 121, 19–32
(2001).
57. J. Friedrich, R. Mix, and S. Wettmarshausen, A new concept of adhesion promotion in
metal-polymer composites by introduction of covalently bonded spacers at interface, J.
Adhesion Sci. Technol., 22, 1123–1143 (2008).
58. R. T. Chen, B. W. Muir, G. K. Such, A. Postma, R. A. Evans, S. M. Pereira, K. M.
McLean, and F. Caruso, Surface “click” chemistry on brominated plasma polymer thin
films, Langmuir, 26, 3388–3393 (2010).
59. V. Gutowski, A. Bilyk, S. Li, M. Espiritu, and I. Burgar, The influence of structure of the
interface and interphase on paint adhesion, Compos. Interface, 12, 817–836 (2005).
60. V. Gutowski, S. Li, L. Russell, P. Lee, H. Con, P. Hoobin, and S. Petinakis, Effect of
surface-grafted molecular brushes on the adhesion performance of bonded polymers and
composite interfaces, Compos. Interface, 9, 89–133 (2002).
61. J. Friedrich, G. Kühn, and R. Mix, Tailoring of polymer surfaces with different mono-
type functional groups of variable density using chemical and plasma chemical pro-
cesses, Progr. Coll. Polym. Sci., 132, 62–71 (2006).
62. Y. Huajie, R. Mix, and J. Friedrich, Influence of differently structured aluminium-
polypropylene interfaces on adhesion, J. Adhesion Sci. Technol., 25, 799–818 (2011).
63. S. T. Amancio-Filho, and J. F. Dos Santos, Joining of polymers and polymer–metal hybrid
structures: Recent developments and trends, Polym. Eng. Sci., 49, 1461–1476 (2009).
64. W. S. Johnson (Ed.), Adhesively Bonded Joints: Testing, Analysis, and Design. American
Society for Testing and Materials (ASTM STP 981) (1986).
65. W. E. Reitz, and R. M. Oman, How to join plastics, Adv. Mater. Process., 158, 49 (2000).
66. J. Rotheiser, Joining of Plastics—Handbook for Designers and Engineers. Carl Hanser,
Munich (1999).
67. A. Baldan, Adhesively-bonded joints in metallic alloys, polymers and composite mate-
rials: Mechanical and environmental durability performance, J. Mater. Sci., 39, 4729–
4797 (2004).
68. E. Mount, Metallized films for food packaging, Converting Magazine, 3, 22–27 (2004).
69. S.-J. Cho, S. P. Shresta, and J.-H. Boo, Surface treatment for Cu metallization on poly-
imide film by atmospheric pressure dielectric barrier discharge plasma system, Curr.
Appl. Phys., 11, S135–S139 (2011).
70. R. Seeböck, H. Esrom, M. Charbonnier, and M. Romand, Modification of Polyimide in
Barrier Discharge Air-Plasmas: Chemical and Morphological Effects, Plasmas Polym.,
5, 103–118 (2000).
71. S.-J. Park, and H.-Y. Lee, Effect of atmospheric-pressure plasma on adhesion characte-
ristics of polyimide film, J. Coll. Interf. Sci., 285, 267–272 (2005).
72. K. Zarnoch, Surface functionalization and metallization of Lexan® polycarbonate, J.
Adhesion Sci. Technol., 8, 501–509 (1994).
73. L. J. Matienzo, and W. N. Unertl, Adhesion of metal films to polyimides, pp. 629–696,
in: Polyimides, Fundamentals and Applications. M. K. Ghosh, and K. L. Mittal (Eds.),
Marcel Dekker, New York (1996).
2 Lateral Adhesion Force
Measurements at the
Drop–Surface Interface
Aisha Leh, Hartmann E. N'guessan,
Prashant Bahadur, Rajiv Jaini, and Rafael Tadmor*
CONTENTS
2.1 Introduction..................................................................................................... 32
2.2 Tribology and Drop–Surface Contact............................................................. 32
2.3 Drops on Surfaces............................................................................................ 32
2.4 Conclusion....................................................................................................... 38
Acknowledgments..................................................................................................... 39
References................................................................................................................. 39
We review the forces at the interface between a liquid drop and a solid sur-
face in light of the newly invented Centrifugal Adhesion Balance (CAB).
Independent measurements of the normal forces, and the lateral adhesion
forces, at a solid-liquid interface made with the CAB allow the decoupling of
lateral and normal forces, and also the measurement of forces of any practical
interest. The CAB is used to measure the impact of two different parameters:
the normal force, and the time the drop rested on the surface prior to sliding
on the lateral force between a drop and a surface. For different normal forces,
the lateral force was found to increase with the time the drop rested on the
surface prior to sliding but plateaus as the resting time approaches infinity.
The experimental results show a complex relation between the normal force
and the lateral force in which the lateral force reaches a minimum around
zero normal force. This result, though seemingly contradictory to tribological
understanding, is in agreement with recent time effect theories.
* Prof. Rafael Tadmor was the lead author for this chapter and can be contacted at rafael.tadmor@
lamar.edu.
31
2.1 INTRODUCTION
In 1974, Dr. Mittal showed that metallic surfaces have inherent organic contami-
nants on them that cannot be cleaned using simple solvents but can be temporarily
cleaned using plasma cleaning [1]. Such surface contamination results in a surface
stiffness modulus that is markedly different than the modulus that is actually being
measured using simple mechanical tests. In this article, we show how surface stiff-
ness, in general, differs from the stiffness of the bulk regions of the same material,
and how this relates to wetting phenomena and to the retention forces of drops on
surfaces, which results in the current tribological understanding as described below.
2.2 TRIBOLOGY AND DROP–SURFACE CONTACT

While Amonton’s law [2] states that the contact area of a rough surface increases
linearly with the load, according to tribological studies, the lateral force required to
slide two surfaces against each other (“friction” force) is proportional to the contact
area [3]. For drop–surface systems, however, the lateral force decreases with the
normal force regardless of the fact that the contact area increases [4].
2.3 DROPS ON SURFACES

Drop–surface theories [5] are widely applicable in various fields ranging from inha-
lation drugs to deformable particles, self-propulsion, and ratcheting [6]. A particular
interesting case of drop–surface studies is the lateral force or retention force, f∥,
associated with the motion of drops on surfaces [7–11]. Different time effect stud-
ies corresponding to different systems using the tilt stage method are presented in
Figure 2.1 [10]. These studies showed a link between the time the drop rests on the
surface prior to sliding (trest) and f∥. The drops used in the experiments in Figure 2.1
started in a horizontal position, but as the tilt was gradually increased, they reached
different combinations of lateral and normal forces. Yet, from Figure 2.1, we see that
f∥ increases with trest and plateaus as trest increases. The unsatisfied normal compo-
nent of the drop–surface tension [12] was theoretically related to this time-dependent
phenomenon of f∥ and the retention force increase at the three-phase contact line [13].
To verify such theories, the conventional tilt stage method [5,10,11,14,15] could not
be used due to the coupling of normal and lateral forces, an intrinsic feature of this
method. Moreover, the tilt stage method is also limited to force measurements between
0 and the weight of the drop. The use of centrifugal forces to induce the motion of
drops on surfaces [16] is a solution to these problems [16,17]. However, although the
solution was known in the past, the visual observation and recording of the drops in
such devices while the system was rotating was compromised. The centrifugal adhe-
sion balance (CAB) (patent pending) [4] is a device that overcomes this obstacle by the
use of a camera that rotates together with the drop, while sending video signals in real
time to a stationary computer nearby with the use of wireless technology.
A more detailed description of the CAB was reported earlier [4]. Here, a short
description is presented. The CAB consists of a centrifugal arm, which can rotate
perpendicular to the gravitational field using a direct current (DC) motor. On one
Lateral Adhesion Force Measurements at the Drop–Surface Interface 33
35 35
(a) (b)
f||/V1/3 (mN/m)
25 25
15 Water/Teflon 15 Water/
OTA-covered glass
5 5
0 25 50 75 100 0 5 10 15
2.5 (c) (d)

4
f||/V1/3 (mN/m)
1.5 Hexadecane/ 3 Hexadecane/

OTA-covered glass OTA-covered mica
1
2
0 10 20 30 40 50 60 0 5 10 15 20
50
(e) 3 (f )
45
f||/V1/3 (mN/m)
2.5
40
35 Water/ 2
OTA-covered mica Tetradecane/
30 1.5 OTA-covered mica
25 1
0 20 40 60 0 5 10 15 20
trest (min) trest (min)
FIGURE 2.1 The lateral force, f∥, required to slide various liquid drops on different surfaces
as a function of the time that the drop was resting on the surface prior to sliding, trest. The f∥ is
normalized by a unit length, V 1/3 (V is the drop volume). These experiments were performed
using the conventional tilt stage method.
end of the arm is a sealed chamber containing a plate fixed with respect to the arm
during the rotation. This plate can be set at any angle, α, between 0° and 360°,
around an axis perpendicular to the centrifugal rotation. A charge-couple device
(CCD) camera, together with the substrate surface on which the drop is placed, is
held stable on the plate. The ability to set α at any angle provides, according to the
following equations, an independent manipulation of normal and lateral forces:
f∥ = m(ω2 L cos α − g sin α) (2.1)
f⊥ = m(ω2 L sin α + g cos α) (2.2)
where f⊥ and f∥ are the normal and lateral forces, respectively; L is the distance from
the center of rotation (of the arm) to the drop; m is the mass of the drop; and ω is the
angular velocity.
Figure 2.2a shows the applied angular velocities, ω, of the CAB as a function of
experiment time for hexadecane drops on a Teflon-coated silicon surface. The plate
tilt values are α = 0° (sessile drop) and α = 180° (pendant drop). The corresponding
pictures of the drop on the surface inside the CAB are shown in Figure 2.2b. The
experiment time consists of trest, the resting time of the drop. trest is defined as the
time from which the drop is placed on the surface until it begins to slide. It is the sum
of tstill, a period of time during which the CAB remains still (ω is 0), and tactive, during
which ω is gradually increased until, at some critical angular velocity, ωc, the drop
(a)
3.5 (iv)
ωc (pendant)
(iii)
3
(iv)
Angular velocity, ω (Rad/s)
ωc (sessile)
2.5 (iii)
2 trest (pendant)
trest (sessile)
1.5 tactive
(pendant)
1 tactive
(sessile)
0.5 tstill
(i) (ii)
0
0 2 4 6 8 10 12 14
Time (min)
(b) (i) – (ii) (iii) (iv)
Sessile
Pendant
FIGURE 2.2 (See color insert.) A single-force datum measurement using a hexadecane
drop on a Teflon-coated silicon surface. The Teflon used was EGC 1700 from 3M. See also the
work of Priest et al. [9]. (a) The applied angular velocity, ω, measured during the experiment as
a function of the measurement time. For a prescribed period, tstill, the drop is left to rest in the
stationary CAB, after which ω is steadily increased until, at a critical angular velocity, ωc, the
drop begins to slide on the surface. The drop is pinned to the surface during tstill and when ω
is increased until just before ωc is reached; this entire time is termed trest. tactive is the time from
when the CAB is started (i.e., at the end of tstill) to when the drop begins to move. During t active,
the applied angular velocity of CAB is gradually increased until the drop begins to move. We
consider pendant and sessile drops. (b) Pictures of the drop taken at different stages of the
experiment. From (i) to (ii), no lateral force is applied, and the drop is symmetric and pinned
to the surface; during the active stage (iii), it is deformed but is still pinned to the surface.
Once ωc is reached, the drop slides, and hence, in (iv), we see only part of it in the frame. We
see that the lateral force required to slide the drop is higher when the normal force is lower.
moves. Throughout this procedure, the drop is resting pinned to the surface, and its
onset of movement signifies the end of the active time. The kinetic stage following
the resting time, where the drop moves, is a subject of other studies [14].
Each point in Figure 2.3 (hexadecane on octadecyl trimethylammonium [OTA]–
treated mica) was obtained following the procedure illustrated in Figure 2.2 (hexa-
decane on Teflon-coated silicon) [9,10]. Figure 2.3 shows the effect of two f⊥ (sessile
“+1g” and pendant “–1g”) cases on f∥ for various resting times. For each f⊥, only the
resting time is varied throughout the experiment. f∥ increases with trest for both ses-
sile and pendant drops. However, surprisingly, f∥ values for pendant drops are higher
than those for sessile ones for the same drop volume, although f⊥ values for pendant
drops are lower than those for their sessile counterparts. This phenomenon is further
emphasized as trest → ∞, where f∥ values are significantly higher for pendant drops
compared to the same-sized sessile drops. Intuitively, this is contrary to what one
might expect, if, for example, the intuition stems from Amonton’s law. (According to
Amonton’s law, the shear forces are expected to be proportional to the normal forces,
while in tribology, one expects that as the load increases, so does the contact area [as
indeed happens in our system], and hence, the shear forces are expected to grow with
contact area [2].) Similarly, the results are contrary to what will be expected from
tribological studies, or even from drop–surface theories [7] where the main pinning
factor of the drop on the surface would be expected to be the three-phase contact line:
for sessile drops, the three phase contact line is longer than it is for pendant drops.
The unsatisfied normal component of the surface tension has been shown theo-
retically by Shanahan and de Gennes [18] to cause surface deformation, which is
proportional to
γ sin θ
(2.3)
r
f||∞
6
f||∞
f||c (µN)
0 3 6 9 12 15
Resting time, trest, (min)
FIGURE 2.3 The drop retention force, f∥c, required for the onset of lateral motion of 3.3 μL
sessile (▴) and pendant (∇) hexadecane drops on an OTA-treated mica surface as a function
of the time, trest, that the drop rested on the surface prior to sliding. We see that the lateral
force required to slide the drop is higher when the normal force is lower (pendant case). f∥∞ is
the value of f∥c in the region of long resting times (when f∥c reaches a plateau).
γ is the surface tension of the liquid, r is the drop radius, and θ is its contact angle with
the surface. Afterward, an experimental demonstration by Carré et al. [19] proved
that the deformation is truly a protrusion from the surface. This deformation is also
associated with a molecular reorientation of the solid surface [20], which strengthens
the liquid–surface interaction (minimizes the free energy associated with the liquid–
surface interaction) [10,11]. Following the Shanahan–de Gennes approach, current
theories relate this stronger liquid–solid interaction to the drop retention on surfaces
[13] by the following equation:
4 γ 2 sin θ
fc = (cos θ R − cos θ A ) (2.4)
G
where G is the “surface protrusion modulus”: a modulus associated with the defor-
mation (primarily protrusion deformation [19]) of the outer surface. In very high-
energy surfaces like metals, we learn from Mittal [1] that this outer surface is a layer
of hydrocarbon contaminants covering the surface. Most rigid surfaces are covered
either with contaminants (unintentional) or methodically by surfactants (intentional).
For rigid surfaces, since the bulk does not deform, only the outer layer of the substrate
molecules will reorient. In any case, at least part (if not all) of the deformation needs
to be molecular or submolecular (only part of the molecule deforms). Additionally,
this modulus is often time dependent; that is, it generally has a plastic component. The
molecular reorientation of the solid surface increases the solid–liquid intermolecular
interaction, that is, decreases the free energy associated with the drop–surface molec-
ular contact. This happens mainly at the three-phase contact line and much less at the
inner drop–surface area. γ sin θ is related to the normal component of the surface ten-
sion, which strengthens the molecular reorientation with time [21]. Since θ is bigger
for pendant drops than for sessile drops, the normal component of the surface tension
has a higher value (Figure 2.4); hence, the molecular reorientation is more significant,
and the resulting intermolecular force is higher. Also, in the case of pendant drops, the
γLV cos θS
γLV sin θS γLV
θS γSV γSL
γSV γSL θP
mg γLV
γLV sin θP
mg
γLV cos θP
θ S < θP
FIGURE 2.4 Comparison of forces acting on sessile and pendant drops. For the pendant
drop (right), θ is bigger than for the sessile drop (left), and therefore, γ sin θP > γ sin θS.
Additionally, γ sin θP acts in the same direction as gravitational force, resulting in increased
solid molecular reorientations and consequential solid–liquid interactions.
gravitational force acts in the same direction as γ sin θ, consequently further increas-
ing the solid molecular reorientation and resultant solid–liquid interactions. In the
case of sessile drops, however, the two forces act in opposite directions, resulting in
weaker deformation; consequently, weaker liquid–solid intermolecular reorientation;
and hence, weaker interactions and smaller retention.
There appears to be an experimental correlation with Equation 2.4, suggesting
that the additional influence of gravity in itself is negligible here. Substituting con-
tact angle values for sessile and pendant in Equation 2.4, and then dividing, we get
the following ratio:
sin θ P (cos θ R ,P − cos θ A,P )

= 1.26, (2.5)
sin θ S (cos θ R ,S − cos θ A,S )
which also is in agreement with the ratio of experimental forces:
f∞ , pendant
= 1.27 (2.6)
f∞ , sessile
f∥∞, pendant corresponds to pendant f∥ values in the plateau region, and f∥∞, sessile cor-
responds to sessile f∥ values in the plateau region. The realization that for drops on
surfaces, the case of smaller normal force, f⊥, and smaller contact area results in
higher lateral force required to slide the drop, thus f∥, is strengthened in agreement
with theory. This, to the best of our knowledge, is the first experimental evidence of a
smaller normal force resulting in a higher lateral force. The normal forces mentioned
here are those for the pendant configuration compared to that of the corresponding
sessile configuration.
Experiments reported previously using the tilt stage method [10,11] indeed showed
the dependence of f∥ on the drop resting time, trest. However, inherent to this method,
an important factor, f⊥ (i.e., the normal force experienced by the drop), changes,
while f∥ is varied at the same time. Ultimately, with the tilt stage, one cannot predict
the effect of f⊥ on f∥ since these forces cannot be manipulated independently. Using
the CAB, f⊥ can be set to a value of any practical interest, while only the lateral force
is varied. In this configuration, we can determine the effect of f⊥ on f∥.
For any given combination of drop size and normal force, we still obtain plots that
are of similar trend to those shown in Figure 2.3, though with different f∥ values. The
variation of the plateau value, f∥ (see Figure 2.3), with the normal force, f⊥, for 0.5 μL
(cf. Equations 2.1 and 2.2) is shown in Figure 2.5. For the same f⊥ , a greater f∥∞ is
measured for the negative normal forces, that is,
( )
f∞ − f⊥ > f∞ f⊥ ( ) (2.7)
However, the more striking feature in this graph is that f∥∞ decreases as the absolute
value of the normal force, f⊥ , decreases, whether f⊥ is positive or negative. Thus, f∥
increases with f⊥ for positive f⊥ but decreases with f⊥ for negative f⊥, and it reaches
a minimum around f⊥ = 0. For this drop size, very close to f⊥ = 0, the drop was
15
f||∞/V1/3 (mN/m)
12
–4 –2 0 2 4
f⊥ (µN)
FIGURE 2.5 The drop retention force after reaching a time plateau, f∥∞ (cf. Figure 2.3),
required for the onset of lateral motion of a 0.5 μL hexadecane drop on an OTA-treated mica
surface, versus the normal force, f⊥, that the drop experiences. The retention force is normal-
ized by unit length. (V is the drop volume.)
unstable, and f∥∞ could not be reached; however, the trend shown in the dashed line
in Figure 2.5 was verified with smaller drop sizes.
The Shanahan–de Gennes approach used above for the truly pendant and truly
sessile configurations can explain the differences between f∥∞ pairs with similar f⊥
in Figure 2.5 but not the decline in f∥∞ as f⊥ → 0. Nonetheless, the explanation should
still relate to the deformation of the surface at the three-phase contact line, which
is apparently enhanced both by pulling normal force and by pushing normal force.
2.4 CONCLUSION
In conclusion, the CAB is used to decouple drop–surface lateral retention forces from
their normal body forces. The strong influence of a drop’s normal body force on the
lateral retention force is theoretically related to the normal component of the surface
tension, which pulls on the solid surface at the three-phase contact line [18]. This phe-
nomenon is time dependent: As the resting time of the drop on the surface increases,
the force required to move the drop increases as well. This time dependence is related
to the plastic change in the surface modulus: For the same force, and therefore the same
stress, the strain is increased, resulting in an effective smaller modulus. The smaller
modulus is associated with higher deformation of the surface molecules, which there-
fore can increase their interactions with the liquid molecules and, hence, the liquid–
solid retention force. Note that in hard surfaces, this pulling results in an insignificant
protrusion (i.e., topographically insignificant) [19], but it is associated with a time-
dependent [4,10,21] molecular reorientation of the solid surface, which results in higher
surface–liquid intermolecular interaction [13] and, hence, in higher retention force.
When the normal force (e.g., gravitational force for a horizontal surface) is in
the same direction as the normal component of the surface tension (pulling on the
surface), then the total pulling force is higher. With a higher pulling force, the solid
surface’s molecular reorientation is further facilitated, and the resultant solid–liquid
intermolecular interaction and associated lateral retention force are stronger for
pendant drops. On the other hand, for sessile drops, the normal component of the
surface tension and normal gravitational forces are in opposite directions, resulting
in lower solid–liquid intermolecular interactions and a lower associated lateral reten-
tion force. If, however, instead of just sessile and pendant drops, we induce gradual
continuous variation of the normal force, we see that the change in the lateral reten-
tion force with the normal force is not monotonous but reaches a minimum close
to zero normal force (as shown in Figure 2.5). This suggests that the solid surface
molecular reorientation is facilitated both by pulling and by pushing body forces.
ACKNOWLEDGMENTS
This study was supported by National Science Foundation grants DMR-0619458 and
CBET-0960229.
REFERENCES
1. D. F. O’Kane and K. L. Mittal, J. Vac. Sci. Tech., 11, 567 (1974).
2. D. Tabor, Proc. I. Mech. Eng. C—J. Mech. Eng. Sci., 205, 365 (1991).
3. D. Tabor, Tribol. Int., 28, 7 (1995).
4. R. Tadmor, P. Bahadur, A. Leh, H. E. N’Guessan, R. Jaini and L. Dang, Phys. Rev. Lett.,
103, 266101 (2009).
5. C. Antonini, F. J. Carmona, E. Pierce, M. Marengo and A. Amirfazli, Langmuir, 25, 6143
(2009); P. Letellier, A. Mayaffre and M. Turmine, J. Coll. Int. Sci., 314, 604 (2007); E.
Bormashenko, J. Phys. Chem. C, 113, 17275 (2009); E. Chibowski and K. Terpilowski,
J. Coll. Int. Sci., 319, 505 (2008); U. Thiele and E. Knobloch, Phys. Rev. Lett., 97,
204501 (2006); D. Quéré, Rep. Prog. Phys., 68, 2495 (2005); P.-G. de Gennes, D. Quéré
and F. Brochard-Wyart, Capillarity and Wetting Phenomena: Drops, Bubbles, Pearls,
Waves. Springer Verlag, New York (2003); P. Brunet, J. Eggers and R. D. Deegan, Phys.
Rev. Lett., 99, 144501 (2007).
6. J. S. Patton and P. R. Byron, Nat. Rev. Drug Discovery, 6, 67 (2007); T. Ohta and
T. Ohkuma, Phys. Rev. Lett., 102, 154101 (2009); T.-I. Kim and K. Y. Suh, Soft Matter,
5, 4131 (2009); J. B. Boreyko and C.-H. Chen, Phys. Rev. Lett., 103, 184501 (2009);
X. Noblin, R. Kofman and F. Celestini, Phys. Rev. Lett., 102, 194504 (2009); N. H. Chen,
T. Kuhl, R. Tadmor, Q. Lin and J. Israelachvili, Phys. Rev. Lett., 92, 024501 (2004).
7. E. B. Dussan, J. Fluid Mech., 151, 1 (1985).
8. H. Kusumaatmaja, J. Leopoldes, A. Dupuis and J. M. Yeomans, Europhys. Lett., 73, 740
(2006).
9. C. Priest, R. Sedev and J. Ralston, Phys. Rev. Lett., 99, 026103 (2007).
10. R. Tadmor, K. Chaurasia, P. S. Yadav, A. Leh, P. Bahadur, L. Dang and W. R. Hoffer,
Langmuir, 24, 9370 (2008).
11. P. S. Yadav, P. Bahadur, R. Tadmor, K. Chaurasia and A. Leh, Langmuir, 24, 3181
(2008).
12. See R. E. Johnson, J. Phys. Chem., 63, 1655–1658 (1959) for an interesting discussion
about the role of gravity and about the normal component being balanced; see also J. R.
Henderson, Phys. Rev., E, 69, 061613 (2004) regarding the balance of the line tension at
various conditions.
13. R. Tadmor, Surf. Sci., 602, L108 (2008).
14. T. Podgorski, J. M. Flesselles and L. Limat, Phys. Rev. Lett., 87, 036102 (2001).
15. M. Sakai, J. H. Song, N. Yoshida, S. Suzuki, Y. Kameshima and A. Nakajima, Langmuir,

22, 4906 (2006).
16. C. W. Extrand and A. N. Gent, J. Coll. Int. Sci., 138, 431 (1990).
17. R. Goodwin, D. Rice and S. Middleman, J. Coll. Int. Sci., 125, 162 (1988).
18. M. E. R. Shanahan and P. G. de Gennes, Comptes Rendus Acad. Sci. Serie II, 302, 517
(1986).
19. A. Carré, J. C. Gastel and M. E. R. Shanahan, Nature, 379, 432 (1996).
20. Such molecular reorientation has been recorded in the literature. See, for example,
T. Yasuda, M. Miyama and H. Yasuda, Langmuir, 8, 1425 (1992). See also Refs. [8] and
references therein.
21. P. J. Steinhardt, Effective Medium Theory for the Elastic Properties of Composites and
Acoustics Applications, JSR-91-112 (1992).
3 Particle Adhesion in the
Pharmaceutical Sciences
Frank M. Etzler*
CONTENTS
3.1 Introduction..................................................................................................... 42
3.1.1 The Role of Surface Free Energy—Adhesion Fundamentals............. 42
3.1.2 Complementary Experimental Methods to Assess Surface
Chemistry............................................................................................46
3.1.2.1 Overview...............................................................................46
3.1.2.2 Flow Microcalorimetry.........................................................46
3.1.2.3 Electron Spectroscopy for Chemical Analysis..................... 47
3.1.2.4 Scanning Probe Microscopy................................................. 48
3.2 Importance of Surface Free Energy and Adhesion to Pharmaceutical
Dosage Forms.................................................................................................. 49
3.2.1 Dry Powder Inhalation Formulations.................................................. 49
3.2.2 Tablets.................................................................................................. 50
3.2.2.1 A Model to Predict the Tensile Strength of Tablets
from Individual Components................................................ 52
3.3 Experimental Evidence Supporting the Role of Adhesion in
Pharmaceutical Formulation............................................................................ 54
3.3.1 Ryshkewitch–Duckworth Equation and Tablet Tensile Strength........ 54
3.3.2 DPIs—Importance of Surface Free Energy and Adhesion................. 57
3.3.2.1 Capsule Surfaces................................................................... 57
3.3.2.2 Lactose Surfaces................................................................... 58
3.3.2.3 Powder Flow......................................................................... 62
3.4 Summary.........................................................................................................64
References.................................................................................................................64
In order to achieve quality by design (QbD) in the pharmaceutical industry, it

is necessary to understand fundamental properties that affect product perfor-
mance and manufacturing processes. Powders are frequently used in the man-
ufacture of pharmaceutical dosage forms. In order to understand the behavior
of powders in the manufacture of pharmaceuticals and the performance of
* Prof. Frank M. Etzler can be contacted at fetzler@lecom.edu.
41
pharmaceuticaldosage forms, it is necessary to understand the adhesion

between particles. In addition, the effects of particle size and shape and their
interplay with the adhesion forces also must be understood. Common exam-
ples of dosage forms that make use of powders include tablets and dry powder
inhalation formulations. In this paper, the role of interparticle adhesion and
surface chemistry in the manufacture and use of tablets and dry powder inha-
lation forms is explored. Surface chemical methods and AFM have been used
to explore the lot-to-lot variability of some common pharmaceutical materials.
It is observed that pharmaceutical materials may indeed exhibit lot–lot varia-
tions in surface chemistry, which, in turn, affect interparticle adhesion and
product performance. For lactose, surface contamination, which presumably
occurs during the crystallization process, appears to be one mechanism that
causes variations in interparticle adhesion.
3.1 INTRODUCTION
Quality by design (QbD) is a goal for pharmaceutical manufacturers that has been
only partially implemented. QbD may allow for the adjustment of manufacturing
processes in order to accommodate changes in material properties that result in cor-
responding changes in product performance. In order for QbD to be fully imple-
mented, a fundamental understanding of the factors affecting the performance and
manufacture of these products must be achieved.
Powders are employed in many pharmaceutical products and manufacturing pro-
cesses. The behavior of powders is complex and is, at the present time, incompletely
understood. Factors that affect the behavior of powders include interparticle adhe-
sion, interparticle friction, particle size, particle shape, and gravitational forces. Two
common types of dosage forms that are affected by the behavior of powders include
tablets and dry powder inhalation dosage forms.
In this paper, the adhesion and surface chemical properties of some common
pharmaceutical powders are explored. Furthermore, the relation of surface chemis-
try and adhesion to product performance is explored in this work.
3.1.1 The Role of Surface Free Energy—Adhesion Fundamentals

The adhesion between materials is determined by the area of intimate contact and
the strength of the interaction between bodies. Factors such as particle shape and
size as well as the degree of compression, in the case of tablets, control the quantity
of interaction. Surface chemistry and surface free energy control the quality of this
interaction.
The thermodynamics of the surface free energy, γ, of solids has been reviewed by
Etzler [1]. The ideal work of adhesion, WA, between materials A and B is defined by
the following relation:
WA = γA + γB − γAB (3.1)
Particle Adhesion in the Pharmaceutical Sciences 43
where the subscripts A and B in the surface free energy terms refer to materials A
and B. Similarly, if materials A and B are identical, the work of cohesion, Wc, can
be defined as
Wc = 2γAB. (3.2)
Fowkes [2,3] suggested that surface free energy and thus the work of adhesion
could be considered as a sum of components resulting from various types of inter-
molecular interactions. At present, it is conventional to express the work of adhesion
as a sum consisting of a term for Lifshitz–van der Waals (LW) interactions and a
second term resulting from Lewis acid–base interactions [1]. Frequently, LW interac-
tions are dominant.
Intermolecular forces between molecules result from interactions between their
corresponding electron orbitals. The principal nonbonding interactions result from
induced dipole–induced dipole (London), dipole–induced dipole (Debye), and dipole–
dipole (Keesom) interactions. The intermolecular potential energy function, U, for
each of these three types of interactions is of the same form. Here, r is the separation
distance between bodies.
−β12
U= (3.3)
r6
If only London dispersion forces are considered, the constant in Equation 3.3, β12,
can be expressed as follows:
2 I I 
( )
1/ 2
β12
d
= 1 2
 β11β 22
d d (3.4)
 I1 + I 2 

Here, the subscripts 11, 22, and 12 refer to interactions between like molecules
(11, 22) and dissimilar molecules (12). β, again, is the coefficient in Equation 3.3. I is
the ionization potential. If I1 ≈ I2, then
( )
1/ 2
β12
d
= β11β 22
d d
(3.5)
Equation 3.5 forms the basis of the Berthelot principle [4,5], which states that dis-
persion interactions between dissimilar molecules can be estimated as the geometric
mean of the interactions between like molecules.
The interaction potentials between molecules have been used to determine the
interactions between macroscopic bodies. In this instance, the Hamaker constants,
A, derived from β, also follow the geometric mean rule [1,5,6]. Thus,
A12 = (A11 A22)1/2 (3.6)

and assuming further that intermolecular distances, d, also follow the geometric
mean rule [1]. Continuing the argument, the work of adhesion due to London disper-
sion forces may be expressed as follows:
( A11 A22 )1/ 2

( ) ( )
1/ 2 1/ 2
W12d ≈ = W11d W22d = 2 γ 11
d d
γ 22 (3.7)
12πd11d 22
It has been further generalized that all of the LW components follow the same
rule; thus,
( )
1/ 2
WALW = 2 γ 1LW γ 2LW (3.8)
Berthelot’s rule (Equation 3.8) is widely accepted and is part of most models used
to explain adhesion and contact angle phenomena.
The Lewis acid–base component of the surface free energy has been modeled in
a number of ways. Etzler [1] has discussed the various models in detail. Here, we
highlight the use of two models, one by van Oss, Chaudhury, and Good and the other
by Fowkes, that are frequently discussed in the literature.
According to the van Oss, Chaudhury, and Good model [7,8], the Lewis acid–
base component is modeled as follows:
( )
1/ 2
γ iAB = 2 γ i+ γ i− , (3.9)
where γ+ is the Lewis acid parameter and γ− is the Lewis base parameter. Van Oss,
Chaudhury, and Good further choose
γ i+ = γ i− ≡ 0 (3.10)
for alkanes, methylene iodide, and α-bromonaphathalene, which presumably interact

only through LW interactions. For water,
γ +H2O = γ −H2O ≡ 25.5 mJ/m 2 (3.11)
Based on these above numerical choices, γ+ and γ− have been experimentally

determined for a variety of liquids. Van Oss [8] has compiled and reviewed the deter-
mination of these values. (Also see the work of Etzler [1].)
Earlier, Owens and Wendt [9] described surface free energy in terms of two com-
ponents, which were called dispersion γd and polar γp. Thus,
γ = γ d + γp (3.12)
While it is generally recognized that γ d ≈ γLW, the meaning of γ p is perhaps hope-
lessly confused in the literature. According to Fowkes [2,3], γ p should refer to dipole–
dipole (Keesom), and dipole–induced dipole (Debeye) interactions. In the van Oss,
Chaudhury, and Good model, such interactions are, however, incorporated into γLW.
Good [10] no longer recommends the use of γ p. See Etzler [1] for a more complete
discussion.
Recalling Equations 3.1, 3.8, and 3.9, together with the relation
γ = γLW + γAB, (3.13)
it follows that
( ) ( ) ( )
1/ 2 1/ 2 1/ 2
Wa = γ L [1 + cos(θ)] = 2 γ lLW γ sLW + 2 γ l+ γ −s + 2 γ l− γ +s (3.14)
If the van Oss, Chaudhury, and Good parameters are known for at least three
liquids and the contact angles of these liquids on a solid are measured, then Equation
3.14 can be used to determine the van Oss, Chaudhury, and Good parameters for the
surface free energy of the solid. Van Oss [8], for instance, has reviewed the numerous
publications that have reported the determination of the van Oss, Chaudhury, and
Good parameters for various liquids.
As discussed above, Fowkes [2,3] first suggested that surface free energy could be
considered as a sum of components resulting from different classes of intermolecu-
lar interactions. The van Oss, Chaudhury, and Good model draws upon the idea of
Fowkes and, as discussed above, uses the geometric mean approximation to model
LW interactions. Fowkes [2,11,12], however, has suggested a different approach to
evaluating the acid–base character of surfaces. (Also see the work of Vrbanac and
Berg [13].) Fowkes has criticized the use of contact angles for determination of inter-
facial properties [14]. Fowkes’ approach is, for experimental reasons, more appli-
cable to powdered samples. As stated previously,
Wa = WaLW + WaAB (3.15)
WAB is then, according to Fowkes, expressed by the following relation:
WaAB = − f ⋅ N ⋅ ∆H aAB , (3.16)
where N is the number of sites per unit area, and
−1
 ∂ ln WaAB 
f = 1 −  , (3.17)
 ∂ ln T 
where
f ≈ 0.2 … 1.0. (3.18)
When using the Fowkes approach, some authors have taken f as unity, although
this does not seem to be a good approximation [13]. Because f and N are generally not
known, direct calculations of the work of adhesion are often not made. Determination
of ∆HAB for multiple probe liquids on a given solid together with models by Drago
[15] or Gutmann [16,17] can be used to assess the acid–base nature of the surface.
(Also see the work of Riddle and Fowkes [18].) Here, we only elaborate further on
Gutmann’s acid–base scale.
Gutmann [16,17] introduced the notion of electron donor numbers (DNs) and
electron acceptor numbers (ANs) [18]. These parameters are similar to the van Oss,
Chaudhury, and Good surface free energy parameters as they both describe the same
molecular parameters but from different points of view. In 1966, Gutmann [16] intro-
duced the DN based on the interaction with SbCl5. A DN has units corresponding to
enthalpy (e.g., kJ/mole). In 1975, Mayer et al. [19] introduced the AN based on the
relative 31P shift induced by triethylphosphine oxide. AN has arbitrary units. In 1990,
Riddle and Fowkes [18] removed the dispersion component from AN. The corrected
AN* values have the usual units of enthalpy.
According to Gutmann’s theory,
ΔHAB = Ka DN + Kd AN*, (3.19)
where Ka and Kd reflect the acceptor and donor characteristics of a solid. Gutmann’s
model works best with hard (low-polarizability) atoms. Gutmann’s approach is
suitable for use with inverse gas chromatography (IGC). It should be noted that
Gutmann’s scale is based on enthalpies, while the van Oss scale is based on free
energies.
IGC is an excellent method used to determine γLW and ∆HAB. Etzler [1] has dis-
cussed, in detail, the use of IGC for determination of surface free energy components.
3.1.2 Complementary Experimental Methods to Assess Surface Chemistry

3.1.2.1 Overview
In the previous section, the theoretical basis for determination of the surface free
energy of solids is addressed. The previous discussion provides methods to deter-
mine numerical values for the surface free energy of solids. In the following section,
methods that do not yield numerical values for surface free energy but that nonethe-
less can shed considerable insight into the nature of the surface are addressed. The
surface chemical methods discussed below may complement the determination of
surface free energy.
3.1.2.2 Flow Microcalorimetry

Flow microcalorimetry for use in surface chemistry has been pioneered by Groszek
and coworkers [20–27]. Fowkes and coworkers [28–31] have also contributed signifi-
cantly to the study of the acid–base properties of surfaces via flow microcalorimetry.
The reader is also directed to the work of other groups [32–34]. The apparatus con-
sists of a pair of syringe pumps, a sensitive calorimeter, and a downstream detector
to monitor the concentration of the adsorbate. The downstream detector monitors
refractive index, ultraviolet absorbance, or other properties sensitive to concentra-
tion. The apparatus measures the heat of adsorption and the amount of material
adsorbed. From the two measured quantities, the molar heat of adsorption can be
determined. Commercial instrumentation is manufactured by Microscal (UK).
Fowkes and coworkers [28–31] have, in particular, addressed the use of flow
microcalorimetry to assess the acid–base properties of surfaces. IGC and flow
microcalorimetry are similar in that both techniques can be used to determine ΔHAB.
Thus, Fowkes’ approach combined with either Drago’s or Gutmann’s model could
be used to determine the acid–base character of materials. IGC and flow microcalo-
rimetry are complementary with respect to determining the acid–base character of
materials. IGC uses the van’t Hoff approach for calculating enthalpies of adsorption,
while microcalorimetry measures the quantity directly.
When using flow microcalorimetry, the heat of adsorption of an acid–base probe
from a neutral solvent such as isooctane is measured. This quantity is essentially
ΔHAB as the heat of adsorption of a neutral solvent from another neutral solvent is
negligible. Fowkes and others have used probes for which Drago parameters are
available to assess the acid–base properties of materials. Such probes usually yield
heats of adsorption that are large enough to be measured with good accuracy.
Flow microcalorimetry has not been used to determine LW contributions to sur-
face free energy.
3.1.2.3 Electron Spectroscopy for Chemical Analysis

One technique for monitoring the chemical nature of surfaces, which is applicable to
a wide variety of materials, is electron spectroscopy for chemical analysis (ESCA).
ESCA can be used to determine the elemental composition and oxidation states of
surfaces. ESCA is also known in the literature as x-ray photoelectron spectroscopy
(XPS). The method has been used extensively in the literature to study adhesion and
other surface chemical problems. ESCA/XPS has been reviewed in the literature by
Barr [35]. Here, we merely highlight the potential for its use to complement surface
free energy calculations.
ESCA is based on the photoelectric effect. When x-ray photons strike a surface,
electrons are emitted. The ejected electrons come from inner electron orbitals (1s, 2s,
etc.). The kinetic energy of the ejected electrons can be related to the electron orbital
energy via the following expression:
1
hν = ⋅ mve2 + Eb + qΦ (3.20)
2
The binding energies, Eb, are determined from the kinetic energy of the emitted
electrons. The electron binding energy for a given electron is characteristic for the
element and the particular electron orbital belonging to that element. qФ is the work
function and is usually characteristic of the apparatus. It represents the extra work to
extract electrons from condensed media. In vacuo, qФ = 0.
In addition to being characteristic for a given element, Eb is also dependent on
the electronegativity of the surrounding atoms; thus, both an elemental analysis and
some information about the chemical environment of each atom under study can be
obtained from ESCA.
The x-ray source typically illuminates a spot on the order of 1–5 mm. For pow-
ders, it is not possible to obtain information about single particles. Electrons emitted
from atoms located within approximately 5 nm of the surface are detectable using
ESCA. Because the electrons come from atoms located near the surface, ESCA is
usually regarded as a surface sensitive technique.
3.1.2.4 Scanning Probe Microscopy

The ability to study the characteristics of a surface down to atomic dimensions allows
one to gain considerable insight into interfacial phenomena. Scanning probe micros-
copy is a powerful technique for studying surfaces down to atomic dimensions [36–38].
A scanning probe microscope (SPM) consists of a needlelike probe attached to a
cantilever spring that can be moved along the surface [36]. Surface chemically modi-
fied tips or particles of interest can be substituted for the usual SPM tip. The practice
of substituting particles for the tip is referred to as the colloidal probe technique. The
lateral (x,y) position of the probe and deflection of the cantilever spring are recorded.
Cantilevers are typically made by microlithography from silicon or silicon nitride. The
dimensions of cantilevers are 100–200 μm long, 10–20 μm wide, and 0.5 μm thick.
The pointed tips are a few micrometers high. Tips may be 30 nm in radius, but when
the contact force is small, the area of contact between the tip and surface is on the
order of atomic dimensions. The forces measured are in the range of 10−13 to 10 −6 N.
SPM contrasts with stylus profilometry, which uses forces of 10 −4 N over submicrom-
eter areas. The deflection of the cantilever has been monitored by a variety of tech-
niques, but laser beam deflection is often used. The force constant of the cantilever can
be selected for the range of forces under study. Once position and deflection data have
been recorded by a computer, a surface map can be constructed. Low deflections are
usually recorded as dark tones (shades) and high deflections as light tones. The range
of forces reflecting discernible tones is selected to be appropriate to the sample.
SPM can be used on a wide variety of surfaces. The surfaces may either be electri-
cal conductors or insulators. The apparatus can be operated using either dry samples
or samples immersed in liquid media. As the exact nature of the sample is relatively
unimportant, SPM has been used on a wide variety of samples.
One of the most straightforward ways to use SPM is for the construction of topo-
graphic maps. The tip and surface experience attractive and repulsive intermolecular
forces whose origins have been discussed above in regard to theories on surface free
energy. In order to construct a topographic map, the SPM tip is brought into close
proximity to the surface, where it experiences a repulsive force. The tip can then be
scanned in a raster pattern at constant force or constant height. Changes in height or
photocurrent are collected for the calculation of topographic information.
Other information about surfaces can also be obtained by applying, for instance,
oscillating (tapping) forces or lateral forces or varying the probe surface chemistry.
These other useful techniques are discussed elsewhere [37]. The present author has,
together with other coworkers, characterized pharmaceutical surfaces using phase,
friction, and adhesion imaging [39–41].
For those wishing to better understand surfaces, SPM provides a number of
techniques that may be used to understand the nature of surfaces, particularly with
regard to surface chemical heterogeneity [42–47].
In phase imaging, the cantilever is oscillated near its resonance frequency. The
oscillation causes the probe to tap along the surface. In general, the oscillation of
the tip will be out of phase with respect to the applied oscillating force. The shift in
phase reflects the mechanical properties of the surface (elasticity, adhesion). A phase
map of the surface can distinguish regions of different natures. In addition to phase
imaging, both adhesion and friction maps may give insight into surface chemical
heterogeneity (Figure 3.7).
SPM tips that have been chemically modified can be used to assess the surface
chemical nature of surfaces by measuring, for instance, the relative adhesion of tips
with different surface chemical modifications.
SPM is a powerful tool for investigation of the surface properties of materials.
Beach and coworkers [48] have recently used SPM to measure surface free energies.
3.2 IMPORTANCE OF SURFACE FREE ENERGY AND

ADHESION TO PHARMACEUTICAL DOSAGE FORMS
3.2.1 Dry Powder Inhalation Formulations
The delivery of aerosols by inhalation may offer advantages over other delivery sys-
tems. The potential advantages may include increased delivery efficiency, increased
bioavailability and decreased systemic toxicity due to the large surface area of thin
alveolar epithelium for rapid absorption, lack of first-pass metabolism, and direct site of
action to the target organ [49]. Delivery of small molecules for the treatment of diseases
such as asthma and chronic obstructive pulmonary disease (COPD) is a major appli-
cation for inhalation technologies [50]. Aerosol technology may also be used to treat
respiratory tract and lung infections using antibiotics or vaccines, or for the delivery
of proteins, peptides, and macromolecules for local and systemic disease therapy [51].
Aerosol drug delivery is implemented in three major classes of devices. These
classes include pressurized metered-dose inhalers (pMDIs), dry powder inhalers
(DPIs), and nebulizers. In recent years, DPI development has had increased interest
due, in part, to the removal of ozone-depleting propellants from the market.
DPIs contain three major components:
1. The device
2. The formulation
3. The dose-metering system
For passive DPIs, the inspiratory flow is the only source of energy for particle dis-
persion from a static powder bed. The inhalation device is designed to allow efficient
resuspension of the powder using a specified fluidization and deaggregation mechanism
[52]. Each commercial DPI device has defined dimensions and internal geometry. Thus,
each of the available commercial devices has its peculiar airflow parameters (e.g., spe-
cific resistance) and combination of aerosol generation mechanisms to enhance drug
delivery. The mechanisms of DPI fluidization include shear, capillary, and mechanical
forces. The mechanisms for particle deaggregation include relative motion, turbulence,
shear stress, and collision [52]. These mechanisms occur to a different extent in each of
the various commercial inhalers [50,52]. The aerodynamic characteristics responsible
for the aerosolization of the particles are usually not completely understood.
Formulations suitable for use in DPIs typically contain respirable drug particles
that have a significant fraction of particles whose aerodynamic diameters are in the
range of 1–5 μm depending on the target site. (Aerodynamic diameter is the diameter
of a spherical particle with a density of 1 gm/cm3 that has the aerodynamic charac-
teristics of the studied particle.) The powder often consists of a respirable drug mixed
with an excipient, which is often α-lactose monohydrate [53]. α-lactose monohydrate,
composed of larger particles (>30 μm [54]) is thought to increase powder flowability,
facilitate drug aerosol formation, and serve as a bulking agent to ensure uniform fill-
ing. (The subject of particle size analysis is complex and beyond the scope of this paper.
Etzler [55] has reviewed this subject in detail. It appears, however, that the develop-
ment of an understanding of the function of components in DPIs is, in part, limited by
particle sizing technologies typically employed by pharmaceutical scientists.)
DPIs use three types of metering systems. These systems include the preme-
tered individual capsules, dose blisters or strips, and multiple-dose reservoirs [56].
The device, the formulation, the patient’s inspiratory rate, and the interparticulate
adhesion of the powder at rest and during resuspension each contribute to the drug
aerosolization efficiency and reproducibility. The physicochemical and aerodynamic
characterization of DPI formulations has been reviewed [57].
Dry powder aerosol performance efficiency and reproducibility can be measured
using the emitted dose, fine particle mass (FPM) or fine particle fraction (FPF), or
mass median aerodynamic diameter (MMAD) [52]. The emitted dose represents the
proportion of the nominal dose that leaves the device. The remaining portion stays
in the device. The FPM or FPF is the mass or proportion with respect to the nominal
dose in a size range considered sufficiently small to enter the lungs. It is frequently the
case that the FPM or FPF is defined with respect to the proportion of mass or fraction
of the total particles whose aerodynamic diameters are below 5 μm (dependent on the
sampler and operating conditions) [52]. The MMAD represents the aerodynamic par-
ticle size for which 50% of the total aerosol by mass has smaller aerodynamic diam-
eter [52]. Both FPM (or FPF) and MMAD reflect the degree of deaggregation, where
larger FPM (or FPF) and smaller MMAD values represent greater deaggregation [58].
In the context of this paper, it clear that factors affecting interparticle adhesion or
adhesion of the formulation to the powder reservoir will affect the performance of
dry powder inhalation formulations. In this chapter, the author considers a formula-
tion contained in gelatin capsules.
It is also important to note that α-lactose monohydrate and many drug substances
adsorb water. If powders are exposed to sufficiently high relative humidity, particle
sintering will occur with a negative effect on FPF or FPM. Relative humidity, hence,
is frequently an important environmental factor affecting the performance of DPIs.
3.2.2 Tablets
Tablets are the most common dosage form employed by the pharmaceutical indus-
try. They are both inexpensive to produce and convenient for patients. Active phar-
maceutical ingredients, particularly those incorporated into innovator company
products, are new chemical substances whose chemical and physical properties are
incompletely known and are sometimes present in large amounts in the manufactured
products. Excipients are often incorporated into tablet formulations to overcome at

least some of the undesirable properties of the active ingredient. Successful formula-
tions, in addition to having desirable medicinal properties and sufficient chemical and
physical stability under environmental stress, must be manufacturable. In order to be
manufacturable, the powder from which tablets are formed must have adequate flow
properties, and tablets must have sufficient tensile strength to survive handling during
manufacturing, packaging, and subsequent handling during shipping and patient use.
Achieving adequate tensile strength can sometimes be a challenge, particularly
when formulations contain large amounts of active ingredient or require special
excipients to enhance dissolution or increase stability. Special excipients may have
undesirable properties with regard to tableting.
During tableting, powder is uniaxially compressed in the die. During the com-
pression process, the powder is densified under the applied pressure. Densification
of the powder occurs through particle rearrangement, particle fracture, and particle
deformation. The relative extent of each of the three identified densification mecha-
nisms that occurs in a specific instance is determined by the applied pressure, the
rate of the increase of pressure, the duration of the applied pressure, and the mechan-
ical properties of the materials that are being compressed.
Adhesion strength depends on both the quality and quantity of intimate contact
between the materials in question. The densification of the powder that occurs dur-
ing tableting results in an increase in the contact area between the particles. The
quality of interaction depends on the specific surface chemical properties of the
materials in question.
The pharmaceutical literature has discussed the mechanical aspects of the tablet-
making process but has rarely discussed the tablet-making process from a surface
chemistry or adhesion science perspective. Below, a model based on surface chemi-
cal principles that can be used to calculate the tensile strength of tablets from con-
stituent components is discussed.
The tensile strength of tablets is routinely tested as a measure of tablet quality [59].
It is possible to consider three types of tensile strength—radial, bending, and axial.
Here, only the radial tensile strength will be discussed as it is most commonly measured.
The radial tensile strength is assessed by measuring the diametrical crushing
force using what is commonly referred to as a “hardness tester.” The diametrical
crushing force is dissimilar to indention hardness, and the use of “hardness” in the
present context is erroneous but nonetheless conventional. Like other mechanical
tests, the diametrical crushing force is dependent on the test speed; thus, results may
vary between testers as pressure to the tablet may not be applied in the same way.
The radial tensile strength, τc, is calculated from the diametrical crushing force,
Fc, for cylindrical tablets using the relation below [59–61].
2 Fc
τc = (3.21)
πdh
Here, d is the tablet diameter, and h the tablet thickness. It is the radial tensile
strength that is commonly determined to assess tablet quality. Procipio et al. [62]
have recently discussed measurement of radial tensile strength.
Ryshkewitch [63] and Duckworth [64] proposed an empirical equation to relate

tablet tensile strength, τ, to porosity of the tablet, ε. Thus,
τ = τ0 e−kε (3.22)
Here, τ0 is the tensile strength of the material at zero porosity (D = 1) and k a

characteristic constant. D is relative density (ρ/ρ 0) where ρ is the bulk density of the
powder and ρ 0 is the material density. Porosity, ε, is 1 – D.
Over the past several decades, the Ryshkewitch–Duckworth equation has been
discussed a number of times. Recently, the Ryshkewitch–Duckworth equation has
been discussed by Doelker [59], Barraleta et al. [65], Nyongesa and Aduda [66],
as well as Tye et al. [67]. The literature suggests that the Ryshkewitch–Duckworth
equation has been successfully applied to a variety of systems that have applications
both to pharmaceutical science and to other disciplines.
The importance of the Ryshkewitch–Duckworth equation is that it suggests that
porosity is a measure of the outcome of the tableting process. Tablets of the same
porosity have the same mechanical properties even though they might have been
made on vastly different tablet presses operating at different speeds.
The speed of the tableting process can be expressed in a number of ways. The
term speed may refer to the speed at which the tablet punches advance, the dwell
time, the total contact time of the punches with the tablet, or the rotational speed
of the press. In a conventional tablet press, all of these terms are interrelated. The
dwell time is usually defined as the time for the compression wheels to pass over
the flat portion of the top of the tablet punch. For an Instron or other universal
testing machine, the time that the tablet punch remains in fixed position can be
considered as a measure of dwell time. This is not dissimilar to the case of the
rotational press.
The Ryshkewitch–Duckworth equation is an empirical equation. Andersson [68]
has, however, offered a theoretical explanation. Knudsen [69] also comments on the
Ryshkewitch–Duckworth equation and associates the increase in strength with an
increase in the surface area of contact and the k parameter to the difficulty of creat-
ing contact area. k is expected to be temperature dependent.
Uniaxially compressed tablets are known to have anisotropic properties. Results
reported by Doelker [59] show that the axial tensile strength of tablets may deviate
from the Ryshkewitch–Duckworth equation at low porosities. Presumably, the vis-
coelastic recovery on decompression results in weakening of the tablet in the axial
direction. It is well known that overcompression of tablets can result in capping and
lamination (i.e., failure in the z direction). The issue of capping and lamination,
although important, will not be discussed further in this paper.
3.2.2.1 A Model to Predict the Tensile Strength of

Tablets from Individual Components
In an earlier paper, Wu et al. [70] proposed a model to calculate the radial tensile
strength of tablets from the Ryshkewitch–Duckworth parameters of the individual
components of the powder being compressed. Tye et al. [67] and Sun [71] have also
commented further on this approach. In particular, Tye et al. [67] have extended
the application to quaternary mixtures. These studies establish the importance of
the Ryshkewitch–Duckworth equation for understanding tensile strength of tablets
composed of multiple components. Furthermore, they reiterate the importance of
porosity as a measure of the outcome of the tableting process. Tablets, even though
produced by vastly different tablet presses operating at different speeds, can be
expected to have the same mechanical properties if they have the same porosity. The
required compaction force or compaction pressure to give identical porosities may,
however, not be the same if different presses are employed.
Here, we propose a model that is similar to that proposed by Wu et al. [70] but
differs from it by drawing more strongly from the principles of adhesion science. The
form of the final equation also differs.
For a given material, the tensile strength, τ, is related to the surface free energy, γ
[6]. Thus, for two particles composed of the same material,
( )
2
τ ∝ γ = γ 1/ 2 , (3.23)
and if the particles are composed of different materials,
( )
γ 12 = γ 11/ 2 γ 12/ 2 (3.24)
or
( )
τ12 = τ11/ 2 τ12/ 2 (3.25)
Berthelot’s rule (see the section The Role of Surface Free Energy—Adhesion
Fundamentals) is thus applied to the tensile strength between particles. We extend
Berthelot’s rule for an ensemble of particles assuming that the geometric mean rule
applies to an ensemble of adhesive contacts between particles and is weighted via the
surface area fraction, φs, thus,
τ= ∏τ i
ϕs
i (3.26)
If the particle sizes of the components are similar, then φs ∝ φv where φv is the
volume fraction. Equation 3.26 is thus rewritten as
τ= ∏τ i
ϕv
i (3.27)
Combining Equation 3.27 and the Ryshkewitch–Duckworth equation (Equation

3.22) yields the following relation.
∏ (τ e ) ∑ϕ ( )
ϕ vi
τ= − ki ε
→ ln(τ) =  ln τ 0 − ki ε  = lnn(τ 0m ) − km ε (3.28)
0i vi  i 
i i
Equation 3.28 indicates that the Ryshkewitch–Duckworth parameters for a pow-

der composed of several components may be calculated from the Ryshkewitch–
Duckworth parameters of the individual components using the following relations.
ln(τ 0 m ) = ∑ ϕ ln ( τ ) (3.29)
i
vi 0i
km = ∑ ϕ k (3.30)
i
vi i
The subscript m refers to the properties of the mixture.
3.3 EXPERIMENTAL EVIDENCE SUPPORTING THE ROLE OF

ADHESION IN PHARMACEUTICAL FORMULATION
3.3.1 Ryshkewitch–Duckworth Equation and Tablet Tensile Strength
In the previous section, a model to calculate the tensile strength of tablets from
the properties of the component materials is described. This model has been
applied to mixtures of common pharmaceutical excipients with sodium dodecyl
sulfate (SDS) by Etzler et al. [73]. In this earlier paper more complete experimen-
tal details are discussed. In the paper by Etzler et al. [73], cylindrical tablets were
prepared at several speeds using an Instron universal testing machine at low speeds
and an MCC Presster compaction simulator at higher speeds. Figure 3.1 shows a
Ryshkewitch–Duckworth plot for dicalcium phosphate. The results suggest that
porosity is, indeed, a principal measure of the outcome of the tableting process as
tablets produced under different conditions of maximum pressure and dwell time
have the same tensile strength if the tablet porosity is the same. Figure 3.2 con-
firms the validity of Equations 3.29 and 3.30 for mixtures of dicalcium phosphate
and SDS. Figure 3.3 shows Ryshkewitch–Duckworth plots for various mixtures of
mannitol and SDS. Figure 3.4 confirms, again, the validity of Equations 3.29 and
3.30 for mixtures of mannitol and SDS. The model suggested in Section 3.2.2.1
appears to be generally adequate for predicting the tensile strengths of tablets
from the Ryshkewitch–Duckworth parameters of the constituent components of
the mixture. Complications that deserve further attention include the effect of wet
granulation on the surface free energy and adhesion of polymeric components
and the effect of a crystal habit on the Ryshkewitch–Duckworth as crystals of
10
Dicalcium phosphate 90 sec
20 sec
500 ms
23 ms
9 ms
1
τ (MPa)
0.1
τ0 = 6.924
k = 0.1325
0.01
0 10 20 30 40
Porosity (%)
FIGURE 3.1 Ryshkewitch–Duckworth plot for dicalcium phosphate at several dwell times.
As suggested in earlier publications [67], no dwell time dependence is noted.
0.16
10
0.12
τ0 (MPa)
0.08
k
0.04
0.1 0
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
SDS (volume fraction) SDS (volume fraction)
FIGURE 3.2 Ryshkewitch–Duckworth parameters for dicalcium phosphate and sodium

dodecyl sulfate mixtures. Results conform to the proposed model and Equations 3.29 and
3.30. The 90% confidence interval of the fit is indicated by the dashed lines. The solid line is
a least-squares fit to the data points.
different habits will present different crystal planes for adhesion. Adsorption of
surface active materials on particle surfaces may also further complicate matters.
Fichtner et al. [74] studied the compaction of lactose and also confirmed the valid-
ity of the Ryshkewitch–Duckworth equation. They studied the compaction of spray-
dried lactose mixed with polysorbate 80 in order to modify the surface free energy
10
Mannitol-SDS mixtures
0% SDS
15% SDS
30% SDS
45% SDS
1
τ (MPa) 100% SDS
0.1
0.01
0 10 20 30 40
Porosity (%)
FIGURE 3.3 Ryshkewitch–Duckworth plots for mannitol mixtures with sodium dodecyl
sulfate (SDS).
10 0.2
0.16
0.12
τ0 (MPa)
1
k
0.08
0.04
0.1 0
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
SDS (volume fraction) SDS (volume fraction)
FIGURE 3.4 Ryshkewitch–Duckworth parameters for mannitol and sodium dodecyl sul-
fate mixtures. Results conform to the proposed model and Equations 3.29 and 3.30. The 90%
confidence interval of the fit is indicated by the dashed lines. The solid line is a least-squares
fit to the data points.
while maintaining constant particle geometry. Figure 3.5, redrawn from Fichtner’s
work, shows the relationship between tablet strength and dispersion component of
surface free energy at various tablet porosities. For a specific tablet porosity, a posi-
tive correlation between the dispersion component of surface free energy and tablet
tensile strength exists. The surface free energy, as suggested by the principles of
adhesion science, is a significant factor determining the tensile strength of powder
compacts. An understanding of surface free energies is thus critical for understanding
8 ε = 0.10
ε = 0.15
ε = 0.20
ε = 0.25
6
τ (MPa)
4
0
46 47 48 49 50
γ LW (mJ/m2)
FIGURE 3.5 Tensile strength of tablets prepared from lactose spray-dried with polysorbate 80
to alter the surface free energy versus Lifshitz–van der Waals (LW) component of surface free
energy. Each curve is for tablets at constant porosity (ε). A correlation between γLW and tensile
strength exists at each porosity. (Data from F. Fichtner et al., Pharm. Res. 25, 2750, 2008.)
tablet tensile strength. Fichtner’s conclusions are consistent with the model presented
in Section 3.2.2.1.
3.3.2 DPIs—Importance of Surface Free Energy and Adhesion

As discussed above, particle adhesion is an important factor contributing to the per-
formance of DPIs. Particle adhesion to the container containing the metered dose of
powder and interparticle adhesion are both important to the performance of DPIs.
DPIs of different designs are commercially available. The DPI type considered
further in this work consists of powder contained in a gelatin capsule. In order to
inhale the powder, the capsule is then placed into a passive-type inhaler device. For
such a DPI system, the respirable mass is determined by the amount of powder leav-
ing the capsule and the dispersion of the powder upon leaving the capsule. In par-
ticular, the adhesion of powder to capsule walls and interparticle adhesion affect the
respirable mass. For the formulations discussed here, storage at high relative humid-
ity is known to cause sintering of particles. Although very important, the effects of
storage at high relative humidity are not discussed in this paper.
3.3.2.1 Capsule Surfaces

Etzler and coworkers [39,40,75,76] have discussed the adhesion of particles to gelatin
capsules. In order to manufacture capsules, cool steel pins of the size and shape of
capsules are dipped into heated gelatin. Upon removal of the pin, a coating of gelatin
remains on the pin. The gelatin coating forms one half of the capsule—either the
top or bottom portion of the capsule. Prior to dipping, the pin is coated with a mold
Acid and base parameters of γ (mJ/m2)

1.6 γ+
γ–
1.2
0.8
0.4
0
0 4 8 12 16
FIGURE 3.6 Acid (γ+) and base (γ–) parameters of surface free energy for several lots of
gelatin capsules determined from contact angle data [75]. The Lifshitz–van der Waals com-
ponent is similar for all lots.
release agent. The interior surface of the gelatin capsule is thus coated with mold
release agent. It was observed that the mold release agent could be removed with an
organic solvent or supercritical carbon dioxide. Furthermore, improved manufactur-
ing processes were able to produce capsules without residual mold release agent.
Figure 3.6 shows the van Oss acid and base parameters of the surface free energy
of capsules. The results suggest that the surface chemistry of capsule surfaces var-
ies between lots. Figures 3.7 and 3.8 show the SPM images of the capsule surfaces.
The figures suggest that capsule surfaces are contaminated with components of the
mold release agent. Furthermore, the figures confirm that these components can be
removed using supercritical carbon dioxide or via improved manufacturing processes
that allow for the production of capsules with more or less clean gelatin surfaces.
Figure 3.9 confirms that lactose particles have lower adhesion to clean capsule
surfaces than to capsules with residual mold release agent. Clean capsule surfaces
are thus essential for the production of DPIs with consistent performance [75].
3.3.2.2 Lactose Surfaces

Dry powder inhalation formulations consist of a powder reservoir (in the present
discussion, this is a gelatin capsule) and powder that is generally composed of a drug
and a carrier, often lactose. Dispersion of the powder in a consistent manner requires
constant particle size and consistent interparticle and particle adhesion forces. In the
previous section, it was shown that the nature of capsule surfaces varies between
lots. This variation is caused by the presence of mold release agent on the capsule
surfaces. Surfaces that have no surface contamination show minimal particle adhe-
sion. In this section, attention is given to the surfaces of lactose.
(a) (b)
(c) (d)
(e) (f )
FIGURE 3.7 Height (left column) and friction SPM images (right column) of gelatin cap-
sules [39]. Images a and b are from capsules produced by an improved process. These
images show that capsules have a uniform surface. Images c and d are from normal cap-
sules. The friction image (d) shows nonuniform friction across the sample. Dark areas are
pits in the capsule surface formed during the manufacturing process. Images e and f are
from normal capsules that have been cleaned with supercritical carbon dioxide. This clean-
ing process removes the residual mold release agent, and thus, the friction across the sample
is uniform.
For powdered samples, IGC is an excellent method to characterize surface free

energy of solids. IGC can be used to determine the LW component of surface free
energy and the acid–base character using Gutmann’s model, as described earlier in
this work (see Section 3.1.1) and reviewed by Etzler [1]. Figure 3.10 shows the results
from IGC measurements made on several lots of lactose (Pharmatose 200M). The
results indicate that the surface chemistry of lactose is not constant between lots. A
tapping mode SPM image of the surface (Figure 3.11) shows surface contamination
(a) (b)
(c) (d)
FIGURE 3.8 Adhesion images (10 µm × 10 μm scan): (a) adhesion and (b) adhesion area
map of normal capsules and (c) adhesion and (d) adhesion area map of cleaned capsules. The
adhesion maps are constructed using the snap-off force. Adhesion area maps are constructed
from the area of the triangular of the force curve representing attractive forces. Light-colored
rings in c are pits in the capsule surface formed during manufacture (see Etzler et al. [76] for
details and also see Willing et al. [40]).
500
Pext
29625ext
72601
29623
400 P
29625
Adhesion force (nN)
300
200
100
0
0 4 8 12 16
Particle diameter (µm)
FIGURE 3.9 Adhesion force of lactose to capsule surfaces as a function of lactose particle
diameter [76]. Linear nature of the curve is consistent with attraction via Lifshitz–van der
Waals forces. Upper curve is for normal capsules and lower curve is for capsules that have the
mold release agent removed.
Pharmatose 200 M lot variability Pharmatose 200 M lot variability

0.6 0.4
Ka
Kd
0.3
0.4
Fraction of lots
Fraction of lots
0.2
0.2
0.1
0 0
38 40 42 44 46 0.3 0.5 0.7 0.9 1.1 1.3
γLW (mJ/m2) Ka or Kd
FIGURE 3.10 The Lifshitz–van der Waals component of the surface free energy for lactose
(Pharmatose 200 mesh), and acid character (Ka) and base character (Kd) of lactose (Pharmatose
200 mesh) are shown. Data indicate that the surface chemistry varies between lots.
of the lactose surface. It is believed that the contamination is likely from protein and/
or lipids found in milk from which the lactose is obtained [75].
The variable surface chemistry of lactose affects the performance of dry
powder inhalation formulations. In particular, the adhesion of lactose to capsule
walls is affected. Figure 3.12 shows the effect of surface contamination on adhe-
sion of lactose to capsule walls. The level of surface contamination was moni-
tored using XPS. It was observed that the ratio of OCO-type carbons to aliphatic
0 1.00 µM
Data type Phase
Z range 2.00 de
FIGURE 3.11 SPM phase image (1 µm × 1 µm) of a Pharmatose 200M surface. Dark areas
and white specks represent surface contaminants. XPS analysis suggests contamination with
lipids and proteins from milk [75].
24 0.75 E03311
0.70
20 620236
∆H pyridine (mcal/g)
620237
492-137 95090101
620237
08109602 0.65
08109603
16 08109601
Ka
E03311 9608101
RM1523 9509240
9608101
0.60 95090101 95070157
024614 620236
E03303 497-013
12
E03256 0.55
E03317 497-013
E03317 95070156
95070156
8 0.50
0 2 4 6 8 10 0 2 4 6 8 10
% OCO/CH % OCO/CH
10
E03311
RM1523
95090101
8
620236
capsule lot 29623
E03256 492-137
95070157
% Retention
9509240
6
620237
95070156
2
0 2 4 6 8 10
% OCO/CH
FIGURE 3.12 The enthalpy of adsorption for pyridine determined by flow microcalorim-
etry (top, left), Gutmann’s Ka parameter (top, right), and retention of lactose in a capsule (bot-
tom) as a function of surface chemical composition determined using XPS. Carbons of the
OCO type are from lactose. Aliphatic carbons (CH) are from surface contaminants. A high
amount of surface contamination corresponds to high adhesion. More detailed information
can be found elsewhere [75].
carbons (CH) provided a good correlation to capsule retention and surface acid-
ity. (See the work of Etzler [75] for a more detailed discussion.) Material retained
on capsule walls is unavailable to the patient and is thus important for product
performance.
3.3.2.3 Powder Flow

The flow of powders is complex and dissimilar to that of simple fluids (gases and
liquids). The flow of powders depends on the interaction of many variables. The
variables include, for instance, particle size, particle shape, friction and adhesion
80 80
60 60
A1 A1
Angle of repose (°)
Angle of repose (°)

A7 A7
40 40
A4 A4
C A5
A6A3 A6A5A3
20 A2 20 A2
0 0
10 100 1000 10,000 0 2 4 6 8 10
Bond number Kd/Ka
10,000
1000 A1
Bond number
A7
A3
A6 A5
100 A4
A2
10
0 2 4 6 8 10
Kd/Ka
FIGURE 3.13 Relations between bond number, angle of repose, and surface basicity for
modified aluminum powders. Bond number is the ratio of cohesion force to particle weight.
Particle weight, in this instance, is based on the volume averaged particle diameter. Surface
basicity (Kd /Ka) is correlated with bond number and angle of repose. (Data redrawn from
Jallo et al., J. Adhesion Sci. Technol. 25, 367, 2011.)
between particles, particle mechanical properties, bulk density of the powder, as

well as the particle density. Particle flow is important to the tableting process as a
consistent amount of material must be delivered to the punch cavity for tableting.
The motion of the powder can be considered to be the motion of two fluids—the
solid particles and the gas phase. In order for flow to occur, the permeability of
the gas phase must be sufficient to allow for the upward motion of the gas phase
and downward motion of the solid particles. Aside from the permeability of the gas
phase, friction and adhesion between particles impede the motion of the particles
downward. Gravity, in contrast, promotes the downward motion of particles. The
balance between adhesion forces and gravitational forces plays a critical role in the
flow of powders. Johanson [72] provides a discussion of the principles of powder flow
from an engineering perspective.
While a complete discussion of powder flow is beyond the scope of this paper, it is
noted that Jallo et al. [77] have discussed the effect of interparticle adhesion on pow-
der flow. Jallo et al. have specifically shown that the angle of repose for silianized
aluminum powders is correlated with the acid–base character of the particle surface
as determined by IGC (Figure 3.13).
3.4 SUMMARY
In order to implement QbD in the pharmaceutical development process, an under-
standing of the relation between material properties and product and process per-
formance must be achieved. In this paper, the relation of surface free energy to the
performance of DPIs and the hardness of tablets is discussed. Some preliminary
work regarding the role of surface free energy in powder flow is also discussed.
With regard to tableting, a model for the calculation of the hardness of tablets using
material properties and the principles of adhesion science is discussed. The model
shows the relation between surface free energy and the resulting tablet hardness.
The surface chemical properties of some of the components used in the formu-
lation of DPIs are also discussed. It is shown that these surface chemical proper-
ties exhibit lot-to-lot variation. In the present examples, the variation appears to be
largely caused by surface chemical contamination. In the case of lactose, the level
of contamination is submonolayer, thus making it undetectable by ordinary bulk
chemical analysis. Extremely small amounts of surface contaminants may thus have
profound implications for product or process performance.
It is rare that materials are characterized for their surface chemical nature. It is
clear from the data presented here that surface free energy, indeed, can play a sig-
nificant role in product and process performance. It will be necessary for the phar-
maceutical industry to increase its understanding of the role of surface properties in
product and process performance in order to fully implement QbD.
REFERENCES
1. F. M. Etzler, in: Contact Angle Wettability and Adhesion, Vol. 3, K. L. Mittal (Ed.),
p. 219. VSP, Utrecht, the Netherlands (2003).
2. F. M. Fowkes, J. Colloid Interface Sci. 28, 493 (1968).
3. F. M. Fowkes, J. Phys. Chem. 66, 382 (1962).
4. D. Berthelot, Compt. Rend. 126, 1857 (1898).
5. W. V. Chang and X. Qin, in: Acid–Base Interactions: Relevance to Adhesion Science
and Technology, Vol. 2, K. L. Mittal (Ed.), p. 3. VSP, Utrecht (2000).
6. A. W. Adamson, Physical Chemistry of Surfaces, 4th ed. John Wiley, New York (1990).
7. R. J. Good, in: Contact Angle, Wettability and Adhesion, K. L. Mittal (Ed.), p. 3. VSP,
Utrecht, the Netherlands (1993).
8. C. J. van Oss, Interfacial Forces in Aqueous Media. Marcel Dekker, New York (1994).
9. D. K. Owens and R. C. Wendt, J. Appl. Polym. Sci. 13, 1741 (1969).
10. R. J. Good, in: Acid–Base Interactions: Relevance to Adhesion Science and Technology,
Vol. 2, K. L. Mittal (Ed.), p. 167. VSP, Utrecht, the Netherlands (2000).
11. F. M. Fowkes and M. A. Mostafa, Ind. Eng. Chem. Prod. Res. Dev. 17, 3 (1978).
12. F. M. Fowkes, J. Adhesion 4, 155 (1972).

13. M. D. Vrbanac and J. C. Berg, in: Acid–Base Interactions: Relevance to Adhesion
Science and Technology, K. L. Mittal and H. R. Anderson Jr. (Eds.), p. 67. VSP, Utrecht,
the Netherlands (1991).
14. F. M. Fowkes, J. Adhesion Sci. Technol. 1, 7 (1987).
15. R. S. Drago, G. C. Vogel and T. E. Needham, J. Amer. Chem. Soc. 93, 6014 (1971).
16. V. Gutmann, A. Steininger and E. Wychera, Montash. Chem. 97, 460 (1966).
17. V. Gutmann, The Donor-Acceptor Approach to Molecular Interaction. Plenum, New
York (1978).
18. F. L. Riddle and F. M. Fowkes, J. Amer. Chem. Soc. 112, 3259 (1990).
19. U. Mayer, V. Gutmann and E. Wychera, Montash. Chem. 106, 1235 (1975).
20. A. J. Groszek and C. E. Templer, Fuel 67, 1658 (1988).
21. A. J. Groszek, Carbon 27, 33 (1989).
22. C. E. Templer, in: Proceedings, Particle Size Analysis, Vol. 26, p. 301. The Society for
Analytical Chemistry (1970), London, England.
23. G. I. Andrews, A. J. Groszek and N. Hairs, ASLE Trans. 15, 184 (1972).
24. I. M. Veiga, A. C. Fernandes and B. S. Alimeda, J. Mater. Sci. Lett. 12, 1206 (1993).
25. A. J. Groszek, S. Partyka and D. Cot, Carbon 29, 821 (1991).
26. A. J. Groszek, in: Physical Chemistry of Colloids and Interfaces in Oil Production,
H. Toulhoat and J. Lecourtier (Eds.). Editions Technip, Paris (1992).
27. A. J. Groszek, Carbon 25, 717 (1987).
28. F. M. Fowkes, Y. C. Haung, B. Shah, M. J. Kulp and T. B. Lloyd, Colloids Surf. 29, 243
(1988).
29. F. M. Fowkes, D. W. Dwight, D. A. Cole, and T. C. Huang, J. Non-Cryst. Solids 120, 47
(1990).
30. F. M. Fowkes, K. L. Jones, G. Li and T. B. Lloyd, Energy Fuels 3, 97 (1989).
31. S. T. Joslin and F. M. Fowkes, I&EC Prod. Res. Dev. 24, 369 (1985).
32. G. W. Woodbury and L. A. Noll, Colloids Surf. 28, 233 (1987).
33. M. F. Finlayson and B. A. Shah, J. Adhesion Sci. Technol. 4, 431 (1990).
34. R. S. Farinato, S. S. Kaminski and J. L. Courter, J. Adhesion Sci. Technol. 4, 633 (1990).
35. T. L. Barr, Modern ESCA: The Practice and Principles of X-ray Photoelectron
Spectroscopy. CRC Press, Boca Raton, FL (1994).
36. J. Frommer, Angew. Chem. Eng. Ed. 31, 1298 (1992).
37. E. zur-Muelen and H. Nieus, in: Particle and Surface Characterization Methods,
R. H. Mueller, W. Mehnert and G. E. Hildebrand (Eds.), p. 99. Medpharm Scientific
Publishers, Stuttgart, Germany (1997).
38. J. Drelich and K. L. Mittal (Eds.), Atomic Force Microscopy in Adhesion Studies. VSP/
Brill, Leiden (2005).
39. T. H. Ibrahim, T. R. Burk, F. M. Etzler and R. D. Neuman, J. Adhesion Sci. Technol. 14,
1225 (2000).
40. G. A. Willing, T. R. Burk, F. M. Etzler and R. D. Neuman, Colloids Surf. A 193, 117 (2001).
41. G. A. Willing, T. H. Ibrahim, F. M. Etzler and R. D. Neuman, J. Colloid Interface Sci.
226, 185 (2000).
42. R. Nessler, Scanning 21, 137 (1999).
43. S. N. Magnov and D. Reneker, Ann. Rev. Mater. Sci. 27, 175 (1997).
44. S. N. Magnov and M.-H. Whangbo, Surface Analysis with STM and AFM. VCH,
Weinheim, Germany (1996).
45. M. Radmacher, M. Fritz, J. D. Cleveland, D. A. Walters and P. K. Hansma, Langmuir 10,
3809 (1994).
46. G. B. Binning, C. F. Quate and C. Gerber, Phys. Rev. Lett. 12, 930 (1986).
47. P. Lemoine, R. W. Lamberton and A. A. Ogwu, J. App. Phys. 86, 6564 (1999).
48. E. R. Beach, G. W. Tormoen and J. Drelich, J. Adhesion Sci. Technol. 16, 845 (2002).
49. A. J. Hickey, Inhalation Aerosols: Physical and Biological Basis for Therapy. Marcel
Dekker, Inc., New York (1996).
50. P. Wright, in: Pharmaceutical Preformulation and Formulation: A Practical Guide
from Candidate Drug Selection to Commercial Dosage Form, Vol. 1, p. 355. Taylor and
Francis Group, CRC Press LLC Boca Raton, FL (2004).
51. M. B. Chougule, B. K. Padhi, K. A. Jinturkar, and A. Misra, Recent Pat. Drug Deliv.
Formul. 1, 11 (2007).
52. H. P. Dunbar, Kona 16, 7 (1998).
53. R. C. Rowe, P. J. Sheskey and S. C. Owen, in: Handbook of Pharmaceutical Excipients,
5th ed., p. 385. Pharmaceutical Press and American Pharmacists Association,
Washington, DC (2006).
54. A. J. Hickey and D. Ganderton, Pharmaceutical Process Engineering. Marcel Dekker,
New York (2001).
55. F. M. Etzler, in: Solid State Characterization of Pharmaceuticals, A. Zakrzewski and
M. Zakrzewski (Eds.), p. 493. ASSA, Inc., Danbury, CT (2006).
56. D. Prime, P. J. Atkins, A. Slater and B. Sumby, Adv. Drug Deliv. Rev. 26, 51 (1997).
57. A. J. Hickey, H. M. Mansour, M. J. Telko, Z. Xu, H. D. Smyth, T. Mulder, R. McLean,
J. Langridge and D. Papadopoulo, J. Pharm. Sci. 96, 1302 (2007).
58. M. D. Louey, M. Van Oort and A. J. Hickey, J. Aerosol. Sci. 37, 1520 (2006).
59. E. Doelker, in: Powder Technology and Pharmaceutical Processes, D. Chulia, M. Deleul
and Y. Pourcelot (Eds.), p. 403. Elsevier, New York (1994).
60. E. Doelker, D. Mordier and S. Kopp, in: Powder and Grains, J. Biarez and R. Gourves
(Eds.), p. 411. Balkema, Rotterdam (1989).
61. J. T. Fell and J. M. Newton, J. Pharm. Sci. 59, 688 (1970).
62. A. T. Procipio, A. Zavaliangos and J. C. Cunningham, J. Mater. Sci. 38, 3629 (2003).
63. E. Ryshkewitch, J. Am. Ceram. Soc. 36, 65 (1953).
64. W. Duckworth, J. Am. Ceram. Soc. 36, 68 (1953).
65. J. E. Barraleta, L. Grovera, T. Gaunta, A. J. Wright and I. R. Gibsonc, Biomaterials 23,
3063 (2002).
66. F. W. Nyongesa and B. O. Aduda, AJST 5, 19 (2004).
67. C. K. Tye, C. Sun and G. E. Amidon, J. Pharm. Sci. 94, 465 (2005).
68. C. A. Andersson, J. Amer. Ceram. Soc. 79, 2181 (1996).
69. F. P. Knudsen, J. Amer. Ceram. Soc. 42, 376 (1959).
70. C.-Y. Wu, S. M. Best, A. C. Bentham, B. C. Hancock, and W. Bonfield, Eur. J. Pharm.
Sci. 25, 331 (2005).
71. C. Sun, Amer. Pharm. Rev. 8, 105 (2005).
72. K. Johanson, in: Encapsulated and Powdered Foods, C. Onwulata (Ed.), p. 331. Taylor
and Francis, New York (2005).
73. F. M. Etzler, T. Bramante, R. Deanne, S. Sienkiewicz and F. J. Chen, J. Adhes. Sci.
Technol. 25, 501 (2011).
74. F. Fichtner, D. Mahlin, K. Welch, S. Gaisford and G. Alderborn, Pharm. Res. 25, 2750
(2008).
75. F. M. Etzler, R. Deanne, T. Burk et al., in: Contact Angle, Wettability and Adhesion,
Vol. 2, K. L. Mittal (Ed.). VSP, Utrecht, the Netherlands (2002).
76. F. M. Etzler, T. H. Ibrahim, and R. D. Neuman, in: Particles on surfaces, Vol. 7, K. L.
Mittal (Ed.). VSP, Utrectht, the Netherlands (2002).
77. L. J. Jallo, J. Bowen, F. Etzler and R. Dave, J. Adhesion Sci. Technol. 25, 367 (2011).
4 The Effects of
Nanostructure and
Composition of Solid
Surfaces on Ice Adhesion
Hanna Dodiuk*, S. Kenig, and A. Dotan
CONTENTS
4.1 Introduction..................................................................................................... 68
4.2 Surface–Liquid Thermodynamics................................................................... 69
4.3 Surface Modeling............................................................................................ 71
4.3.1 Smooth Surfaces.................................................................................. 71
4.3.2 Rough Surfaces.................................................................................... 73
4.4 Experimental................................................................................................... 75
4.5 Results and Discussion.................................................................................... 77
4.6 Conclusions...................................................................................................... 81
References................................................................................................................. 82
Ice adhesion to surfaces is detrimental to electrical and communication cables

and to air and ground transportation vehicles. Hence, anti-icing materials and
mechanical and heating means have been used for de-adhesion of ice to various
substrates. Alternatively, icephobic surface treatments have been developed
to reduce the adhesion of ice to solid surfaces. However, the understanding
of ice anti-adhesion phenomenon is lacking with respect to the effects of sur-
face roughness and chemical composition. Consequently, the effects of rough-
ness on the nanoscale and microscale as well as of the chemical composition
of surface treatments on ice repellency were investigated. In the first part of
the study, the thermodynamics of water wettability of coated surfaces was
theoretically evaluated. Accordingly, relationships were derived between the
interfacial adhesion strength of a liquid drop to a polymer surface of a given
composition, the mass of the drop, the measured contact angles, and the slid-
ing angle. To verify the proposed model, various hydrophobic coatings having
* Prof. Hanna Dodiuk was the lead author for this chapter and can be contacted at Hannad@shenkar.ac.il.
67
different surface nano- and microroughnesses were prepared using polycar-

bonate as a substrate.
The surface treatment was based on fluoroalkylsilane. Roughness was
introduced by means of silica (microparticle) and polyhedral oligomeric
silsesquioxane (POSS; hybrid inorganic–organic nanoparticle) using a double
morphology topography. In the case of optimal compositions and processing
conditions, transparent ultrahydrophobic (superhydrophobic) characteristics
were obtained. In these cases, the water contact angle was above 160° and the
sliding angle was below 5°. In the second part of the investigation, the relation-
ship between water adhesion and ice repellency represented by ice adhesion
to the coated surfaces was studied. Consequently, the surfaces were tailored
by nanotechnology approach for their chemistry as well as their nanostructure
in order to obtain fully wetting (hydrophilic) having a contact angle close to
zero and the fully dewetting (ultrahydrophobic; self-cleaning surfaces) where
the contact angle approaches 180°. Ice adhesion test results have shown that
the lower the surface energy, the lower the ice adhesion. The lowest adhesion
values were obtained where ultrahydrophobicity was present, a combination of
very low sliding angle with very high contact angle.
These novel ultrahydrophobic surface coatings and the associated model
were used to develop and design surfaces for a variety of applications such as
self-cleaning, glazing, and anti-icing surfaces.
4.1 INTRODUCTION
Ice accretion and its adhesion to structure surfaces are detrimental to telecommunica-
tions cables, electrical wires, and air and ground transportation vehicles. Specifically,
the aviation industry has been facing difficulties related to loss in radar sensitivity
and aerodynamic control when aircraft fly through supercooled liquid, that is, cloud,
rain, or drizzle [1]. The severity of icing depends on the supercooled water content, its
temperature, and the size of the droplets. As the rain strikes a surface at subzero tem-
perature, it freezes, and glazed ice accumulates at a rate reaching 1 cm per 4 miles.
Ice accretion on the aircraft control surfaces not equipped with anti-icing means can
cause problems such as weight increase and change in weight distribution, which
result in loss of control. Additionally, visibility may be impaired as ice adheres to the
windshield, and thrust may be reduced due to icing of the engine blades [2]. There are
two recognized sources of ice accretion. The first is in-cloud icing, in which super-
cooled water droplets float in the air and contact a surface due to air movement. The
second is precipitational icing, where the droplets are massive enough to fall from the
atmosphere onto the structure. A variety of approaches are used to prevent ice accre-
tion, among them, anti-icing methods to prevent accretion and deicing to remove ice
once it has formed. Deicing methods require mechanical or thermal energy or the use
of liquids designed to reduce the freezing point of water [3]. The ideal solution to icing
is to devise a material that will reduce the adhesion of ice to aircraft surfaces. The
so-called “icephobic” material should be solid, durable, easy to apply, and inexpen-
sive [1]. A variety of low-wettability oils, greases, and permanent coatings have been
evaluated and applied, due to their hydrophobicity, to reduce ice adhesion.
Effects of Nanostructure and Solid Surfaces on Ice Adhesion 69
At a molecular scale, three phenomena contribute to the adhesion of ice: electro-

static attraction, covalent bonding, and the van der Waals forces. The electrostatic
force is considered to be significant since water and ice are polar materials and the
alignment of their molecules is influenced by an electric field. On freezing, water
molecules reorganize from a liquid to a crystalline structure that remains polar.
Thus, ice adheres to the metal surface for the same reason that water does [4,5].
Ice adhesion has been evaluated for many solid materials showing icephobic
properties [6]. The common hypothesis is that any material having poor chemical
affinity with water should also have weak ice adhesion. However, this hypothesis
is not unanimously accepted among researchers. Another factor studied is the sur-
face finish and morphology [1,7]. Despite the considerable number of studies related
to icephobic materials, the understanding regarding antiadhesion properties is still
lacking, at times even controversial. Consequently, no material has yet been identi-
fied as efficient to assure protection against ice accumulation [8]. Under freezing
conditions at –10°C, the icephobic materials studied were found to be inefficient at
preventing ice accumulation and adhesion [6,7]. In fact, under this condition, water
droplet nucleation on the surface is so fast that droplets freeze before sliding on
the inclined hydrophobic surfaces. Furthermore, it was identified that ice adhesion
increases with surface roughness due to mechanical anchoring of the supercooled
water drops. As a result, the greater the number of pores, the greater is the possi-
bility of anchoring, hence an increase in adhesion. Polishing of solid surfaces was
shown to reduce adhesion but was not sufficient to eliminate it [1]. However, it was
shown that the greater the surface roughness of hydrophilic materials, the better the
adhesion. In contrast, the larger the surface roughness of hydrophobic coatings, the
poorer the adhesion. This difference is attributed to the difference in water penetra-
tion in the cases of hydrophilic and hydrophobic surfaces [9–11].
4.2 SURFACE–LIQUID THERMODYNAMICS

The thermodynamics between a solid and a liquid was described by Young [12]
in 1805 for smooth surfaces. Generally, when the contact angle of water is less
than 90°, the surface is called hydrophilic; when the contact angle is greater than
90°, the surface is hydrophobic. A surface having a water contact angle greater
than 150° is usually classified as ultra-hydrophobic, that is, a water-repellent sur-
face. Wenzel [11,13] recognized the importance of surface roughness and proposed
a modification to Young’s equation, which included a roughness factor defined
as the ratio between the actual rough surface area and the geometric projected
area. Young and Wenzel considered chemically homogeneous surfaces. Cassie and
Baxter [14,15] extended Wenzel’s work to non-homogeneous surfaces. Cassie’s
equation (Equation 4.1) relates the contact angle θ′ for a chemically heterogeneous
surface composed of a fraction f1 of chemical type 1 and a fraction f 2 of chemi-
cal type 2, where θ1 and θ2 are the contact angles measured on type 1 and type 2
surfaces, respectively.
cos θ′ = f1 cos θ1 + f 2 cos θ2 (4.1)

f1 + f 2 = 1 (4.2)
Cassie and Baxter extended their model to porous and rough surfaces. In this
case, f1 is the solid–liquid interface fraction, and f 2 the air–liquid interface fraction.
For the air–liquid interface, θ2 = 180°; then by substituting θ2 for 180° in Equation
4.1, Equation 4.3 is obtained for porous surfaces, where θ′ is the contact angle for the
porous surface, f is the solid fraction of the porous surface, and θ is the contact angle
for the smooth solid surface.
cos θ′ = f cos θ + f − 1 (4.3)
On a hydrophobic rough surface, the liquid repellency impedes the liquid

from fully penetrating into the depressions of the roughness morphology. Using
Equation 4.3, Equation 4.4 can be derived, where θ′ is the contact angle for the
rough surface, fC (according to Cassie) is the contact area fraction between the
liquid and the rough hydrophobic surface, and θ is the contact angle for the smooth
solid surface [16].
cos θ′ + 1
f = fC = (4.4)
cos θ + 1
Nanostructured surfaces characterized by nanoscale roughness require special

attention. The work by Israelachvili and Gee [17] deals with chemical heterogene-
ities of atomic or molecular scale. Based on theories on intermolecular and surface
forces, they derived the following equation:
(1 + cos θ′)2 = f1 (1 + cos θ1)2 + f 2 (1 + cos θ2)2 (4.5)
In order to follow the same line of reasoning used to obtain Equation 4.3, it
is necessary to bring in a quantity analogous to the fC used by Cassie and Baxter
but applicable to the nanoscale. Rios et al. [18,19] suggested to define f R, the
contact fraction between the liquid and the nanorough hydrophobic surface, and
introduced a novel relationship, as defined in Equation 4.6, where θ′ is the contact
angle for the nanorough surface; f R, according to Rios et al. [18,19], is the contact
fraction between the liquid; and θ is the contact angle for the original smooth
solid surface.
2
 cos θ′ + 1 
f = fR =  (4.6)
 cos θ + 1 
It could be concluded that for a hydrophobic surface, a high degree of roughness

is needed to achieve ultra-hydrophobicity. For instance, to increase the contact angle
from 120° to 150°, fC = 0.27, while f R = 0.07; that is, the liquid drop makes contact
with only 7% of the solid surface.
Nishino et al. [20] demonstrated that the highest water contact angle could
be obtained for a regularly aligned and close-packed surface composed of CF 3
groups. Accordingly, the highest theoretical water contact angle on a flat and
smooth solid surface is 120°. Further increase of the contact angle can be obtained
only by varying the surface roughness. It is commonly accepted that ultra-hydro-
phobicity can be obtained by a proper combination of surface chemistry and sur-
face roughness.
From a practical point of view, the contact angle is not the only significant param-
eter for defining hydrophobicity. For ultra-hydrophobic surfaces, self-cleaning
is of importance. In this application, a low level of water drop adhesion is of sig-
nificance. The adhesion of a water drop to a surface can be characterized by the
critical tilting angle on the surface at which a liquid drop, with a certain weight,
begins to slide down the tilted plane. Consequently, a more comprehensive defini-
tion of an ultra-hydrophobic surface should specify the highest possible contact
angle (as close to 180° as possible) and the lowest possible sliding angle (as close to
0° as possible).
4.3 SURFACE MODELING

Different models have been proposed to correlate contact angles, sliding angles, and
interaction energies between a liquid and a smooth or rough surface. The existing
models and the newly proposed one are discussed and compared in light of well-
designed experimental results.
4.3.1 Smooth Surfaces
The radius R of a liquid drop on a smooth surface can be calculated from its density,
ρ, mass, m, and the contact angle of the liquid with the solid, θ. Assuming that the
drop is a perfect sphere, the radius of the drop can be calculated [18,19] according
to Equation 4.7:
1
 3m 3
R=  (4.7)
 ρπ(2 − 3 cos θ + cos θ) 
3
The radius, r, of the contact area between the drop and the solid is given by
Equation 4.8:
r = R sin θ (4.8)
When the horizontal smooth plane is tilted to an angle α, the contact area is
assumed to remain circular with the radius, r, even if the drop is deformed by tiling.
At the onset of drop motion, the forces acting on the drop will be at equilibrium
according to Equation 4.9:
FA = mg sin α (4.9)
Model A: If the adhesion of the liquid drop to the solid is assumed to be the result
of the forces acting at the contact periphery between the drop and the solid (as in
Young’s equation), then
FA = K A 2πr, (4.10)
where K A is a constant with units of surface tension (N/m) or energy (J/m2) and is a
measure of the adhesion energy between the two phases.
From Equations 4.9 and 4.10, the following relationship is obtained:
K A 2πr
sin α = (4.11)
mg
Substituting r from Equations 4.7 and 4.8 into in Equation 4.11 yields the follow-
ing expression:
1
2
K 2π  3 3 −
sin α = A  sin θm 3
(4.12)
g  ρπ(2 − 3 cos θ + cos3 θ) 
According to Equation 4.12, for a given liquid and solid surface (ρ and θ con-
stant), the sine of the sliding angle α depends on the mass of the drop to the –2/3
power. K A represents the interaction energy (constant for a given surface chemistry).
Model B: Model A is based on the assumption that the adhesion of the liquid drop
to the solid surface is the result of the forces developed at the periphery of the contact
area between the liquid drop and the solid. However, when evaluating the sliding
angle, a different approach should be taken. This approach should include the adhe-
sion phenomenon of the liquid drop to the solid, resulting from the intermolecular
forces developed in the interfacial contact area. The interfacial forces attract the
liquid molecules to the solid molecules also in the contact area of the drop with the
solid and not only along the periphery. Consequently, the proposed model hypoth-
esizes that the force of adhesion is proportional to the contact area between the drop
and the solid:
FA = KB πr 2, (4.13)
where KB is a constant with units of force/area (N/m2 = Pa) and is a measure of the
adhesion strength between the liquid and the solid. Combining Equations 4.9 and
4.13, Equation 4.14 is obtained:
K πr 2
sin α = B (4.14)
mg
Substituting r in Equation 4.14 using Equations 4.7 and 4.8 yields the following
expression:
2
K π 3  3 2 − 13
sin α = B  sin θm (4.15)
g  ρπ(2 − 3 cos θ + cos3 θ) 
Hence, for a given liquid and solid surface (i.e., ρ and θ constant) sin α depends
on the mass of the drop to the –1/3 power.
4.3.2 Rough Surfaces
On a hydrophobic rough surface, the liquid is prevented from penetrating the rough-
ness depressions. Assuming an idealized rough surface, the representation of the
contact angle changes from θ for a smooth surface to θ′ for a rough surface. As a
result, r and R also change to r′ and R′, respectively. Both Equations 4.7 and 4.8 hold
and, for a rough surface, become
r′ = R′sin θ′ (4.16)
Furthermore, provided that f is the interface contact fraction between the liquid
and the solid, the effective interface contact perimeter will be 2πr′f, and the effective
interface contact area will be πr′2f. Based on these principles, the following models
are suggested.
Model A′: The interaction energy constant K A depends on the chemistry of the
solid surface. However, the introduction of roughness may change the surface chem-
istry by changing the molecular configuration at the surface. Therefore, we assume
that K A for a smooth surface becomes K ′A for a rough surface. Thus, FA becomes FA′ :
FA′ = K ′A 2πr ′f (4.17)
and
K ′A 2πr ′f
sin α′ = (4.18)
mg
Substituting r′ in Equation 4.18 from Equations 4.16 and 4.17,
1
2
K ′ 2πf  3 3 −
sin α′ = A   sin θ′ m 3 (4.19)
 ρπ(2 − 3 cos θ′ + cos θ′) 

3
g
Model B′: For a rough surface,
FA′ = K B′ πr ′ 2 f (4.20)
Thus,
K B′ f πr ′ 2
sin α′ = (4.21)
mg
Substituting r′ in Equation 4.21 from Equations 4.16 and 4.17,
2
1
K ′ πf  3 3 2 −
sin α′ = B   sin θ′ m 3
(4.22)
g  ρπ ( 2 − 3 cos θ′ + cos 3
θ′ ) 
It can be concluded that the rougher the surface becomes, the lower is the sliding
angle. Furthermore, in the case of hydrophobic surfaces, where the liquid does not
penetrate the roughness depressions, small-scale roughness should lead to a reduc-
tion in the actual contact area between the drop and the solid surface (as with a
fakir’s bed of nails) and to a reduction of the adhesion strength. In addition, the slid-
ing angle, which depends strongly on the contact angle, falls sharply as the contact
angle increases. Figure 4.1 depicts the sliding angle for a 5 μL water drop on a rough
surface as a function of the contact angle for models A′ and B′. It is assumed that the
“smooth” contact angle is 120°, and the 5 μL drop will start rolling off a totally verti-
cal surface (α = 90°). As can be seen, for a “rough” contact angle of more than 155°,
the predicted sliding angles are less than 5°. Thus, when the contact angle reaches
155° the surface will become increasingly self-cleaning. At a contact angle of 165°,
the sliding angle approaches 0°. Among the calculated angles, α′CR is the fastest-
decreasing sliding angle, reaching less than 5° at a contact angle slightly over 140°.
90
75
Sliding angle (deg)
60
45
30
15
0
120 130 140 150 160 170 180
Contact angle (deg)
α´ (A´B) α´ (A´R) α´ (B´B) α´ (B´R)
FIGURE 4.1 Sliding angle on a rough surface as a function of contact angle for a 5 μL water
drop (smooth contact angle is 120°): for Model A′, according to fC (A′B) or f R (A′R), and for
Model B′, according to fC (B′B) or f R (B′R). (From Rios, P.F. et al., J. Adhes. Sci. Technol.,
20, 563, 2006.)
According to the proposed model, the lowest sliding angle will be reached when
nanoscale roughness is achieved.
4.4 EXPERIMENTAL
In the previous sections, the thermodynamics of liquids in contact with surfaces and
especially the effect of liquid adhesion to surfaces, which are governed by the sur-
face chemistry and roughness, were reviewed and developed. It was hypothesized that
water adhesion to surfaces, at room temperature, is the decisive phenomenon that
affects the adhesion of ice to surfaces. Consequently, the experimental study com-
prised two parts. In the first part, a variety of methods were developed to synthesize
hydrophobic surfaces having various chemical compositions and roughnesses, on
smooth and clear polycarbonate (PC) specimens. These surfaces were characterized
with respect to their adhesion to room-temperature water by means of the resultant
contact angles and sliding angles. In the second part, a variety of PC coated samples
were prepared using the best hydrophobic surface treatment developed in the first
part, in addition to commercially available hydrophobic and hydrophilic coatings.
These samples were used to investigate the adhesion of ice to the various coatings.
In the first part, PC samples were coated with either one or two layers. In all
cases, the outer layer contained hydrophobically functionalized polyhedral oligo-
meric silsesquioxane (POSS) compounds. POSS is an intermediate between silica
(SiO2) and silicone (R2SiO) and has a cage-like structure of 1.5 nm in size. Thus,
POSS can be defined as an intrinsically nanostructured organic–inorganic sub-
stance. Unlike silica and modified clays, each POSS molecule contains covalently
bonded functionalities. POSS functionalities include alcohols and phenols, amines,
halides, acrylates and methacrylates, epoxides, esters, nitriles, olefins, phosphines,
thiols, and fluoroalkyls. To provide both nanoroughness and hydrophobic sur-
face chemistry, fluoro-functionalized POSS (FPOSS) was used. Previous authors
[11,13,21] have found that both the length and flexibility of the fluoroalkyl molecule
have an effect on the surface energy decrease (and, therefore, contact angle increase)
due to enhanced segregation of the CF3 chain ends on the surface. Hence, POSS
molecules with long pendant fluoroalkyl chains were chosen: trifluoro cyclopentyl
POSS, C50H93F39O12Si10 (FL0590, Hybrid Plastics, Hattiesburg, MS, U.S.A.; here
after FPOSS1) and fluoro disilanol isobutyl POSS, C38H75F13O12Si8 (FL0569, Hybrid
Plastics, Hattiesburg, MS, U.S.A.; hereafter FPOSS2).
The two-layer structure was adopted with the objective of mimicking the lotus
effect. The inner layer of micrometric roughness was produced using hydrophilic
fumed silica (Aerosil 200, Evonik Industries, Frankfurt, Germany), and the outer
layer of nanometric roughness was produced using hydrophobic FPOSS.
In a second configuration, the silica and the FPOSS were mixed together in the
same liquid and coated as a single layer onto the PC substrate. Thus, it was pos-
sible to compare results of the same composition having one-layer or two-layer
configurations.
The silica and FPOSS were mixed in various ratios in appropriate solvents accord-
ing to the material type. The PC samples were first cleaned with isopropyl alcohol
(IPA) then dipped in the corresponding mixture, dried, cooled, and conditioned in
controlled laboratory conditions (25°C, 60% Relative Humidity) for 24 h. In the case
of two layers, this procedure was adopted for both the first and second layers.
In the second part for the study of ice adhesion [22] to different surfaces hav-
ing high (hydrophilic) and low (hydrophobic) energies, the following samples were
prepared:
• NEAT—Neat PC (PALRAM, Israel)

• UHFL—Ultra-hydrophilic coating (commercial antifog coating)
• ECHAS—Hydrophobic coating (commercial antiscratch coating based on
interpenetrating network of epoxy silicone)
• HFBDS—Hydrophobic coating (based on fluorosilane)
• UHFB—Ultra-hydrophobic coating (based on silica + FPOSS2—two-layer
morphology)
The contact angles and sliding angles were measured with a Dataphysics OCA
20 instrument using 5 μL water drops for the contact angle and 30 μL for the sliding
angle. Surface morphology was analyzed from the tapping mode topography using
an atomic force microscope (Auto Probe Research). A silicon probe with a spring
constant of 40 N/m and resonance frequency of 300 kHz was used.
Ice accretion was measured (Anti-Icing Laboratory [AMIL], Chicoutimi, Canada).
The icephobic character of the surface coating was evaluated by measuring the adhe-
sion of ice compared to an uncoated aluminum surface (reduction factor). The sam-
ples were iced with polycrystalline ice having a thickness of 0.7 cm that covered an
area of 11 cm2. The icing was carried out in cloud icing conditions and performed
in a refrigerated wind tunnel. A Centrifuge Adhesion Test was used to characterize
the iced samples. The test arrangement consisted of three small aluminum beams
covered with the samples, compared with three bare aluminum beams. Each bal-
anced sample beam was rotated in a centrifuge. The rotation speed increased with a
constant acceleration rate until the centrifugal force resulting from rotation reached
the adhesion strength of the ice, indicated by its detachment from the sample.
From the centrifuge test, the force was determined according to Equation 4.23:
F = mrω2, (4.23)
where F = centrifugal force (N), m = mass of ice (kg), r = radius of the beam (m), and
ω = speed of rotation (rad/s).
The shear stress τ was calculated according to Equation 4.24:
τ = F/A, (4.24)
where A is the ice-covered area.

The adhesion reduction factor (ARF) was calculated using the average adhesion
shear stress measured from the three coated beams normalized by the average stress
measured from three untreated bare beams, as expressed in Equation 4.25.
ARF = τbare/τcoated (4.25)

4.5 RESULTS AND DISCUSSION

To meet the objective of the current investigation, a variety of process parameters
and materials were studied. Table 4.1 shows [23] the optical and surface properties
for selected samples. Sample 1 is the uncoated PC reference sample. Sample 2 is a
homogeneous single layer produced by coating the PC substrate with a solution of
3 wt% FPOSS1 in α,α,α-trifluorotoluene (TFT). The PC substrate was dipped for
10 s and dried at room temperature. Samples 3, 4, and 5 were produced by double-
coating the PC substrate first with a layer of silica in IPA and then with a second layer
of FPOSS2 also in IPA. The PC substrate was dipped in the silica mixture for 1 min,
dried in a drying oven for 1 h at 120°C, and left for conditioning at room tempera-
ture for 24 h. The procedure was repeated for the FPOSS2 mixture. Sample 6 is a
heterogeneous single layer produced by mixing both silica and FPOSS2 together in
IPA and applying them as a single layer onto the PC substrate. Coating component
concentrations of each sample are shown in Table 4.1.
Different types of one-layer approaches were attempted. Best results were obtained
for sample 2 (3 wt% FPOSS1 in TFT). This sample showed good transparency, an
increased contact angle, and a reduced sliding angle compared to the untreated PC.
However, the improvement was limited, and the sample did not exhibit ultra-hydro-
phobic characteristics. Sample 4 showed the optimal balance of these two properties.
As could be concluded from Table 4.1, ultra-hydrophobic surfaces with potential
self-cleaning applications were obtained when silica and FPOSS2 were applied in a
two-layer configuration using IPA as solvent (Samples 3 and 4). It was noticed that
silica and FPOSS in high concentrations tended to agglomerate, reducing transpar-
ency. Consequently, for transparent surfaces, the concentrations of both components
were reduced to the minimum requirement (Sample 4). Further reduction of concen-
tration decreased the ultra-hydrophobic effect. Sample 4 (0.5% silica inner layer and
TABLE 4.1
Light Transmission (LT), Haze, Contact Angle (θ), and Sliding Angle (α) for
Different Samples
θ (5 µL) α (5 µL)
Sample No. Coating LT (%) Haze (%) (degree) (degree)
1 Uncoated PC 92.6 0.45 81.3 67
2 3 wt% FPOSS1 in TFT (one-layer) 95.1 5.9 110.0 40
3 1 wt% silica + 3 wt% FPOSS2 in 91.4 30.4 >165 <1
IPA (two-layer)
4 0.5 wt% silica + 1.5 wt% FPOSS2 90.0 5.0 >165 <1
in IPA (two-layer)
5 0.25 wt% silica + 0.75 wt% 93.0 4.5 135 >90
FPOSS2 in IPA (two-layer)
6 0.5 wt% silica + 1.5 wt% FPOSS2 92.0 7.5 130.5 64
in IPA (mixed one-layer)
Source: Rios, P.F. et al., J. Adhes. Sci. Technol., 21, 399–408, 2007.
1.5% FPOSS outer layer) was found to present the optimal balance of properties. At
higher concentrations, ultra-hydrophobicity was obtained at the expense of transpar-
ency (Sample 3). In these ultra-hydrophobic samples, the contact angle is very high,
and the sliding angle, very low. It is extremely difficult to measure the contact angle
since water drops are hardly stationary and slide from the horizontal solid surface at
the slightest disturbance. Water drops will slide before the tilting unit reaches 1° and
the measured contact angle is higher than 165°. When the particle concentration was
decreased, the sample lost its ultra-hydrophobicity (Sample 5). Sample 5, while still
hydrophobic, shows a significantly lower contact angle, and a 5 μL water drop will
not slide from its surface even when the surface is completely vertical (α = 90°). For
comparison, Table 4.1 also describes the results for a single-layer treatment having
the same optimal concentration as the ultra-hydrophobic sample 4 (Sample 6). As
shown, the heterogeneous one-layer coating is hydrophobic but not ultra-hydrophobic,
and a significant but higher sliding angle is observed (64°). It was concluded that the
two-layer lotus-like morphology configuration was the optimal one.
Figure 4.2 shows the contact angles of 5 μL sessile water drops on (1) sample 1
(uncoated PC), (2) sample 6 (0.5 wt% silica + 1.5 wt% FPOSS2 heterogeneous one-
layer), and (3) sample 4 (0.5 wt% silica + 1.5 wt% FPOSS2 as two layers).
Figure 4.3 shows Atomic Force Microscope (AFM) topographic images of
sample 1 (uncoated PC), sample 2 (homogeneous single layer of 3 wt% FPOSS1
in TFT), and sample 4 (0.5 wt% silica + 1.5 wt% FPOSS2 as two layers). Figure
4.4 shows an AFM image of a two-layer coating showing a very similar surface
roughness compared to the lotus leaf. The R rms roughness calculated for uncoated
PC is 1.5 nm, indicating a high degree of smoothness of the starting PC substrate.
After coating the PC surface with a single layer of FPOSS1, the R rms roughness
increased to 14.5 nm, and for the two-layer coating, the R rms roughness increased
to 60 nm. The R rms roughness values of these three samples correlate to the mea-
sured values of light transmission and haze. Since these values are smaller than
the light wavelength (400 to 700 nm), the surface roughness does not disturb
significantly the passage of the light, and the samples remain reasonably clear.
Sample 2 is hydrophobic due to the fluorinated functionalities of the POSS, and
its surface is rough in the nanometric range. However, this degree of roughness
is not enough, and ultra-hydrophobicity is not achieved. Sample 4 has a similar
(a) (b) (c)
FIGURE 4.2 Sessile drops for contact angle measurements on (a) uncoated PC and PC
coated with (b) 0.5 wt% silica + 1.5 wt% FPOSS (mixed) sample or (c) 0.5 wt% silica + 1.5
wt% FPOSS sample (two-layer coating). (From Rios, P.F. et al., J Adhes. Sci. Technol., 21,
5–6, 399–408, 2007.)
(a)
µm/div
0.064
0
2.0 µm/div
2.0 µm/div
(b)
µm/div
0.076
2.0 µm/div
2.0 µm/div
(c)
µm/div
0.068
2.0 µm/div
2.0 µm/div
FIGURE 4.3 AFM topography images of (a) uncoated PC and PC coated with (b) single
layer of 3 wt% FPOSS in TFT or (c) 0.5 wt% silica + 1.5 wt% FPOSS (two-layer coating).
(From Rios, P.F. et al., J. Adhes. Sci. Technol., 21, 5–6, 399–408, 2007.)
surface chemistry, but the higher level of nanometric roughness gave rise to ultra-
hydrophobicity. Moreover, because of the fakir bed–like morphology of this
sample, as seen in the AFM image, water should not penetrate significantly the
surface valleys, and hence, the effective contact area will comprise only the nano
roughness crests, leading to a very low contact area fraction f R; thus, the sliding
angle approaches 0° (Equation 4.3).
In the second stage, the relationship between water wetting and ice adhesion was
studied. Except for the laboratory-developed ultra-hydrophobic coating (sample 4),
commercial coatings were used. The coatings were characterized as hydrophilic, in
the case of UHFL coating, with a contact angle of 20° and sliding angle greater than
443.26 nm
221.63 nm
0 nm
10 µm
10 µm 5 µm
5 µm
0 µm 0 µm
FIGURE 4.4 AFM topography image of two-layer coating. (From Dotan, A., J. Adhes. Sci.
Technol., 23, 1907–1915, 2009.)
90°. Distinctively, the ultra-hydrophobic coating UHFB (sample 4) was achieved

when the contact angle obtained was approximately 160°. In this case, the sliding
angle was close to 0°. The microstructure morphology of the ultra-hydrophobic coat-
ing can be seen in Figure 4.3c. The AFM topography image shows the nanosize and
microsize surface roughness, obtained by the use of nanoparticles and micropar-
ticles, resembling the lotus leaf morphology. The contact angle, sliding angle, and ice
adhesion results can be seen in Table 4.2.
The ice reduction factor (compared with aluminum) was calculated for in-cloud
icing. Figure 4.5 indicates that a very good correlation exists between the contact
angle and the ice reduction ability for the different surface coatings.
As is evident from the above results, the more effective coating is the ultra-
hydrophobic one, leading to an ice reduction factor of 18. The less effective coating
was the hydrophilic one.
TABLE 4.2
Ice Adhesion Compared to Contact and Sliding Angles
Centrifuge Ice Adhesion Test
In-Cloud Icing –15°C
Contact Sliding
Coating Angle (°) Angle (°) Shear Stress (MPa) ± 20% Adhesion Reduction Factor
NEAT 78 >90 0.129 4.4
UHFL 20 >90 0.177 1.4
ECHAS 105 20 0.141 4
HFBDS 107 >90 0.099 3.9
UHFB 160 0 0.074 18.3
Source: Dotan, A. et al., J. Adhes. Sci. Technol., 23, 1907–1915, 2009.

Centrifuge ice adhesion test -
Iced reduction factor (comp. tp Al)

wind tunnel ice
20
NEAT
15
UHFL
10 ECHAS
NFBDS
5
UHFB
0
0 50 100 150 200
Contact angle
FIGURE 4.5 Ice reduction factor measured in centrifuge ice adhesion test as a function of
the contact angle. (From Dotan, A., J. Adhes. Sci. Technol., 23, 1907–1915, 2009.)
4.6 CONCLUSIONS
It is widely accepted that ultra-hydrophobicity can be obtained by a proper com-
bination of surface chemistry and surface roughness. In this work, transparent PC
substrates were coated with either one or two layers. One-layer coatings were pro-
duced taking advantage of intrinsically nanostructured POSS particles functional-
ized with hydrophobic fluorine. These coatings exhibited good optical clarity and
improved hydrophobicity (increased contact angle and reduced sliding angle) com-
pared to the uncoated PC sample. However, ultra-hydrophobicity was not achieved
with these one-layer POSS coatings. Ultra-hydrophobic and transparent surfaces
with potential self-cleaning applications were achieved with a two-layer configu-
ration comprising a base layer of hydrophilic fumed silica and an upper layer of
hydrophobic POSS.
The self-cleaning surfaces were studied for their icephobic properties. The pres-
ent comprehensive study should shed light on the mechanism of ice adhesion. The
higher the contact angle, the lower the ice adhesion. In the case of ultra-hydrophilicity
(low contact angle), the adhesion is high. In the case of high contact angles (hydro-
phobicity), the adhesion of ice is weak. The lowest ice adhesion values were obtained
for ultra-hydrophobic surfaces. This unique surface energetics was obtained using a
combination of low surface energy and microroughness and nanoroughness. Though
in both extreme cases—hydrophobic and hydrophilic surfaces—the supercooled
water drops crystallized upon contact with the cold surfaces, the chemical as well as
the mechanical interactions that developed were different with respect to the type of
surface. A low level of interaction in the case of the hydrophobic surface led to low
adhesion of ice, and in contrast, high interaction levels led to high ice adhesion in the
case of hydrophilic surfaces.
Finally, it should be emphasized that the roughness level of the solid surface prob-
ably affects the crystallization and, especially, the nucleation of the ice. This is a
subject for future research.
REFERENCES
1. P. R. Louchez, J. L. Laforte, G. Bouchard and M. Farzaneh, “Laboratory evaluation
of aircraft ground del antiicing products,” International Journal of Offshore and Polar
Engineering, vol. 5, no. 1, pp. 479–484 (1995).
2. S. P. Otta and A. P. Rothmayer, “Instability of stagnation line icing,” Computers &
Fluids, vol. 38, no. 2, pp. 273–283 (2009).
3. C. J. Whitaker, Standard Handbook of Broadcast Engineering, McGraw Hill, U.S.A. (2005).
4. Air Force Aircraft and Airfield Deicing/Anti-icing—The Role of Deicing and Anti-icing
in the Air Force, Fact Sheet Update, May 1998, http://www.p2pays.org/ref/01/00593.pdf.
5. V. Petrenko, US Patent 6,027,075 (1998).
6. V. K. Croutch and R. A. Hartley, “Adhesion of ice to coatings and the performance of ice
release coating,” Journal of Coating Technology, vol. 64, no. 815, pp. 41–52 (1997).
7. L. O. Andersson and C. G. Golander, “Ice adhesion to rubber materials,” Journal of
Adhesion Science and Technology, vol. 8, no. 2, pp. 117–132 (1994).
8. C. Laforte and A. Beisswenger, “Icephobic material centrifuge adhesion test,” IWAIS
XI, June (2005).
9. H. Saito, K. Takai and G. Yamauchi, “Water and ice repellent coatings,” Surface Coatings
International, vol. 80, no. 4, pp. 168–171 (1997).
10. S. R. Coulson, I. Woodward and J. P. S. Badyal, “Super-repellent composite fluoropoly-
mer surfaces,” Journal of Physical Chemistry, Part B, vol. 104, pp. 8836–8840 (2000).
11. R. Wenzel, “Resistance of solid surfaces to wetting by water,” Industrial and Engineering
Chemistry, vol. 28, no. 8, p. 988 (1936).
12. T. Young, “An essay on the cohesion of fluids,” Philosophical Transactions of the Royal
Society, London, vol. 95, p. 65 (1805).
13. R. N. Wenzel, “Surface roughness and contact angle,” Journal of Physical Chemistry,
vol. 53, p. 1466 (1949).
14. A. B. D. Cassie and S. Baxter, “Wettability of porous surfaces,” Transactions of the
Faraday Society, vol. 40, p. 546 (1944).
15. A. B. D. Cassie, “Contact angle,” Discussions of the Faraday Society, vol. 3, pp. 11–16
(1948).
16. P. G. De Gennes, F. Brochard-Wyart and D. Quere, Capillarity and Wetting Phenomena.
Springer, New York, p. 222 (2002).
17. J. N. Israelachvili and M. L. Gee, “Contact angles on chemically heterogeneous sur-
faces,” Langmuir, vol. 5, p. 288 (1989).
18. P. F. Rios, H. Dodiuk, S. Kenig, S. McCarthy and A. Dotan, “The effects of nanostructure
and composition on the hydrophobic properties of solid surfaces,” Journal of Adhesion
Science and Technology, vol. 20, p. 563 (2006).
19. P. F. Rios, H. Dodiuk, S. Kenig, S. McCarthy and A. Dotan, “The effect of polymer sur-
face on the wetting and adhesion of liquid systems,” Journal of Adhesion Science and
Technology, vol. 21, p. 227 (2007).
20. T. Nishino, M. Meguro, K. Nakamae, M. Matsushita and Y. Ueda, “The lowest surface
free energy based on -CF3 alignment,” Langmuir, vol. 15, p. 4321 (1999).
21. R. Claessen, K. Reihs, P. Figueiredo, O. Stahlschmidt and A. Duparre, “Towards a true
“no-clean” property: highly durable ultra-hydrophobic coatings for optical applica-
tions,” Proc. Europ. Coat. Conf. Smart Coatings, p. 181 (2002).
22. A. Dotan, H. Dodiuk, S. Kenig and C. Laforte, “The relationship between water wetting
and ice adhesion,” J Adhesion Science and Technol, vol. 23, pp. 1907–1915 (2009).
23. P. F. Rios, H. Dodiuk, S. Kenig and A. Dotan, “Durable ultra-hydrophobic surfaces for self-
cleaning applications,” Polymers for Advanced Technologies, vol. 19, pp. 1684–1691 (2008).
24. P. F. Rios, H. Dodiuk, S. Kenig and A. Dotan, “Transparent ultra-hydrophobic surfaces,”
Journal of Adhesion Science and Technology, vol. 21, nos. 5–6, pp. 399–408 (2007).
5 A Review on the
Fabrication of
Nanostructured
Superhydrophobic
Aluminum Surfaces
Dilip K. Sarkar* and N. M. Saleema
CONTENTS
5.1 Introduction.....................................................................................................84
5.1.1 Superhydrophobicity in Nature............................................................ 87
5.1.2 Mimicking Superhydrophobic Surfaces.............................................. 88
5.2 Experimental................................................................................................... 89
5.2.1 Characterization.................................................................................. 89
5.3 Results and Discussion: Fabrication of Nanostructured
Superhydrophobic Aluminum Surfaces...........................................................90
5.3.1 Superhydrophobic Properties of Aluminum Surfaces Prepared
by rf-Sputtered Teflon Coatings........................................................... 91
by Plasma-Enhanced Chemical Vapor–Deposited Fluorinated-
Hydrocarbon Coatings......................................................................... 93
5.3.3 Superhydrophobic Aluminum Alloy Surfaces Produced via
a One-Step Process Using Fluoroalkylsilane in an Alkaline
Medium................................................................................................96
5.3.4 Fabrication of Superhydrophobic Aluminum by
Electrodeposition of Copper Followed by Electrochemical
Modification....................................................................................... 104
5.3.5 Superhydrophobic Aluminum Surfaces by Coating with
Monodispersed Fluorinated Silica Nanoparticles............................. 109
5.3.6 Application of Superhydrophobic Aluminum Surface in Ice
Adhesion Reduction........................................................................... 116
5.4 Conclusions.................................................................................................... 121
References............................................................................................................... 121
* Prof. Dilip K. Sarkar was the lead author for this chapter and can be contacted at dsarkar@uqac.ca.
83
Superhydrophobic surfaces are of tremendous importance in fundamental

research due to their potential usefulness in a range of industries. Due to their
unique water-repellency and self-cleaning abilities, the applications of super
hydrophobic surfaces are diverse and include areas such as corrosion resistance,
stain-resistant textiles, drag-reduction, inhibition of snow or ice adhesion, bio-
medical applications, and many others. Owing to their importance in today's
emerging technologies, many efforts have been made to replicate nature’s way
of making surfaces superhydrophobic as seen on lotus leaves and many other
plants and animal bodies. In this review, we have described various meth-
ods of rendering aluminum surfaces superhydrophobic via one-step as well as
two-step procedures. We have employed methods such as chemical etching of
aluminum surfaces followed by rf-sputtered Teflon coating; electrochemical
deposition of copper followed by electrochemical modification using stearic
acid molecules, in a two-step procedure; functionalization of silica nano
particles using low surface energy fluoroalkylsilane (FAS-17) molecules and
their deposition on aluminum surfaces by spin coating; immersion of alumi-
num substrates in FAS-17 in an alkaline medium, etc., in a one-step procedure.
Various surface characterization techniques were used to study superhydro-
phobic aluminum surfaces prepared by the various methods. The creation of
a rough micronanostructure on the treated surfaces was revealed by Scanning
Electron Microscopy (SEM) and the root mean square (rms) roughness was
measured by an optical profilometer. X-ray Photoelectron Spectroscopy (XPS)
and Infrared Spectroscopic studies confirmed the presence of the respective
surface components such as ZnO, TiO2, etc., as well as low surface energy
functional groups of fluorinated carbon on the superhydrophobic surfaces. The
resulting surfaces modified using low surface energy components via both
one-step and two-step procedures demonstrated water contact angles greater
than 150° and contact angle hystereses less than 5°.
5.1 INTRODUCTION
Wettability is one of the most important properties of a solid surface. Many practical
applications depend on the wetting or nonwetting characteristics of the solid surface.
When a water drop comes in contact with a surface, the surface is either wetted or
not wetted depending on the hydrophilicity or hydrophobicity of the surface, respec-
tively. The degree of wetting is characterized by the angle of contact (θ) the water
drop makes with the surface, which is the angle at which the liquid–vapor (LV)
interface meets the solid–liquid (SL) interface and the solid–vapor (SV) interface.
If the surface is hydrophilic, the water drop is strongly attracted to the surface
and will spread on the surface, providing a contact angle (CA) close to 0°. Less
stronger hydrophilic surfaces can have a CA up to 90°. If the surface is hydropho-
bic, the interaction forces between the water drop and the solid are weaker, and the
water drop will have a tendency to form a compact droplet on the surface. The CA
on such surfaces is usually greater than 90°. The CA, therefore, is determined by the
strength of the interfacial forces between the water drop and the solid surface. These
Fabrication of Nanostructured Superhydrophobic Aluminum Surfaces 85
interfacial forces are sometimes denoted as interfacial tensions or interfacial surface

free energies. Young describes the connection between the surface free energies of
the SL (γSL), SV (γSV), and LV (γLV) interfaces and the CA θ between the tangent
drawn to the spherical surface of the water drop from any point on the boundary
where all three phases meet [1], and this is expressed in the form of the so-called
Young’s equation:
γ SV − γ SL
cosθ = . (5.1)
γ LV
Figure 5.1 shows a schematic of a liquid drop resting on a solid surface with the
various parameters governing the CA θ of the liquid drop. Young’s equation is a
result of the thermodynamic equilibrium of the three interfaces. Depending on the
value of the CA of the water drop with the surface, the surface is categorized as
hydrophilic (θ < 90°), hydrophobic (θ > 90°), or superhydrophobic (θ > 150°). These
values are static CAs, which are obtained by sessile drop measurements using a
goniometer in which a water drop is deposited on the surface.
Another method used in characterizing the wetting characteristics is to measure
the dynamic CA at which a water drop is allowed to advance in one direction, as
shown in Figure 5.2, and the difference between the advancing CA θA and the reced-
ing CA θR is measured. This parameter is called the contact angle hysteresis (CAH).
The CAH value can reveal the ability of a water drop to stick or slide on a surface.
On a surface with a very low CAH, the water drop would roll off easily with a slight
tilt of the surface, and on a surface with a high CAH, the water drop would stick to
the surface.
Superhydrophobic surface treatments of various substrates have proved to be of
great interest in recent years [2]. To reach extreme values of the CA, greater than
150°, surface roughness is often added to amplify the hydrophobicity of the coating
materials. These CAs are sometimes addressed in the literature as apparent CAs
since the CA obtained does not represent the real CA on the corresponding flat sur-
face. The surface topography effects have been mathematically described by the
Wenzel [3] model (Figure 5.3) and the Cassie–Baxter model (Figure 5.4) [3,4].
γLV
Vapor
Liquid
γSV γSL
θ
Solid
FIGURE 5.1 Interfacial forces acting on a water drop in contact with a solid surface and its
contact angle with the solid surface.
θR
θA
FIGURE 5.2 Contact angle hysteresis (θA – θR) on an inclined surface.
θ΄
FIGURE 5.3 Schematic of the Wenzel model.
θ΄
FIGURE 5.4 Schematic of the Cassie–Baxter model.
In the Wenzel [3] model, the water drop penetrates the surface irregularities, as
shown in Figure 5.3. Mathematically, the Wenzel equation is written as
cos θ′ = Rw cos θ, (5.2)
where the roughness factor Rw is the ratio of the true and apparent (geometric) sur-
face areas. According to the Wenzel model, when the true CA θ of water on a smooth
surface is less than 90°, the apparent CA θ′ is less than the true CA θ on a rough
surface, and when the true CA θ is greater than 90°, the apparent CA θ′ is greater
than the true CA θ on a rough surface. It has been shown that the CA and the CAH
increase as the roughness factor increases [5]. However, it has also been demon-
strated that the CA continues to increase when the roughness factor exceeds a certain
level (~1.7), whereas the hysteresis starts to decrease [5]. This decrease in the CAH
is attributed to a transition from the Wenzel regime to the Cassie–Baxter regime,
where the water drop does not penetrate the surface irregularities; rather, it stays
suspended on the tips of the rough crests due to the entrapment of air pockets in those
crests (Figure 5.4).
The Cassie–Baxter model [4] describes the effect of roughness on chemically

heterogeneous structures where the apparent CA is mathematically derived from the
Cassie equation as follows:
cos θ′ = f1 cos θ1 + f 2 cos θ2, (5.3)
where θ′ is the apparent CA of the composite coating consisting of two components

with CAs θ1 and θ2 and corresponding area fractions f1 and f 2. In such a composite
system, f1 is assumed to be the solid surface, and f 2 is assumed to be air, where θ2 is
180°. As f1 + f 2 = 1, Equation 5.3 can be written as
cos θ′ = f1 (cos θ1 + 1) − 1. (5.4)
This equation explains that on a rough surface with large amount of air entrap-
ment in the surface irregularities, one can achieve a highly superhydrophobic surface
with a very small area fraction f1 of the surface in contact with the water drop. The
configuration leads to a very high CA and a very low CAH, leading to the rolling off
of water drops from the contacting solid.
The water drops with reduced contact area with the solid are termed “nonstick”
drops. Quéré deduced a condition for a drop of radius 0.5 mm to stick or slide by
relating the CA and CAH on a surface inclined at an angle α = 90° [6].
πlγ(cos θR − cos θA) ≥ ρgΩ sin α (5.5)
Here, πl is the contact line where all three interfaces meet; θR and θA are the
receding and advancing CAs, respectively; ρ, the density of liquid; g, the gravita-
tional acceleration; Ω, the volume of drop; and α is the angle of inclination of the sur-
face. Equation 5.5 provides the condition for a drop to remain stuck on the surface,
that is, a condition for which the adhesion forces between the drop and the surface
are greater than the gravitational force, tending to make the drop slide. It is obvious,
however, from Equation 5.5 that for a negligibly small CAH (<4°–5°), the condition
expressed can never be satisfied, and the drop will slide.
The author plotted the CAH against the CA using Equation 5.5 and showed theore
tically that nonstick drops may be achieved even with CAs as low as <20° provided
that the CAH is very low (<5°). However, such an ideal surface is practically difficult
to achieve. The author also showed that water drops must slide on a surface if the CA
is greater than 160° even if the CAH is higher due to the lower contact area between
the water drop and the solid. It can be inferred that the ideal conditions to obtain
a superhydrophobic surface with nonstick water drops would be to obtain a CA of
>160° as well as a CAH <10°.
5.1.1 Superhydrophobicity in Nature
Nature’s wonder inspires not only poetry but also science, in an attempt to learn
more about the world and the way it works [7]. Technology is frequently the applica-
tion of science, which is, in turn, inspired by nature. The technology used by many
researchers in making superhydrophobic surfaces is one such example. Numerous

plant surfaces and body parts of certain insects and animals exhibit superhydro-
phobic properties, evolved to survive many of nature’s hazards by repelling dirt and
dust particles in rain or snow. To start with, the leaves of the sacred lotus, scien-
tific name Nelumbo nucifera, have this inherent property of water repellency due to
which water drops roll off their surfaces, carrying away dirt and leaving the surfaces
of the leaves clean [8]. This property is also referred to as self-cleaning or the “lotus
effect.” Barthlott and Neinhuis [8] showed in their study that the key element in this
water repellency is the presence of hydrophobic surface components (epicuticular
wax crystalloids) on the microscopically rough structure of the leaf’s surface. Gould
[9] reports that a leaf surface is covered with tiny bumps, 5–10 μm high and about
10–15 μm apart. This uneven surface is itself covered with waxy, hydrophobic crys-
tals, measuring around 1 nm in diameter [9].
Barthlott and Neinhuis [8] reported a water CA of 160.4° ± 0.7° on the surface of
the lotus leaves.
Among plant leaves, there exist many other examples that exhibit the lotus effect
[10]. Feng et al. [10] showed that the water CAs on the surfaces of rice leaves, taro
leaves, and canna leaves are, respectively, 157° ± 2°, 159° ± 2°, and 165° ± 2°. Certain
leaves also have been found to demonstrate the lotus effect on both the frontal and
rear faces. For example, purple setcreasea and perfoliate knotweed leaves showed
water CAs of 167° ± 2° and 165° ± 2° on the frontal and rear faces, respectively [10].
In all these cases, Feng et al. [10] state that a rough microstructure and a hydrophobic
compound on the surface are the reasons for water repellency [10].
Nature also produces this remarkable property on the surfaces of certain insect
body parts. A popular example is the water strider’s legs. Water striders effortlessly
walk and stand on the surface of water, showing that that their legs are strikingly
water repellent. According to Gao and Jiang [11], the ability of the insect to stand
or walk easily on the water surface is due to the special hierarchical structure of the
legs, which are covered with a large number of oriented tiny hairs (microsetae) and
fine nanogrooves coated with waxy materials.
Research on butterfly wings indicates that water droplets easily roll off their sur-
faces, a special ability resulting from the direction-dependent arrangement of flex-
ible nanopatterns on ridging nanostripes and overlapping microscales on the wings,
where two distinct contact modes of a droplet with the orientation-tunable micro-
structures occur and thus produce different adhesion forces [12]. Zheng et al. [12]
measured a water CA of 152° ± 1.7° on the surfaces of the butterfly wings.
Superhydrophobicity in nature is not restricted to the examples mentioned above,
and many more natural candidates demonstrate this property. However, a few impor-
tant examples have been reviewed here to emphasize the importance to water repel-
lency of the rough microstructure and the presence of a hydrophobic coating.
5.1.2 Mimicking Superhydrophobic Surfaces

Due to the tremendous importance of superhydrophobic surfaces in today’s emerg-
ing technologies, many efforts have been made to replicate nature. The term “bio
mimicking” is commonly used to denote the artificial production of superhydrophobic
surfaces as the technology to make them involves copying nature’s solutions. The
applications of superhydrophobic surfaces are diverse due to their unique water-
repellency and self-cleaning abilities. The most common areas where superhydro-
phobic surfaces attract attention include antibiofouling paints for boats [13], biochips
[14], biomedical applications [15], microfluidics [16], corrosion resistance [17], eye-
glasses, self-cleaning windshields for automobiles [6], stain-resistant textiles [18],
antisticking of snow to antennas and windows [19], expected inhibition of adhesion
of snow, oxidation, current conduction [20], and many others.
A literature survey on the various attempts to make a superhydrophobic surface
has shown that the methods can be basically divided into two categories: (1) a one-
step approach in which a low-surface-energy material is roughened and (2) a two-
step procedure in which a surface is first roughened and then modified chemically
or physically to lower the surface energy [21–24]. In the one-step approach, these
studies only refer to roughening a low-surface-energy material; however, rendering a
metal surface such as aluminum superhydrophobic in one step has not been realized
so far. Both approaches mentioned here, however, emphasize the importance of the
combination of roughness and low surface energy. The field is vast, and the methods
reviewed here may not constitute an exhaustive list; however, in this review, we have
reviewed the methods that we have employed in achieving superhydrophobicity on
various surfaces, more specifically on aluminum surfaces, via both one-step and
two-step procedures.
5.2 EXPERIMENTAL
Various methods of making the materials to render the aluminum surfaces super-
hydrophobic have been described in the respective paragraphs detailing results and
discussions.
5.2.1 Characterization
The presence of a rough nanostructure on aluminum surfaces following chemical
etching and various other methods was confirmed by examination of images obtained
using field emission scanning electron microscopy (FESEM, LEO 1525). Image
analysis was used in the case of copper-coated aluminum surfaces to investigate the
interparticle distances and to determine the number density of copper particles. Ten
scanning electron microscope (SEM) images of each copper-coated sample prepared
under specific potentials were analyzed using Clemex software (CLEMEX JS-2000,
PE4.0). X-ray photoelectron spectroscopy (XPS; ESCALAB 220iXL, INRS-ÉMT,
Varennes, Canada) was used for surface chemical analyses of the different surfaces
prepared. Fourier transform infrared spectroscopy (FTIR; Perkin-Elmer Spectrum
One) was used to evaluate the atomic bonding at the surfaces prepared. Infrared
reflection–absorption spectroscopy (IRRAS) was also employed to characterize the
surface chemistry of certain sets of sample surfaces. IRRAS (Nicolet 6700 FT-IR)
is equipped with a mid-IR MCT-A N2-cooled detector and a KBr beam splitter. The
Smart SAGA (Specular Apertured Grazing Angle) accessory was used to analyze
samples at an average incidence angle of 80° relative to surface normal. The spectra
were recorded from 4000 to 650 cm−1 with a resolution of 4 cm−1, and 120 scans were
collected. The IR radiation was p-polarized, and a background spectrum taken from
a clean gold-coated reference sample was subtracted from the resulting spectrum.
The x-ray diffraction (XRD) analyses of the samples were carried out using a Bruker
D8 Discover System. The roughness of surfaces was measured using an optical pro-
filometer (MicroXAM-100 HR 3D Surface Profilometer).
The wetting properties of the surfaces coated with the three dielectrics were inves-
tigated using a CA goniometer (Krüss GmbH, Germany). A water drop of volume
~5 μL was suspended from the syringe needle and brought into contact with the
superhydrophobic surface using a computer-controlled device provided by Krüss
GmbH. The advancing and receding CAs were measured by holding the needle sta-
tionary and moving the sample in one direction. The static CA data were acquired
by fitting the symmetric water drops using the Laplace–Young equation, and the
advancing and receding CAs were determined from the asymmetric water drops
using the tangent-2 method as provided by the Krüss software. The wetting charac-
teristics of the surfaces were also examined using a First Ten Ångstroms CA goni-
ometer for certain sets of experiments.
Icephobicity tests were performed by exposing the sample surfaces to a freezing
drizzle of supercooled water droplets at –10°C in a wind tunnel at a fixed wind speed
of 10 m/s in order to simulate atmospheric icing conditions in natural outdoor situa-
tions. The samples mounted on untreated supports were placed vertically in the wind
tunnel perpendicular to the direction of flow of the freezing drizzle. Following the
accumulation of ice on the sample surfaces, the adhesion strength of ice was deter-
mined using a centrifugal ice adhesion test (CAT) apparatus as a standard method
for testing various surfaces for ice adhesion strength. In this method, an aluminum
beam (32 mm wide and 30 cm long) with the test sample (32 × 50 mm) and accu-
mulated ice attached to one end, and a counterweight attached to the other end to
balance the beam, is fixed in the centrifuge test chamber maintained at –10°C. The
beam is then rotated at increasing speeds, resulting in a controlled ramp of the cen-
trifugal force. When this force reaches the adhesion force of ice, the ice detaches
from the sample surface. The exact rotation speed at the time of ice detachment is
determined from the computer software developed in-house. The adhesion force F
is then determined using the formula F = mrω2, where m is the mass of ice, r is the
radius of the beam, and ω is the rotation speed. The shear adhesion strength (τ = F/A)
of ice is then determined from the apparent area A of the sample surface that was in
contact with ice.
5.3 RESULTS AND DISCUSSION: FABRICATION

OF NANOSTRUCTURED SUPERHYDROPHOBIC
ALUMINUM SURFACES
We have reported several superhydrophobic surfaces prepared via two-step proce-
dures where rough micro/nanostructures were created via techniques such as chemi-
cal bath deposition (CBD), substrate chemical etching, galvanic exchange reactions,
electrochemical deposition, and so forth. The passivation of these surfaces was
carried out using organic molecules such as stearic acid, using fluoroalkylsilane
(FAS-17) molecules, or by coating with rf-sputtered Teflon as well as fluorinated-
hydrocarbon thin films [25–29]. Furthermore, we have also reported superhydropho-
bic surfaces prepared in two steps by depositing via spin-coating fluorinated silica
nanoparticles [30,31]. All these studies emphasized the importance of the coexis-
tence of both surface roughness and the low-surface-energy coating in order for the
surface to exhibit superhydrophobicity.
On the other hand, we have also reported the preparation of superhydrophobic
surfaces of aluminum alloy, silver nanostructures, and so forth, via a one-step proce-
dure in which both the creation of a rough micro/nanostructure and the lowering of
the surface energy take place simultaneously in a single step [32–35]. Additionally,
we have also reported the ice adhesion reduction on superhydrophobic aluminum
surfaces [36,37]. Some of our findings in this field related to the aluminum substrates
are described in this review.
5.3.1 Superhydrophobic Properties of Aluminum Surfaces

Prepared by rf-Sputtered Teflon Coatings
Chemical etching in combination with radio frequency (rf) sputtering of Teflon has
been employed to render aluminum surfaces superhydrophobic [26]. In this work,
rolled sheets of aluminum alloy (AA6061) of dimensions 25 × 25 × 1.58 mm were
used as substrates. The substrates were etched with 14.8 wt% hydrochloric acid
(HCl) for different times ranging from 1 to 5 min. The etched clean samples were
dried and further coated with ultrathin Teflon films by rf sputtering using Ar plasma
in an inductively coupled plasma reactor.
Figure 5.5 shows FESEM images of rf-sputtered Teflon-coated as-received alumi-
num substrate (Figure 5.5a) and aluminum substrate etched with 14.8 wt% HCl for
2.5 min (Figure 5.5b). It is clear from Figure 5.5b that etch pits are formed on the alu-
minum surface after etching with HCl acid, which effectively modifies the surface
morphology [38,39]. XPS analysis confirms the presence and chemical composition
of the rf-sputtered Teflon films. Figure 5.6 shows the C1s spectrum of an rf-sputtered
(a) (b)
20 µm 20 µm
FIGURE 5.5 FESEM images of rf-sputtered Teflon-coated (a) as-received aluminum sub-
strate and (b) aluminum substrate etched with 14.8 wt% HCl for 2.5 min. (From Sarkar, D.K.
et al., Mater. Lett., 62, 1226, 2008.)
1.2
-CF2
1.0
-C-CFn
-CF-CFn
-CF3
0.8
Intensity (a.u.)
-C-F
0.7
-C-C
0.5
0.3
0.2
296 292 288 284

Binding energy (eV)
FIGURE 5.6 (See color insert.) C1s XPS spectrum of an rf-sputtered Teflon film. (From
Sarkar, D.K. et al., Mater. Lett., 62, 1226, 2008.)
30
1.6 (a) (c)
Contact angle hyst. (°)
20
Thickness (mm)
1.5
1.4 10
1.3
0
180 2 3 4 5
(b) Etching time (min.)
160
Contact angle (°)
140
120
0 1 2 3 4 5
Etching time (min.)
FIGURE 5.7 (a) Change in thickness of aluminum substrates with etching time; (b) contact
angle of water, and (c) contact angle hysteresis of water on rf-sputtered Teflon-coated alumi-
num substrates with different etching times. (From Sarkar, D.K. et al., Mater. Lett., 62, 1226,
2008.)
Teflon film. Figure 5.7a shows changes in thickness of the aluminum substrates in
relation to etching time. The as-received aluminum substrates are 1.58 ± 0.006 mm
thick. However, after 2.5 min of etching, the thickness is reduced to 1.46 ± 0.014 mm
and then further reduced to 1.25 ± 0.012 mm when etching is prolonged to 5 min.
The change in thickness of the aluminum substrates, etched between 2 and 5 min,
follows a first-order exponential decay law with time as presented in Equation 5.6:
d = 122 + 1.74 exp(−t/1.28), (5.6)
where d is the thickness of the aluminum substrate and t is the etching time.
Figure 5.7b shows the CA data of water on etched aluminum (EA) substrates thinly
coated with rf-sputtered Teflon. As-received aluminum substrates coated with rf-
sputtered Teflon showed a CA of ~121° ± 1.3°. The CA increased to 164° ± 3° for the
aluminum substrate that was etched for 2.5 min.
It is interesting to note that the CA remains constant on the aluminum surfaces
that were etched for longer than 2.5 min, as shown in Figure 5.7b.
An average CAH of 2.5° ± 1.5° was achieved on Teflon-coated EA surfaces, as
shown in Figure 5.7c. CA values similar to our observations or even higher have been
reported in recent literature [40].
The insets of Figure 5.7b show the shapes of water drops on 1 min and 2.5 min
etched surfaces, to illustrate CA values. The existence of a large number of CF3
groups along with CF2 groups makes the EA surfaces highly superhydrophobic, pro-
viding a water CA of ~164° ± 3° with a very low CAH below 2.5° ± 1.5°. Water drops
simply roll off such surfaces. This simple, low-cost, and efficient way of preparing
substrates with superhydrophobic properties has promising industrial applications.

by Plasma-Enhanced Chemical Vapor–Deposited
Fluorinated-Hydrocarbon Coatings
In this method, superhydrophobic aluminum surfaces are created by coatings of
thin fluorinated-hydrocarbon films on aluminum substrates [27]. Hydrocarbon and
fluorinated-hydrocarbon coatings were performed using inductively coupled plasma-
enhanced chemical vapor deposition (PECVD), applying a power of 100 W in the
rf source using a mixture of Ar and CH4 for hydrocarbon and Ar, CH4, and C2F6 for
fluorinated-hydrocarbon coatings [27,41]. The hydrocarbon and fluorinated-hydro-
carbon coatings were deposited on smooth Si (100) surfaces as well as on chemically
etched 6061 aluminum alloy surfaces.
Figure 5.8 shows the XPS survey spectra of hydrocarbon and fluorinated-
hydrocarbon coatings deposited on smooth Si (100) surfaces. The survey spectra
confirm the presence of carbon in hydrocarbon coatings and carbon and fluo-
rine in fluorinated-hydrocarbon coatings. A trace of oxygen was observed in both
hydrocarbon and fluorinated-hydrocarbon coatings. The additional peak of F1s at
698.2 eV in the survey spectrum of fluorinated-hydrocarbon coating (absent in the
case of hydrocarbon coating) is an excellent indication of the presence of fluorine
4 (b) F-hydrocarbon
(a) Hydrocarbon
F1s
F KLL
3
Intensity (a.u.)
2
C1s
(b)
O1s
1
F1s
(a)
0
1200 1000 800 600 400 200 0
Binding energy (eV)
FIGURE 5.8 XPS surface survey spectra of hydrocarbon and fluorinated-hydrocarbon

coatings. (From Sarkar, D.K. et al., Appl. Surf. Sci., 256, 3698, 2010.)
in the coating, originating from the combination of Ar and C2F6 gases used in the
PECVD coating process.
Figure 5.9 shows the high-resolution C1s spectra of hydrocarbon and fluorinated-
hydrocarbon coatings on smooth Si (100). It can be seen that the C1s spectrum of
fluorinated-hydrocarbon has been resolved into five different components, namely,
‑CF3, ‑CF2, ‑C ‑F, ‑C ‑CFn and –C–C at binding energies of 292.8, 290.5, 288.4,
287, and 285.0 eV, respectively, as compared to only one component, –C–C at 285.0
eV, in the C1s spectrum of hydrocarbon coating. The various fluorocarbon radicals in
fluorinated-hydrocarbon as seen in Figure 5.9 obviously arise from the fragmentation of
(a) (b)
1.2
-C-C
C1s (Hydrocarbon) C1s (F-hydrocarbon)

2.5
1.0
2.0 0.8
Intensity (a.u.)
Intensity (a.u.)
-C-CFn
1.5 0.6
-C-F
-CF2
1.0 0.4
-CF3
0.5 0.2
0.0 0.0
294 292 290 288 286 284 282 294 292 290 288 286 284 282
Binding energy (eV) Binding energy (eV)
FIGURE 5.9 XPS high-resolution C1s spectra of (a) hydrocarbon (Ar-CH4) and (b) fluorinated-
hydrocarbon (Ar-C2F6) coatings on Si (100) surfaces. (From Sarkar, D.K. et al., Appl. Surf. Sci.,
256, 3698, 2010.)
C2F6 in the presence of Ar. The presence of a low-surface-energy component such

as CF3 and CF2 groups in the fluorinated-hydrocarbon–coated surface helps to lower
the surface energy.
The morphological characterization of both hydrocarbon and fluorinated-hydro-
carbon coatings was performed using atomic force microscopy (AFM) as shown in
Figures 5.10a and b and 5.11a and b. A slight variation in the morphological features of
hydrocarbon and fluorinated-hydrocarbon coatings has been observed as the number
density of the networks in fluorinated-hydrocarbon coating (Figure 5.11a and b) is com-
parably higher than that observed in case of hydrocarbon coating (Figures 5.10a and b).
A change in roughness, although minor, has also been observed in the two coatings
as the root mean square (rms) roughness of the hydrocarbon coating is ~6 nm and that
of fluorinated-hydrocarbon is ~9 nm, as evaluated using AFM analyses. However,
the water CAs measured on the two surfaces showed considerable difference. The
CA on the hydrocarbon-coated surface was found to be ~72° (Figure 5.10c), whereas
the fluorinated-hydrocarbon–coated surface showed a CA of ~105° (Figure 5.11c).
The higher CA value in the case of the fluorinated-hydrocarbon–coated surface
is due to the presence of a greater number of low-surface-energy CF3, CF2, and
other components, as evident from XPS analyses (Figure 5.9). The concentration of
(a) (b)
200.000
(c) CA 72°
µm
1.5
1.0
500 nm 0.5
FIGURE 5.10 (a) 2-D and (b) 2 × 2 μm 3-D tapping-mode AFM images of hydrocarbon
coating on smooth Si (100) surfaces. (c) Image of a water drop on a hydrocarbon-coated
smooth Si surface. (From Sarkar, D.K. et al., Appl. Surf. Sci., 256, 3698, 2010.)
(a) (b)
200.00
(c) CA 105°
µm
1.5
1.0
500 nm 0.5
FIGURE 5.11 (a) 2-D and (b) 2 × 2 μm 3-D tapping-mode AFM images of fluorinated-
hydrocarbon coating on smooth Si (100) surfaces. (c) Image of a water drop on a fluorinated-
hydrocarbon–coated smooth Si surface. (From Sarkar, D.K. et al., Appl. Surf. Sci., 256, 3698,
2010.)
(b)
(a)
(c)
200 nm
FIGURE 5.12 (a) SEM image of micro/nanorough aluminum (6061) surface. (b) Image of
a water drop suspended from a syringe needle during CAH measurements on hydrocarbon-
coated micro/nanorough aluminum and (c) on fluorinated-hydrocarbon–coated micro/nanor-
ough aluminum. (From Sarkar, D.K. et al., Appl. Surf. Sci., 256, 3698, 2010.)
fluorine on the surface of fluorinated-hydrocarbon coatings was found to be ~36

at %. The lower CA value in the case of the hydrocarbon-coated surface is due to the
absence of such low-surface-energy groups (CF3, CF2, etc.).
To understand the behavior of the water drops on a micro/nanorough surface,
hydrocarbon and fluorinated-hydrocarbon coatings were deposited on chemically
etched micro/nanorough aluminum surfaces. The CA of the hydrocarbon-coated EA
surface increased to ~135°, and the water drop stuck to these surfaces, as demonstrated
by Figure 5.12b, making it impossible to measure the CAH. However, when coated
with fluorinated-hydrocarbon, the micro/nanorough aluminum surfaces demonstrated
superhydrophobic properties with CA values of ~165° and CAH of <2° (Figure 5.12c).
The water drops on these surfaces were found to roll off easily even with a slightest tilt
of the surfaces as observed in our previously reported rf-sputtered Teflon-coated EA
surfaces [26]. This observation complements the fact that the higher CA value on the
fluorinated-hydrocarbon–coated micro/nanorough aluminum surfaces as compared
to those coated with hydrocarbon is due to the presence of a large number of low-
surface-energy fluorinated carbon compounds in the fluorinated-hydrocarbon coatings.
5.3.3 Superhydrophobic Aluminum Alloy Surfaces Produced via a

One-Step Process Using Fluoroalkylsilane in an Alkaline Medium
In the present study, aluminum alloy surfaces are rendered superhydrophobic using
a simple one-step process in which the aluminum coupons are simply immersed in
an aqueous solution containing FAS-17 molecules and sodium hydroxide (NaOH)
[32,33]. The chemical reaction of NaOH with aluminum in the presence of FAS-17
molecules results in an etching process due to the presence of NaOH in the solution,
leading to a rough microporous micro/nanostructure on the surfaces, as shown in
the FESEM images of Figure 5.13. Figure 5.13b through d shows aluminum surfaces
(a) CA–73° (b) CA>160° CAH <5°
50 µm 50 µm
(c) (d)
10 µm 5 µm
FIGURE 5.13 FESEM images of (a) untreated aluminum and (b) aluminum treated with NaOH
and FAS-17 with a FAS-17/NaOH molar ratio of 0.2 for 15 min. (c–d) Higher-magnification
images of (b). (From Saleema, N. et al., ACS Appl. Mater. Interfaces, 3, 4775, 2011.)
treated with NaOH in the presence of FAS-17 molecules at a FAS-17/NaOH ratio of

0.2, revealing craterlike microfeatures of ~10 μm in size, evolved following treat-
ment, as compared to the as-received clean aluminum surface shown in Figure 5.13a.
The microcraters are, in addition, decorated with nanometer-sized fibers, as is evi-
dent from the higher-magnification images shown in Figure 5.13c and d, which are
present all over the surface, providing the system with a two-tier micro/nano binary
structure. The presence of such a rough binary structure on a surface is one of the two
important requirements for superhydrophobicity [8,25–28,32,34,35]. Further SEM
analyses also revealed that these morphological features remained similar on all sur-
faces treated with FAS-17 molecules. We have previously reported microcrater for-
mation on surfaces treated with NaOH and FAS-17 [32]. Fu and He [42] also reported
a binary structure composed of microscale craterlike pits and nanoscale reticula on
aluminum surfaces, but via a two-step roughening procedure that involved metal-
lographic abrasion followed by etching in a combination of nitric acid and copper
nitrate. The formation of a rough binary microcratered nanofibrous structure on the
surface in combination with a modified chemistry arising from the adhesion of low-
surface-energy FAS-17 molecules contributes to the evolution of superhydrophobic
properties.
IRRAS spectral analysis of the aluminum surfaces treated with NaOH and FAS-
17 at different FAS-17/NaOH molar ratios and for different treatment times, recorded
in the wave number range of 4000–650 cm–1, revealed the presence of the fluorinated
functional groups, which are mainly responsible for the lowering of surface ener-
gies. The –OH stretching absorption band centered in the wave number range of
3600–3000 cm–1 arising from aluminum hydroxide formed on the surface follow-
ing the reaction during etching was observed in the full-range IR spectra [43,44].
Al Abadleh and Grassian [43] have presented similar hydroxide IR peaks for water
adsorption on alumina powder. Figure 5.14a shows the IR spectra of aluminum alloy
surfaces treated with a FAS-17/NaOH ratio of 0.2 for different treatment periods in
the wave number range of 1800–650 cm–1. The appearance of a peak belonging to
–CF2– functional groups arising from C‑F stretching vibration is evident in all these
spectra between wave numbers 1120 and 1350 cm–1 [45–47]. It can be seen that the
intensity of this peak is at its maximum at 15 min of treatment time. Therefore, a
critical etching time of 15 min has been found, after which the peak intensity slightly
decreased with further increase in the time of treatment.
Figure 5.14b shows the IR spectra of the surfaces treated with varying molar ratios
of FAS-17 to NaOH (0.05 to 0.4) for the critical etching time of 15 min obtained
from the time-dependent studies. The peaks belonging to –CF2‑functional groups
emerge between wave numbers 1120 and 1350 cm–1 [45–47]; however, these peaks
are present only on the surfaces treated with FAS-17/NaOH ratios of 0.4 and 0.2,
with increased intensity of the peaks for the surface treated with the higher FAS-17/
NaOH ratio of 0.4. These C‑F stretching vibrations disappear as the FAS-17/NaOH
ratio further decreases to 0.1 and 0.05. Therefore, a higher relative concentration of
NaOH in the solution slows down the deposition of the FAS-17 molecules onto the
aluminum alloy surface during the reaction.
In both time-dependent and concentration-dependent studies, a few other promi-
nent IR peaks are also seen, which have a tendency to increase with increasing etch-
ing time and decreasing FAS-17/NaOH molar ratio. These peaks are situated around
700–1000, 1370, and 1580 cm–1. The peaks appearing around 1370 and 1580 cm–1
may be assigned to the bending mode of molecular water coordinated to octahedral
(a) (b)
1.5
0.25
25 min 0.05
1.2 15 min 0.1
5 min 0.20 0.2
CF2 0.4
CF2
Absorption (a.u.)
Absorption (a.u.)
0.9 OH 0.15
OH OH
0.6 0.10 OH
0.3 0.05
0.0 0.00
1750 1500 1250 1000 750 1750 1500 1250 1000 750
Wavenumber (cm–1) Wavenumber (cm–1)
FIGURE 5.14 IRRAS spectra acquired on aluminum alloy surfaces treated (a) with a FAS-
17/NaOH molar ratio of 0.2 for different treatment times and (b) with different FAS-17/NaOH
molar ratios for 15 min treatment time. (From Saleema, N. et al., ACS Appl. Mater. Interfaces,
3, 4775, 2011.)
and tetrahedral aluminum ion sites, respectively, on alumina according to Satoh and
Nakazumi [18]. Other studies indicate that these peaks may originate from adsorbed
water on an aluminum oxide surface [44]. The peaks observed between 700 and 1000
cm–1 may also be associated with the hydroxide formed as a result of water adsorp-
tion on the aluminum alloy surfaces during the etching reaction. All these peaks
are found to increase in intensity with decreasing FAS-17/NaOH ratio. Therefore,
with increased NaOH concentration in the aqueous mixture of NaOH and FAS-17
and with increased treatment time, the water adsorption is greater, increasing the
hydroxide group concentration with reduced concentrations of fluorinated functional
groups on the surface. Increased amounts of hydroxides and reduced or negligible
amounts of low-surface-energy fluorinated groups on the surface may not lead to
superhydrophobic properties.
XPS investigations of the aluminum alloy surfaces treated with NaOH and FAS-
17 at different FAS-17/NaOH ratios for 15 min revealed the presence of C, F, O, and
Si with no trace of Na present in the survey spectra. (Figure not shown.) Figure 5.15a
and b shows the high-resolution C1s core-level spectra with the corresponding O1s
spectra in the insets, acquired from the aluminum surface treated with a FAS-17/
NaOH ratios of 0.4 and 0.1, respectively. The C1s spectrum of the surface treated
with a FAS-17/NaOH ratio of 0.4 was resolved into seven components, namely, ‑CF3
(293.8 eV), ‑CF2 (291.2 eV), ‑CH2‑CF2 (288.8 eV), ‑C‑O (286.1 eV), ‑C‑C (284.7
eV), ‑C‑Si (281.8 eV), and –C‑metals (280.8 eV). The C1s spectrum of the surface
treated with a FAS-17/NaOH ratio of 0.1 could also be resolved into seven compo-
nents, namely, ‑CF3 (293.6 eV), ‑CF2 (291.1 eV), ‑CH2‑CF2 (288.7 eV), ‑C‑O (286.5
eV), ‑C‑C (284.5 eV), ‑C‑Si (281.7 eV), and –C‑metals (279.5 eV). The F1s peak
in both cases was observed at 688.5 eV. The O1s binding energies observed on the
surfaces treated with FAS-17/NaOH ratios of 0.4 and 0.1 were 531.8 and 531.1 eV,
respectively. The binding energies reported here are consistent with our previous
reports [32,49]. The CF3 and CF2 concentrations from the C1s spectra of the surface
treated with a FAS-17/NaOH molar ratio of 0.4 are 6.2% and 47.7% as we reported
previously [32,49]. These values as previously reported are slightly higher than the
theoretical values obtained from the molecular structure of FAS-17, which are 6%
and 41%, respectively, for CF3 and CF2. The higher concentrations of CF3 and CF2
observed on the surface indicate that these low-surface-energy components com-
prise the outermost surface, contributing to superhydrophobic properties. However,
with a reduced FAS-17/NaOH molar ratio of 0.1, the CF3 and CF2 concentrations as
obtained from the C1s spectral analysis decreased to 3.6% and 21.8%, respectively.
The reduced concentrations of these low-surface-energy components on the surface
treated with a FAS-17/NaOH ratio of 0.1 are consistent with the FTIR analysis.
From the FTIR and XPS investigations, a possible mechanism for the treat-
ment with NaOH and FAS-17 molecules with varying FAS-17/NaOH ratios can be
outlined. Figure 5.16 shows a schematic presentation of possible reaction mecha-
nisms leading to a superhydrophobic surface and a hydrophilic surface. On a sur-
face treated with higher FAS-17/NaOH ratios (for example, a FAS-17/NaOH ratio of
0.2), the formation of aluminum hydroxide and the incorporation of CF2 functional
groups originating from FAS-17 molecules on the surface, as revealed by the FTIR
and XPS measurements may be schematized as shown in Figure 5.16a. The C2H5
(a)
0.7 2.0
C1s O1s
Intensity (a.u.)
1.6
0.6 1.2
0.8
-CF2
0.4
Intensity (a.u.)
0.5 538 536 534 532 530 528 526
-CH2-CF2
Binding energy (eV)
-C-O
-CF3
0.4
-C-C
-C-Metals
-C-Si
0.3
0.2
295 290 285 280
Binding energy (eV)
(b)
2.5
O1s C1s
Intensity (a.u.)
2.0
0.6
1.5
-C-C
1.0
0.5 0.5
Intensity (a.u.)
538 536 534 532 530 528 526

Binding energy (eV)
-C-O
-CH2-CF2
0.4
-CF2
-CF3
-C-Metals
-C-Si
0.3
0.2
295 290 285 280
Binding energy (eV)
FIGURE 5.15 (See color insert.) High-resolution C1s core-level spectra of aluminum surfaces
treated with FAS-17/NaOH ratio of (a) 0.4 and (b) 0.1. Insets show the corresponding high-
resolution O1s spectra. (From Saleema, N. et al., ACS Appl. Mater. Interfaces, 3, 4775, 2011.)
component is removed from FAS-17 molecules in the hydrolysis process and the
silicon bonds with the oxygen on the surface so that the C-F functional groups are
oriented outward from the surface. Such bonding does not take place throughout the
surface, as evidenced by the observation of only a few OH groups on the surface. The
number of FAS-17 molecules adsorbed on the surface with the low-surface-energy
C-F functional groups oriented outward on the resulting surface may be sufficient
to provide superhydrophobic properties. However, when the FAS-17/NaOH ratio is
lower, the etching of the surface dominates over the incorporation of FAS-17 mol-
ecules. Therefore, the resulting surface is characterized by a large number of OH
(a) F F
F C F F C F
[F C F]7 [F C F]7
[H C H]2 [H C H]2
O Si O Si O
OH O O OH
Al2O3
Aluminum substrate
(b)
OH OH OH OH OH OH OH OH OH OH OH OH
Al2O3
Aluminum substrate
FIGURE 5.16 Schematic of bonding mechanism of FAS-17 molecules to render (a) super
hydrophobic and (b) hydrophilic aluminum alloy surfaces. (From Saleema, N. et al., ACS
Appl. Mater. Interfaces, 3, 4775, 2011.)
groups with a negligible concentration of FAS-17 molecules present as revealed by

IRRAS measurements, schematized in Figure 5.16b. The presence of a large number
of OH groups on the surface may lead to hydrophilic behavior as the affinity of water
is higher with hydroxides.
Figure 5.17a and b shows the variation in the water CA and rms roughness,
respectively, with etching time for surfaces treated with FAS-17/NaOH ratio of 0.2.
It is clear from Figure 5.17a that with increasing etching time, the CA increases up
to a critical etching time of 15 min and then begins to decrease. A water CA value
of only 146° ± 8° at 5 min etching time increased to 161° ± 6° at 15 min etching
time, combined with a very low CAH of 5° ± 3°, at which point the water drops
started to roll off the surface. Such a high water CA can be attributed to the pres-
ence of a high concentration of low-surface-energy fluorinated molecules on these
surfaces, as revealed by both IRRAS and XPS investigations (Figures 5.14 and 5.15).
The CAH for surfaces etched for less than 15 min was not possible to measure as
the water droplet stuck to those surfaces, demonstrating very low water CA values.
At an etching time of 25 min, the surface exhibited a water CA of 154° ± 2° and a
CAH of 14°, remaining in the superhydrophobic zone (>150°). Figure 5.17b shows
that the rms roughness of the surface increases with etching time. The rms rough-
ness increased to 0.7 ± 0.07 μm at 15 min etching time from a value of 0.5 ± 0.06
μm obtained at 5 min etching time. Along with increasing rms roughness, the CA
also increased up to 15 min etching time. With further increase in etching time, the
(a) (b)
170 1.6
1.4 FAS-17/NaOH 0.2
rms roughness (µm)

160
1.2
CA (°)
150
1.0
140 0.8
0.6
130
5 10 15 20 25
5 10 15 20 25 Time of etching (min)
Time of etching (min)
FIGURE 5.17 (a) Water contact angle measured on aluminum alloy surfaces and (b) surface
roughness of aluminum alloy surface as a function of the etching time, treated with a FAS-
17/NaOH molar ratio of 0.2. (From Saleema, N. et al., ACS Appl. Mater. Interfaces, 3, 4775,
2011.)
rms roughness further increased to a value of 1.2 ± 0.1 μm at 25 min etching time;
however, the water CA value started to decrease. This reduction in water CA in spite
of higher rms roughness may be due to the loss of CF2 fragments at longer etching
times, eventually leading to more hydroxide on the surface, as revealed by FTIR
studies (Figure 5.14a). Therefore, it may be concluded that a critical time of etching
of 15 min is necessary to obtain superhydrophobic properties, after which the super-
hydrophobic behavior is found to gradually deteriorate.
Figure 5.18a and b shows the variation in the water CA and rms roughness, respec-
tively, on the aluminum alloy surfaces treated with different FAS-17/NaOH ratios
for an etching period of 15 min. Figure 5.18a shows an increase in the water CA
(a) (b)
180
2.4
150
rms roughness (µm)
2.0
120
CA (°)
1.6
90 1.2
60 0.8
0.4
30 0.0 0.1 0.2 0.3 0.4
0.0 0.1 0.2 0.3 0.4 (FAS/NaOH) molar ratio
(FAS/NaOH) molar ratio
FIGURE 5.18 (a) Water contact angle measured on aluminum alloy surfaces and (b) surface
roughness of aluminum alloy surfaces treated with different FAS-17/NaOH molar ratios for
15 min etching time. (From Saleema, N. et al., ACS Appl. Mater. Interfaces, 3, 4775, 2011.)
with increasing FAS-17/NaOH ratio. Below a FAS-17/NaOH ratio of 0.2, the surfaces
were found to be hydrophilic, with the water CA values below 90°. Superhydrophobic
behavior was obtained when the FAS-17/NaOH ratio was 0.2, with a very high water
CA of 161° ± 6° and very low CAH of 5° ± 3°. With further increase in the FAS-17/
NaOH ratio to 0.4, the water CA further increased to 166° ± 4.5°, and the CAH fur-
ther decreased to 4° ± 0.5°. Therefore, a FAS-17/NaOH ratio of 0.2 is found to be the
critical concentration necessary to obtain superhydrophobic properties on the alumi-
num alloy surfaces. The increase in water CAs, again, can be attributed to the pres-
ence of low-surface-energy C‑F functional groups on the surface oriented outward
from the surface (Figure 5.16), as revealed by both IRRAS and the XPS investiga
tions. However, Figure 5.18b shows that the rms roughness decreases with increasing
FAS-17/NaOH ratio. Below the critical FAS-17/NaOH ratio of 0.2, the rms roughness
varied between 1.6 and 2.3 μm. The rms roughness remained similar with values of
0.75 ± 0.07 μm and 0.7 ± 0.06 μm, respectively, on surfaces treated with FAS-17/
NaOH ratios of 0.2 and 0.4. It is clear from these values that, in this particular case,
with higher rms roughness, the CA values are lower and in the hydrophilic zone,
and with lower roughness, the CA values are higher and in the superhydrophobic
zone. The lower CAs on the surfaces with higher rms roughness can be attributed to
the presence of a higher concentration of hydroxides on these surfaces, as is evident
from the IRRAS studies (Figure 5.14b). Lower ratios of FAS-17 to NaOH lead to a
surface modification dominated by roughening of the surface and also result in the
fragmentation of FAS-17 molecules, as seen by XPS, leading to a negligible coverage
by low-surface-energy fluorinated compounds on the surface. However, the higher
CAs on the surfaces with a lower rms roughness of ~0.7 μm treated with higher
FAS-17/NaOH ratios indicate that this roughness is sufficient to obtain superhydro-
phobicity provided that a sufficient number of low-surface-energy components are
correctly oriented on the surface. Generally, an increase in roughness of a surface
while maintaining a constant low surface energy leads to large amount of air entrap-
ment, resulting in an increase in water CA values, according to the Cassie–Baxter
[4] model. We have confirmed this behavior in our recent works on thin films with
FAS-modified silica nanoparticles as well as on aluminum surfaces with electro-
chemically stearic acid–modified copper microdots [29,30]. Brassard et al. [30]
showed increased water CA values on low-surface-energy aluminum surfaces with
FAS-modified silica nanoparticles where the surface roughness increased due to the
increase in particle size of the silica nanoparticles. Similarly, Huang et al. [29] cre-
ated the surface roughness by depositing copper microdots on an aluminum surface
and further modified these copper microdots electrochemically using stearic acid to
maintain similar surface chemical composition. Both works showed that an increase
in surface roughness leads to superhydrophobicity provided that the chemical com-
position of the surface remained the same. Contrary to our present work, a higher
roughness resulting from an increased NaOH quantity in the solution of FAS-17
and NaOH leads to reduced water CA values. In this present case, not only do we
increase the surface roughness, but a change in FAS-17 molecules with loss of CF3
and CF2 groups also occurs, as evidenced by XPS. Although higher surface rough-
ness is created in this process, due to the loss of the low-surface-energy components,
we obtain lower water CAs, which contradicts the conventional thinking.
We have previously reported similar high water CAs and low CA hystereses on
various surfaces produced via one-step as well as two-step processes [25,26,32,35].
Sarkar et al. [26] reported a water CA of 164° ± 3° and a hysteresis of 2.5° ± 1.5° via
a two-step process on chemically etched aluminum surfaces coated with rf-sputtered
Teflon. Fu and He’s [42] study on superhydrophobic aluminum surfaces produced by
a combination of mechanical abrasion and chemical etching followed by passivation
with decyltriethoxysilane demonstrated a similar water CA of 159.7°. Song and Shen
[48] reported a CA of 150° on aluminum sheets exhibiting a petallike microstruc-
ture via a two-step procedure for surface roughening by immersion in a solution of
NH2(CH2)6NH2 followed by passivation using perfluorodecyltriethoxysilane. All these
authors, however, reported superhydrophobic aluminum surfaces produced via a two-
step procedure involving surface roughening methods to obtain a rough micro/nano-
structure and surface passivation methods to lower the surface energy. In the present
study, we have reported very high water CAs (>160°) and very low CA hystereses
(<5°) via a simple one-step procedure in which the creation of micro/nanoroughness
as well as lowering of the surface energy take place simultaneously. Further, the water
CA obtained on an as-received aluminum surface of only 73° ± 3° shows that these
superhydrophobic surfaces (FAS-17/NaOH > 0.2, rms roughness < 0.7 μm) follow the
Cassie–Baxter [4] model where, due to the presence of air in the gaps of the rough
surface, therefore forming a composite surface of air and the solid, the water CA is
enhanced to values greater than 150°. On the other hand, on the hydrophilic surfaces
(FAS-17/NaOH < 0.1, rms roughness > 1.6 μm), the water CAs drop below 70°, indi-
cating that these surfaces follow the Wenzel [3] regime. Therefore, a transition from
Wenzel regime to Cassie–Baxter regime has been found depending on the FAS-17/
NaOH ratio and the corresponding roughness resulting from a certain etching time.
5.3.4 Fabrication of Superhydrophobic Aluminum by Electrodeposition

of Copper Followed by Electrochemical Modification
Considering the importance of the coexistence of the micro/nanostructure and low

surface energy in obtaining a superhydrophobic surface, we have demonstrated
recently that it was possible to obtain such a surface even in a one-step process
[32,34,35,49]. In one of the one-step methods, we have reported the fabrication of
superhydrophobic copper surfaces by electrochemical modification of a copper elec-
trode in an ethanolic stearic acid solution [34]. However, with a similar procedure as
reported for copper surfaces [34], aluminum surfaces failed to show superhydropho-
bic properties. Therefore, in this study, the aluminum surfaces were initially elec-
trodeposited with copper films and further electrochemically modified by ethanolic
stearic acid solution to obtain superhydrophobic properties on aluminum surfaces.
The morphological, structural, and wetting characteristics of the composite films on
these surfaces have been discussed [29].
The cleaned aluminum substrates of AA6061 alloy were coated with copper films
using an aqueous solution of 0.1 M CuSO4 and 0.1 M NaClO4 as used by Sarkar et
al. [50] under potentials ranging from 0 to –1 V for a deposition time of 10 min. The
copper-coated aluminum alloy substrates were electrochemically modified in etha-
nolic stearic acid solution for 30 min at 30 V.
The SEM images of copper films deposited on aluminum substrates are illus-
trated in Figure 5.19a, c, e, and g. Furthermore, the SEM images of the stearic acid–
modified copper surfaces are shown in Figure 5.19b, d, f, and h. It has been found
that the electrodeposited copper films are composed of microdots of copper, which
are identified by energy dispersive X-ray (EDX) (spectrum not shown) as well as
by XRD analysis. By means of an image analysis technique, the surface density of
copper microdots (shown in Table 5.1) is found to be 4.5 × 104 cm–2, and the aver-
age distance between the microdots (nearest neighboring particles) is 26.66 μm for
(a) CA~ 24° (b) CA~ 115°
10 µm 10 µm
(c) CA~ 51° (d) CA~ 130°
10 µm 10 µm
(e) CA~ 50° (f ) CA~ 157°
10 µm 10 µm
(g) CA~ 51° (h) CA~ 157°
10 µm 10 µm
FIGURE 5.19 (Left) SEM images of copper films deposited on aluminum substrates at
(a) –0.2 V, (c) –0.4 V, (e) –0.6 V, and (g) –0.8 V for 10 min. (Right) The SEM images of
the same copper films after electrochemical modification in ethanolic (b, d, f and h) are the
stearic acid solution at 30 V for 30 min. The insets show the images of water drop on the
respective surfaces [29].
TABLE 5.1
Number Density and Distance between Microdots of Copper Film
Deposited on Aluminum Substrates at Different Potentials
Number Density
Applied DC of Cu Particles Interparticle Contact Angle
Voltage (V) (Number/cm2) Distance (μm) Roughness (μm) (Degree)
0.0 1.1 × 104 ± 2.5 × 103 51.31 ± 22.67 2.39 ± 0.23 115 ± 1
–0.2 4.5 × 104 ± 4.9 × 103 26.66 ± 9.71 3.15 ± 0.65 130 ± 5
–0.4 8.3 × 104 ± 4.6 × 103 18.69 ± 6.62 4.24 ± 0.46 143 ± 4
–0.6 2.0 × 105 ± 7.1 × 103 13.15 ± 4.00 6.20 ± 0.97 157 ± 1
–0.8 2.9 × 105 ± 2.8 × 104 11.03 ± 3.33 6.77 ± 0.70 157 ± 1
–1.0 2.4 × 105 ± 2.5 × 104 8.66 ± 2.59 7.02 ± 0.33 157 ± 1
Note: The roughness and the contact angle are also presented.
the film deposited at –0.2 V. Table 5.1 also shows that an increase in the negative
potential to –0.4 V resulted in an increase in the number density to 8.3 × 104 cm–2
and a decrease in the distance between neighboring particles to 18.69 μm. However,
the sharp increase in the number density of copper particles on aluminum substrate
appeared at –0.6 V, where the number density and interparticle distance on cop-
per film are 2 × 105 cm–2 and 13.15 μm, respectively. The particle density further
increases with an increase in negative potential, and at –0.8 V, the number density
of particles is found to be 2.9 × 105 cm–2. Moreover, the sizes of the microdots are
found to decrease with the increase in the negative deposition potential. According
to the Volmer–Weber [51] growth mode, in case of an electrochemical deposition
process, the number of the nucleation sites increases with the increase in the negative
deposition potential. Therefore, it can be concluded that the increase in the negative
potential is responsible for the increase in the number density of copper microdots
and, consequently, the decrease in the distances between them in the present study.
We have observed that the aluminum surfaces modified by ethanolic stearic acid
solution under the influence of direct current (DC) voltage did not show the super-
hydrophobic properties as we observed in the case of copper surfaces [34]. In order
to obtain superhydrophobic properties on aluminum surfaces, the copper-coated alu-
minum surfaces were further electrochemically modified in ethanolic stearic acid
solution at 30 V for 30 min. The stearic acid–modified copper microdots were dec-
orated with nanofibers resembling micro/nanostructured flowerlike particles after
stearic acid modification, as shown in Figure 5.19 (b, d, f and h). The flowerlike mor-
phologies formed on copper microdot–coated aluminum substrates are very simi-
lar to those obtained from the electrochemically modified pure copper surfaces we
reported previously [34]. Further, it is interesting to note that the surface coverage on
the stearic acid–modified surfaces is higher, covering nearly the whole area at higher
negative potential as compared to a lower coverage at lower negative potential. The
higher coverage might be due to the shorter interparticle distances between the cop-
per microdots and the coalescence of the branches of copper stearate, a composition
of copper and stearic acid, initiating from individual copper microdots in the empty
space. The insets in Figure 5.19 show the image of water drops on the respective
surfaces.
Figure 5.20a shows the XRD patterns of the copper films deposited on the aluminum
substrates under the application of different deposition potentials of –0.2 V, –0.4 V,
–0.6 V, and –0.8 V in the 2θ scan range of 35°–55°. The XRD patterns show two
characteristic peaks at 38.47° and 44.72° of Al (111) and Al (200) planes, respectively,
arising from the aluminum substrate [52]. The XRD patterns also show the two char-
acteristic peaks of Cu (111) and Cu (200) planes at 43.32° and 50.45°, respectively
(a)
1800
1600
Cu (111)
Cu (200)
Al (111)
Al (200)
1400
–0.8 V
1200
Intensity (a.u.)
1000 –0.6 V
800
–0.4 V
600
400 –0.2 V
200
Substrate
0
36 38 40 42 44 46 48 50 52 54
2θ (º)
(b)
700
600
500 –0.8 V
Intensity (a.u.)
400 –0.6 V
300 –0.4 V
–0.2 V
200
Substrate
100
CuSA[18]
0
4 6 8 10 12 14
2θ (º)
FIGURE 5.20 (a) High-angle XRD patterns of copper films deposited on aluminum sub-
strates for the duration of 10 min under the application of various potentials followed by elec-
trochemical modification in ethanolic stearic acid solution and (b) low-angle XRD patterns
of (a). CuSA is the XRD pattern of copper stearate films, as we reported. (From Huang, Y.,
Nano-Micro Lett., 3, 160, 2011; Huang, Y., Mater. Lett., 64, 2722, 2010.)
[53], arising from copper films. The observation of two distinct Cu (111) and Cu (200)
planes confirms copper deposition on aluminum substrates. The tiny peak at 43.32°
on the aluminum substrate resulted from the small amount of copper in AA6061
aluminum alloy. Figure 5.20a also shows that the peak intensity of Cu (111) increases
and that of Al (111) decreases with the increase in negative potential, showing that the
amount of copper deposition is higher at larger negative potential. This observation is
consistent with the SEM images (Figure 5.19), which show the deposition of a large
number of copper microdots with the increase in negative potential.
Figure 5.20b shows the low-angle (2θ = 3° – 11°) XRD patterns of the electro-
chemically modified copper films composed of microdots of copper on aluminum
substrates in ethanolic stearic acid solution. Figure 5.20b also displays the low-
angle XRD patterns of the aluminum substrate as well as that of copper stearate
film (CuSA) prepared by electrochemically modified copper substrate in ethanolic
stearic acid solution in our previous study [34]. The four distinct peaks present in
Figure 5.20b arise from the copper stearate [(CH3(CH2)16COO)2Cu] formation from
a reaction between stearic acid [CH3(CH2)16COOH] and copper under the applica-
tion of DC voltage. On the other hand, appearance of the broad peak at 9.45° on the
aluminum substrate might be due to the presence of some precipitate in the AA6061
aluminum alloy. Similarly, the XRD patterns of the four copper films deposited at
–0.2, –0.4, –0.6, and –0.8 V on aluminum substrates with further modification by
stearic acid also show the four characteristic peaks of copper stearate. As the cop-
per stearate consists of methylated (CH2 and CH3) components, the electrochemical
modification by stearic acid of the copper films reduces the surface energy of the
copper-coated aluminum surfaces.
Figure 5.21a and b depicts the variation of surface roughness and water CA of
electrodeposited copper film followed by electrochemical modification in ethanolic
stearic acid solution. The surface roughness of electroless deposited copper films
(applied potential equal to zero) after electrochemical modification with stearic acid
is found to be 2.39 μm, and the water CA on the surface is found to be 115°. The
surface roughness and the water CA were found to increase with an increase in the
negative deposition potential. The roughness and the water CA of the copper films
increased to 3.15 μm and 130°, respectively, for the copper films deposited at – 0.2V.
Similarly, at – 0.4 V, the roughness and water CA further increased to 4.24 μm and
143°, respectively. At – 0.6V, the roughness further increased linearly with a slope of
–2.45 μm/V to 6.20 μm, and the water CA increased to 157°, demonstrating super-
hydrophobicity on these surfaces. It can also be observed in Figure 5.21a that the
slope is reduced to –2.33 μm/V between – 0.6 and –1.0 V. The roughness obtained at
– 0.8 V was 6.77 μm, and the water CA on this surface remained the same, 157°. The
roughness increased slightly to 7.02 μm on the surface of copper films deposited at
–1.0 V; however, no further change in water CA was observed. It has been observed
that the water CA increases with an increase in the surface roughness; however, a
critical roughness of 6.2 μm was essential in our study to obtain a surface with a
water CA of 157° to exhibit superhydrophobicity. It is worthy to mention that we
have obtained superhydrophobic surfaces having roughness ranging between 6 and
7 μm (Figure 5.21b).
(a)
7
Roughness (µm) 5
2
–1.0 –0.8 –0.6 –0.4 –0.2 0.0
DC voltage (V)
(b)
160
150
Water contact angle (º)
140
130
120
110
2 3 4 5 6 7 8
Roughness (µm)
FIGURE 5.21 (a) The variation of roughness vs. deposition potential and (b) variation of
water contact angle vs. surface roughness on the stearic acid–modified copper film deposited
on aluminum surfaces. The insets of show (a) the 3-D images of the rough surfaces and (b) the
images of water drops (–0.2 V and –0.8 V). (From Huang, Y., Nano-Micro Lett., 3, 160, 2011.)
5.3.5 Superhydrophobic Aluminum Surfaces by Coating with

Monodispersed Fluorinated Silica Nanoparticles
Due to their water-repellent property, superhydrophobic surfaces have several emerg-
ing applications in a large number of fields, such as self-cleaning fabrics, anticorro-
sive industrial parts, and even surface drag reduction to reduce energy consumption
in transport systems [34,54,55]. Different techniques such as chemical etching,
CBD, electrodeposition, photolithography, plasma processes, and so forth are cur-
rently being used to create superhydrophobic surfaces [32,34,35,37]. The most sig-
nificant barrier caused by these techniques is the difficulty of large-scale depositions
in a cost-effective way. Solgel techniques together with spray coatings or commercial
paintbrush coatings might be one approach to overcome such a drawback. Silica
nanoparticles with varying sizes were prepared in the laboratory using a standard
Stöber process and further functionalized with fluoroalkylsilane (FAS-17) in an
ethanolic solution to obtain fluorinated silica nanoparticles. These fluorinated silica
nanoparticle dispersions were spin-coated on aluminum alloy substrates to obtain
superhydrophobic thin films [30,31].
The dispersions containing fluorinated silica nanoparticles of varying sizes were
prepared in the laboratory using the standard Stöber process [56]. Initially, at a hot
plate temperature of 50°C, ethanol and ammonium hydroxide (NH4OH) were mixed
in a beaker using a magnetic stir bar. While stirring, a solution of tetraethoxysilane
[TEOS; Si(OC2H5)4] is added drop by drop to the mixed solution. The transparent
mixed solution turns opaque, confirming the formation of silica nanoparticles [57].
The synthesized silica nanoparticles are further functionalized in an ethanolic fluo-
roalkylsilane or FAS-17 (C16H19F17O3Si) solution. The experimental details are pre-
sented in Table 5.2. The AA-6061 aluminum alloy and silicon substrates were coated
with the fluorinated silica nanoparticles by a spin-coating process. The nanoparticle-
coated films were dried at 70°C on a hot plate to allow the ethanol and excess water
to evaporate from the films.
FTIR analyses were carried out on the thin films deposited on silicon substrates
to develop an understanding of the interatomic bonding in the fluorinated silica
nanoparticles. Figure 5.22a and b shows the FTIR spectra of the thin films pre-
pared with silica nanoparticles and fluorinated silica nanoparticles, respectively. The
spectrum in Figure 5.22a shows only three significant peaks of silica. The strongest
peak is at 1100 cm–1, with an accompanying asymmetric broad shoulder extended to
1250 cm–1. This is due to the asymmetric stretching vibration of Si-O-Si bonds in the
silica nanoparticles in the thin films [58,59]. Another small peak appearing around
800 cm is associated with the bending mode of Si–O–Si bonds. A weak absorption
peak, as compared to the peak of Si-O-Si at 1100 cm–1, appears at 950 cm–1 due to
the presence of Si-OH bonds. Figure 5.22c shows the reaction and the model for the
formation of a spherical oligomer of TEOS molecules in the ethanol solutions in
TABLE 5.2
Chemicals for Preparation of Fluorinated Silica Nanoparticles, Their
Diameter, Roughness of Prepared Thin Films, and Water Contact
Angles on Films
NH4OH / TEOS
(Molar Ratio) with Constant Molar Diameter Roughness Contact
Test # Ratio of FAS-17/TEOS = 0.07 (nm) (µm) Angle (°)
1 2 40 ± 22 0.542 122 ± 3
2 6 69 ± 21 0.612 129 ± 3
3 8 101 ± 14 0.678 138 ± 3
4 10 119 ± 12 0.697 151 ± 4
5 12 169 ± 08 0.726 162 ± 6
6 15 300 ± 07 0.733 165 ± 5
(c) OH
OH OH
OEt
NH4OH
C-F EtO Si OEt OH Si O Si OH
Ethanol
Si-O-Si OEt
OH
Absorbance (a.u.)
OH
OH
C-F
Si-O-C TEOS
C-H
• 1100 cm–1 • 950 cm–1
C-F C-F • 800 cm–1
(b) (d)
Si-OH Si-O-Si FAS17
FAS17 FAS17
(a) H F F
FAS17 Si O Si O Si O C C C F
1200 1000 800 600
H F F
Wavenumber (cm–1) FAS17 FAS17
2 7
FAS17
• 610 cm–1
• 1145 cm–1 • 900 cm–1 • 730 cm–1
• 960 cm–1
• 1250 cm–1
FIGURE 5.22 FTIR spectra of the thin films prepared with (a) silica nanoparticles and
(b) fluorinated silica nanoparticles. (c) The reaction mechanism and the schematic model of
the formation of silica nanoparticles, the vibrational bonds, the corresponding wave numbers,
and (d) the functionalization of silica nanoparticles by FAS-17 molecules are shown, as well
as the vibrational bonds and the corresponding wave numbers. (From Brassard, J.D. et al.,
ACS Appl. Mater. Interfaces, 3, 3583, 2011.)
the presence of an NH4OH catalyst. The vibrational bands of Si-O-Si with Si-OH
and their corresponding wave numbers are also presented in the model. The FTIR
spectrum of the films prepared using fluorinated silica nanoparticles shown in
Figure 5.22b reveals distinctly the emergence of some additional new peaks as com-
pared to those prepared using only silica nanoparticles, shown in Figure 5.22a. Two
tiny peaks superimposed between 1130 cm–1 and 1250 cm–1 on the left shoulder of
the Si-O-Si peak at 1100 cm–1 are due to the stretching vibration of C-F bonds antici-
pated due to the fluorination of the silica nanoparticles by the FAS-17 molecules [45].
The existence of C-F bonds in the form of CF, CF2, or CF3 is also seen at 575, 610,
730, and 960 cm–1 [45]. A reasonably broad peak around 900 cm–1 is assigned to the
C-H bonds, arising from FAS-17 molecules [60]. A peak approximately at 1145 cm–1
is due to the Si-O-C bond that arises due to the link between FAS-17 molecules and
the silica nanoparticles [61], as shown in our model in Figure 5.22d. The presence of
additional new peaks arising from the fluorinated functional groups present in the
films confirms that the nanoparticles of silica are functionalized by fluorine from the
FAS-17 molecules. Figure 5.22d shows the schematic model of the functionalization
and the formation of fluorinated silica nanoparticles as decoded from the FTIR spec-
trum of Figure 5.22b. Every H atom from each OH bond around the silica nanopar-
ticles of Figure 5.22c is replaced by a Si atom of the FAS-17 molecule, as presented
in the model of Figure 5.22d. The new vibrational bands of Si-O-C, C-H, and C-F
and their corresponding wave numbers are also presented in the figure.
Table 5.2 shows the different NH4OH/TEOS ratios used in the preparation of the
thin films of fluorinated silica nanoparticles with a constant FAS-17/TEOS molar
ratio of 0.07 using the standard Stöber process to prepare silica nanoparticles fol-
lowed by modification using FAS-17 molecules. Table 5.2 also summarizes the infor-
mation on the size of the prepared nanoparticles, which varies from 40 to 300 nm
with an increase in the molar ratio from 2 to 15. The size of the silica nanopar-
ticles is found to increase with the increase in the molar ratio of NH4OH/TEOS.
Also, it is clear from Table 5.2 that both roughness and water CAs increase with an
increase in the size of the fluorinated silica nanoparticles in the films. Figure 5.23a
through c shows the SEM images of three different films prepared from three dif-
ferent sizes of monodispersed fluorinated silica nanoparticles on aluminum alloy
surfaces. Figure 5.23a shows that the size of the nanoparticles prepared with a molar
ratio of NH4OH/TEOS of 6 is 69 ± 22 nm. Figure 5.23b shows that the particles size
increased to 119 ± 12 nm with the increase in NH4OH to an NH4OH/TEOS molar
ratio of 10. Subsequently, Figure 5.23c shows that the size of the nanoparticles is
further increased to 169 ± 8 nm with an increase in the molar ratio of NH4OH/TEOS
to 12. Nozawa et al. [62] obtained 110 ± 20 nm silica nanoparticles using the Stöber
process at an NH4OH/TEOS molar ratio of 10; however, in our study, the size of the
nanoparticles obtained under the same conditions is ~10 nm larger (119 ± 12 nm)
than that obtained by Nozawa et al. The larger size in our case is attributed to the
presence of an outer shell of fluorinated silane bonded around the silica spheres as
(a) (b)
500 nm 500 nm
(c) 180 (d)

Constant angle (º)
165
150
135
120
0 80 160 240 320
500 nm Diameter (nm)
FIGURE 5.23 SEM images of fluorinated silica nanoparticles of sizes (a) 69 ± 22 nm,
(b) 119 ± 12 nm, and (c) 169 ± 8 nm. (d) The water contact angles on the thin films prepared
from fluorinated silica nanoparticles of varying sizes. (From Brassard, J.D. et al., ACS Appl.
Mater. Interfaces, 3, 3583, 2011.)
a result of fluorination using FAS-17 molecules as presented in the schematic model

shown in Figure 5.22d.
Figure 5.23d shows the variation in water CA with the size of the fluorinated silica
nanoparticles in the thin films. Figure 5.23d shows that the water CAs of the films
prepared with fluorinated silica nanoparticles of sizes 40 ± 22 nm and 69 ± 21 nm
are 122° ± 3° and 129 ± 3°, respectively. These values are slightly higher than the
CA value of 108° obtained on a FAS-17 passivated flat glass surface [63]. The slight
increase, although not great, in the CA values can be attributed to the presence of
low-surface-energy fluorinated species arising from the functionalization of the sil-
ica nanoparticles. We have observed that the water CA reached 138° ± 3° on the sur-
face of the thin films having fluorinated silica nanoparticles larger than 101 ± 14 nm.
Superhydrophobicity is defined here as a CA of more than 150°. This experimental
series showed that superhydrophobicity (151° ± 4°) is reached on the films having flu-
orinated silica nanoparticles of a critical size of 119 ± 12 nm. The water CA increased
to 162° ± 6° by increasing the size of the fluorinated silica nanoparticles to 169 ± 8
nm in the films. As the size of the fluorinated silica nanoparticles increased further
to 300 ± 7 nm, the water CA further increased to 165° ± 5°. Yang et al. [64] show
a similar tendency in the variation of water CAs of CH3-terminated (using methyl-
triethoxysilane) silica nanoparticles. These authors reported a water CA of 135° on
the surface of the films having CH3-terminated silica nanoparticles of sizes between
30 and 50 nm. In their study, the critical size of the CH3-terminated silica nanopar-
ticles to obtain superhydrophobic films was reported to be 200 nm. Furthermore,
their study shows that increasing the size of the CH3-terminated silica nanoparticles
to 300 nm did not further influence the water CA. Cao et al. [65] reported the behav-
ior of water CA of the thin polymer films prepared using organosilane-modified
silica particles of various diameters (20 nm, 50 nm, 100 nm, 1 μm, 10 μm, and 20
μm). These authors reported a water CA greater than 150° for the polymer films
loaded with organosilane-modified silica particles of initial size of only 20 nm.
With an increase in their particle size to 50 and 100 nm, the CA increased to ~158°
and remained similar in both cases. However, with further increase in their particle
size to more than 100 nm, the authors reported a decrease in CA. However, these
authors did not show microscopic evidence to confirm the size of their organosilane-
modified silica particles in the films. The sizes of their organosilane-modified silica
nanoparticles in their films might actually be much larger than reported, leading to
sufficient surface roughness to provide high water CAs. Nevertheless, our observa-
tion of increase in water CA with an increase in size of the functionalized silica
nanoparticles is consistent with the observations by Yang et al. as well as Cao et
al. Such an increase in the water CA with the increase in the size of the spherical
nanoparticles could be attributed to a large amount of air entrapment between the
particles forming a composite microstructure of solid particles and air that reduces
the effective contact area of water to the surfaces [66].
A model has been presented in Figure 5.24 to explain the increase in rough-
ness with the increase in the size of the particles. Figure 5.24a shows a schematic
of a one-dimensional growth of spherical particles on a surface. In this model, we
consider these particles as perfect circles. The shaded areas between the circles are
responsible for providing roughness to the films due to the spherical form of particles
(a) R = 0 Rn R2 R1 R (b)
50
y z
40
x (II)
(I)
RMS (nm)
30
20
y z 10
(I) (II) (III) 0

x 0 80 160 240 320
Radius Diameter (nm)
(c)
(d) (e)
1.5
1.0
Depth (µm)
0.5
100 µm
0.0
–0.5
–1.0
0 20 40 60 80 100
Position (µm)
FIGURE 5.24 (a) A schematic model of 1-D growth of perfectly spherical particles on a
surface. The shaded area between particles (I), (II), and (III) is responsible for the entrapment
of air as well as for the roughness. Particles (I) and (II) have been redrawn in larger diameters
on the top part of (a), and several vertical lines are drawn with varying lengths from R to 0.
(b) Plot of the calculated roughness with the increase in the diameter of the spherical particles
as in (a). (c) More realistic model of our thin films. (d) 3-D image as obtained from optical
profilometry. (e) 1-D cross section of the image. (From Brassard, J.D. et al., ACS Appl. Mater.
Interfaces, 3, 3583, 2011.)
on the surface. Figure 5.24a (top) also shows a shaded area between the two circles.
A horizontal line (tangent to both circles) is drawn on the top of the two circles. This
shaded area is separated into two parts, each associated with a circle. Considering
the left circle and the shaded part associated with it, we can calculate the roughness
arising from this part. Several vertical lines have been drawn from the horizontal
line, as shown in Figure 5.24a. It is clear from the figure that the lengths of the verti-
cal lines vary between R and 0 from right to left. In this figure, R is the radius of the
circle (or sphere). In our calculation, the left vertical length decreases by one unit
from the adjacent right vertical length for each step. Therefore, we generate several
vertical lengths, Ri, between R and 0. The generated verticals lengths are R, R-1,
R-2, ....., 0, and their average length, R, is R/2. Utilizing the statistical analysis as
shown in Equation 5.7, we have determined the rms of these numbers. This calcu-
lated rms is considered as the roughness of the surface in our model.

1
n ∑[R − R ]
i =1
i
2
(5.7)
Using this calculation method, we have derived the roughness of all surfaces
based on their varying particle sizes measured experimentally. Figure 5.24b shows
a linear relation between the roughness and the diameter of the particles. The value
of the calculated roughness is much smaller than the value obtained experimentally
(Table 5.2). Figure 5.24c shows a more realistic model surface that is composed
of particles, clusters, and cracks on the films prepared with a spin-coating process
using the spherical nanoparticles. This model surface is comparable with the SEM
images of Figure 5.23a through c showing agglomerated spherical nanoparticles as
well as several cracks. Therefore, the measured roughness of the prepared thin film
surfaces is much higher than that from our one-dimensional model of film surfaces
with spherical particles.
It has to be emphasized that the value of the rms roughness is associated with the
area between two circles (contour XYZ). It is obvious that the area of this region also
increases with an increase in the particle size. According to the Cassie–Baxter equa-
tion for an air trapped rough surface, cos (θ′) = f1 cos (θ1) + f 2 cos (θ2), where f1 and
f 2 are, respectively, a fraction of solid and fraction of air, and similarly, θ1 and θ2 are,
respectively, CAs of water with solid and air. Assuming that f1 = 1– f 2, and θ2 is 180°
in air, this equation can be rewritten in the form of cos (θ′) = (1 − f 2) cos(θ1) − f 2 [4].
It is evident from this equation that the larger the fraction of air ( f 2), the greater the
CA. According to our experimental results and mathematical calculation, thin films
with larger particles provide higher roughness and higher area of air entrapment ( f 2),
and thus, an increase in surface roughness and CA with an increase in size of the
fluorinated silica nanoparticles in our films. A similar observation has been reported
by Synytska et al. [67] on polymer particles dispersed thin films.
The profilometry data have been used to determine the fraction of solid to explain
the experimentally observed CAs. Figure 5.24d shows a three-dimensional image
of the thin films obtained by the optical profilometry technique; the thin films were
prepared using 119 ± 12 nm fluorinated silica nanoparticles. (A SEM image of the
film is shown in Figure 5.23b.) Obviously, the resolution of this image, due to light
source, is much lower than the SEM image. Figure 5.24e shows a one-dimensional
cross-sectional image of Figure 5.24d. A line parallel to the x-axis has been drawn
in Figure 5.24d to show the position of the water drop on the surface. The fraction
of the solid ( f1 in the Cassie–Baxter model) has been calculated as the sum of
the lengths of the lines between the individual peaks of Figure 5.24d. Ten cross-
sectional data plots have been used to determine the fraction of the solid ( f1), and it
has been found to be 0.18 ± 0.05. This f1 value along with the θ1 value, which is con-
sidered to be equal to 108° for a smooth FAS-17–coated surface [28], has been put in
the Cassie–Baxter model to calculate the CA. This calculated CA has been found to
be 151° ± 8°, which matches very well with our observed CA for the films prepared
with 119 ± 12 nm fluorinated silica nanoparticles. Though most of the literature on
superhydrophobic surfaces has used the Cassie–Baxter model to explain wetting
properties, recently, Lui et al. [68] proposed a mathematical model to explain the
superhydrophobic phenomenon on high-energy surfaces considering a sinusoidal
shape of the surface. This model has been used by Sahoo et al. [69] to explain the
superhydrophobic behavior of their GaN thin film surface prepared using plasma
enhanced chemical vapor deposition.
We have demonstrated that the solgel technique is an appropriate method to pre-
pare fluorinated silica nanoparticles and fabricate superhydrophobic surfaces by a
spin-coating process. The same fluorinated silica nanoparticles can be used on a
large scale to prepare superhydrophobic surfaces by a spray-coating method with the
proper control to achieve similar roughness for industrial applications.
5.3.6 Application of Superhydrophobic Aluminum

Surface in Ice Adhesion Reduction
The accretion of ice on the surfaces of power network systems, aircraft, commu-
nication networks, and so forth is known to cause serious problems that often lead
to costly safety issues. The strong adhesion of ice to materials is due mainly to the
properties of the ice–solid interface, where polar water molecules strongly interact
with the solid. Among the various processes involved in ice adhesion, it is the elec-
trostatic forces between charges at the ice surface and mirror charges induced in
the solid substrate that have been found to largely dominate the adhesion of ice to
surfaces [70]. The use of materials with a very low dielectric constant would reduce
the electrostatic interactions at the ice–dielectric interface and result in significant
reduction in the adhesion strength of ice. The interactions of both water and ice with
a surface should be similar; therefore, the adhesion strength of ice would be expected
to be lower against surfaces that water would not wet. In other words, if a surface is
fabricated in such a way that a water droplet does not wet it (i.e., it is a superhydro-
phobic surface), ice also would not be expected to adhere to it.
In this section, we report that EA surfaces rendered superhydrophobic by the
application of a Teflon coating by rf sputtering show icephobicity with zero ice adhe-
sion under freezing conditions. The effect of the dielectric constant was also inves-
tigated by coating the nanostructured aluminum surfaces with materials of different
dielectric constants [36].
The presence of TiO2, ZnO, and rf-sputtered Teflon on the nanopatterned alumi-
num surfaces was confirmed from the XPS spectra. Figures 5.25 and 5.26 show the
survey spectra of TiO2 and ZnO solgel coatings, respectively. The survey spectrum
of TiO2 (Figure 5.25) shows strong signals of Ti2p and O1s as well as strong Ti LMM
Auger peak with Ti2p3/2 and Ti2p1/2 peaks at 458.85 eV and 464.53 eV, respectively,
and the O1s corresponding to TiO2 at 530.1 eV.
Similarly, the survey spectrum of ZnO in Figure 5.26 shows strong signals of
Zn2p and O1s as well as a strong Zn LMM Auger peak with the Zn2p3/2 peak at
1021.6 eV and the O1s peak of ZnO at 530.28 eV.
8.0
7.0
6.0
5.0
Intensity (a.u.)
O KLL
O1s
Ti LMM
4.0
Ti2p
3.0
Ti2s
2.0
C 1s
1.0
0.0
1200 1000 800 600 400 200 0
Binding energy (eV)
FIGURE 5.25 XPS survey spectrum of solgel TiO2 coating. (From Saleema, N. et al.,
J. Adhesion Sci. Technol., 25, 27, 2011.)
6.0
Zn2p3/2
5.0
Zn2p1/2
Intensity (a.u.)
4.0
Zn LMM
3.0
O KLL
2.0
O1s
1.0
0.0
1200 1000 800 600 400 200 0
Binding energy (eV)
FIGURE 5.26 XPS survey spectrum of solgel ZnO coating. (From Saleema, N. et al.,
J. Adhesion Sci. Technol., 25, 27, 2011.)
Wetting properties were investigated using CA measurements. Water droplets

were found to spread completely on the rough coral-like nanostructure of the EA
surfaces, demonstrating hydrophilicity. A similar behavior was observed on ZnO-
coated EA (ZCEA) surfaces. A small CA of ~25° was measured on the TiO2-coated
EA (TCEA) surfaces, still showing hydrophilic behavior. The reason for TCEA pro-
viding a certain value for CA could be the presence of some contamination on the
TCEA surface or the difference in dielectric constant as compared to ZnO. However,
with a CA value well below 90°, all these surfaces fall into the hydrophilic regime.
All these surfaces presenting certain nanoroughness exhibited superhydrophobicity
when coated with low-surface-energy rf-sputtered Teflon, providing CAs >160° and
CAH values <5°. Due to the very low CAH, the water droplets were found to roll off
the superhydrophobic EA surfaces, even at the slightest tilt of the sample. A water
CA of >160° and CAH of <5° were obtained on Teflon-coated EA. There was no
significant change in the CA data or the water roll-off behavior when the EA surfaces
were coated with ZnO or TiO2 prior to rf-sputtered Teflon coating. These observa-
tions demonstrate the importance of the coexistence of surface roughness and low-
surface-energy coating in obtaining superhydrophobic properties, as emphasized in
previous studies [25,26,28,49,71].
Hydrophilic and superhydrophobic samples were tested for icephobicity by
exposing the surfaces to simulated freezing rain in a wind tunnel. Visual analyses
showed that the ice formed and continued to accumulate on the hydrophilic samples
as soon as the supercooled water droplets came into contact with their surfaces. The
superhydrophobic surfaces, either with or without the oxide coatings, showed no ice
formation and, therefore, no accumulation of ice from the freezing drizzle. Petrenko
and Peng [72], in their study on the adhesion of ice to hydrophilic and hydrophobic
self-assembled monolayers (SAMs) on gold surfaces, showed that the shear adhesion
strength of ice decreased with an increase in hydrophobicity. The study of Somlo and
Gupta [73] on ice adhesion to various SAMs coated onto aluminum surfaces found a
reduction in the tensile adhesion strength compared to the bare aluminum surfaces,
emphasizing the importance of low-surface-energy materials, one of the two impor-
tant factors in enhancing hydrophobic properties. Surface roughness, being the other
important factor in enhancing the water CA on surfaces, has also been investigated
for its role in ice adhesion. However, as reported by Saito et al. [74], an increase in
surface roughness led to an increase in ice adhesion. Recent studies have shown that
a combination of surface roughness and a low-surface-energy coating leads to very
high water CAs and greatly reduces the adhesion strength of ice [37]. However, so
far, there are no reports claiming the existence of an icephobic surface on which ice
would not stick at all. Figure 5.27 shows photographs of our superhydrophobic (with
rf-sputtered Teflon coating) and hydrophilic (with no rf-sputtered Teflon coating)
dielectric-coated EA surfaces (ZCEA in figure) before and after ice accumulation
Superhydrophobic Hydrophilic Superhydrophobic Hydrophilic

nanorough dielectric nanorough dielectric nanorough dielectric nanorough dielectric
FIGURE 5.27 (Left) Sample surfaces (ZCEA in figure) before ice accumulation.
(Right) Superhydrophobic and hydrophilic sample surfaces after ice accumulation in the
wind tunnel. (From Saleema, N. et al., J. Adhesion Sci. Technol., 25, 27, 2011.)
in the icing wind tunnel. It is clear from Figure 5.27 that ice hardly forms on super-
hydrophobic surfaces as the quantity of ice accumulated is much smaller on the
superhydrophobic surfaces than on the hydrophilic surfaces for the same period of
accumulation under the same conditions. The smaller quantity of ice seen stuck at
the corners may be due to surface defects or due to the untreated sample supports
favoring ice formation, thereby providing nucleation sites for ice growth. Since the
evidence suggests that ice does not adhere to these superhydrophobic surfaces under
the conditions investigated, we can tentatively conclude that the strength of ice adhe-
sion to such surfaces as evaluated here is effectively zero.
The hydrophilic samples were further tested using the CAT apparatus to evaluate
the shear adhesion strength of ice on these surfaces. Bare aluminum was tested sev-
eral times with the CAT apparatus before testing the samples to compare with litera-
ture values, and the measured adhesion strength of 420 ± 27 kPa was found to be in
good agreement with values reported in the literature [75]. The ice adhesion strength
obtained on the nanorough aluminum surface was found to be 466 ± 122 kPa.
Similarly, the ice adhesion strengths on ZnO (ε = ~8) and TiO2 (ε = ~80) coated
on EA (~482 and ~480 kPa, respectively) are similar and comparable to the value
obtained on bare aluminum with the measurement errors taken into consideration
(Table 5.3). These values obtained using CAT tests and the visual investigation in
the wind tunnel show that the adhesion of ice is stronger on hydrophilic surfaces and
weaker on superhydrophobic surfaces, as evident from Table 5.3, which summarizes
the ice adhesion strength values along with water CA and CAH values. The “***”
in Table 5.3 indicates that these data were unobtainable due to experimental limi-
tations. Weaker adhesion strength of ice on superhydrophobic surfaces shows that
TABLE 5.3
Contact Angle Data, Adhesion Strength, and Dielectric Constant Values
Static Water Contact Angle Ice Adhesion Dielectric
Sample Contact Angle (°) Hysteresis (°) Strength (kPa) Constant
BA ~74 *** 420 ± 27 –
ZCBA ~31 *** 450 ± 101 ~8
TCBA ~43 *** 460 ± 75 ~80
PCBA ~110 *** 188 ± 12 ~2
EA 0 *** 466 ± 122 –
ZCEA 0 *** 482 ± 125 ~8
TCEA ~25 *** 480 ± 60 ~80
PCEA, PZCEA, PTCEA >160 <5 *** ~2
Note: Abbreviations: BA, bare aluminum; EA, etched aluminum; PCBA, polytetrafluoroethelyne or
Teflon (PTFE)-coated bare aluminum; PCEA, PTFE-coated etched aluminum; PTCEA,
PTFE-coated TCEA; PZCEA, PTFE-coated ZCEA; TCBA, TiO2-coated bare aluminum;
TCEA, TiO2-coated etched aluminum; ZCBA, ZnO-coated bare aluminum; ZCEA, ZnO-
coated etched aluminum.
*** These data were unobtainable.
on these surfaces, not only the water droplets at room temperature are repelled but
also the supercooled water droplets. Sarkar et al. [26] have previously shown that at
room temperature, water droplets roll off the superhydrophobic aluminum surfaces
prepared via a similar chemical etch process followed by rf-sputtered Teflon coating.
The effect of the dielectric constant on ice adhesion was also evaluated in this
study by comparing the ice adhesion on EA surfaces coated with three different
dielectrics: TiO2 (ε = ~80), ZnO (ε = ~8), and Teflon (ε = ~2). Studies indicate that
materials with a lower dielectric constant will have a lower adhesion to ice due to
screening of mirror charges, expressed as
ε −1
q′ = q (5.8)
ε +1
where q′ is the image charge produced on the dielectric surface by a charge q on
the ice surface and ε is the dielectric constant of the insulator [70]. TCEA (ε = ~80)
showed a very high ice adhesion strength of 480 ± 60 kPa. When the EA was coated
with the lower-dielectric-constant material ZnO (ZCEA; ε = ~8), the adhesion strength
(482 ± 125 kPa) remained high. However, when the EA was coated with rf-sputtered
Teflon (PCEA), with the lowest dielectric constant (ε = ~2), the ice adhesion as dis-
cussed above was found to be almost zero. Similar behavior was found when the
dielectric-coated EA surfaces were further coated with rf-sputtered Teflon.
In order to eliminate the effect of roughness on the surfaces tested, we performed
the same experiments on flat (bare) aluminum surfaces coated with ZnO, TiO2, or
rf-sputtered Teflon. From Table 5.3, it can be seen that flat ZnO (ZCBA) and TiO2
(TCBA) surfaces showed a high and comparable adhesion strength of 450 ± 101 kPa
and 460 ± 75 kPa, respectively, both comparable to 420 ± 27 kPa obtained on bare
aluminum. Therefore, varying ε from ~80 (TiO2) to ~8 (ZnO) showed no effect on
ice adhesion; however, the adhesion strength was considerably lowered (~188 kPa)
when bare aluminum surfaces were coated with rf-sputtered Teflon, with a dielectric
constant as low as ε = ~2. Therefore, the dielectric constant of a material is found
to have an influence on ice adhesion; however, a change in the dielectric constant
from 80 to 8 does not have a significant effect, although a reduction in the dielectric
constant to 2 does have effect. The experimental observation is complemented by
the theoretical prediction in the literature [70] based on mirror charges. The mir-
ror charges q′ on ZnO (ε = ~8) and TiO2 (ε = ~80) are found to be 0.78q and 0.98q,
respectively, based on theoretical predictions. However, for a dielectric with ε = ~2
(Teflon), the image charge q′ is predicted to be only 0.33q, which is one-third of the
charge on ice as compared to the other two dielectrics studied, for which q′ is nearly
the same as q.
Another interesting observation in this study is that an EA surface becomes
superhydrophobic once a coating of Teflon is applied over it, irrespective of the pres-
ence or absence of an intermediate layer of TiO2 or ZnO. Teflon has a very low
surface energy of ~18mJ/m2 [4,8] and, therefore, has very low affinity to water. The
interaction depth of water with low-surface-energy materials as reported by Bain
and Whitesides is only ~0.5 nm [76]. Therefore, water and, hence, ice do not “sense”
the substrate material underlying the Teflon layer, which nearly completely masks
the substrate in the XPS spectra, indicating a thickness of only a few nanometers.
5.4 CONCLUSIONS
In this review, we have addressed the various methods for fabricating superhydro-
phobic aluminum surfaces, which comprise (1) the traditional two-step method,
where a certain roughness is first created on the surface by different means, for
example, chemical etching, which is then followed by surface modification by pas-
sivation with low-surface-energy molecules or by rf sputtering of Teflon; and (2) the
new one-step method where the surface roughness and low surface energy are simul-
taneously acquired on the surface by a one-step modification. The different surfaces
prepared have been characterized using various surface characterization tools for
morphological and chemical analyses as well as for wetting analysis. The surface
analyses of the various superhydrophobic aluminum surfaces revealed the creation
of a morphology and chemistry that was comparable to that of lotus leaves’ surfaces,
where the superhydrophobic properties are owed to their specific rough structure
with entrapped air covered with low-surface-energy waxy coatings. The Cassie–
Baxter and Wenzel models, which describe the importance of roughness mathemati-
cally for obtaining superhydrophobic surfaces, have been addressed and discussed
under different cases. In this review, we have reported that surfaces with varying
surface chemistries (i.e., surface energies) obtained as a result of the use of differ-
ent passivation methods, such as stearic acid or FAS-17 molecules, require different
surface roughnesses to result in superhydrophobic properties with water CA values
greater than 150°. This indicates that there is no specific roughness that should be
claimed as a critical value to obtain superhydrophobicity as it depends on the surface
energy. The lower the surface energy, the lower the surface roughness sufficient to
obtain superhydrophobicity. The applications of such highly water-repellent surfaces
are vast, with corrosion resistance, aerodynamic drag reduction, and ice repellence
being commonly cited by the superhydrophobic community. We have presented
one of these issues and showed promises of superhydrophobic surfaces in ice adhe-
sion reduction by bringing the importance of low dielectric constant of the surface.
However, ice adhesion reduction depends on factors more than just dielectric con-
stant, the most important parameters being surface roughness and surface energy,
which also relate to superhydrophobicity. More work is underway in addressing other
issues such as corrosion and drag reduction. In summary, in this review, we have
provided a flavor of different feasible methods for fabrication of superhydrophobic
aluminum surfaces, pointing out the parameters influencing the wetting behavior on
different surfaces with different surface properties.
REFERENCES
1. T. Young, Phil. Trans. R. Soc. Lond. 95, 65 (1805).
2. R. Blossey, Nat. Mater. 2, 301 (2003).
3. R. N. Wenzel, Ind. Eng. Chem. 28, 988 (1936).
4. A. B. D. Cassie and S. Baxter, Trans. Faraday Soc. 40, 546 (1944).
5. R. E. Johnson Jr. and R. H. Dettre, Adv. Chem. Ser. 43, 112 (1964).
6. D. Quéré, Rep. Prog. Phys. 68, 2495 (2005).
7. S. T. Raphail, Master of Science in Media Arts and Sciences, Massachusetts Institute of

Technology.
8. W. Barthlott and C. Neinhuis, Planta 202, 1 (1997).
9. P. Gould, Mater. Today 6, 44 (2003).
10. L. Feng, S. Li, Y. Li, H. Li, L. Zhang, J. Zhai, Y. Song, B. Liu, L. Jiang and D. Zhu, Adv.
Mater. 14, 1857 (2002).
11. X. Gao and L. Jiang, Nature 432, 36 (2004).
12. Y. Zheng, X. Gao and L. Jiang, Soft Matter 3, 178 (2007).
13. A. Scardino, R. De Nys, O. Ison, W. O’Connor and P. Steinberg, Biofouling 19, 221
(2003).
14. J. D. J. S. Samuel, P. Ruther, H.-P. Frerichs, M. Lehmannc, O. Paulb and J. Ruhe, Sensor
Actuator B 110, 218 (2005).
15. A. Singh, L. Steely and H. R. Allcock, Polym. Prepr. (ACS, Div. Polym. Chem.) 46, 599
(2005).
16. H. Gau, S. Herminghaus, P. Lenz and R. Lipowsky, Science 283, 46 (1999).
17. T. Liu, Y. Yin, S. Chen, X. Chang and S. Cheng, Electrochim. Acta 52, 3709 (2007).
18. K. Satoh and H. Nakazumi, J. Sol–Gel Sci. Technol. 27, 327 (2003).
19. T. Kako, A. Nakajima, H. Irie, Z. Kato, K. Uematsu, T. Watanabe and K. Hashimoto,
J. Mater. Sci. 39, 547 (2004).
20. A. Nakajima, K. Hashimoto and T. Watanabe, Monatsh. Chemie 132, 31 (2001).
21. M. Ma and M. R. Hill, Curr. Opin. Colloid Interface Sci. 11, 193 (2006).
22. X. Zhang, F. Shi, J. Niu, Y. Jiang and Z. Wang, J. Mater. Chem. 18, 621 (2008).
23. X.-M. Li, D. Reinhoudt and M. Crego-Calama, Chem. Soc. Rev. 36, 1350 (2007).
24. J. Genzer and K. Efimenko, Biofouling 22, 339 (2006).
25. N. Saleema and M. Farzaneh, Appl. Surf. Sci. 254, 2690 (2008).
26. D. K. Sarkar, M. Farzaneh and R. W. Paynter, Mater. Lett. 62, 1226 (2008).
27. D. K. Sarkar, M. Farzaneh and R. W. Paynter, Appl. Surf. Sci. 256, 3698 (2010).
28. A. Safaee, D. K. Sarkar and M. Farzaneh, Appl. Surf. Sci. 254, 2493 (2008).
29. Y. Huang, D. K. Sarkar and X.-G. Chen, Nano-Micro Lett. 3, 160 (2011).
30. J. D. Brassard, D. K. Sarkar and J. Perron, ACS Appl. Mater. Interfaces 3, 3583 (2011).
31. J. D. Brassard, D. K. Sarkar and J. Perron, Adv. Mater. Res. 409, 469 (2012).
32. N. Saleema, D. K. Sarkar, R. W. Paynter and X.-G. Chen, ACS Appl. Mater. Interfaces
2, 2500 (2010).
33. N. Saleema, D. K. Sarkar, D. Gallant, R. W. Paynter and X.-G. Chen, ACS Appl. Mater.
Interfaces 3, 4775 (2011).
34. Y. Huang, D. K. Sarkar and X.-G. Chen, Mater. Lett. 64, 2722 (2010).
35. D. K. Sarkar and N. Saleema, Surf. Coat. Technol. 204, 2483 (2010).
36. N. Saleema, M. Farzaneh, R. W. Paynter and D. K. Sarkar, J. Adhesion Sci. Technol. 25,
27 (2011).
37. D. K. Sarkar and M. Farzaneh, J. Adhesion Sci. Technol. 23, 1215 (2009).
38. J. M. Montero-Moreno, M. Sarret and C. Müller, Surf Coat. Technol. 201, 6352 (2007).
39. V. Kapaklis, A. Georgiopoulos, P. Poulopoulos and C. Politis, Physica E 38, 44 (2007).
40. E. Honoso, S. Fujihara, I. Honma and H. Zhou, J. Am. Chem. Soc. 127, 13458 (2005).
41. D. K. Sarkar and M. Farzaneh, Appl. Surf. Sci. 254, 3758 (2008).
42. X. Fu and X. He, Appl. Surf. Sci. 255, 1776 (2008).
43. H. A. Al-Abadleh and V. H. Grassian, Langmuir 19, 341 (2003).
44. J. van den Brand, S. Van Gils, P. C. J. Beentjes, H. Terryn and J. H. W. de Wit, Appl. Surf.
Sci. 235, 465 (2004).
45. A. Hozumi and O. Takai, Appl. Surf. Sci. 103, 431 (1996).
46. O. Takai, A. Hozumi and N. Sugimoto, J. Non-Cryst. Solids 218, 280 (1997).
47. H.-J. Jeong, D.-K. Kim, S.-B. Lee, S.-H. Kwon and K. Kadono, J. Colloid Interface Sci.
235, 130 (2001).
48. H.-J. Song and X.-Q. Shen, Surf. Interface Anal. 42, 165 (2010).
49. D. K. Sarkar and R. W. Paynter, J. Adhesion Sci. Technol. 24, 1181 (2010).
50. D. K. Sarkar, X. J. Zhou, A. Tannous and K. T. Leung, J. Phys. Chem. B 107, 2879
(2003).
51. M. Volmer and A. Z. Weber, Z. Phys. Chem. 119, 277 (1926).
52. Joint Committee on Powder Diffraction Standards Al (01-085-1327).
53. Joint Committee on Powder Diffraction Standards Cu (01-085-1326).
54. G. McHale, N. J. Shirtcliffe, C. R. Evans and M. I. Newton, Appl. Phys. Lett. 94, 6
(2009).
55. N. J. Shirtcliffe, G. McHale, S. Atherton and M. I. Newton, Adv. Colloid Interface Sci.
161, 124 (2010).
56. W. Stöber, A. Fink and E. Bohn, J. Colloid Interface Sci. 26, 62 (1968).
57. J. Bravo, L. Zhai, Z. Wu, R. E. Cohen and M. F. Rubner, Langmuir 23, 7293 (2007).
58. D. K. Sarkar, D. Brassard, M. A. El Khakani and L. Ouellet, Thin Solid Films 515, 4788
(2007).
59. A. Hozumi and O. Takai, Thin Solid Films 303, 222 (1997).
60. S. S. Latthe, H. Imai, V. Ganesan and A. V. Rao, Appl. Surf. Sci. 256, 217 (2009).
61. B. Stuart, Infrared Spectroscopy: Fundamentals and Applications. p. 224, John Wiley,
Chichester (2004).
62. K. Nozawa, H. Gailhanou, L. Raison, P. Panizza, H. Ushiki, E. Sellier, J. P. Delville and
M. H. Delville, Langmuir 21, 1516 (2004).
63. H. Li, X. Wang, Y. Song, Y. Liu, Q. Li, L. Jiang and D. Zhu, Angew. Chem. Int. Ed. 40,
1743 (2001).
64. H. Yang, P. Pi, Z.-Q. Cai, X. Wen, X. Wang, J. Cheng and Z.-R. Yang, Appl. Surf. Sci.
256, 4095 (2010).
65. L. Cao, A. K. Jones, V. K. Sikka, J. Wu and D. Gao, Langmuir 25, 12444 (2009).
66. A. Carré and K. L. Mittal (Eds.), Superhydrophobic Surfaces. VSP/Brill, Leiden (2009).
67. A. Synytska, L. Ionov, S. Minko, M. Motornov, K.-J. Eichhorn, M. Stamm and K.
Grundke, Polym. Mater. Sci. Eng. 90, 624 (2004).
68. J. L. Liu, X. Q. Feng, G. Wang and S. W. Yu, J. Phys.: Condens. Matter 19, 356002
(2007).
69. P. Sahoo, S. Dhara, S. Dash, B. Raj, I. Manna and A. K. Tyagi, Appl. Phys. Lett. 98,
043103 (2011).
70. I. A. Ryzhkin and V. F. Petrenko, J. Phys. Chem. B 101, 6267 (1997).
71. P. F. Rios, H. Dodiuk, S. Kenig, S. McCarthy and A. Dotan, J. Adhesion Sci. Technol. 21,
399 (2007).
72. V. Petrenko and S. Peng, Can. J. Phys. 81, 387 (2003).
73. B. Somlo and V. Gupta, Mech. Mater. 33, 471 (2001).
74. H. Saito, K. Takai and G. Yamauchi, Surf. Coat. Intl. 80(4), 168 (1997).
75. M. Javan-Mashmool, C. Volat and M. Farzaneh, Hydrol. Process. 20, 645 (2006).
76. C. D. Bain and G. M. Whitesides, J. Am. Chem. Soc. 110, 5897 (1988).
Section II
Surface Modification of Polymers
6 Surface Modification
of ABS for Electroless
Copper Plating without
Chromic Acid Etching
Norihiro Inagaki*
CONTENTS
6.1 Introduction................................................................................................... 128
6.2 Experimental................................................................................................. 129
6.2.1 Materials............................................................................................ 129
6.2.2 Hydrolysis of ABS Plates and Electroless Copper Plating Process........ 130
6.2.3 IR Spectra of the Hydrolyzed ABS Material Surfaces...................... 130
6.2.4 XPS Spectra of the Hydrolyzed ABS Surfaces................................. 130
6.2.5 Contact Angle of Hydrolyzed ABS Surfaces.................................... 131
6.3 Results and Discussion.................................................................................. 131
6.3.1 Hydrolysis of ABS Surfaces by Aqueous NaOH Solution................ 131
6.3.2 Electroless Copper Plating on Hydrolyzed ABS Surfaces................ 135
6.3.3 Acceleration of Hydrolysis Process of ABS Surfaces....................... 136
6.4 Conclusions.................................................................................................... 138
References............................................................................................................... 138
Hydrolysis conversion of nitrile groups at ABS surfaces into amide and car-
boxylate groups was investigated in order to seek a new recipe for the electro-
less copper plating process of ABS surfaces without etching using chromic
acid. Hydrolysis conversion of nitrile groups was successfully carried out in
a NaOH (70 wt%) solution at 80°C for more than 72 h, and nitrile groups
were modified into amide and carboxylate groups. The hydrolysis modification
was accelerated by addition of dioxane as a supplement to the NaOH solution.
The hydrolysis modification when a solution of NaOH (35 wt%) and dioxane
(10 wt%) was used as a reagent for the modification was accomplished at 65°C
* Prof. Norihiro Inagaki can be contacted at CBL06881@nifty.com.
127
in 30 min. The hydrolyzed ABS plate surfaces were successfully metallized

by electroless copper plating. Adhesion between the deposited copper metal
and ABS plate surface passes the crosscut test. Consequently, we can propose
a new recipe for electroless copper plating process without etching process
using chromic acid.
6.1 INTRODUCTION
ABS, acrylonitrile–butadiene–styrene copolymer, is widely used in automobiles (door
mirrors, console box, front grille, etc.) and electrical household appliances (housing
for TV and DVD players, enclosures for vacuum cleaners and air conditioners, etc.)
due to its good mechanical properties such as high impact resistance and strength and
thermal resistance. Such usage especially in automobiles demands a metal-like appear-
ance from a viewpoint of decorativeness. Surfaces of door mirrors and front grilles
are already metallized by a combination of electroless and electro plating techniques.
ABS surfaces are easy to be metallized by the electroless plating technique. Firstly,
ABS surfaces are etched by a mixture of chromic acid and sulfuric acid to roughen the
surfaces, and then a palladium (Pd) catalyst for electroless plating process is adsorbed
on the roughened surfaces. Afterward, the Pd catalyst-containing ABS surfaces are
immersed in an aqueous solution of copper sulfate, formaldehyde, and sodium hydrox-
ide to deposit copper metal on the ABS surfaces [1]. Such a copper metal-deposited
ABS surface is applicable to the conventional electroplating process for other metals
such as Au, Ag, Cr, Ni metal, etc. [1]. Therefore, the electroless copper plating is a basic
process in metallization of ABS surfaces. There are two important aspects for success-
ful performance of the electroless copper plating process:
(1) How to adsorb Pd catalyst on ABS surfaces?

(2) What reducing reagent is effective in reduction reaction of Cu2+ to Cu0?
In the conventional electroless copper plating process, ABS surfaces are treated
with a mixture of chromic acid and sulfuric acid, and the butadiene component in the
ABS surface is oxidized and removed from the ABS surface (reaction 6.1).
Chromic acid/
sulfuric acid
CH2–CH=CH–CH2 CH2–CH CH–CH2
Butadiene component O O
in ABS
O O
CH–CH + CH–CH2
Bond scission
As a result, the ABS surfaces become rough, and many pits are created on the
surfaces. Pd catalyst is easily absorbed on the roughened surfaces and gets in pits
on the ABS surfaces. The etching process by a mixture of chromic acid and sulfuric
acid is an easy and time-saving process. However, the usage of chromic acid leads
Surface Modification of ABS for Electroless Copper Plating 129
to a new problem, RoHS (Restriction of the Use of Certain Hazardous Substances in

Electrical and Electronic Equipment). A new electroless plating technique without
chromic acid is earnestly required from an environmental point of view. An answer
to this request is the graft copolymerization of nitrogen-containing groups on poly-
mer surfaces. The nitrogen-containing groups such as vinylimidazole, vinylpyridine,
etc. are graft-copolymerized on polymer surfaces and act as adsorption sites for a Pd
catalyst [2–5]. However, the graft-copolymerization technique involves troublesome
and time-consuming processes, and this is not a replaceable process for the etching
process by chromic acid.
Nitrile groups can be hydrolyzed with hydroxyl ions into amide and carboxylate
groups (reaction 6.2) [6,7].
OH OH
H2O
R–C N: + OH– H
C C
R N– R N
O O
H3O+ + NH3
C H C
R N in NaOH soln. R OH
H
If nitrile groups in the acrylonitrile component of ABS can be hydrolyzed into

amide and carboxylate groups, these groups would be expected to act as adsorption
sites for Pd catalyst. Electroless copper plating may be possible on the hydrolyzed
ABS surfaces without an etching process using chromic acid.
In light of the above discussion, we investigated a new etching technique for cop-
per electroless plating process without chromic acid. The following questions are
addressed:
(1) Can nitrile groups on ABS surfaces be hydrolyzed into amide and carboxy
late groups?
(2) What is the important factor for the hydrolysis reaction? Do the NaOH con-
centration in NaOH solution, hydrolysis temperature, and hydrolysis time
influence the hydrolysis reaction?
(3) Is there any difference in adhesion of the deposited copper metal/ABS sys-
tem between the new process (hydrolysis of nitrile groups) and the conven-
tional process (etching using chromic acid)?
6.2 EXPERIMENTAL
6.2.1 Materials
The ABS used for the experiments was Cycolac 3001M, provided by UMG ABS
Ltd., Japan, and was fabricated into plates (dimensions of 50 × 90 × 5 mm) using an
injection molding technique. The plates were rinsed with ethanol in an ultrasonic
washer in order to remove contaminants from the plate surface and were used as
specimens for the hydrolysis experiments. Commercial chemical kits for the electro-
less copper plating process were used: “OPC-750M-A,” “OPC-750M-B,” and “OPC-
750M-C” (Okuno Chemical Industries Co. Ltd., Japan). A commercial Pd catalyst kit
for electroless copper plating process containing “OPC-80 Catalyst M” and “OPC-
SAL M” (Okuno Chemical Industries Co. Ltd., Japan) was used.
6.2.2 Hydrolysis of ABS Plates and Electroless Copper Plating Process

The ABS plates were hydrolyzed in NaOH solution (7.5–70 wt%) at 65°C–80°C for
1–96 h. The NaOH solution was stirred with a magnetic stirrer during the hydrolysis
process. After finishing the hydrolysis process, the hydrolyzed ABS surfaces were rinsed
with water to remove completely NaOH from the ABS surfaces and then immersed in
dilute (1 M) HCl aqueous solution to convert sodium carboxylate to carboxylic acid.
The hydrolyzed ABS surfaces were immersed in a solution of OPC-80 Catalyst
M and OPC-SAL M at room temperature for 30 min in order to adsorb Pd catalyst at
carboxylic acid sites on the ABS surfaces, and then electroless copper plating was car-
ried out using the mixture solution of OPC-750M-A, OPC-750M-B, and OPC-750M-C
at room temperature. After finishing the electroless copper plating process, the ABS
surfaces were rinsed with water and dried and stored in vacuum at room temperature.
The cross-cut test was used for evaluation of the adhesion between the depos-
ited copper metal and the NaOH-hydrolyzed ABS. One hundred crosshatch patterns
(1 mm spacing) were made through the deposited copper metal up to the ABS sur-
face. A Scotch adhesive tape was then placed over the crosshatch patterns and was
rubbed with a fingertip to ensure good contact. The tape was removed by pulling
it off. The number of crosshatch patterns not pulled off from the ABS surface was
counted in the percentage for assessing the adhesion.
6.2.3 IR Spectra of the Hydrolyzed ABS Material Surfaces

ATR-IR spectra of the hydrolyzed ABS surfaces were recorded in the attenuated
total reflectance mode using a universal ATR attachment with a diamond prism
(refractive index = 2.4) (PerkinElmer Inc., U.S.A.). A Spectrum One Fourier infra-
red spectrometer (PerkinElmer Inc., U.S.A.) was used for the measurement of IR
spectra. The incident angle of the infrared light was 45°. From the incident angle and
refractive index ratio of the diamond prism (refractive index = 2.4) and hydrolyzed
ABS materials (assumed to be 1.55), the analytical depth for the ATR-IR spectra was
estimated to be 0.48–2.97 μm in the wavenumber range of 4000–680 cm−1.
6.2.4 XPS Spectra of the Hydrolyzed ABS Surfaces

XPS (C1s, N1s, and O1s) spectra of the hydrolyzed ABS surfaces were obtained on an
ULVAC-PHI Quantera SXM spectrometer (ULVAC-PHI Inc., Japan) using a mono-
chromatic AlKα photon source at 400 W at a pressure of 3.2 × 10−8 Pa. To avoid
charging of the specimen surface, a neutralizer was operated during the XPS measure-
ments. XPS spectra were recorded at takeoff angles of 20° and 45° against the sample
surface. The obtained C1s, N1s, and O1s spectra were decomposed using the curve-
fitting program (PeakFit) supplied by Sea Solve Software Inc., Framingham, MA.
6.2.5 Contact Angle of Hydrolyzed ABS Surfaces

Contact angles of water on the hydrolyzed ABS surfaces were measured at 20°C with
the sessile drop method [8,9]. A contact angle meter (Erma Co. Ltd., Japan, model G-1)
was used for the measurements. Water used for the measurement was twice-distilled
water, and the distillation was done just before the contact angle measurement.

6.3.1 Hydrolysis of ABS Surfaces by Aqueous NaOH Solution
The ABS surfaces were hydrolyzed in NaOH solutions (7.5–70 wt%) at 70°C or
80°C for 1–96 h. The contact angle of water for the NaOH-hydrolyzed ABS surfaces
was measured as functions of the NaOH concentration, the hydrolysis temperature,
and the hydrolysis time. Figure 6.1 shows typical water contact angle results for the
ABS plates hydrolyzed in 46 and 70 wt% NaOH solutions at 80°C as a function of
the hydrolysis time. If some of nitrile groups on the ABS surface were hydrolyzed
into amide and carboxylate groups as shown in reaction 6.2, the ABS surfaces would
be changed from hydrophobic to hydrophilic. As shown in Figure 6.1, the water
contact angle decreased gradually with increasing hydrolysis time: for example, the
water contact angle when the ABS surface was hydrolyzed with a NaOH (70 wt%)
solution at 80°C for 72 h decreased from 76° to 53°. Such decrease in water contact
angle indicates the possibility that some of nitrile groups on the ABS surfaces were
hydrolyzed into amide and carboxylate groups. Figure 6.2 shows typical ATR-IR
spectra for the NaOH-hydrolyzed ABS and the pristine ABS. The IR spectrum
for the pristine ABS showed intense and characteristic absorptions at 2223, 1450,
100
Hydrolyzed at 80˚C
80
Contact angle of water (˚)
NaOH conc.
46 wt%
60
70 wt%
40
20
0
0 20 40 60 80
Hydrolysis time (h)
FIGURE 6.1 Contact angle of water on the NaOH-hydrolyzed ABS surfaces as functions of
the NaOH concentration and hydrolysis time.
Styrene and
Acrylonitrile butadiene
Styrene and Acrylonitrile

butadiene
Pristine ABS
Hydrolyzed
for 24 h
C(O)NH2 and
Hydrolyzed
C(O)Na
for 72 h
Hydrolyzed
for 92 h
4000 3000 2000 1000

Wavenumber (cm–1)
FIGURE 6.2 ART-IR spectra of the ABS surfaces hydrolyzed in NaOH (70 wt%) at 80°C
as a function of the hydrolysis time.
1068, and 1028 cm−1 due to a acrylonitrile component, at 3026, 3003, 2927, 967,
and 908 cm−1 due to a butadiene component, and at 3026, 3003, 2927, 2850, 1603,
1495, and 1450 cm−1 due to a styrene component [10,11]. On the other hand, the
NaOH-hydrolyzed ABS surfaces showed new absorptions at 1666 and 1560 cm−1,
which were assigned to C = O stretching vibrations due to C(O)NH2 and C(O)ONa
groups, respectively (Figure 6.2). These changes in IR spectra show that some of
nitrile groups were hydrolyzed in the NaOH solution and converted into amide and
carboxylate. To what nitrile groups were hydrolyzed was evaluated from changes in
the absorption intensity at 2223 cm−1 (nitrile groups). Figure 6.3 shows the intensity
ratio of the absorption at 2223 cm−1 (C≡N groups in the acrylonitrile component) and
1603 cm−1 (phenyl groups in the styrene component) as a function of the hydrolysis
time in the NaOH solution (70 wt%) at 80°C. The intensity ratio shows only a small
decrease until a hydrolysis time of 48 h and then showed a large decease. The large
decrease continues up to a hydrolysis time of 72 h. Therefore, the hydrolysis process
of nitrile groups in the NaOH (70 wt%) solution was a slow one, and the hydrolysis
modification of the ABS surfaces requires a long time, at least 72 h.
Further, in order to investigate what modification occurred on the ABS surfaces,
C1s, O1s, and N1s spectra were measured at different takeoff angles of 20° and 45°
against the ABS specimen surface. The analytical depths at takeoff angles of 20°
and 45° are 3.1 and 6.4 nm, respectively, when the inelastic free path of electrons
is assumed to be 3 nm. Table 6.1 shows the atomic composition of the NaOH-
hydrolyzed ABS surfaces as a function of the takeoff angle. The N/C atomic ratio
at a takeoff angle of 20° was zero. On the other hand, the N/C atomic ratio at 45°
was 0.03. This difference in the N/C atomic ratio indicates that the acrylonitrile
2.0
Hydrolyzed in NaOH solution
Intensity ratio of CN (at 2223 cm–1)/

(70 wt%) at 80˚C
1.5
Ph (at 1603 cm–1)

1.0
0.5
0.0
0 20 40 30 80 100
Hydrolysis time (h)
FIGURE 6.3 Intensity ratio of CN and phenyl absorptions for ABS surfaces hydrolyzed in
NaOH (70 wt%) solution at 80°C as a function of the hydrolysis time.
TABLE 6.1
N/C and O/C Ratios for ABS Surfaces Treated with NaOH Aqueous Solution
NaOH Treatment Takeoff Angle 20° Takeoff Angle 45°
N/C O/C N/C O/C
NaOH Treatment Treatment Atomic Atomic Atomic Atomic
Concentration (%) Temperature (°C) Time (h) Ratio Ratio Ratio Ratio
None None None 0.00 0.09 0.03 0.08
46 80 72 0.01 0.18 0.02 0.12
70 80 72 0.05 0.24 0.06 0.21
component in ABS did not appear as the topmost layer but was distributed in inner
layers as deep as 3.2 nm from the surface. The NaOH-hydrolyzed ABS surfaces
showed a large amount of oxygen (O/C ratio = 0.12–0.24) and a small amount of
nitrogen (N/C ratio = 0.01–0.02). This high O/C atomic ratio indicates that the ABS
surfaces were intensely oxidized during the hydrolysis treatment.
Figure 6.4a and b shows C1s and O1s spectra for the pristine ABS and the ABS
surfaces hydrolyzed in a NaOH (70 wt%) solution at 80°C for 72 h. Their N1s spectra
are not shown here because of weak intensity and poor resolution of these spectra. The
pristine ABS showed widely distributed C1s spectrum from 284 to 292 eV. The C1s
spectrum at a takeoff angle of 20° was decomposed into six C1s components, which
appeared at 284.7 eV due to CH groups in aromatic carbons (C#1), at 285.4 eV due
to CH2 and CH groups in aliphatic carbons (C#2), at 286.5 eV due to C–O and C–N
groups (C#3), at 287.6 eV due to C = O groups (C#4), at 289.4 eV due to C(O)–O groups
(C#5), and at 291.3 eV due to π–π* shake-up satellite (C#6) [12–14]. On the other hand,
the C1s spectrum at a takeoff angle of 45° for the pristine ABS was decomposed into
four C1s components, which appeared at 284.7 eV due to CH groups in aromatic car-
bons (C#1), at 285.4 eV due to CH2 and CH groups in aliphatic carbons (C#2), at 286.5
(a) (b)
O1s spectra
O#1
C1s spectra O#2 O#3
Takeoff Takeoff ABS surface etched (iv)
(ii) angle 45˚ (ii) angle 20˚ with chromic acid
O#1
C#1 C#1
C#2 ABS surface hydrolyzed O#2 O#3
C#2
C#3 C#3 in a mixture of
NaOH and dioxane (iii)
C#4 C#4
ABS surface C#5
hydrolyzed C#6
for 72 h O#1
O#2 O#3
(i) (i)
ABS surface hydrolyzed
in NaOH solution (ii)
C#1 C#1
C#2 C#2
C#3 C#3 O#1
C#4 C#4
Pristine O#2
ABS surface C#6 C#5
Pristine ABS surface (i)
290 285 280 290 285 280 535 525

FIGURE 6.4 (a) C1s spectra of pristine ABS surface (i) and ABS surfaces hydrolyzed in
NaOH (70 wt%) solution at 80°C for 72 h (ii) as a function of takeoff angle. (b) ABS surface
etched with chromic acid (iv), and in a mixture of NaOH (35 wt%) and dioxane (10 wt%) at
65°C for 30 min (iii), and ABS surfaces hydrolyzed in NaOH (70 wt%) at 80°C for 72 h (ii),
and O1s spectra of pristine ABS surface (i).
eV due to C–O and C–N groups (C#3), and at 287.6 eV due to C = O groups (C#4)
[12–14]. The components C#5 and C#6 were not observed in the C1s spectrum at 45°.
On the other hand, there was less difference in O1s spectra between takeoff angles of
20° and 45°. Therefore, Figure 6.4b shows the O1s spectra only at a takeoff angle of 20°.
The O1s spectrum for the pristine ABS was decomposed into two O1s components at
532.3 eV due to O = C and O = C–O (O#1) and at 533.6 eV due to O–C groups (O#2).
The NaOH-hydrolyzed ABS surfaces, as shown in Figure 6.4a, showed six C1s
components (C#1–C#6), which were at the same peak positions as those for the pris-
tine ABS. The relative concentrations of these C1s components, especially oxygen-
containing components, were different from those for the pristine ABS. The C#4 and
C#5 components due to C = O and C(O)–O groups [12–14], respectively, increased
from 3% to 6%–8%. On the other hand, the C#3 component due to C–O and C–N
groups decreased from 19% to 10%. The O1s spectrum for the NaOH-hydrolyzed
ABS surfaces, as shown in Figure 6.4b, showed a new component (O#3 component)
at 531.5 eV in addition to the O#1 and O#2 components [12–14]. The new O1s compo-
nent (O#3) is assigned as an oxygen atom of amide groups, and their concentrations
were 24% and 15% at takeoff angles of 20° and 45°, respectively. The appearance of
the new O1s component (O#3) indicates that some of nitrile groups were hydrolyzed
into amide groups. The amide group, as shown in reaction 6.2, is a precursor of the
carboxylate group in hydrolysis reactions of nitrile groups.
From the results of ATR IR spectra and XPS spectra for the hydrolyzed ABS
plate surfaces, we can conclude that some of nitrile groups at the ABS surfaces were
surely hydrolyzed in NaOH solution into amide and carboxylate groups.
6.3.2 Electroless Copper Plating on Hydrolyzed ABS Surfaces

The NaOH-hydrolyzed ABS surfaces were immersed in a solution of OPC-80 Catalyst
M and OPC-SAL M at room temperature for 30 min in order for the Pd catalyst to
adsorb. Afterward, copper metallization was carried out in an electroless plating solu-
tion (a mixture solution of OPC-750M-A, OPC-750M-B, and OPC-750M-C) at room
temperature. The NaOH-hydrolyzed ABS surfaces were amenable to copper metal-
lization. The ABS surfaces succeeded in copper metallization. The adhesion between
the deposited copper and the hydrolyzed ABS was evaluated by the cross-cut test, and
the number of crosshatch patterns not pulled off from the ABS surface was counted
in percentage for assessing for evaluation of the adhesion. The cross-cut results are
summarized in Figure 6.5a and b as functions of the NaOH concentration, hydrolysis
temperature, and hydrolysis time. The adhesion between the deposited copper metal
and the hydrolyzed ABS surface, as shown in Figure 6.5a and b, was closely related
to the NaOH-hydrolysis conditions. The ABS surfaces hydrolyzed at 70°C even when
hydrolyzed in a NaOH solution of 70 wt% for 96 h failed in the cross-cut test (Figure
6.5a). The adhesion in the cross-cut test was 25%. On the other hand, the ABS sur-
faces hydrolyzed at 80°C in NaOH solution of 70 wt% for 72 h or longer showed
strong adhesion (Figure 6.5b). The cross-cut result was 100%.
From these experimental results, we believe that NaOH hydrolysis of ABS surfaces
is an effective technique for electroless copper plating. This modification process is a
technique without chromic acid, which is very advantageous. However, the hydrolysis
modification is a slow process. Reduction for the hydrolysis time is urgently needed.
(a) (b)
100 100
Percentage of patterns remaining
Hydrolysis temp. 70˚C Hydrolysis temp. 80˚C

after cross-cut test (%)
Hydrolysis time 96 h Hydrolysis time 96 h

72 h
50 48 h 50 72 h
24 h
48 h
24 h
0 0
7.5 15 23 46 70 7.5 15 23 46 70
NaOH conc. (%) NaOH conc. (%)
FIGURE 6.5 (See color insert.) (a) Cross-cut test results for electroless Cu-metallized ABS
hydrolyzed at 70°C as functions of the NaOH concentration and hydrolysis time. (b) Cross-cut
test results for electroless Cu-metallized ABS hydrolyzed at 80°C as functions of the NaOH
concentration and hydrolysis time.
6.3.3 Acceleration of Hydrolysis Process of ABS Surfaces

As discussed in the previous section, the NaOH-hydrolysis modification of ABS
surfaces is an effective technique for electroless copper plating. Adhesion of the
deposited copper metal was comparable to that for the ABS surfaces metallized by a
conventional electroless copper plating process using chromic acid. From this view-
point, the NaOH-hydrolysis modification is practically applicable as a new etching
process without chromic acid for electroless copper metallization of ABS surfaces.
However, the NaOH-hydrolysis modification is a slow process and needs long time,
72 h. In order to reduce the NaOH-hydrolysis time, hydrolysis reactions occurring on
the ABS surfaces were reinvestigated.
The N/C atomic ratio for the pristine ABS surfaces, as shown in Table 6.1, indi-
cates an important aspect of the chemistry of ABS surfaces. The N/C atomic ratio
analyzed at a takeoff angle of 20° was zero, and that at a takeoff angle of 45° was
0.03. The analytical depths at takeoff angles of 20° and 45° were 3.1 and 6.4 nm
from the topmost layer, respectively. This difference in the N/C atomic ratio between
takeoff angles of 20° and 45° indicates that the topmost layer of the ABS surfaces
was completely covered with butadiene and styrene components of the ABS plates,
and acrylonitrile component did not appear at the ABS surface but existed at 3.1 nm
from the ABS surface. Therefore, in order to accelerate NaOH-hydrolysis reactions
of nitrile groups, the butadiene and styrene components that aggregate at the ABS
surfaces should be removed from the surfaces. To remove the surface-aggregated
butadiene and styrene components from the ABS surface, dioxane was added to the
aqueous NaOH solution. Dioxane (10.0 cal1/2/cm3/2) is close in the solubility param-
eter to polybutadiene (8.4–8.6 cal1/2/cm3/2) and polystyrene (8.56–10.3 cal1/2/cm3/2)
[15]. Therefore, we can expect that dioxane will diffuse deeply into the polybutadi-
ene and styrene components aggregated at the ABS surfaces, and these components
will be dissolved or swelled in the dioxane.
Mixtures of NaOH (35 wt%) and dioxane (5–10 wt%) were used as hydrolysis
media, and the hydrolysis modification was carried out at 65°C and 70°C for 30 min.
Figure 6.6 shows a typical ATR IR spectrum for the ABS surfaces that were hydro-
lyzed in a mixture of NaOH (35 wt%) and dioxane (10 wt%) at 65°C for 30 min.
There appeared intense absorptions at 1666 and 1560 cm−1 due to C = O stretching
vibration of C(O)NH2 and C(O)ONa groups [10,11], respectively, on the ATR IR
spectrum for the ABS surface hydrolyzed with the NaOH and dioxane mixture. The
absorption at 2223 cm−1 due to C≡N groups became weak in absorption intensity.
These changes in IR spectrum are similar to those for the ABS surfaces, which were
hydrolyzed in a NaOH solution (70 wt%) at 80°C for 72 h (Figure 6.6), indicating that
that nitrile groups on the ABS surface were successfully modified into amide and
carboxylate groups as far as dioxane was used as a supplement reagent, even though
the NaOH concentration in the hydrolysis process lowered from 70 to 35 wt% and the
hydrolysis temperature also dropped from 80°C to 65°C.
Finally, copper metallization of ABS surfaces hydrolyzed with a mixture of 35
wt% NaOH and 10 wt% dioxane was carried out using a mixture solution of OPC-
750M-A, OPC-750M-B, and OPC-750M-C. The copper metallization was success-
ful, and the adhesion between the deposited copper metal and the ABS surfaces
Styrene and
butadiene
Acrylonitrile
Styrene and
Acrylonitrile
butadiene
Pristine ABS (a)
ABS hydrolyzed C(O)NH2 and

in NaOH solution C(O)Na
(b)
ABS hydrolyzed
in mixture of
NaOH and dioxane
(c)
4000 3000 2000 1000

Wavenumber (cm–1)
FIGURE 6.6 ATR-IR spectra of pristine ABS surface (a), ABS surface hydrolyzed in NaOH
(70 wt%) at 80°C for 72 h (b), and ABS surface hydrolyzed in a mixture of NaOH (35 wt%)
and dioxane (10 wt%) at 65°C for 30 min (c).
hydrolyzed with the NaOH–dioxane mixture was evaluated to be 100% by the

cross-cut test. Figure 6.7 shows typical photographs of pristine ABS and electroless
copper-metallized ABS surface, which was hydrolyzed in NaOH (35 wt%)–dioxane
(10 wt%) mixture at 65°C for 30 min, and then copper-metallized using a mixture of
OPC-750M-A, OPC-750M-B, and OPC-750M-C. Therefore, we can conclude that a
(a)
Not-metallized
(b)
Cu metallized
FIGURE 6.7 Photographs of pristine ABS surface (a) and electroless copper-metallized ABS
surface hydrolyzed in a mixture of NaOH (35 wt%) and dioxane (10 wt%) at 65°C for 30 min (b).
new hydrolysis process using the mixture of NaOH (35 wt%) and dioxane (10 wt%)
is practically applicable as an etching technique for electroless plating.
6.4 CONCLUSIONS
The hydrolysis modification of nitrile groups at ABS surfaces was investigated in
order to seek a new recipe for copper metallization of ABS surfaces without etching
treatment using chromic acid. Results are summarized as follows:
(1) Nitrile groups of ABS surfaces were hydrolyzed into amide and carboxy
late groups with NaOH solutions (46–70 wt%) at 70°C–80°C for more than
24 h.
(2) These NaOH-hydrolyzed ABS surfaces could be metallized by the electro-
less copper plating technique.
(3) Adhesion between the hydrolyzed ABS surface and deposited copper metal
strongly depends on the hydrolysis conditions. ABS surfaces hydrolyzed in
an aqueous NaOH solution of 70 wt% at 80°C for more than 72 h succeeded
in the cross-cut test. No square was removed by the tape. However, ABS
surfaces modified in a NaOH solution of lower than 70 wt%, at 70°C, and
for less than 72 h failed in a cross-cut test.
(4) The hydrolysis modification was accelerated by the addition of dioxane as
a supplement to the NaOH solution. ABS surfaces were completely hydro-
lyzed with a mixture of NaOH (35 wt%) and dioxane (10 wt%) at 65°C in
30 min.
(5) ABS plate surfaces modified in the mixture of NaOH and dioxane suc-
ceeded in copper metallization. Adhesion between the deposited copper
metal and the hydrolyzed ABS surface passed the cross-cut test.
From these experimental results, we can propose a new recipe for electroless cop-
per metallization of ABS surfaces without chromic acid.
REFERENCES
1. S. Haruyama, Electrochemistry, p. 185, Maruzen, Tokyo (2005).
2. G. H. Yang, E. T. Kang, and K. G. Neoh, Appl. Surf. Sci., 178, 165 (2001).
3. W. C. Wang, E. T. Kang, and K. G. Neoh, Appl. Surf. Sci., 199, 52 (2002).
4. G. H. Yang, D. Lim, Y. P. Tan, Y. Zhang, E. T. Kang, and K. G. Neoh, Eurp. Polym. J.,
38, 2153 (2002).
5. W. C. Wang, R. K. H. Vora, E. T. Kang, and K. G. Neoh, Macromol. Mater. Eng., 288,
152 (2003).
6. J. McMurry, Fundamentals of Organic Chemistry, p. 848, Cole Publishing Co.,
Monterey, CA (2007).
7. M. L. Gupta, B. Gupta, W. Oppermann, and G. Hardtmann, J. Appl. Polym. Sci., 91,
3127 (2004).
8. J. D. Malcolm, and H. M. Paynter, J. Colloid Interface Sci., 82, 269 (1981).
9. L. M. Lander, L. M. Siewierski, W. J. Brittain, and E. A. Voger, Langmuir, 9, 2237
(1993).
10. L. J. Bellamy, The Infrared Spectra of Complex Molecules, Wiley, New York (1966).
11. Sigma-Aldrich Database; http//www.sigmaaldrich.com/Area_of_Interest/chemistry/
Materials_Science.html.
12. G. Beamson, and D. Briggs, High Resolution XPS of Organic Polymers: The Scienta
ESCA 300 Database, Wiley, Chichester (1992).
13. NIST Standard Reference Database; http://srdata.nist.gov/xps/.
14. Thermo Scientific Database; http://www.lsurface.com/accueil/index.php.
15. The Society of Polymer Science, Japan. Ed., Handbook of Polymer Data, p. 591,
Baifukan Co. Ltd, Tokyo (1986).
7 Achieving Nanoscale
Surface Structure
on Polyester Film by
Atmospheric Pressure
Plasma Treatment
W. Y. I. Tsoi, C. W. Kan*, C. W. M. Yuen,
T. M. Choi, and T. B. Tang
CONTENTS
7.1 Introduction................................................................................................... 142
7.2 Experimental................................................................................................. 143
7.2.1 Materials............................................................................................ 143
7.2.2 Plasma Treatment.............................................................................. 143
7.2.3 Surface Morphology.......................................................................... 143
7.2.4 Wettability......................................................................................... 144
7.2.5 Surface Chemical Analysis................................................................ 144
7.3.1 Etching and Hydrophilization........................................................... 145
7.3.2 Relations to Individual Operating Parameters.................................. 149
7.3.2.1 Variation with Treatment Time........................................... 149
7.3.2.2 Variation with Ignition Power and Oxygen Concentration.... 151
7.3.2.3 Variation with Jet Distance................................................. 152
7.4 Conclusions.................................................................................................... 154
Acknowledgments................................................................................................... 154
References............................................................................................................... 154
* Dr. C.W. Kan was the lead author for this chapter and can be contacted at tccwk@inet.polyu.edu.hk.
141
A nanoscale etching and hydrophilization of poly(ethylene terephthalate)

(PET) film were demonstrated by an atmospheric pressure He–O2 plasma
treatment with an atmospheric pressure plasma jet. Optimum conditions for a
nanostructure development were found to be (1) treatment time = 3 s, (2) igni-
tion power = 80 W, (3) O2 concentration = 1%, and (4) jet distance = 5 mm.
The resultant modification of the plasma-modified PET film was characterized
by scanning electron microscopy (SEM), static contact angle goniometry, and
X-ray photoelectron spectroscopy (XPS). The modified PET film was textur-
ized with nanospot (surface spot texture) in less than 100-nm diameter under
the optimum conditions. This paper provides an analysis of the correlation
between plasma operating parameters and nanoscale surface treatment operat-
ing under atmospheric pressure. The polar surface energy, as calculated from
the results of contact angle goniometry, of the modified PET film was ampli-
fied by more than 15 times, which was confirmed with XPS analysis showing
incorporation of additional hydrophilic groups on the substrate surface. The
nanospot density and the surface energy of the film were shown to be depen-
dent on the operating parameters. This study exhibits the efficacy of He–O2
atmospheric pressure plasma in achieving a nanoscale surface hydrophiliza-
tion of PET substrates.
7.1 INTRODUCTION
Nanostructures exhibit distinctive properties because of their high area-to-volume
ratio [1]. Plasma, in general, is capable to alter the surface layer of a substrate while
leaving the desirable bulk properties unaffected under the controlled operating
conditions. However, plasma induces a milder surface modification compared to
the other dry surface treatments such as electromagnetic irradiations [2]. Plasma
is capable to achieve nanoscale surface modification compared to other conven-
tional treatments. In addition to physical ablation, plasma is able to functionalize
a substrate simultaneously in a single-step treatment using various reactive non
polymerizing and polymerizing gases and precursors at low temperature. Because
of this aspect, the development of plasma nanofabrication has attracted considerable
attention [3–5].
Atmospheric pressure plasma (APP) recently has drawn increased interest owing
to its outstanding attributes such as being continuous and having a low ion flux com-
pared to its counterpart operated in reduced pressure [4–8]. This kind of plasma
offers a potential way to achieve a nanoscale surface modification. With the con-
tinual expansion of the applications of APP, it is important to control the degree
of modification to an appropriate extent. Hitherto, quantitative investigation of the
formation of topographical structures in relation to the plasma operating parameters
is seldom reported. A quantitative dimensional analysis of the plasma-induced struc-
tures on a substrate is essential for the manipulation of topographical development
and to design desirable surface properties.
Synthetic polymeric materials with superior mechanical properties and chemical
inertness are popular in various industrial sectors. However, the original smooth
Nanoscale Surface Structure on Polyester Film by APP Treatment 143
substrate surface could be texturized and functionalized to offer improved surface

performances such as wettability, adhesion, antistaticity, etc. utilizing plasma tech-
niques [4,8–10]. In this study, surface analysis was carried out on the nanoscale
texturized PET film.
7.2 EXPERIMENTAL
7.2.1 Materials
The biaxially oriented poly(ethylene terephthalate) (PET) film (0.1 mm thick) sup-
plied by Goodfellow Ltd., Cambridge, was used as the substrate. The PET film
was successively rinsed with acetone and absolute ethanol to remove grease and
dirt. The cleaned PET film was air-dried completely and conditioned under stan
dard conditions of 65 ± 2% relative humidity and 21 ± 1°C for at least 24 h prior to
experiments.
7.2.2 Plasma Treatment
APP treatment was conducted using an APP jet (APPJ, AtomfloTM 200) developed
by Surfx Technologies LLC (Redondo Beach, CA, U.S.A.) with a rectangular nozzle.
The substrate moved at a constant speed of 2 mm/s and was exposed to the afterglow
plasma generated from RF of 13.56 MHz. Helium (He) and oxygen (O2) were used
as carrier and reactive gases, respectively. A schematic diagram of the experimental
setup is shown in Figure 7.1. Four operating parameters: (1) treatment time, (2) igni-
tion power, (3) O2 concentration, and (4) jet distance, were studied for evaluating
their effect on nanoscale surface modification. He flow rate was fixed at 15 L/min.
The four variable parameters are listed in Table 7.1.
7.2.3 Surface Morphology
Surface topographical modification of plasma-modified PET film was analyzed
using a scanning electron microscope (SEM) (Model JSM-6490, JEOL Ltd., Japan).
He O2 Plasma
nozzle
RF
power Jet distance
Substrate
FIGURE 7.1 Schematic diagram of experimental setup (He and O2 were used as carrier and
reactive gas, respectively).
TABLE 7.1
Operating Parameters Studied in the Plasma Treatment
Operating Parameter Value
Treatment time (s) 1, 2, 3, 4, and 5
Ignition power (W) 60, 70, 80
O2 concentration (%)* 1, 2
Jet distance (mm) 3, 5, 7
*O2 concentration refers to the O2 to He ratio in which 1% and 2% O2 concen-

trations mean the flow rates of O2 were 0.15 and 0.30 L/min, respectively.
Images of plasma-modified PET films were captured with 20 kV accelerating volt-

age at a magnification of 15,000× for the nanostructure investigation. From the SEM
image, the nanospot density was calculated, which was defined as the number of
nanospots (spots with size smaller than 100 nm diameter) per unit area (µm2).
7.2.4 Wettability
Wettability modification of the plasma-modified PET film was characterized by
static contact angle goniometry using the sessile drop technique with a contact angle
meter (Tantec Inc., Schaumburg, IL, U.S.A.) in accordance with the ASTM D 5946.
Two probe liquids were used, distilled water (72.8 mN/m) and glycerol (63.4 mN/m),
and the drop size was 3 μL. The contact angle was reported as the average of 10 mea-
surements on an individual sample. Surface energies of samples, based on contact
angle values, were determined with Kaelble’s two-liquid method using the following
equations [11]:

( ) ( )  (7.1)
1/ 2 1/ 2
γ l (1 + cos θ) = 2  γ sp γ pl + γ sd γ dl 

γ s = γ sp + γ sd (7.2)
where θ is the static contact angle between the film and the probe liquid; γl is the sur-
face tension of the probe liquid; γ pl and γ dl are the polar and dispersion components
of the probe liquid, respectively; γs is the overall surface energy of the PET film; and
γ sp and γ sd are the polar and dispersion components of the PET film, respectively.
7.2.5 Surface Chemical Analysis

The surface chemistry of PET film was investigated by x-ray photoelectron spec-
troscopy (XPS). XPS analysis was carried out with an SKL-12 spectrometer
(Leybold Heraeus-Shengyang, Sengyang, China) modified with a VG CLAM 4
multichannel hemispherical analyzer equipped with Al/Mg twin anode. The spec-
trometer was operated with nonmonochromatic Mg Kα (1253.6 eV) radiation for the
characterization of the plasma-modified substrate under a pressure of 8 × 10 –8 Pa.

Spectra were analyzed using XPSPEAK 4.1 (Chinese University of Hong Kong,
Hong Kong).

7.3.1 Etching and Hydrophilization
The plasma treatment, using He and O2 gases, is a physicochemical modification
capable of etching and hydrophilization of the PET film surface. The etching and
hydrophilization rely on two main processes, physical ablation and chemical etch-
ing. Physical ablation is induced by the inert helium bombardment, while chemical
etching is driven by the reactive oxygen. Surface etching is predominantly caused
by physical ablation. In addition, surface etching forms a great deal of volatile low
molecular weight oxidized fragments (LMWOFs), which are easily removed from the
substrate surface via sublimation in the open system. These reactions simultaneously
occurring on the substrate elicit a distinctive surface topography. Nanostructures
on the plasma-modified PET film were revealed using a SEM at the magnification
of 15,000×. The original PET film was smooth without distinguishable features on
the surface. Nanospots were found at different stages of the treatment illustrated in
Figure 7.2, and the film was consequently texturized by the plasma.
(a) (b)
20 kV ×15,000 1 µm 09 30 SEI 20 KV ×15,000 1 µm 11 30 SEI
(c) (d)
95 nm 60 nm
20 kV ×15,000 1 µm 11 30 SEI 20 kV ×15,000 1 µm 10 30 SEI
FIGURE 7.2 SEM micrographs illustrating time-dependent nanospot formation on the

surface of PET film exposed to He–O2 APP (magnification 15,000×): (a) control; (b) 1 s
exposure; (c) 2 s exposure; (d) 3 s exposure (ignition power: 80 W; O2 concentration: 1%; jet
distance: 5 mm).
In addition to etching, He–O2 plasma is capable of hydrophilizing the PET film.

On one hand, physical ablation cleaving polymeric chains into smaller fragments
increases the order of surface randomness. On the other hand, oxidation of the PET
film introducing more polar carboxylic and hydroxyl groups increases surface polar-
ity [2,4,12]. The plasma-modified film showed improved wettability. Contact angle
goniometry was employed to evaluate the degree of hydrophilization of the plasma-
modified PET film. Contact angles were measured with two polar probe liquids,
water and glycerol. A reduction in contact angles (ΔCA) was observed, and the sur-
face energy was calculated from the contact angle results. Alteration of the surface
energy (γs) of the modified film was revealed based on the calculation using Kaelble’s
two-liquid method [11]. The calculation gives the surface energy components, that
( ) ( )
is, polar γ s and dispersion γ s components, as shown in Figure 7.3. A drastic
p d
increase in the polar component of the plasma-modified film was discovered. The
original PET film possessed low polar surface energy γ s about 4 mJ/m2. The
p
( )
polar component increased to more than 20 times, that is, 80 mJ/m2, after exposure
to the plasma for 3 s. This explained that ΔCA in the goniometry was caused by the
polar interaction at the liquid–solid interface. Another piece of information obtained
from the goniometry was that plasma hydrophilization was predominant in the ini-
tial stage without significant change in the surface morphology of the film as shown
in Figure 7.3. The initial increment of the polar component of surface energy at 1 s
was found to be the most significant. Further increasing the treatment time did not
show significant effect on the enhancement of surface wettability. Moreover, the
plasma-modified PET film did not show complete wetting to the probe liquids for the
operating parameters studied.
In order to confirm the experiential results obtained from the contact angle goni-
ometry, XPS analysis was conducted to evaluate the surface chemical composition
of the plasma-modified PET films. The control PET film was composed of three
elements, namely, carbon (C), oxygen (O), and silicon (Si), as shown in Figure 7.4.
100
90
80
70
60
50
γsp
40
γds
30
20 γs
10
0
0 1 2 3 4 5
–10
Treatment time (s)
FIGURE 7.3 Time-dependent variation of surface energy and its components for the
plasma-modified film (ignition power: 80 W; O2 concentration: 1%; jet distance: 5 mm).
Control Wide scan

60,000
50,000
O1s
40,000
Count C1s
30,000
20,000
10,000
Si2p
0
0 200 400 600 800 1000 1200
Binding energy (eV)
C = 0.68 O/C ratio = 0.47
O = 0.32
Plasma-modified Wide scan

60,000
50,000
O1s
40,000
Count
30,000
C1s
20,000 N1s
10,000 Si2p
0
0 200 400 600 800 1000 1200
C = 0.68
Binding energy (eV)
O = 0.32 O/C ratio = 0.62
N = 0.06
FIGURE 7.4 Surface elemental composition of control and plasma-modified PET films
(ignition power: 80 W; O2 concentration: 1%; jet distance: 5 mm).
It is because small amounts of Si surface additives were incorporated during the

production of polymeric PET films [13,14]. Therefore, a trace amount of Si is present
in the control PET film.
XPS results indicated an increase in O content and a reduction in C content of
the plasma-modified PET film as presented in Table 7.2. In general, O-containing
functional groups are hydrophilic and have a strong affinity toward polar wetting
agents, and the improved surface wettability is essential for further wet processing of
TABLE 7.2
Surface Elemental Composition of the He–O2 Plasma-Modified
PET Film
C O N O/C
Control PET film 0.68 0.32 N/A 0.47
Modified PET film 0.58 0.36 0.06 0.62
Note: Ignition power, 80 W; O2 concentration, 1%; jet distance, 5 mm; treatment time, 3 s.
textile materials. The ratio of O and C atomic contents (O/C ratio) is an indicator of
the surface polarity of a substrate. This ratio indicates the relative quantity of polar
O-containing functionalities present on the PET film surface. The higher the O/C
ratio, the more hydrophilic a PET film surface will be. The O/C ratio of the original
PET film surface so obtained was 0.47. The He–O2 APP incorporated an additional
O into the PET film via surface oxidation. The O/C ratio of the plasma-modified
PET film surface was increased to 0.62. A small amount of N was detected on the
plasma-modified PET films, which may originate from the ambient environment
during plasma treatment. Polar N-containing functional groups also contribute to the
increment of PET film surface wettability. Additional polar functionalities incorpo-
rated onto the PET film surface indicated that plasma treatment would assist further
wet processing of textile materials. The XPS results confirmed with the observation
in the contact angle goniometry.
The exact surface composition of the plasma-modified PET films was evalu-
ated by deconvolution of individual elemental peaks via XPSPEAK 4.1 software.
Deconvolution of PET films revealed that C1s atomic peak is assigned to be C–H,
C–O, and C═O bonds as shown in Figure 7.5. No new bonds were detected in the
present study. However, the atomic content of individual bonds varies with plasma
operating parameters.
From Figure 7.6, the concentrations of individual surface functionalities were
found to be time dependent. From the deconvolution of C1s peak, C–H was found
to be reduced with treatment time, while C–O and C═O increased with time. This
opposite trend with treatment time indicated that plasma hydrophilization was a pro-
gressive and continuous reaction via surface oxidation. Deconvolution analysis of
XPS data complied with the variation of polar and dispersion components of surface
energy of the plasma-modified PET films as illustrated in Figure 7.3. Deconvolution
analysis of C1s peak indicated the reaction mechanism of surface hydrophilization
(a) C1s (b) C1s

9000 9000
8000 C–H 8000
7000 7000 C–H
6000 6000
Count
Count
5000 5000
4000 4000
C–O
3000 3000 C–O
–
C–O –
C–O
2000 2000
1000 1000
0 0
282 284 286 288 290 292 294 280 285 290 295 300
FIGURE 7.5 XPS deconvoluted C1s peaks of (a) control film and (b) plasma-modified PET
film using XPSPeak 4.1 software film (ignition power: 80 W; O2 concentration: 1%; jet dis-
tance: 5 mm; treatment time: 3 s).
0.5
Relative concentration (%)

0.4
0.3 C–H
C–O
0.2 –
C–O
0.1
0
0 1 2 3 4 5
Treatment time (s)
FIGURE 7.6 Time-dependent surface functionalities of plasma-modified PET film deter-

mined by deconvolution of C1s peak (ignition power: 80 W; O2 concentration: 1%; jet dis-
tance: 5 mm; treatment time: 3 s).
of the PET film. He–O2 plasma induced surface oxidation predominantly in 1 to 3 s
treatment time. A C–H reduction with a C–O increment indicated that oxidation
occurred on the ethylene units on the PET polymer chains. Increased C–O bonds
improved surface polarity of plasma-modified PET films. After the optimum treat-
ment time, that is 3 s, C–H content increased while C–O reduced. Beyond the opti-
mum treatment time, the change of individual functionality was saturated with time.
This indicated that the equilibrium between etching and oxidation was reached.
7.3.2 Relations to Individual Operating Parameters

In the course of the plasma treatment, both the surface topography and chemistry
varied with the four operating parameters: (1) treatment time, (2) ignition power,
(3) O2 concentration, and (4) jet distance. Plasma modification is a progressive reac-
tion with respect to the concentration of plasma active species on the PET film
surface. The treatment time is required to accumulate sufficient amount of active
species on the substrate for the reaction [4]. This variation is an indicator represent-
ing the equilibrium shifting between the physical ablation and chemical etching in
the plasma treatment. Variation with respect to two interactive parameters, namely,
ignition power and O2 concentration, determines the stability and effective energy
content of the plasma gaseous admixture. The jet distance representing the proxim-
ity of the nozzle to the substrate surface demonstrates how effective transfer and
accumulation of the active species can affect the overall PET film surface etching–
hydrophilization reaction.
7.3.2.1 Variation with Treatment Time

Surface topographical modification is a time-dependent process. By operating under
80 W plasma with 1% O2 at 5 mm jet distance, a progressive topographical alterna-
tion is illustrated in Figure 7.2. For 1 s exposure, plasma induced only microcracks
(ranging from 125 to 250 nm in size) on the PET film. After 2 to 3 s exposure,
observable nanospots were developed on the PET film surface. The degree of etching
was expressed in terms of the nanospot density (number of nanospots per unit area
[μm–2]) in an SEM micrograph. The diameter of the majority of nanospots ranged
from 60 to 95 nm. Overall, 3 s exposure to the plasma generated maximum nanospot
density as shown in Figure 7.7. Insufficient treatment time was not able to generate
observable nanospots in 1 to 2 s of treatment. Overall surface energy (γs), however,
increased rapidly to almost twice as shown in Figure 7.3. The enhancement in wet-
tability of the modified film was due to the increase in the polar functionalities on
the substrate surface. In the initial stage, chemical etching versus physical ablation
was predominant, and surface wettability increased without any observable nano-
structures generated. Nanospot density with a maximum of 76 μm–2 was achieved
at 3 s for the highest surface energy of 84 mJ/m2 attained using 80 W at 5 mm jet
distance of treatment as calculated from experimental data. Three-second exposure
was found to be the optimum treatment time in the present study. Nanospot density
reduced with time after achieving a maximum at 3 s, indicating that plasma would
ablate the newly formed nanostructures and polar LMWOFs. The degree of ablation
was depicted by both the variation of nanospot density and dispersion component of
surface energy of the modified PET film. Dispersion component of surface energy,
as a result, symbolized the plasma-induced chain scission on the PET film surface.
This energy constantly remained at a low value during the plasma treatment for 1 to
5 s and indicated that plasma-induced etching would not impose severe degradation
of the bulk polymer film. Meanwhile, with longer treatment time, the concentration
of active species from the plasma jet accumulating on the surface of the PET film
increased. Once the concentration of the active species increased to a critical level,
the reaction between the active species and PET film surface would be saturated.
80
70
Average nanospot density (µm–2)
60
50
1% O2
40
2% O2
30
20
10
0
0 1 2 3 4 5
Treatment time (s)
FIGURE 7.7 Time-dependent in the nanospot density on the plasma-modified PET film
with respect to O2 concentration of plasma (ignition power: 80 W; jet distance: 5 mm).
The excess active species would then be neutralized by the surrounding air spe-
cies. Therefore, the resulting chemical (surface energy) and physical (morphological
change) changes would be reduced as shown in Figures 7.3 and 7.7, respectively.
7.3.2.2 Variation with Ignition Power and Oxygen Concentration

Ignition power directly controls the plasma power in a plasma discharge. A higher
power would be capable of attaining nanospots at a higher rate as the plasma active
species possess higher energy when bombarding the substrate. Physical ablation was
comparatively more rapid at 70 to 80 W than at 60 W. During the study duration
of 1 to 5 s, 60 W was not able to induce the observable nanostructures on the PET
film subjected to 1% O2 plasma exposure at a jet distance of 5 mm. In contrast, both
70 and 80 W generated some distinctive nanospots within 3 s of exposure. Ignition
power did play an important role in the initial stage of nanospot formation. There
were no significant changes in the dimension and density of nanospot in the course
of the treatment once nanostructures were formed. Likewise, the surface energy of
the plasma-modified PET film demonstrated a similar relationship as the nanospot
density with respect to the power. This indicated that ignition power was the prime
driving force for both physical ablation and chemical etching.
The effect of ignition power on the surface chemistry was investigated by XPS.
Increasing ignition power can promote surface hydrophilization of PET substrate
as shown in Table 7.3. It is obvious that the O/C ratio increases with ignition power.
A higher ignition power is capable of producing more active species, resulting in
higher concentrations of He and O active species present in the plasma afterglow.
As a result, both the etching and oxidation reactions are more effective. Based on
the calculation of ΔO/C with respect to the original film, the increment of O/C with
respect to the ignition power shows a linear relationship. Eighty watts was found to
be the optimum power in this study.
Generally speaking, a higher O2 concentration should be able to incorporate
more polar functionalities on a substrate. The experimental result, however, was the
contrary that could be explained by the plasma chemistry. O active species are not
directly formed by the electrical discharge in the APPJ. O active species are gen-
erated mainly by the Penning reaction with He metastables [15,16]. A higher O2
TABLE 7.3
Atomic Ratio with Respect to Ignition Power
Ignition Power (W) C O N* O/C ΔO/C
Control 0.68 0.32 N/A 0.47 0
60 0.61 0.33 0.06 0.54 +0.07
70 0.59 0.34 0.07 0.58 +0.11
80 0.58 0.36 0.06 0.62 +0.15
Note: O2 concentration, 1%; jet distance, 5 mm; treatment time, 3 s.

*N originated from the ambient environment during plasma treatment.
90
80
70
60
50 1% O2
40 2% O2
30
20
10
0
0 1 2 3 4 5
Treatment time (s)
FIGURE 7.8 Time-dependent surface energy of the plasma-modified PET film with respect
to O2 concentration in the plasma (ignition power: 80 W; jet distance: 5 mm).
concentration consumes more He metastables, and the reduced concentration of He

metastables hinders the sequential O active species production. Experimental results
supported the above proposed explanation. On one hand, the improvement in surface
energy of the PET film modified with 1% O2 was found to be greater than that with
2% O2 plasma as illustrated in Figure 7.8. When the concentration of O2 increased
during the plasma treatment, the supply of active species for reaction on the PET film
surface would also increase. However, if a large amount of oxygen is continuously
supplied, the active species might react with the oxygen instead of with the PET
film surface. The amount of active species would be reduced, and hence, the surface
reaction would be lowered. As a result, 1% O2 plasma was found to be comparatively
more effective for etching and hydrophilization.
The ignition power and the ratio of O2 to He not only affect the etching and hydro-
philization power but also determine the stability of the plasma. Plasma was found to
be more stable when operated at a higher ignition power with a lower O2 concentra-
tion. For a given ignition power, a lower O2 concentration in plasma offered a more
stable plasma modification of the substrate. As seen in Figures 7.7 and 7.8, both the
nanospot density and surface energy gave correlated quadratic relationships with the
treatment time using 1% O2 plasma at 80 W. The correlation is proposed as an indi-
cator of the plasma stability, and the results deviated from the quadratic relationship
when lower ignition power and higher percentage of O2 were used in the treatment.
Although 70 W induced an appreciable amount of nanospots similar to that at 80 W,
the correlation deviated from the quadratic relationship. Reducing ignition power
at a given O2 concentration could not provide sufficient energy for the formation
of stable plasma for surface etching. The stability of plasma generated is the prime
attribute for the experimental design and prediction of the resultant modification.
7.3.2.3 Variation with Jet Distance

APPJ is a downstream treatment. Jet distance is the perpendicular distance between
the plasma nozzle and the substrate located below as shown in Figure 7.1. The
80
Average nanospot density (µm–2)

70
60
50
3 mm
40 5 mm
7 mm
30
20
10
0
0 1 2 3 4 5
Treatment time (s)
FIGURE 7.9 Time-dependent nanospot density on the plasma-modified PET film with dif-
ferent jet distances (ignition power: 80 W; O2 concentration: 1%).
distance between the plasma source and the substrate governs the efficacy of the
active species in etching a surface. In the APP treatment, the active plasma species
experience a severe collision with air molecules when travelling toward the substrate
surface. Velocity and energy content decrease with respect to the time and distance
transported. The rate of nanospot formation at 5 mm was much faster than the modi-
fication at 3 and 7 mm as shown in Figure 7.9.
The amount of additional O incorporated into the substrate surface was also found
to be distance dependent. A closer proximity allowed effective energy transfer from
plasma active species to the polymeric substrate surface. As a result, 5 mm jet distance
could achieve the largest increment in O/C ratio of +0.15 as compared to the other
two distances 3 and 7 mm as shown in Table 7.4. As a result, 5 mm plasma treatment
was found to be more effective and efficient to achieve a nanoscale etching and hydro
philization on the substrate. When the distance between the plasma jet and the PET film
surface was too small, that is, 3 mm, the flow of the gas from the jet was almost blocked
by the film, and the gas could only be bounced off the surface and flew out more paral-
lel to the film surface direction, which greatly reduced the effectiveness of the plasma
TABLE 7.4
O/C Atomic Ratio with Respect to Jet Distance
Jet Distance (mm) C O N* O/C ΔO/C
Control 0.68 0.32 N/A 0.47 0
3 0.64 0.32 0.04 0.50 +0.03
5 0.58 0.36 0.06 0.62 +0.15
7 0.60 0.34 0.06 0.57 +0.10
Note: Ignition power, 80 W; O2 concentration, 1%; treatment time, 5 s.

*N originated from the ambient environment during plasma treatment.
treatment. On the other hand, when the distance reached 7 mm, the velocity and the
activity of the active species in the plasma jet would be greatly reduced when reaching
the surface of the PET film and thus did not effectively affect the surface of the film.
7.4 CONCLUSIONS
Nanoscale He–O2 APP etching and hydrophilization on PET film surface was achiev-
able under controlled conditions. Nanosurface structure density was found to vary with
the treatment time as a quadratic relationship. Ignition power and jet distance related
to nanospot density and surface hydrophilicity in a linear manner. Oxygen concentra-
tion contributed to the correlated factor in the treatment. In investigating optimum
conditions for plasma treatment, 3 s was determined to be the optimum treatment time
achieving the development of the nanospot in the range of 60 to 95 nm diameter on the
PET film with the highest overall surface energy 84 mJ/m2 using the APPJ operating
at 80 W under 5 mm jet distance by feeding 1% O2/He gaseous admixture.
ACKNOWLEDGMENTS
This work was supported by a grant from the Research Grants Council of the Hong Kong
Special Administrative Region, China, under the project PolyU 5192/08E and research
grant from the Hong Kong Polytechnic University. The XPS analysis was conducted
in the Centre for Surface Analysis and Research of the Hong Kong Baptist University.
REFERENCES
1. K.K. Samanta, M. Jassal and A.K. Agrawal, Surface Coatings Technol., 203, 1336–1342
(2009).
2. Y. Shin and D.I. Yoo, J. Appl. Polym. Sci., 108, 785–790 (2008).
3. S.R. Matthews, Y.J. Hwang, M.G. McCord and M.A. Bourham, J. Appl. Polym. Sci., 94,
2383–2389 (2004).
4. Y. Ren, C.X. Wang and Y.P. Qiu, Surface Coatings Technol., 202, 2670–2767 (2008).
5. C. Zhang and K. Fang, Surface Coatings Technol., 203, 2058–2063 (2009).
6. C.W. Kan and C.W.M. Yuen, Nucl. Instrum. Methods Phys. Res. B, 266, 127–132 (2008).
7. M. Šimor, J. Ráhel, M. Černák, Y. Imahori, M. Štefečka and M. Kando, Surface Coatings
Technol., 172, 1–6 (2003).
8. T. Wang, C. Wang and Y. Qiu, J. Appl. Polym. Sci., 108, 25–33 (2008).
9. C. Borcia, G. Borcia and N. Dumitrascu, Appl. Phys. A, 90, 507–515 (2008).
10. A. Vesel, I. Junkar, U. Cvelbar, J. Kovac and M. Mozetic, Surface Interface Anal., 40,
1444–1453 (2008).
11. D.H. Kaelble, J. Adhesion, 2, 66–81 (1970).
12. B. Gupta, J. Hilborn, C.H. Hollenstein, C.J.G. Plummer, R. Houriet and N. Xanthopoulos,
J. Appl. Polym. Sci., 78, 1083–1091 (2000).
13. B.D. Beake, J.S.G. Ling and G.J. Leggett, J. Mater. Chem., 8, 1735–1742 (1998).
14. B.D. Beake, J.S.G. Ling and G.J. Leggett, J. Mater. Chem., 8, 2845–2854 (1998).
15. A. Schütze, J.Y. Jeong, S.E. Babayan, J. Park, G.S. Selwyn and R.F. Hicks, IEEE Trans.
Plasma Sci., 26, 1685–1693 (1998).
16. J.Y. Jeong, J. Park, I. Henins, S.E. Babayan, V.J. Tu, G.S. Selwyn, G. Ding and R.F.
Hicks, J. Phys. Chem. A, 104, 8027–8032 (2000).
8 A Simple Way to
Aerosol-Based DBD
Provide Polymers with

Functional Groups for
Adhesion Promotion
Renate Mix,* H. Yin, and Jörg F. Friedrich
CONTENTS
8.1 Introduction................................................................................................... 156
8.2 Experimental................................................................................................. 158
8.2.1 Materials............................................................................................ 158
8.2.2 DBD Treatment.................................................................................. 158
8.2.3 XPS Characterization........................................................................ 159
8.2.4 Contact Angle Measurements............................................................ 159
8.2.5 Metal Deposition............................................................................... 159
8.2.6 Peel Test Measurements..................................................................... 159
8.3 Results............................................................................................................ 160
8.3.1 Influence of Aerosols of Pure Liquids on Surface Properties
and Peel Strength of DBD-Treated PP in Comparison to Air
Corona–Treated Samples................................................................... 160
8.3.2 Introduction of Aerosols of Aqueous Solutions of
Functionalized Substances into the DBD.......................................... 164
8.4 Conclusions.................................................................................................... 167
References............................................................................................................... 169
* Dr. Renate Mix was the lead author for this chapter and can be contacted at renate.mix@bam.de.
155
Polypropylene (PP) foils were exposed to the plasma of an atmospheric dielec-

tric barrier discharge (DBD) in presence of different aerosols. Pure liquids
such as water, ethanol, and 1/1 mixtures ethanol/water (v/v) or aqueous solu-
tions of acetic acid, poly(acrylic acid), poly(vinyl alcohol), and ethylene glycol–
poly(vinyl alcohol) copolymer were introduced into the DBD. Surface
composition, the number of functional groups per 100 C atoms, and water con-
tact angles were correlated with the energy density of DBD plasma exposure.
The wettability of the treated surfaces was investigated for different storage
times. Peel strength of the thermally deposited aluminum layer to modified
polymer surfaces revealed that the aerosol introduction strongly improves the
adhesion between aluminum and polypropylene.
8.1 INTRODUCTION
Generally, polymers and, in particular, polyolefins are characterized by chemical
inertness coupled with hydrophobicity. The superior bulk properties, the low pro-
duction costs, and the easy manufacturing of polyolefins are not adequately utilized
because of the absence of functional groups on the surface. Therefore, polyolefins
are principally difficult to print, to metallize, or to bond adhesively to other materi-
als. Different methods are developed to overcome this drawback by equipping the
polymer surfaces with functional groups.
Atmospheric pressure plasma techniques are often used for cleaning, activating,
and coating of polymers, paper, metal foils, glass, or ceramic sheets. Nowadays,
life sciences need also often functionalized (polyolefin) surfaces but prepared with
low costs. Several plasma methods for producing the desired functionalized polymer
surfaces were inspected and compared [1]. The dielectric barrier discharge (DBD)
treatment is widely used for preparation of polyolefin surfaces for printing, laminat-
ing, or coating. The ordinary DBD treatment (often also called “corona”) is per-
formed in air. Other gas atmospheres are also used such as nitrogen, argon, etc.
[2–6]. Fundamental studies of the DBD modification of polyolefins were published
by Strobel and coworkers [7,8] and by Novak and colleagues [9–11]. It was found that
air-DBD generates all types of oxygen-containing functional groups at polyolefin
surfaces such as OH, C–O–C–, epoxy, >C═O, CHO, COOH, COOR, C(O)O–OH,
C(O)O–OR, and CO3. At the same time, however, the plasma particle shower and the
action of plasma radiation destroy the topmost layer of the polyolefin surface. Low
molecular weight oxidized materials (LMWOMs) are formed, which are composed
of oligomers and other low molecular weight products. Cross-linking also occurs.
The polymer debris hinders the adhesion to the intact polyolefin macromolecules
because it is no longer bonded to the bulk polymer and, therefore, produces a weak
boundary layer. However, these degradation products do not influence the adhesion
when they can rapidly diffuse into coatings or inks. Otherwise, the degraded and
oxidized material can be removed by washing with water or other polar solvents [12].
A noticeable problem from the technical point of view is the “surface dynamics”
at polyolefin surfaces causing the “hydrophobic recovery” characterized by reori-
entation of functionalized polymer segments toward polymer bulk [13–15]. This
effect is responsible for the rapid loss in adhesion ability of DBD-treated polyolefin
Aerosol-Based DBD 157
surfaces during storage and exposure to the ambient air. Guimond and Wertheimer
[12] compared the atmospheric pressure glow discharge in nitrogen (N2 APGD) and
the corona discharge treatment of low-density polyethylene (PE) and biaxially ori-
ented polypropylene (PP) in air. They found that the hydrophobic recovery after
corona treatment was a long-lasting process while that after N2 APGD exposure was
more rapidly proceeding within a few days [12].
Another interesting idea was the introduction of monomer aerosols into the DBD
as claimed by Goodwin et al. [16]. Different monomers such as pure acrylic acid
also in combination with antimicrobial substances [17,18], hexamethyldisiloxane,
other siloxanes or silanes [19–21], and recently allylamine [22] were introduced into
the DBD plasma to form thin polymer coatings with special properties. Helium,
argon, or nitrogen was used as carrier gas. Vangeneugden et al. [23] introduced some
acrylates and methacrylates into the DBD plasma for producing silicon-free release
liners for pressure-sensitive adhesive tapes. Mix et al. [24] injected simple liquids
(water, ethanol) as well as 1 wt.% aqueous solutions of functionalized polymers such
as poly(vinylamine), poly(acrylic acid), and poly(vinylpyrrolidone) into the DBD
plasma using air as carrier gas. The aim of these different preparation methods was
to equip PE surfaces with different functional groups either by oxidation or by poly-
mer coating [24].
For several purposes, especially for grafting, the introduction of monotype func-
tional groups is necessary at the polymer surface. Only a few plasma processes are
able to produce exclusively one sort of functional group. Among such processes,
the polymer deposition from aerosols of polymer solutions with assistance of atmo-
spheric DBD plasma is preferred because the process is rapid and efficient as needed
for industrial applications. In the presence of the air DBD plasma degradation of the
injected polymer occurs [24], but sufficient functional groups survive the plasma
deposition process and produce the desired more or less monotype-functionalized
polymer surface [25].
This paper deals with DBD treatment of PP foils in the presence of aerosols
generated from different liquids, such as water, ethanol, or aqueous solutions of
functionalized film-forming macromolecular or oligomeric substances. Both pro-
cesses should introduce adhesion-promoting functional groups onto the polyole-
fin surface. However, the aerosols made from liquids produce a broad variety of
O-functional groups linked to the polyolefin molecules (functionalizing mode),
maybe with dominance of OH groups as can be derived from the structure of
precursor molecules (H2O, C2H5OH). The polymer and copolymer precursors
(film-forming mode) used here contain only one sort of functional group in their
molecules in stoichiometric concentration. Therefore, polymer layers deposited
from such aerosols possess the given monotype functional groups. Other types
of functional groups can only be produced by fragmentation or other side reac-
tions depending on the intensity and duration of plasma exposure. The aim of this
work was to increase the concentration of OH or COOH groups and adjust their
dominance among all other produced groups using the water and ethanol addi-
tion. By depositing the nebulized polymers in the DBD, a feasible way should be
found between maximal adhesion of this layer and the retention of their functional
groups in the layer.
8.2 EXPERIMENTAL
8.2.1 Materials
PP (Goodfellow, U.K.) of 100 μm thickness and PE (Bernburg Folien GmbH,
Germany, Type HBM 35, thickness 35 μm) were ultrasonically cleaned in diethyl
ether for 15 min. The X-ray Photoelectron Spectroscopy (XPS) inspection of the
two types of cleaned foils showed oxygen concentrations at the surface of <0.5 at%.
As functionalizing aerosols, the following substances were used: water (bi-
distilled), ethanol (Merck, Germany, >99%), and 1% aqueous solutions of acetic acid
(Th. Geyer GmbH, Germany, >99%); and as film-forming aerosols, the following
oligomers/polymers were tested: Kollicoat IR (copolymer made from ethylene glycol
and vinyl alcohol, MW = 45.000 g/mol, BASF, Germany), poly(acrylic acid) (PAA)
(MW = 1.800 g/mol, Aldrich, USA), and poly(vinyl alcohol) (PVA) (Celvol 205S,
Celanese, USA, MW = 30.000–50.000 g/mol).
For determination of retained or produced OH-groups at the PP surface, their
derivatization with trifluoroacetic anhydride (TFAA, >99%, Merck, Germany) was
performed in combination with XPS measurement of the thus introduced F-atoms.
Carboxylic groups were labeled with trifluoroethanol (TFE, >99%, Merck, Germany)
in the presence of pyridine (>99%, Merck, Germany) as a catalyst and 1,3-di-tert-
butylcarbodiimide (>99%, Fluka, Switzerland) as a water trapping substance.
8.2.2 DBD Treatment
The experiments were performed using the DBD System TG 101 (Ahlbrandt GmbH,
Germany), shown in Figure 8.1, consisting of a powered ceramic-coated electrode
of 51 cm length and a ceramic-coated drum of 16 cm diameter as counter electrode.
The maximum energy input was 1 kW.
Aerosols were generated by a pneumatic nebulizer producing a drop size of about
1 μm (McGaw Air, Drägerwerk, Germany) using a nitrogen stream at 2.6 bar and an
air stream at 0.9 bar. The nebulized substances were injected into the plasma-filled
electrode gap at a concentration of 30–40 g/h estimated by weighing the paper filters,
which were treated for 1 min by nebulizing a 10% NaCl solution and subsequent drying.
Ceramic wrapped electrode(s)
DBD plasma
Aerosol
Cylindrical
Nebulizer
electrode
Polymer foil
FIGURE 8.1 Schematic of the DBD-aerosol system used.

The energy density Ed is given by the following equation:
P
Ed = * n [ J/cm 2 ] (8.1)
w* S
where P is power (in W), w is the length of the discharge zone (48 cm), n is the
number of DBD passes, corresponding to treatment time, and S is the rim speed of
drum electrode (in cm/s). The samples were treated at different energy inputs but at
constant rim speed of 5 m/min (8.3 cm/s). The number of passes through the DBD
zone was varied from 1 to 8 with power of 250 W and 1 to 6 with power of 500 W.
The plasma residence time (tres [in s]) was defined as the ratio of the width of the
discharge zone b, which was determined from photographs of the working treatment
process, and the rim speed S:
b
tres = (in S) (8.2)
S
8.2.3 XPS Characterization
XPS spectra were acquired with a SAGE 150 spectrometer (Specs, Germany) using
nonmonochromatic MgKα radiation (11 kV, 18 mA) at a pressure of <1 × 10 –7 mbar in
the analysis chamber. The angle between the axis of x-ray source and analyzer lenses
was 54.9°. The analyzer was mounted at 18° to the surface normal. XPS spectra were
measured in the constant analyzer energy (CAE) mode. The analyzed surface area was
about 3 mm × 4 mm. The measured elemental concentrations were referenced to 100 C
atoms in %. The derivatization procedures with TFAA or TFE for characterization and
quantification of the hydroxyl and carboxylic groups were reported elsewhere [26–29].
8.2.4 Contact Angle Measurements

For characterization of wettability, the static water contact angle (sessile drop) was
determined with the Automatic Contact Angle System G2 (Krüss, Germany). The
drop volume was 3 μl. Usually contact angles of 4 drops were measured. If the
standard deviation exceeded 1°, further 4 drops were measured and the result was
reported as an average of 6 drops.
8.2.5 Metal Deposition
For Al evaporation, the DBD-treated polymer foils were transferred to a metallizer
(Auto 306, Edwards, U.K.). The thickness of all thermally deposited aluminum lay-
ers was adjusted to 100 nm as determined by a quartz microbalance.
8.2.6 Peel Test Measurements

The metal–polymer adhesion was tested using the 90° peel test. The samples were
prepared by bonding the aluminum side of the laminates onto a rubber support by
means of a double-sided adhesive tape (TESA 4965, Beiersdorf, Germany). The 90°
peel test was carried out at a peel speed of 25 mm/min. The width of the samples was
10 mm. For estimation of the average peel force, the force–distance curve was evalu-
ated for a distance of 45 mm beginning after 20 mm peeling. The adhesion strength
was averaged from three to five tested specimens. The standard deviation in the peel
strength varied between 10% and 15%.
8.3 RESULTS
8.3.1 Influence of Aerosols of Pure Liquids on Surface
Properties and Peel Strength of DBD-Treated PP in
Comparison to Air Corona–Treated Samples
Oxygen incorporation and water contact angle are plotted in Figure 8.2a–c for air-
DBD treated PP measured within one day after modification. O (12%–17%) and 3
to ~5 OH groups (referred to 100 C atoms) at the surface were detected. The water
contact angle decreased from 85° to 67°–73° with increasing energy density using
freshly produced samples. Storage at room temperature increased the contact angle
(a) (b)
20 6
5
16
O-groups/100 C
4
O (atom %)
12
3
8 250 W
500 W 2
4 1 250 W
PP untreated 500 W
PP untreated
0 0
0 1 2 3 4 5 6 7 8 0 1 2 3 4 5 6 7 8
Energy density (J/cm2) Energy density (J/cm2)
(c)
110
PP untreated
100
90
80
100 d
70 28 d
0–1 d
60
50
0 1 2 3 4 5 6 7 8
Energy density (J/cm2)
FIGURE 8.2 Oxygen content of air-DBD treated PP in dependence on applied energy den-
sity (a), number of OH groups/100 C atoms (b), and water contact angles of DBD-treated PP for
different storage periods (c) for treatments using 250 W (filled symbols) or 500 W (open symbols).
within 4 weeks to 72°–80° and after 3 months to 76°–82° caused by the hydrophobic
recovery.
Water-aerosol DBD-modified PP showed in comparison to air-DBD treated only
faintly changed surface composition and functionality. Most noteworthy findings
were the decreasing trend of the OH group concentration during storage if produced
with energy densities > 2 J/cm2 and the constant contact angle at storage times ≥ 1
month (see Figure 8.3a–c).
Ethanol–aerosol DBD modified PP (Figure 8.4a–c) shows a similar oxygen incor-
poration and formation of OH-groups in dependence on energy density as found for
air-DBD treatment. Contact angles of ethanol-aerosol DBD modified PP at 250 W
decreased if energy densities < 4 J/cm2 were applied. Samples produced at 500 W
showed contact angles between 65° and 69° weakly increasing with energy den-
sity. After 1 month storage, these differences disappeared and the measured contact
angles were found to be between 78° and 80°. The oxygen incorporation in the 30
days stored samples confirmed that the decrease in wettability was due to the hydro-
phobic recovery as evidenced by disappearance of 10%–20% of oxygen-containing
functional groups (Figure 8.4a).
(a) (b)
20 5
16 4
O (atom %)
OH/100 C
12 3
8 2
250 W
500 W 1 250 W
4 500 W
0 0
0 1 2 3 4 5 6 7 8 0 1 2 3 4 5 6 7 8
(c)
110
100 Untreated PP
90
80
28 d/100 d
70
1d
60
50
0 1 2 3 4 5 6 7 8
FIGURE 8.3 Oxygen incorporation by water-aerosol DBD treatment of PP in dependence

on applied energy density (a), number of OH groups/100 C atoms (b), and water contact angles
for different storage times (c). Treatment at 250 W (filled symbols) or 500 W (open symbols).
(a) (b)
20 6
5
0–1 d 0–1 d
15
30 d 4
O (atom %)
OH/100 C
10 3
2
5 250 W
, 250 W 1 500 W
Untreated PP , 500 W
Untreated PP
0 0
0 1 2 3 4 5 6 7 8 0 1 2 3 4 5 6 7 8
(c)
110
100 , 250 W
Untreated PP , 500 W
90
80
30 d
70 0–1 d
60
50
0 1 2 3 4 5 6 7 8
FIGURE 8.4 Oxygen concentration (a), OH-functionality (b), and water contact angles
(c) of PP surfaces modified with ethanol–aerosol DBD in dependence on applied energy den-
sity measured directly after treatment (0–1 day) or after 30 days storage at r.t.
800 250 W
500 W
600
Peel strength (N/m)
400
200
Al–PP (untreated)
0
0 1 2 3 4 5 6 7 8
FIGURE 8.5 Peel strength of Al–air DBD treated PP.

Figures 8.5 through 8.8 illustrate the effect of the aerosol-assisted DBD surface
modification of PP on the peel strength of Al–PP laminates in comparison to air
DBD. Additional to water and ethanol, also a 1:1 mixture (v/v) of water and ethanol
was nebulized into the discharge.
Using the DBD treatment in air, peel strength between 300 and 400 N/m was
achieved. The peeled Al and polymer surfaces showed interfacial failure. Similar
results were found when applying water aerosol. Most samples exhibited peel strengths
of about 300 N/m. Ethanol and water/ethanol (1:1) aerosol DBD were more efficient
in adhesion promotion. The aluminum layer could not be peeled, that is, the peel front
propagated between aluminum and the double-sided adhesive tape, which was used to
fix the samples onto the support. Therefore, the peeling between the supporting tape
and the rubber support was measured, and the Al–PP laminates must possess higher
peel strength than the measured 800 N/m. There was only one sample (produced at
500 W power) where the aluminum layer partially could be removed (cf. Figure 8.8).
800
600
Peel strength (N/m)
400
200
Al–PP (untreated)
0
0 1 2 3 4 5 6 7 8
FIGURE 8.6 Peel strength of Al–water aerosol DBD treated PP.
1200
1000
Peel strength (N/m)
800
600 No aluminum
was removed!
400
200 Al–PP (untreated)
0
0 1 2 3 4 5 6 7 8
FIGURE 8.7 Peel strength of Al–ethanol aerosol DBD treated PP.

1200
1 sample with partially
1000 removed Al
Peel strength (N/m)

800
600
400
0
0 1 2 3 4 5 6 7 8
FIGURE 8.8 Peel strength of Al–ethanol/water (1:1) aerosol DBD treated PP.
8.3.2 Introduction of Aerosols of Aqueous Solutions

of Functionalized Substances into the DBD
Recent works have shown that, in particular, hydroxyl and carboxyl groups are able
to promote the adhesion in PP–Al laminates [30,31]. Therefore, 1% aqueous solu-
tions of OH- and COOH-containing precursors such as acetic acid (CH3COOH),
PAA, PVA, and ethylene glycol/vinyl alcohol copolymer (Kollicoat IR) were nebu-
lized and introduced into the DBD discharge as aerosols.
Figures 8.9 through 8.12 demonstrate the efficiency of these aerosols in DBD
treatment on adhesion of Al to PP. The measured peel strength is depicted versus
energy density.
In Figure 8.9, the influences of nebulized 1% acetic acid solution introduced into
DBD on oxygen content and COOH concentration are plotted; additionally, both
parameters are plotted versus the plasma residence time tres (Figure 8.9b and d). One
pass for the sample through the DBD plasma zone lasted about 0.4 s.
Nebulizing an aerosol of 1% acetic acid into the DBD produced oxygen concen-
trations of 12–18 at%; however, only about 1.6 COOH groups/100 C atoms could be
incorporated into the surface layer of PP. The Ototal/COOH ratio indicated that the
acetic acid molecules were strongly fragmented and decarbonylized/decarboxylized
in the DBD plasma, thus explaining the low COOH concentration.
An analogous behavior was observed for RF-plasma polymerized acrylic acid
layers [32].
Nevertheless, the Al peel strength exceeded 800 N/m, and the metal could not be
peeled. The peel strength was found to be considerably higher than as shown before
with water (Figure 8.6). Only in one sample the aluminum layer could be partially
removed.
Comparable peel test results were obtained using the PAA aerosol (Figure 8.10c).
Although the oxygen incorporation in the PP surface was limited to 4–7 at% and
less than 1 COOH group/100 C was detected (Figure 8.10a and b), showing a low
(a) (b)
20 20
15 15
O (atom %)
O (atom %)
10 10
250 W
5 5 500 W
250 W
500 W
0 0
0 1 2 3 4 5 6 7 8 0 1 2 3 4
Energy density (J/cm2) Residence time in plasma (s)
(c) (d)
2.0 2.0
1.5 1.5
COOH/100 C
COOH/100 C
1.0 1.0
250 W
0.5 500 W 0.5
250 W
0.0 0.0 500 W
0 1 2 3 4 5 6 7 8 0 1 2 3 4
Energy density (J/cm2) Residence time in plasma (s)
(e)
1200
1000
Peel strength (N/m)
800
600 Al partially
removed
400
Al–PP (untreated)
200
0
0 1 2 3 4 5 6 7 8
FIGURE 8.9 Oxygen incorporation (a, b) and COOH concentration (c, d) of PP produced by
DBD modification using nebulized 1% acetic acid solution and peel strength (e) of appropri-
ately modified Al–PP laminates.
coverage of the PP surface and/or extensive polymer fragmentation [33], a strong

positive effect on peel strength (>700 N/m, Al not peelable) was observed.
The measured water contact angles of the modified PP were found to be about 90°
independent of energy density indicating the existence of a polyolefin-like surface
with a low concentration of polar groups.
(a) (b)
10 2.0 250 W
250 W
500 W 500 W
8
1.5
OH groups/100 C
O (atom %)
6
1.0
4
0.5
2
0 0.0
0 1 2 3 4 5 6 7 8 0 1 2 3 4 5 6 7 8
(c)
1000
800
Peel strength (N/m)
600
400
0
0 1 2 3 4 5 6 7 8
FIGURE 8.10 Influence of PAA aerosol on oxygen content (a), COOH concentration of
DBD-modified PP (b), and peel strength of Al–PP laminates (c).
This observation confirms the significant role of COOH-groups in Al adhesion

promotion also in very low concentrations, which do not notably influence the water
contact angle.
The effect of hydroxyl groups on the adhesion was proved by applying aerosols of
PVA and the copolymer vinyl alcohol/ethylene glycol (Kollicoat IR). Both polymers
were introduced as 1% aqueous solution. Figures 8.11 and 8.12 demonstrate these
types of surface modifications and depict the peel strength values. While the PVA-
PEG modification produced only moderate peel strength between 300 and 400 N/m
and the interfacial failure dominated (Figure 8.11c), the pure PVA aerosol generated
Al peel strengths > 700 N/m (Figure 8.12c).
The mechanism of Al interaction with COOH groups was discussed in terms
of acid–base interactions [34] or more specifically in terms of chemical bonds
[35]. Formation of Al–monodentate bonds (Al–O–CO–R) between AlOOH cov-
ered Al surfaces formed in air is weakly favored in comparison to bidentate bonds
((Al–O)2C–R). The evaporation of Al under high-vacuum conditions and the interac-
tions with carboxylic groups confirm the bonding of more than one Al atom with
one COOH group, indicating monodentate and bidentate formation [36]. Moreover,
(a) (b)
10 5
250 W 250 W
500 W 500 W
8 1000 W 4 1000 W
OH groups/100 C
O (atom %)
6 3
4 2
2 1
0 0
0 2 4 6 8 10 12 14 16 0 2 4 6 8 10 12 14 16
(c)
600
250 W
500 500 W
1000 W
Peel strength (N/m)
400
300
200
Al–PP (untreated)
100
0
0 2 4 6 8 10 12 14 16
FIGURE 8.11 Oxygen content (a) and number of OH groups/100C (b) for PP modified
by DBD in presence of an aerosol of a solution of vinyl alcohol/ethylene glycol copolymer
(Kollicoat IR) and peel strength (c) of Al-modified PP laminates.
individual evaporating Al atoms are very reactive toward all oxygen (or fluorine)
containing groups in the polymer molecules. Therefore, the hydrolysis-sensitive
metal–oxygen–carbon bond is strongly metastable, and more likely, the reaction
of Al atoms with carboxylic groups proceeds to a complete redox reaction with
formation of aluminum oxide (2 Al + HOOC-R → Al2O3 + degraded polymer)
[37,38]. The redox reaction produces also an adjustment of the different surface
energies of metal and polymer, which is advantageous for adhesion promotion by
providing more thermodynamic compatibility [39].
8.4 CONCLUSIONS
1. Aerosol-assisted DBD modification of PP is an interesting way to equip
chemically inert polyolefin surfaces with functional groups using common
DBD equipment under atmospheric pressure conditions in air. The use of
expensive rare gases is not required, but a common nebulizer for aerosol
application is additionally necessary.
(a) (b)
20 8
15 6
OH groups/100 C
O (atom %)
10 4
5 2 250 W
250 W 500 W
500 W
0 0
0 1 2 3 4 5 6 7 8 0 1 2 3 4 5 6 7 8
(c)
1000
800
Peel strength (N/m)
600
400
Al–PP (untreated)
200
0
0 1 2 3 4 5 6 7 8
FIGURE 8.12 Oxygen incorporation (a) and OH concentration per 100 C atoms (b) of PP
modified by DBD in the presence of an aerosol of 1% solution of PVA in water and measured
peel strength (c) of Al evaporated PP laminates
2. Simple liquids, appropriate monomers, and polymers with functional

groups can be applied as aqueous solutions. Special attention was given to
the use of environment-friendly chemicals to avoid any additional health
and safety measures.
3. Both, hydroxyl and carboxyl groups containing precursors, are very efficient
in adhesion promotion between aluminum and PP. The adhesion-promoting
effect of acetic acid, PAA, and PVA on the adhesion of aluminum–PP lami-
nates was well demonstrated. Obviously, the success is the consequence of
the combined action of functional groups and surface activation. It could be
shown that even small concentrations of functional groups can improve the
adhesion strongly.
4. Furthermore, it is speculated that the aerosol introduction cools the poly-
mer surface by removal of the evaporation heat and decreases the surface
degradation and therefore the formation of a weak boundary layer. The
measured high peel strength or the impossibility of peeling was not studied
using washing procedures of the polymer surfaces after DBD treatment.
REFERENCES
1. J. F. Friedrich, R. Mix, R.-D. Schulze, A. Meyer-Plath, R. Joshi, and S. Wettmarshausen,
Plasma Proc. Polym. 5, 407 (2008).
2. A. Villermet, P. Cocolios, G. Rames-Langlade, F. Coeuret, J.-L. Gelot, E. Prinz, and F.
Förster, Surf. Coat. Technol. 174–175, 899 (2003).
3. N. De Geyter, R. Morent, T. Desmet, M. Trentesaux, L. Gengembre, P. Dubruel, C.
Leys, and E. Payen, Surf. Coat. Technol. 204, 3272 (2010).
4. G. Borcia, C. A. Anderson, and N. M. D. Brown, Surf. Coat. Technol. 201, 3074 (2006).
5. N. De Geyter, R. Morent, C. Leys, L. Gengembre, and E. Payen, Surf. Coat. Technol.
201, 7066 (2007).
6. K. Wang, W. Wang, D. Yang, Y. Huo, and D. Wang, Appl. Surf. Sci. 256, 6859 (2010).
7. V. Jones, M. Strobel, and M. J. Prokosch, Plasma Proc. Polym. 2, 547 (2005).
8. M. Strobel, C. Dunatov, J. M. Strobel, C. S. Lyons, S. J. Perron, and M. C. Morgen,
J. Adhesion Sci. Technol. 3, 321 (1989).
9. I. Novak, V. Pollak, and I. Chodak, Plasma Proc. Polym. 3, 355 (2006).
10. I. Novak, and S. Florian, Polymer Int. 50, 49 (2001).
11. I. Novak, V. Pollak, and I. Chodak, Angew. Makromol. Chem. 260, 47 (1998).
12. S. Guimond, and M. R. Wertheimer, J. Appl. Polym. Sci. 94, 1291 (2004).
13. T. Yasuda, T. Okuno, K. Yoshida, and H. Yasuda, J. Polym. Sci. B: Polym. Phys. 26, 1781
(1988).
14. J. Behnisch, A. Holländer, and H. Zimmermann, Surf. Coat. Technol. 59, 356, (1993).
15. F. Truica-Marasescu, S. Guimond, P. Jedrzejowski, and M. R. Wertheimer, Nucl.
Instrum. Methods Phys. Res. B 236, 117 (2005).
16. A. J. Goodwin, P. J. Merlin, J. P. S. Badyal, and L. J. Ward, PCT Patent WO 0228548
(2002).
17. L. J. Ward, W. C. E. Schofield, J. P. S. Badyal, A. J. Goodwin, and P. J. Merlin, Chem.
Mater. 15, 1466 (2003).
18. L. O’Hare, L. O’Neill, and A. J. Goodwin, Surf. Interface Anal. 38, 1519 (2006).
19. L. J. Ward, W. C. E. Schofield, J. P. S. Badyal, A. J. Goodwin, and P. J. Merlin, Langmuir
19, 2110 (2003).
20. B. Twomey, M. Rahman, G. Byrne, A. Hynes, L.-A. O’Hare, L. O’Neill, and D. Dowling,
Plasma Process. Polym. 5, 737 (2008).
21. J. Bour, J. Bardon, H. Aubriet, D. Del Frari, B. Verheyde, R. Dams, D. Vangeneugden,
and D. Ruch, Plasma Proc. Polym. 5, 788 (2008).
22. J. Petersen, T. Fouquet, M. Michel, V. Toniazzo, A. Dinia, D. Ruch, and J. A. S. Bomfim,
ACS Appl. Mater. Interfaces 4, 1072 (2012).
23. D. Vangeneugden, M. Dubreuil, E. Bongaers, and W. Hendrix, Contrib. Plasma Phys.
49, 648 (2009).
24. R. Mix, J. Friedrich, and A. Rau, Plasma Proc. Polym. 6, 566 (2009).
25. J. Friedrich, R. Mix, R.-D. Schulze, and A. Rau, J. Adhesion Sci. Technol. 24, 1329
(2010).
26. A. Chilkoti, B. D. Ratner, and D. Briggs, Chem. Mater. 3, 51 (1991).
27. F. Fally, I. Virlet, J. Riga, and J. J. Verbist, J. Appl. Polym. Sci. 59, 1569 (1996).
28. M. R. Alexander, P. V. Wright, and B. D. Ratner, Surf. Interface Anal. 24, 217 (1996).
29. A. P. Ameen, R. J. Ward, R. D. Short, G. Beamson, and D. Briggs, Polymer 34, 1795,
(1993).
30. J. F. Friedrich, R. Mix, and G. Kuehn, Surf. Coat. Technol. 200, 565 (2005).
31. R. Mix, G. Kuehn, and J. F. Friedrich, in: Adhesion Aspects of Thin Films, K. L. Mittal
(Ed.), Vol. 2, p. 123, VSP, Utrecht (2005).
32. R. Mix, J. Falkenhagen, and J. Friedrich, in: Proc. MACRO 2010: 43rd IUPAC World
Polymer Congress, B7_P06, Glasgow, (2010).
33. M. Nitschke, S. Zschoche, A. Baier, F. Simon, and C. Werner, Surf. Coat. Technol. 185,
120 (2004).
34. S. R. Leadly, and J. F. Watts, J. Electr. Spectr. Rel. Phenom. 85, 107 (1997).
35. M. R. Alexander, G. Beamson, C. J. Blomfield, G. Leggett, and T. M. Duc, J. Electr.
Spectr. Rel. Phenom. 121, 19 (2001).
36. G. L. Fisher, A. E. Hooper, R. L. Opila, D. L. Allara, and N. Winograd, J. Phys. Chem.
B 104, 3267 (2000).
37. J. Friedrich, I. Loeschcke, and J. Gähde, Acta Polymerica 37, 687 (1986).
38. J. Friedrich, in: Polymer-Solid Interfaces, J. J. Pireaux, P. Bertrand, and J. L. Bredas
(Eds.), pp. 443–454, Institute of Physics Publishing, Bristol, (1991).
39. G. Habenicht, Kleben. Springer, Berlin (2009).
Section III
Adhesion Aspects of Bio-Based
Materials and Bio-Adhesion
9 Wood Joints Adhesion
and Performance in
Mechanical Friction
Welding of Wood
without Adhesives
A. Pizzi*
CONTENTS
9.1 Different Types of Wood Welding Methods.................................................. 174
9.1.1 Linear Vibration Welding.................................................................. 174
9.1.2 Rotational Dowel Welding................................................................. 177
9.2 Conclusion..................................................................................................... 179
References............................................................................................................... 180
The equipment used for mechanical friction welding of wood in absence of an

adhesive is the same type of equipment used for metals and yield wood joints
of considerable strength. Mechanical friction welding of wood gives bonds
satisfying the relevant standards. The influence on the bond of friction welding
time, contact holding time, pressure, and shift amplitude of one surface rela-
tive to the other are of importance. The mechanism of mechanically induced
wood vibration welding has been shown to be due mostly to the melting and
flowing of wood intercellular polymer material. This causes the “ungluing”
of long wood fibers and the formation of a fiber entanglement network in the
matrix of solidified molten cell-interconnecting material. Thus, a high-density,
high-strength wood cell/fiber entanglement network composite having a mol-
ten lignin polymer matrix is formed. Parameters influencing the water resis-
tance of welded wood joints are also discussed.
* Prof. A. Pizzi can be contacted at antonio.pizzi@univ-lorraine.fr.
173
9.1 DIFFERENT TYPES OF WOOD WELDING METHODS

9.1.1 Linear Vibration Welding
Thermoplastic welding techniques that are widely used in the plastic and car indus-
tries have recently been applied also to joining wood by melting a thermoplastic
polymer between the two wood surfaces to be joined. A variety of techniques such as
ultrasound, mechanical friction, and others have been used to melt the thermoplastic
polymer in situ.
However, the same mechanical techniques at the interface of two solid wood
pieces in the absence of any thermoplastic material, or any other binder, yield joints
of considerable strength [1,2]. The equipment used for the mechanical vibration
welding of wood in absence of an adhesive is the same type of equipment as used
for frictional welding of metals. Figure 9.1 shows the characteristic linear vibrational
movement of the type of industrial metal welding machine used as well as the fric-
tional shift and force applied to the two pieces of wood during welding.
The mechanism of mechanically induced wood vibration welding has been shown
to be due mostly to the thermoplastic behavior of some amorphous, cell-intercon-
necting polymer material in the structure of wood, mainly lignin, but also hemicel-
luloses. This causes partial detachment, the “ungluing” of long wood cells, wood
fibers, and the formation of a fiber entanglement network in the matrix of thermo-
plastic cell-interconnecting material, which then solidifies. Thus, a wood cell/fiber
entanglement network composite having a thermoplastic matrix of lignin polymer
and hemicelluloses, mainly xylans, is formed. Scanning electron micrographs show
the detail of the type of composite formed in the bond line of a solid wood joint
[1] (Figure 9.2). The two characteristic modes of welding are shown for scots pine
(Figure 9.2a) and for oak (Figure 9.2b).
Linear welding of wood can give bonding results satisfying the relevant standards,
while orbital welding gives much lower values. Some of the parameters that influ-
ence welding of metals with the same type of equipment also influence wood weld-
ing. Thus, the influences on the final bond of the vibration welding time, the contact
holding time after the welding vibration has stopped, the welding pressure exerted
on the surfaces, the holding pressure after the welding vibration has stopped, and the
amplitude of the shift imparted to one surface relative to the other during vibrational
welding are of importance. Welding frequencies of 100 Hz have been used. The
newer technology uses 150 Hz [3,4]. The joint tensile strength depends on vibration
amplitude, showing some good bond strength for 3 mm vibrational amplitude; on
Applied force
Wood Linear shift
FIGURE 9.1 Linear frictional movement for wood welding.

Mechanical Friction Welding of Wood without Adhesives 175
(a) 2 µm
(b) 10 µm
FIGURE 9.2 Two types of bond lines obtained in linear friction (Scot pine). Images from
a scanning electron microscope with variable pressure. (a) Late wood tracheid entangled in
the molten network of the bond. (b) Direct welding of the cell walls showing intact tracheid
structure and fused intercellular material.
welding pressure, with values of 2 to 2.3 MPa giving the best results; and on welding
time, but less markedly than on welding pressure. In general, combinations of 1.5 to
3 s welding time and 4 to 5 s holding time give strong joints presenting strength in
excess of 10–12 MPa and sometime of the order of 15 MPa. The relevant European
norm for these types of joints requires strengths equal to or higher than 5 MPa.
The strong joints obtained are not capable of satisfying specifications for exte-
rior joints as they show very poor resistance to water. These joints can then only be
considered for interior applications such as for furniture and for interior grade wood
joints. However, recent findings have shown that resistance to water immersion for
1 year or longer is possible for linear vibration welded wood joints [5,6]. Furthermore,
the technique at this stage is only usable for solid wood joints and for joints between
premanufactured panels presenting the same type of characteristics as solid wood,
such as oriented strand board (OSB). The technique, however, has considerable inter-
est for its low cost and in the implementation of totally environment-friendly wood
joints in joinery and furniture manufacturing.
During the welding period, some of the detached wood fibers that are no longer
held by the interconnecting material are pushed out of the joint as excess fibers.
Cross-linking chemical reactions also have been confirmed to occur (the most likely
of these identified by nuclear magnetic resonance appears to be a cross-linking reac-
tion of lignin with carbohydrate-derived furfural and furfural self-polymerization)
[1,7,8]. These reactions, however, are relatively minor contributors during the very
short welding period. Their contribution increases after welding has finished,
explaining why some holding time under pressure after the end of welding contrib-
utes strongly to obtaining a good bond.
As the equipment is expensive and the length of joints that can be welded is lim-
ited to a maximum of 1.5 m, one application that is particularly successful with lin-
ear welding is the preparation of butt joints without adhesives and without the need
for expensive finger jointing. Even more interesting is the preparation of butt joints
using high-density Australian eucalyptus species. These have shown good perfor-
mance (Table 9.1) without the need for finger jointing or gluing them [9].
The appearance of these end-grain joints is similar to finger jointing at the ana-
tomical rather than macro level (Figure 9.3). End-grain-to-end-grain welding gave
butt joints of good strength when using three high-density Australian eucalyptus
woods, namely, Sidney blue gum (Eucalyptus saligna), Spotted gum (Eucalyptus
maculata, Corymbia maculata spp.), and Black butt gum (Eucalyptus pilularis).
The absence or limited tendency to defibration in end-grain-to-end-grain welding
for wood densities as high as 800–900 kg/m 3 indicates that end-grain-to-end-
grain welding is possible and yields good joint strengths, but with a different
appearance of the welded interphase. At the anatomical level, three features were
noticeable:
1. The paintbrush-like appearance of the cell tips bent by the frictional move-
ment during welding. This leaves these cell tips free to intertwine with each
other from the surface of the opposite wood piece.
2. The absence of a great amount of molten material between the straight
shafts of the parallel bundles of cells, showing that the greater part of the
TABLE 9.1
Strength Results of Welded Australian Eucalyptus Wood Joints
Welding Cycle Average Welded
Time (s) Pressure (MPa) Displacement (mm) Joint Strength (MPa)
Sidney blue gum 1/1/4 0.25/0.25/1.75 1/2/2 6.6 + 0.4
Spotted gum 1/1/6 0.25/0.25/1.75 1/2/2 8.6 + 0.5
Black butt gum 1/1/4 0.25/0.25/1.75 1/2/2 5.3 + 0.4
FIGURE 9.3 Finger joint–like appearance at the wood anatomical level of a high-density
Australian eucalyptus wood butt joint obtained by end-grain linear wood welding. Note the
separation of fibers at the end of the fibers and the molten material in between.
intercellular material has melted and flowed away from the cells toward the
interphase.
3. The amount of molten intercellular material found on the upper surface of
the cell tips, giving to the joint the appearance of an anatomical finger joint.
9.1.2 Rotational Dowel Welding

High-speed rotation-induced wood dowel welding (Figure 9.4), without any adhe-
sive, has been shown to rapidly yield wood joints of considerable strength [10–15].
The mechanism of mechanically induced high-speed rotation wood welding has
been shown to be due, as already observed in vibration welding, to the temperature-
induced softening and flowing of some amorphous, cell-interconnecting polymer
material in the structure of the wood interphase, mainly lignin, but also of hemicel-
luloses and consequent high densification of the bonded interface. Wood species,
Wooden dowel
Predrilled
hole
Wood
substrate
FIGURE 9.4 Schematic representation of rotational wood dowel welding.

relative diameter difference between the dowel and the receiving hole (Figure 9.4),
and pressing time were shown to be the parameters yielding significant strength dif-
ferences; while relative orientation of the fiber grain of the dowel in relation to the
fiber grain of the substrate, relative rate of rotation within a limited range, and the
use of rough or smooth dowels did not have any significant influence. The welded
contact area is sufficient to yield strength results comparable to or even slightly
higher than those obtained by poly(vinyl acetate) (PVAc) adhesive bonding, but in a
fraction of the time (Table 9.2).
High-speed dowel rotation welding was used to manufacture a full-scale 400 cm ×
400 cm × 22 cm raised access floor, that is, an applicable civil engineering struc-
ture, to demonstrate that scaling up of the welding technique was feasible. This was
coupled with obtaining a more lightweight floor assembly at equal stiffness by maxi-
mizing the rigidity of the suspended floor while minimizing the number of timber
boards used to build it, and maintaining its vibration frequency high and its level
of vibration low. Several assembly and connection combinations of two and three
boards linked through welded wood dowels were tried to determine the resistance
of the crossover joints that had to be used in the building of the floor. Deformation
under 4-point static load on the floor was carried out to determine displacement
under load, and the floor vibration behavior was determined by the use of acceler-
ometers. The fundamental first natural frequency measured does satisfy well the
requirements specified by Eurocode 5.
The recently developed rotational dowel welding techniques for structural appli-
cations [14,15] such as for suspended housing floors [15] can be applied to the manu-
facture not just of normal furniture but also of furniture that otherwise could not
have been manufactured without metallic or other special supports. The example
at hand and presented here is the Z-chair devised by Dutch architect Gerrit T.
Rietveld (1888–1964). For the first time, it was possible to build this type of chair
without metallic or angle supports just by using rotationally welded wooden dowels
(Figure 9.5). Equally possible is the preparation of panels with good performance
through edge doweling (Figure 9.6) [16].
TABLE 9.2
Average Tensile Strength Results for Dowels Inserted
to 20 mm Depth in Single Beech Blocks, Tested Dry
and after 24 h Cold Water Soak, and Comparison
with PVAc Glued Dowels
Dry (N) 24 h Cold Water Soak (N)
Welded dowels a 1979 ± 103 1746 ± 153
PVAc-glued dowelsb 1844 ± 177 1286 ± 224
a After 3 s welding.
b After 24 h clamping.
FIGURE 9.5 Rietveld’s Z-chair manufactured by dowel welding alone. Note the two differ-
ent angles of dowel insertion to avoid a corkscrew failure of the joint.
FIGURE 9.6 Blockboard prepared by rotational dowel welding.
9.2 CONCLUSION
Wood linear welding has been used for several applications giving good performance
of the welded joints. The high price of the machines for wood linear welding and the
short length of the wood joints that can be manufactured with this technique limit
the number of applications for which this technique can be used. However, welding
of high-density, difficult-to-glue hardwoods is possible with this technique, and even
more interesting, preparation of butt joints having equivalent or better strength than
glued finger joints appears possible. Conversely, wood rotational dowel welding uses
easily available inexpensive machinery. It allows a greater flexibility of the type of
joints that can be assembled by this technique. Furniture impossible to prepare with
other techniques has been shown to be possible using this technique. Furthermore,
this technique has been shown to be well adapted to the manufacture of wood struc-
tures of great dimensions.
REFERENCES
1. B. Gfeller, M. Zanetti, M. Properzi, A. Pizzi, F. Pichelin, M. Lehmann and L. Delmotte,
J. Adhesion Sci. Technol., 17: 1425–1590 (2003).
2. J.-M. Leban, A. Pizzi, S. Wieland, M. Zanetti, M. Properzi and F. Pichelin, J. Adhesion
Sci. Technol., 18: 673–685 (2004).
3. H.R. Mansouri, P. Omrani and A. Pizzi, J. Adhesion Sci. Technol., 23: 63–70 (2009).
4. P. Omrani, A. Pizzi, H. Mansouri, J.-M. Leban and L. Delmotte, J. Adhesion Sci.
Technol., 23: 827–837 (2009).
5. H.R. Mansouri, A. Pizzi, J.-M. Leban, L. Delmotte, O. Lindgren, M. Vaziri and P.
Omrani, J. Adhesion Sci. Technol., 25: 1987–1995 (2011).
6. A. Pizzi, H.R. Mansouri, J.-M. Leban, L. Delmotte, P. Omrani and F. Pichelin,
7. L. Delmotte, C. Ganne-Chedeville, J.-M. Leban, A. Pizzi and F. Pichelin, Polymer
Degrad. & Stabil., 93: 406–412 (2008).
8. C. Ganne-Chedeville, M. Properzi, J.-M. Leban, A. Pizzi and F. Pichelin, J. Adhesion
Sci. Technol., 22: 761–773 (2008).
9. M. Mansouri, J.-M. Leban and A. Pizzi, Wood Sci. Technol., 44: 399–406 (2010).
10. A. Pizzi, J.-M. Leban, F. Kanazawa, M. Properzi and F. Pichelin, J. Adhesion Sci.
Technol., 18: 1263–1278 (2004).
11. F. Kanazawa, A. Pizzi, M. Properzi, L. Delmotte and F. Pichelin, J. Adhesion Sci.
Technol., 19: 1025–1038 (2005).
12. C. Ganne-Chedeville, A. Pizzi, A. Thomas, J.-M. Leban, J.-F. Bocquet, A. Despres and
H.R. Mansouri, J. Adhesion Sci. Technol., 19: 1157–1174 (2005).
13. A. Pizzi, A. Despres, H.R. Mansouri, J.-M. Leban and S. Rigolet, J. Adhesion Sci.
Technol., 20: 427–436 (2006).
14. J.-F. Bocquet, A. Pizzi and L. Resch, J. Adhesion Sci. Technol., 20: 1727–1739 (2006).
15. J.-F. Bocquet, A. Pizzi, A. Despres, H.R. Mansouri, L. Resch, D. Michel and F. Letort,
16. B. Belleville, C. Segovia, A. Pizzi, T. Stevanovic and A. Cloutier, J. Adhesion Sci.
Technol., 25: 2745–2753 (2011).
10 Effect of Pretreatment
of Rice Straw Used as a
Bio-Filler in Reinforced
Polypropylene
Composite
A. Karim, M. Masudul Hassan,*
and Mubarak A. Khan
CONTENTS
10.1 Introduction................................................................................................... 182
10.2 Experimental................................................................................................. 183
10.2.1 Materials............................................................................................ 183
10.2.2 Mercerization Treatment................................................................... 183
10.2.3 Acetylation Treatment....................................................................... 183
10.2.4 Composite Fabrication....................................................................... 184
10.2.5 Determination of Mechanical Properties of Composites.................. 184
10.2.6 Scanning Electron Microscopy Characterization.............................. 184
10.2.7 Water Absorption Test....................................................................... 184
10.2.8 Simulated Weathering Test................................................................ 185
10.3.1 TS and Elongation at Break of Composites....................................... 185
10.3.2 Bending and Impact Strength............................................................ 186
10.3.3 Effect of Surface Treatment on RSPP Composites............................ 187
10.3.3.1 Mercerization Effect........................................................... 187
10.3.3.2 Acetylation Effect............................................................... 188
10.3.4 Scanning Electron Microscopy......................................................... 188
10.3.5 Water Uptake..................................................................................... 188
10.3.6 Simulated Weathering Effect............................................................. 188
* Prof. M. Masudul Hassan was the lead author for this chapter and can be contacted at msdhasan@
yahoo.com.
181
10.4 Conclusion..................................................................................................... 191

Acknowledgment.................................................................................................... 191
References............................................................................................................... 191
Rice straw (RS) fiber can be utilized as a bio-filler in polypropylene (PP) matrix
due to its lignocellulosic characteristics. In this study, rice straw–polypropylene
composites were prepared in different ratios 15:85, 25:75, 35:65, 45:55, and
50:50 (RS wt%/PP wt%) by extrusion molding process. Mechanical proper-
ties of the prepared composites such as tensile strength (TS), bending strength
(BS), impact strength (IS), and elongation at break (Eb%) were investigated.
The composite with a ratio of 35:65 (RS/PP) was found to be optimum, which
showed good performances with TS = 24 MPa, BS = 52 N/mm2, IS = 1.8 kJ/
mm2, and Eb = 27%. For further improvement, RS was pretreated for surface
modification such as mercerization and acetylation to prepare composites with
enhanced mechanical properties. From the experimental observations, it was
noticed that mercerized rice straw–polypropylene (MRSPP) exhibited superior
mechanical properties (TS = 31 MPa and Eb = 37%) than the acetylated RSPP
composites. Water uptake, simulated weathering, and soil degradation tests on
composites were also carried out.
10.1 INTRODUCTION
In recent years, there has been a growing interest in agricultural waste as a substi-
tute for wood-based raw materials. Rice straw (RS), among the various agricultural
wastes, can be a very interesting material as a filler in biodegradable polymer com-
posites, due to its good thermal stability, competitive specific mechanical properties,
availability, low cost, and light weight compared to other agricultural wastes [1].
Chemically, lignocellulosic RS fiber has similar composition as other natural fibers
(NFs) used in thermoplastics. RS mainly consists of carbohydrate components such
as hemicellulose, cellulose, and lignin [2]. On the other hand, high content of silica
represents an additional potential benefit regarding flame retardancy when used in
the building industry [3]. Worldwide, about 580 million tons of straw is left behind
in rice fields annually, representing a large, untapped resource of cellulose fibers
[4]. Global ecological issues and energy shortage, biodegradability, and lignocellu-
losic characteristics of RS have led to exploration for its utilization in recent decades
[5,6]. Currently many research groups have directed their work toward improving
mechanical and biodegradable properties of RS–polymer composites to achieve
products with lower cost [7,8].
Some of the disadvantages and limitations of NFs, when used as reinforcement
for composites, are related to the lack of proper interfacial adhesion, poor resis-
tance to moisture absorption, limited processing temperature to about 200°C, and
low dimensional stability (shrinkage, swelling). Surface modification of NFs prior
Pretreatment of Rice Straw Used as a Bio-Filler in PP Composite 183
to their use in composite materials is also needed to facilitate fiber dispersion and
induce bond formation between the fiber and the polymer matrix [9–11]. The fiber-
to-fiber interaction in NFs, resulting from intermolecular hydrogen bonding, usually
restricts the dispersion of fibers in the polymer matrix. Alkali treatment of NFs,
also called mercerization [12], is the usual method to produce high-quality fibers by
removing the natural and artificial impurities. Acetylation of NFs is a well-known
esterification method that is based on the reaction of cell wall hydroxyl groups of lig-
nocellulosic materials with acetic or propionic anhydride at elevated temperatures.
As a result of acetylation treatment, the following reaction takes place [10]:
Fiber–OH + (H3C–CO)2O
→ Fiber–O–CO–CH3 + CH3COOH
Our earlier work [13–16] reported that RS effectively reinforces the PP matrix,
and the prepared composite with a certain content of RS showed better mechanical
properties. The results presented in these works also indicate that it is possible to
enhance mechanical properties of hybrid filler reinforced composites by hybridiza-
tion of RS and other fillers (such as seaweed and jute) with PP matrix by extrusion
and injection molding processes. Encouraged by our previous work, the pretreat-
ment of RS (surface treatment, e.g., mercerization and acetylation) for suitable com-
patibility and better adhesion between filler (RS) and matrix (polypropylene [PP])
was attempted to achieve improved mechanical performance of the prepared RS–
polypropylene composite.
10.2 EXPERIMENTAL
10.2.1 Materials
RS was collected from Bangladesh Rice Research Institute (BRRI), Regional Centre,
Habigonj, Bangladesh. RS was dried at 80°C in a vacuum oven for 24 h prior to prep-
aration of the composites. PP (Melt Flow Index [MFI], 230, 2.16 kg = 47 g/10 min)
was used as the matrix polymer and was supplied by Polyolefin Co. Ltd. Singapore.
10.2.2 Mercerization Treatment
For this purpose, 20% NaOH solution was prepared and chopped RS fibers were
soaked in the solution for 1 h and heated for 1 h at 80°C. The weight of RS was deter-
mined before soaking. After neutralizing the NaOH with acetic acid, the RS fibers
were washed with distilled water and dried in an oven for 24 h.
10.2.3 Acetylation Treatment
RS was treated in glacial acetic acid for 1 h at 20 ± 2°C. It was further treated with
acetic anhydride containing concentrated H2SO4 as a catalyst for 5 min. RS was then
washed with distilled water and dried for composite preparation.
TABLE 10.1
Relative Amount (wt%) of Reinforcing Materials and Polymer
Matrix in Composites
Fillers as Reinforcing Materials (wt%) Polymer Matrix (wt%) Composites
Rice straw: 15 PP: 85 RS 15
10.2.4 Composite Fabrication

Composites were prepared by compounding using a single-screw extruder (Axon
AB Plastics Machinery, Sweden) and a hot-press machine. The extrusion process
was carried out at a screw speed of 15 rpm. The extruded compound was then com-
pression molded by hot pressing under 10 MPa for 5 min at 185°C into sheets of
suitable thickness for making the specimens as per ASTM standard. The relative
amounts (wt%) of reinforcing materials in the composites are presented in Table 10.1.
10.2.5 Determination of Mechanical Properties of Composites

The tensile properties of the composites were determined using a Universal test-
ing machine (model 1011, Lloyd Instruments, U.K.). Tensile strengths (TSs) were
measured and three-point bending tests were carried out in accordance with ASTM
D-53455 and ASTM D-53452, respectively. All the tests were performed in a dis-
placement control mode on a closed-loop servohydraulic MTS testing machine. The
crosshead speed was 1 mm/min. Impact strength was measured according to ASTM
D-53433. All tests were carried out at room temperature (20°C). At least five samples
were tested for each composition, and results were averaged.
10.2.6 Scanning Electron Microscopy Characterization

Scanning electron microscopy (SEM) micrographs were taken using an S-520 equip-
ment (Hitachi, Japan) under 20 kV accelerating voltage to analyze the fracture mor-
phology of the composites. A mechanical test bar was fractured in liquid nitrogen,
and the fracture surfaces were coated with gold before characterization.
10.2.7 Water Absorption Test

After oven drying at 105°C for 24 h, the composite specimens were kept in desicca-
tors at room temperature. The specimens were weighed before being immersed in
distilled water. They were periodically taken out of the water, surface dried with an
absorbent paper, reweighed, and immediately put back into the water. Water absorp-
tion was calculated according to ASTM D-5229.
10.2.8 Simulated Weathering Test

The composites were tested using a simulated weathering tester from Q-Lab,
Cleveland, Ohio (model Q.U.V.). The weathering test was performed in alternating
cycles of sunshine for 4 h (65 ± 2°C) and dew and condensation for 2 h (65 ± 2°C).
This was carried out for a period of about 30 days.

10.3.1 TS and Elongation at Break of Composites
The tensile properties such as TS and elongation at breaks of RS–PP (RSPP) com-
posites containing 15, 25, 35, 45, and 50 wt% filler were determined, and the results
are shown in Figures 10.1 and 10.2, respectively. It is observed that with an increase
in filler content from 15 to 35 wt%, the TS gradually increased, but it decreased with
30
Tensile strength (MPa)
25
20
15
10
5 15 25 35 45 55
RS content (wt%) in PP composites
FIGURE 10.1 Tensile strength (TS) of RSPP composites.
35
30
Elongation at break (%)
25
20
15
10
5
5 15 25 35 45 55
FIGURE 10.2 Elongation at break (%) of RSPP composites.

increasing filler loading from 40 to 50 wt%. This may be due to the lack of stress
transfer from the PP matrix to the filler [17]. The elongation at the break of the com-
posites shows a similar trend as shown by TS, and maximum elongation at the break
is obtained for 35 wt% RS sample. An increase in the elongation at the break of a
composite increases the toughness and ductility of the composite [18].
10.3.2 Bending and Impact Strength

The results on bending and impact strengths of the composites are shown in
Figures 10.3 and 10.4, in which an increasing trend with increasing RS content from
15 to 35 wt% is observed, followed by a decreasing trend. It is envisaged that as
the size of the filler becomes smaller, greater interaction between the filler and the
matrix could result in better and more efficient stress transfer, which, in turn, could
increase both bending and impact strengths of the composites [19]. From the results,
it can be noted that the composite with 35 wt% RS content exhibited better mechani-
cal properties.
Bending strength (N/mm2)
55
45
35
25
5 15 25 35 45 55
FIGURE 10.3 Bending strength (BS) of RSPP composites.
1.9
Impact strength (kJ/mm2)
1.8
1.7
1.6
1.5
1.4
5 15 25 35 45 55
FIGURE 10.4 Impact strength (IS) of RSPP composites.

10.3.3 Effect of Surface Treatment on RSPP Composites

10.3.3.1 Mercerization Effect
For better compatibility between RS and PP, a mercerization treatment was used to
modify the surface of the RS. After mercerization treatment, the weight of RS was
determined; it was found that the weight had changed and the weight loss was 30%.
This could be due to removal of certain impurities and chemical changes in the com-
position of RS. The changes in main chemical constituents of RS due to merceriza-
tion are shown in Table 10.2, which have also been mentioned elsewhere [20].
The effects of mercerization treatment on the RSPP composites are shown in
Figure 10.5, which represents comparative mechanical properties of the mercerized
(MRSPP) and untreated RS composites. It may be noted that the mercerization treat-
ment improves the fiber surface adhesion characteristics by removing natural and
artificial impurities, thereby producing a rough surface topography. It also changes
the form of the cellulose crystallites, increasing the amount of amorphous cellulose
at the expense of crystalline cellulose. Thus, mercerization separates the composite
TABLE 10.2
Changes in Main Components of RS after
NaOH Pretreatment
Content (%)
Component Untreated (Raw) NaOH Treated
Cellulose 33.4 29.9
Hemicellulose 28.2 19.1
Lignin 7.4 5.7
Ash 12.8 20.8
45
40
TS (MPa)
Eb (%)
Numerical values
35
30
25
20
RSPP MRSPP AcRSPP
RS content (35 wt%) in PP composites
FIGURE 10.5 Comparison of the tensile properties of various composites.

fiber bundle into smaller fibers, increasing the effective surface area available for
contact with PP polymer matrix during composite fabrication [21,22].
10.3.3.2 Acetylation Effect

The acetylation treatment resulted in the highest elongation at the break of the
AcRSPP composite, which is represented in Figure 10.5. The composite ability to
withstand the applied stress is manifested by higher elongation value (40%), which
indicates an elastic nature of the material. The hydroxyl groups that react with the
treatment reagent are those of lignin and hemicellulose, while the hydroxyl groups of
cellulose prevent the diffusion of the reagent. It has also been shown that acetylation
improves the dispersion degree of lignocellulosic materials in a polymer matrix PP,
as well as the dimensional stability and interface of the composites, which might be
responsible for the elongation enhancement of the composites [10,23,24].
10.3.4 Scanning Electron Microscopy

Scanning electron micrographs of untreated RS composite (RSPP), mercerized
RS composite (MRSPP), and acetylated RS composite (AcRSPP) are presented in
Figure 10.6a, b, and c, respectively. The image in Figure 10.6a shows a random distri-
bution of RS filler in the PP matrix, and it can also be seen that the RSs are completely
pulled out of the matrix where voids are visible. A better dispersion and pullout of the
fibers from the matrix are observed in the image of MRSPP composite shown in Figure
10.6b. Acetylation fills the gaps between the fibrils and ultimately smooths the surface
of the fiber as seen in the image of the AcRSPP composite presented in Figure 10.6c.
10.3.5 Water Uptake

Water uptake capacity is an important characteristic of a composite material,
which was studied by immersing in distilled water at 28°C. The water uptakes of
the untreated RSPP composite, mercerized RSPP composite, and AcRSPP compos-
ite are shown in Figure 10.7. It is observed that initially the uptake is linear, then
becomes gradual, and finally reaches a plateau. The highest water uptake was found
for untreated RSPP composite, and the lowest water uptake was found for acetylated
RSPP composite. The lowest water uptake of the acetylated composite indicates that
more OH groups of cellulose in the fibers of the composites are blocked by their
interaction with the PP matrix; also these are hindered from being accessed by water.
10.3.6 Simulated Weathering Effect

The three types of composites, namely, RSPP, MRSPP, and AcRSPP composites,
were subjected to accelerate weathering test for a period of about 30 days to study the
degradation of properties. The weight loss, TS, and Eb of the samples due to weath-
ering are shown in Table 10.3. The loss in TS of the RSPP composite is about 6%,
while that of MRSPP is about 5% and for AcRSPP is about 3%. Similarly, Eb loss is
24% for the RSPP sample, and for the MRSPP and AcRSPP samples, Eb losses are
22% and 18%, respectively.
(a) (b)
(c)
FIGURE 10.6 (a) SEM micrograph of the RSPP composite with 35 wt% rice straw content.
(b) SEM micrograph of the MRSPP composite with 35 wt% rice straw content. (c) SEM
micrograph of AcRSPP composite with 35 wt% rice straw content.
0.5
0.45
0.4
0.35
Water uptake (%)
0.3
0.25
0.2 RSPP
0.15 MRSPP
AcRSPP
0.1
0.05
0
0 20 40 60 80 100 120 140
Soak time (day)
FIGURE 10.7 Water uptake of various composites with 35 wt% rice straw content.
190
TABLE 10.3
Loss in Weight and Mechanical Properties of Composites due to Simulated Weathering
Weight Loss (%) Loss in TS (%) Loss in Eb (%)
Weathering Time (day) RSPP MRSPP AcRSPP RSPP MRSPP AcRSPP RSPP MRSPP AcRSPP
1 3 2.5 2 1 0.9 0.9 2 2 1.5
2 5 4 1.5 2 1.2 1 4 3 2
3 7 6 5 3.5 1.8 1.2 7 6 5
4 9 7 6 4 2 1.9 10 10.2 8
5 12 9 6.5 4.2 2.6 2 15 16 13
10 13 11 8 4.5 3.5 3 20 19 16
20 15 14 9 6.5 4.5 4.2 24 22 18
30 18 16 10 8 7 5 26 24 20
10.4 CONCLUSION
From the results of this work, it is evident that after a certain filler content, the tensile
properties decreased as the percentage of filler increased. However, the results of this
study demonstrate that the composite prepared with 35 wt% content of RS showed
better mechanical performance. The results presented in this work also indicate that
it is possible to enhance the properties of fiber-reinforced composites through fiber
surface modification (e.g., through mercerization and acetylation). Composites based
on the surface modified fiber of RS fibers have shown, in general, superior mechani-
cal properties than the composites containing unmodified RS fibers.
ACKNOWLEDGMENT
The authors are thankful to Diana J. Tartakowska of Berlin Technical University,
Germany, for the SEM micrographs of the composites.
REFERENCES
1. A. Buzarovska, G. Bogoeva-Gaceva, A. Grozdanov, M. Avella, G. Gentile and M.
Errico, Australian J. Crop Sci. 1, 37–42 (2008).
2. D. N. Saheb and J. P. Jog, Adv. Polym. Technol. 18, 351–363 (1999).
3. H. S. Yang, H. J. Kim, J. Son, H. J. Park, B. J. Lee and T. S. Hwang, Compos. Struct. 63,
305–312 (2004).
4. R. Narendra and Y. Yiqi, J. Agric. Food Chem. 54, 8077–8081 (2006).
5. A. V. Bridgwater, D. Meier and D. Radlein, Org. Geochem. 30, 1479–1493 (1999).
6. S. Yaman, Energy Convers. Manage. 45, 651–671 (2004).
7. F. Yao, Q. Wu, H. Liu, Y. Lei and D. Zhou, J. Appl. Polym. Sci. 119, 2214–2222 (2011).
8. F. Yao, Q. Wu, Y. Lei and Y. Xu, Ind. Crops Prod. 28, 63–72 (2008).
9. S. G. Lee, S. S. Choi, W. H. Park and D. Cho, Macromol. Symp. 197, 89 (2003).
10. S. Mishra, M. Misra, S. S. Tripathy, S. K. Nayak and A. K. Mohanty, Polym. Compos.
23, 164 (2002).
11. M. Avella, C. Bozzi, R. D. Erba, B. Focher, A. Marzzetti and E. Martuscelli, Angew.
Makromol. Chem. 233, 149 (1995).
12. G. Bogoeva-Gaceva, M. Avella, M. Malinconico, A. Buzarovska, A. Grozdanov, G.
Gentile and M. E. Errico, Polym. Compos. 28, 98–107 (2007).
13. M. M. Hassan, M. Mueller, D. J. Tartakowska and M. H. Wagner, J. Appl. Polym. Sci.
120, 1843–1849 (2011).
14. M. M. Hassan, M. H. Wagner, H. U. Zaman and M. A. Khan, J. Adhesion Sci. Technol.
25, 615–626 (2011).
15. M. M. Hassan, M. Mueller, D. J. Tartakowska, M. H. Wagner and M. A. Khan,
J. Adhesion Sci. Technol. 25, 1961–1971 (2011).
16. M. M. Hassan, M. H. Wagner, H. U. Zaman and M. A. Khan, J. Nat. Fibers 7, 165–177
(2010).
17. X. Colom, F. Carraco, P. Pages and J. Canavate, Compos. Sci. Technol. 63, 161 (2003).
18. S. Sultana, M. M. Huque and M. M. Helali, Polym. Plast. Technol. Eng. 46, 385–391
(2007).
19. H. D. Rozman, L. Musa and A. Abubakar, Polym. Plast. Technol. Eng. 44, 489–500
(2005).
20. H. Yanfeng, P. Yunzhi, L. Yanping, L. Xiujin and W. Kuisheng, Energy Fuels 22, 2775–
2781 (2008).
21. M. M. Hassan, M. R. Islam, S. Shehrzade and M. A. Khan, Polym. Plast. Technol. Eng.
42, 515–531 (2003).
22. M. A. Khan and M. M. Hassan, in: Polymer Surface Modification, Relevance to
Adhesion, Vol. 3, K. L. Mittal (Ed.), pp. 263–283 VSP, Utrecht (2004).
23. R. M. Rowell, in: Science and Technology of Polymers and Advanced Materials, P. N.
Prasad, J. E. Mark, S. H. Kendil and Z. H. Kafafi, (Eds.), p. 717, Plenum Press, New
York (1998).
24. K. Joseph, L. H. C. Mattoso, R. D. Toledo, S. Thomas, L. H. Carvalho, L Pothen, S. Kala
and B. James, in: Natural Polymers and Agrofibers Composites, E. Frallini, Leao, A. L.
Mattoso and M. H. C. (Eds.), p. 159, Embrapa, San Carlos, Brazil, (2000).
11 “Green” Composites
Based on Bacterial
Cellulose Produced
Using Novel Low-Cost
Carbon Source and
Soy Protein Resin
Kaiyan Qiu and Anil N. Netravali*
CONTENTS
11.1 Introduction................................................................................................... 194
11.2 Materials and Methods.................................................................................. 196
11.2.1 Microorganism and Culture Media................................................... 196
11.2.2 SF Powder Treatment......................................................................... 196
11.2.3 Sugar Consumption of SFE Medium during Culture........................ 196
11.2.4 BC Production................................................................................... 197
11.2.5 Preparation of SF and MSF Resin Sheets.......................................... 197
11.2.6 Fabrication of BC-Based “Green” Composites with SF and MSF
Resins................................................................................................. 197
11.2.7 Characterization................................................................................ 198
11.3.1 Influence of Autoclaving on the SFE Medium.................................. 198
11.3.2 Sugar Consumption in SFE Medium during the Culture.................. 199
11.3.3 BC Yield in SFE Medium..................................................................200
11.3.4 Microstructure of BC–Soy Resin Composite and Its Fabrication
Mechanism.........................................................................................202
11.3.5 Tensile Properties of Resins and Green Composites......................... 203
11.3.6 Thermal Stability of BC–SF and BC–MSF Resin Composites.........205
* Prof. Anil N. Netravali was the lead author for this chapter and can be contacted at ann2@cornell.edu.
193
11.4 Conclusions....................................................................................................206
Acknowledgments...................................................................................................206
References...............................................................................................................206
A novel low-cost carbon source, soy flour extract (SFE), was developed from
defatted soy flour (SF) to produce bacterial cellulose (BC). The enriched pro-
tein remaining after extracting sugars was used as a resin to form BC-based
thin membrane-like “green” composites. The results of this study showed that
SFE consists of five sugars: fructose, glucose, sucrose, raffinose, and stachy-
ose. The study indicated that Acetobacter xylinum, the bacterium used in this
study, metabolized sugars in the following decreasing order: fructose and glu-
cose, sucrose, raffinose, and stachyose during the culture process. However,
the consumption rates of raffinose and stachyose were extremely low. Results
also indicated that autoclaving process resulted in hydrolyzing sucrose to
fructose and glucose. Based on the same concentration of sugars, BC yield
achieved using SFE medium (based only on the concentration of fructose,
glucose, and sucrose) was close to or even higher than the yields obtained
using expensive conventional carbon sources such as fructose, mannitol, and
glucose. Modified soy flour (MSF), the residual protein after removing sug-
ars from SF, was successfully used as a resin to fabricate “green” BC–MSF
membrane-like thin composites. These composites had excellent tensile and
thermal properties that were better than the BC–SF resin composites. This is
primarily due to the higher protein content in MSF compared to SF. The sug-
ars in SF also plasticized the protein, thus reducing its modulus and increasing
the fracture strain.
11.1 INTRODUCTION
Bacterial cellulose (BC) produced by Acetobacter xylinum is a promising, sustain-
able, and biodegradable nanofibrous material that has the same chemical structure
as the plant-based cellulose. However, BC fibers have diameters in the range of a
few nanometers and display many unique characteristics including high purity, high
degree of polymerization, high crystallinity, high tensile strength, high modulus, and
strong biological adaptability [1–5]. The BC material is already being used in many
applications including artificial skin and blood vessels, binding agent for fibers and
other materials, loud speaker diaphragms, high-quality paper, foods, textiles, com-
posite membranes, and so forth [5–10]. Many pure sugars, such as glucose, sucrose,
fructose, and so forth, and sugar alcohols, such as mannitol, xylitol, sorbitol, and so
forth, have been used as carbon sources for BC culture [11–15]. Among them, glu-
cose, fructose, and mannitol are the most common and have shown excellent results
in terms of BC production [11–17]. However, the cost of these sugars or sugar alcohols
is high, and hence, they are not considered to be ideal for large-scale, inexpensive BC
production. As a result, many researchers have attempted to obtain higher BC yields
as well as to reduce the cost of the carbon sources. Some of these efforts have been
“Green” Composites Based on Bacterial Cellulose 195
successful. For example, konjac powder hydrolyzate [11], sugarcane molasses [18],
beet molasses [19], and processed rice bark [20] have been found to be useful in BC
production. While some of these sources may be used for industrial BC production
in the near future, there is significant scope to further reduce the cost of BC produc-
tion by using inexpensive waste products and expand its use in many mass volume
applications. This chapter describes a novel and inexpensive source derived from
soybeans for BC production.
Defatted soy flour (SF) is obtained as a by-product after extracting oil from
soybeans. It is commercially available and consists mainly of protein (52%–54%),
sugars(30%–32%), dietary fiber (2%–3%), minerals and ash (3%–6%), and mois-
ture (6%–8%). SF is also very inexpensive, about $0.25/lb. The soybean is a legume
species native to East Asia and is classified as an oilseed. It is an annual and eco-
nomic crop and has been abundantly produced and used in many countries for over
5000 years [21]. Currently, it is an important global crop and provides a major
amount of edible oil and protein [22]. Soybeans contain decent amounts of sugars,
including fructose, glucose, sucrose, raffinose, and stachyose [23]. Fructose, glu-
cose, and sucrose have been used routinely as carbon sources for BC production
[24]. It has also been reported that raffinose and stachyose can be metabolized by
lactic acid bacteria [25]. To obtain higher protein content from SF, the sugars are
removed in the form of soy flour extract (SFE), a by-product. The present research
discusses the use of SFE, a mixture of five sugars mentioned earlier, as an inexpen-
sive carbon source to produce BC.
Because of the excellent mechanical properties of BC, some research on fabrica-
tion of BC-based composites with petroleum-based resins has already been reported
[26,27]. High-strength composites using BC sheets impregnated with phenolic resin
or acrylic resin have been developed [24,25]. Although useful, these resins are not
biodegradable, and as a result, composites are not biodegradable or environment
friendly. It should, however, be possible to fabricate completely degradable BC-based
“green” composites using sustainable and biodegradable resins.
As mentioned earlier, SF contains sugars and protein. It is possible to dissolve and
extract water-soluble sugars from SF, termed soy flour extract (SFE). After remov-
ing SFE, the residual protein product, called modified soy flour (MSF), has a higher
protein content of 65%–70%. MSF-based resin has been shown to have better tensile
properties than SF resin [28]. MSF resin also has a higher interfacial bonding with
ramie fiber compared to SF resin [28]. As a result, composites using MSF resin could
be expected to have much better mechanical properties than those using SF resin.
In the present research, the sugar-containing SFE has been successfully devel-
oped for BC production as a high-yield and low-cost carbon source [29]. The con-
sumption of different sugars by A. xylinum and the compositional changes of sugars
in the SFE medium during autoclaving were analyzed as well. The residual insoluble
protein, after SFE extraction, called MSF, was successfully used as resin to fabricate
BC-based “green” composite. Thus, both fibers and resins were produced from defat-
ted SF as the sole feedstock. The tensile and thermal properties of the composites
were characterized and were found to be comparable to or even better than those of
traditional plastic materials.
11.2 MATERIALS AND METHODS

11.2.1 Microorganism and Culture Media
A. xylinum, ATCC 23769, obtained from the American Type Culture Collection
(ATCC, Manassas, VA), was used as the model strain and maintained on agar plates
containing 25 g/L d-mannitol, 5 g/L yeast extract and 5 g/L tryptone, and 20 g/L
agar. The SFE medium used for BC production consisted of 5 g/L yeast extract,
5 g/L tryptone, and the autoclaved SFE as the sole carbon source [29]. Other culture
media used for comparison of BC yields consisted of 25 g/L carbon sources raffi-
nose, glucose, sucrose, fructose, and mannitol, individually, and 5 g/L yeast extract
and 5 g/L tryptone, which have a similar function for BC production as Hestrin–
Schramm (HS) medium [17].
11.2.2 SF Powder Treatment

The SF (product no. 7B) powder obtained from ADM Co. (Decatur, IL) was mixed
with deionized water to obtain SFE. SF powder was initially soaked in deionized
water in a ratio of 3:17, and the pH of the mixture was adjusted to 4.5, its isoelec-
tric point, by adding hydrochloric acid. The mixture was kept at 50°C in a water
bath for 1 h. After that, the mixture was filtered to remove the solid content (MSF),
mostly the insolubilized protein. Part of the filtrate, containing the soluble sugars
(SFE), was then allowed to evaporate to obtain the desired sugar concentration for
BC culture [29].
11.2.3 Sugar Consumption of SFE Medium during Culture

The concentrations of sugars in the SFE, including fructose, glucose, sucrose, raf-
finose, and stachyose, were determined before and after autoclaving using high-
performance liquid chromatography (HPLC) (UltiMate 3000 LC system, Dionex,
Sunnyvale, CA) equipped with a refractive index (RI) detector (RI-101, Ecom,
Purage, Czech Republic). Autoclaving of the SFE was carried out at 121°C and about
0.1 MPa pressure in a sterilizer (Market Forge, Alfa Medical, Westbury, NY) for
25 min. After autoclaving, the SFE was filtered to remove the remaining solid pro-
tein deposits and used for BC culture. Sugar concentrations in the SFE during 10
days of culture were determined on a daily basis using HPLC to obtain a quantita-
tive measure of the sugar consumption by the bacteria. After filtering the SFE cul-
ture medium samples through a Teflon filter (0.45 μm pore size) and removing tiny
BC fibrils and other impurities, concentration for each sugar was analyzed using
a SUPELCOSIL LC-NH2 column (25 cm × 4.6 mm inner diameter (ID) and 5 μm
particles, Supelco, Bellefonte, PA) and the RI detector. The HPLC column was used
at 30°C. The mobile phase was a mixture of acetonitrile and deionized water (3:1,
v/v) and was kept at a flow rate of 1 mL/min. It is important to note that while all
sugar concentrations were determined individually, it was not possible to measure
the fructose and glucose concentrations separately, and hence, they were measured
together and reported as “fructose plus glucose” [29].
11.2.4 BC Production
The Acetobacter strain was inoculated into a conical flask containing the prepared
SFE culture medium as the seed culture. The initial pH value of the medium was
adjusted to 5.0 and was not regulated during the culture. The seed culture was incu-
bated at 30°C and 130 rpm on a rotary shaker for 2 days, and 6 mL of this was inocu-
lated into a 100 mL culture medium in a 600 mL conical flask for production of BC.
The cultivation was carried out initially at a pH of 5.0 and 30°C in a static incubator
for 10 days. Samples of the culture medium and BC were extracted every day during
the 10-day culture period to measure consumption of individual sugars and BC yields.
The BC pellicles taken out from the medium were washed successively with water and
1% (w/v), aqueous NaOH at 90°C for 15 min each and then washed with deionized
water to remove all microbial product contaminants. The purified cellulose pellicles
were finally dried at 105°C on a Teflon plate until constant weight was reached [29].
In another set of experiments, BC pellicles produced using other culture media
containing individual sugars (mentioned in Section Microorganism and Culture
Media) were harvested every day. The BC pellicles were then washed and dried
using the same procedure mentioned earlier. BC pellicles cultured in SFE medium
and other culture media (mentioned in Section Microorganism and Culture Media)
were compared for their yields. Dried BC specimens were conditioned at American
Society for Testing and Materials (ASTM) conditions of 21°C and 65% relative
humidity (RH) for 3 days before tensile testing [29].
11.2.5 Preparation of SF and MSF Resin Sheets

The SF powder and MSF obtained during SFE production were individually mixed
with deionized water at a weight ratio of 1:15. Glycerol was added (15% by weight) as
a plasticizer; the pH value of both solutions were adjusted to 10 by addition of sodium
hydroxide [22]. The solutions were maintained at 75°C while stirring continuously
for 30 min to obtain precured SF resin and MSF resin. This “precuring” process
helps denature the globular protein by opening up the molecules. Precured SF resin
and MSF resin were individually cast on a Teflon-coated glass plate and dried in a
35°C air-circulated drying oven for 16 h. A dried SF resin sheet and MSF resin sheet
were cured using a Carver hydraulic hot press (model 3981-4PROA00, Wabash, IN)
at 120°C for 25 min under a pressure of 7 MPa. The thickness of all resin sheets
was in the range of 0.2 mm. The cured SF and MSF resin sheets were conditioned at
ASTM conditions for 3 days prior to characterizing their tensile properties.
11.2.6 Fabrication of BC-Based “Green” Composites

with SF and MSF Resins
BC-based membrane-like “green” composites with SF and MSF resins were pro-
duced by using BC pellicles impregnated individually with precured SF and MSF
resins. Resin impregnation into the BC pellicle was achieved using ultrasonication
for 30 min. The wet BC–SF and BC–MSF resin composites were dried in an air-
circulating oven at 35°C for 8 h to obtain prepregs. The BC content in the BC–SF
resin composite and BC–MSF resin composite was kept around 50% in order to
achieve tensile properties that are comparable to traditional nonbiodegradable plas-
tics. While the BC content in both composites could be easily adjusted by varying
the resin concentration, it was difficult to obtain uniform distribution of resin with
higher BC concentration. The prepregs were then cured by hot-pressing at 120°C
under a pressure of 7 MPa. The thickness of all composites was in the range of
0.2 mm. The cured composites were conditioned at ASTM conditions for 3 days
prior to characterizing their tensile properties.
11.2.7 Characterization
Freeze-dried specimens of BC and BC–MSF resin composites were sputter-coated
with gold, and their surface topographies were observed with a scanning electron
microscope (SEM; LEO 1550 FESEM) at an accelerating voltage of 15 kV.
Tensile testing was performed using an Instron tensile testing machine (Instron,
model 5566). The test specimens were prepared by cutting the BC membranes and
green composites into 10-mm-wide and 60-mm-long strips using a precise cutter.
The Young’s moduli of the specimens were determined from the tensile test results
conducted according to ASTM D882-02. Two ends of the specimens were fixed
between the upper and lower jaws of the Instron, leaving a gauge length of 30 mm.
Crosshead speed during the tensile tests was maintained at 0.6 mm/min to obtain a
strain rate of 0.02/min.
Thermogravimetric analysis (TGA; TA Instruments, model no. 2050) was carried
out to analyze the thermal properties of both BC–SF and BC–MSF resin composites.
TGA runs were performed using aluminum pans between 25°C and 600°C under a
nitrogen environment. The scan rate was 20°C/min, and the nitrogen purge flow rate
was maintained at 10 mL/min.

11.3.1 Influence of Autoclaving on the SFE Medium
The HPLC analysis of the as-obtained SFE used in this study showed that it consisted
of 1.92 g/L fructose and glucose (combined), 21.21 g/L sucrose, 1.59 g/L raffinose,
11.92 g/L stachyose, water, and other components, including proteins. The concen-
tration of total sugars was over 36 g/L. After autoclaving (sterilizing) at 121°C and
pressure of 0.1 MPa for 25 min, however, the HPLC analysis showed a different
composition of sugars in the SFE medium. Therefore, the influence of autoclaving
on the SFE medium was further explored.
It has been reported that sucrose is prone to partial hydrolysis during autoclav-
ing, and hence, a sucrose-containing sterilization medium will result in a mixture of
d-glucose, d-fructose, and sucrose [30]. In another study it was shown that 15%–25%
of the sucrose may hydrolyze to glucose and fructose during autoclaving at the ele-
vated temperature [31,32].
Table 11.1 presents the HPLC data of various sugar concentrations in SFE before
and after autoclaving. Table 11.1 also gives adjusted values for all sugars after taking
TABLE 11.1
Effect of Autoclaving on the Concentrations of Various Sugars in the
SFE Medium
Concentrations of Sugars in SFE Medium
Fructose + Sucrose Raffinose Stachyose Total Sugars
Glucose (g/L) (g/L) (g/L) (g/L) (g/L)
SFE before autoclaving 1.92 21.21 1.59 11.92 36.64
SFE after autoclaving 7.54 17.54 1.58 9.92 36.58
SFE after autoclaving 7.16 16.66 1.50 9.42 34.74
(Adjusted data by
considering water
evaporation)
into consideration the water evaporation during autoclaving. Before autoclaving, the
freshly made SFE medium had concentrations of 1.92 g/L for fructose plus glucose,
21.21 g/L for sucrose, 1.59 g/L for raffinose, and 11.92 g/L for stachyose. After auto-
claving, the concentrations changed to 7.54 g/L for fructose plus glucose, 17.54 g/L
for sucrose, 1.58 g/L for raffinose, and 9.92 g/L for stachyose. The concentration of
total sugars was 36.58 g/L, and the concentration of three traditional carbon sources
(fructose, glucose, and sucrose) for BC production was approximately 23–25 g/L,
which was almost the same as the regular concentration of the carbon source used
for BC production by others [11].
The data in Table 11.1 clearly indicate that there was a significant (about 20%)
decrease in sucrose concentration and a corresponding increase in the concentration
of fructose plus glucose after autoclaving. As a result, there was no significant change
in the combined concentration of fructose, glucose, and sucrose, which remained in
the range of 23–25 g/L, before and after autoclaving. As discussed earlier, this was
mainly due to the degradation of sucrose [30–32]. During the autoclaving process,
our data also indicated a small decrease in the concentrations of both stachyose and
raffinose (Table 11.1). While the reasons for this are not well understood, this may be
either due to hydrolysis of raffinose and stachyose similar to that of sucrose or due
to side reactions such as caramelization or the Maillard reaction [33]. The splitting
of sucrose into glucose and fructose is, in fact, better as it will be seen later that the
A. xylinum bacteria metabolize glucose and fructose much more easily than sucrose.
11.3.2 Sugar Consumption in SFE Medium during the Culture

Our preliminary study had indicated that all five sugars in SFE, fructose, glucose,
sucrose, raffinose, and stachyose, could be used as carbon sources separately for BC
culture by A. xylinum, and different sugars had different effectiveness for BC yields.
A detailed discussion of this is presented later in this chapter. To measure the actual
consumption of individual sugars in the SFE culture medium, specimens were ana-
lyzed for sugar content every day.
20
18
Concentration (g/L)
16
14
12
10
8
6
4
2
0
0 1 2 3 4 5 6 7 8 9 10 11
Culture time (days)
FIGURE 11.1 Consumption of sugars in SFE medium during BC culture by Acetobacter

xylinum as a function of time in days. Symbols: ♦, fructose plus glucose; ◾, sucrose; ▴, raf-
finose; ×, stachyose.
Figure 11.1 shows plots of changes in the concentrations of all five sugars as
a function of culture time in days. As can be seen from the plots, concentration
of fructose plus glucose decreased steadily and almost linearly until day 7. During
that period, concentrations of the other three sugars remained more or less stable.
After the sixth day, the sucrose concentration started to decrease. This suggests
that when the concentration of fructose plus glucose decreased to an extremely low
value (around 1.09 g/L) from the initial 7.54 g/L, the A. xylinum started to consume
sucrose. These results indicate that the A. xylinum preferred to consume fructose and
glucose before the other three sugars present in the SFE medium. During the entire
10-day culture time, very little or no raffinose and stachyose were consumed, and as
a result, no significant change was noticed in their concentrations.
11.3.3 BC Yield in SFE Medium

It has been reported that BC yield in a fructose or glucose medium was higher than in
a sucrose medium [24]. Based on our preliminary experiments, fructose and glucose
were better carbon sources for BC production compared to a combination of sucrose,
raffinose, and stachyose, if the pH value of the medium was kept constant. As men-
tioned earlier, after autoclaving, the concentration of fructose plus glucose in the
SFE medium reached 7.54 g/L from the initial concentration of 1.92 g/L. This higher
concentration of fructose plus glucose was obviously beneficial for the BC culture.
It is important to note that the total concentration of the three sugars (fructose, glu-
cose, and sucrose) in the SFE medium that A. xylinum mainly consumed during the
10-day culture was around 25 g/L, which was almost the same as the concentration
of conventional carbon sources used by other researchers for BC production [11].
Figure 11.2 shows BC yields in SFE (blue), fructose (brown), mannitol (green),
glucose (purple), and sucrose (light blue) media as a function of culture time in days.
As seen in the blue curve in Figure 11.2, BC yield in SFE medium increased rapidly
during the initial 3–4 days. However, after 7 days of culture, the yield growth sig-
nificantly decreased. The main reason for this was that the preferred carbon sources,
fructose and glucose, were used up almost fully at this time. BC yield, however,
continued to increase during the period from days 7 to 10 but with a relatively lower
300
250
BC yield (mg)
200
150
100
50
0
0 1 2 3 4 5 6 7 8 9 10 11
Culture time (days)
FIGURE 11.2 BC yields in SFE (♦), fructose (◾), mannitol (▴), glucose (×), and sucrose (*)
media as a function of culture time.
rate partially because A. xylinum started to consume sucrose and other sugars, which
were not as suitable or metabolizable as fructose and glucose for BC production.
The results in Figure 11.2 also indicate that BC yield in an SFE medium can reach
255 mg after 10 days of culture, which is close to or even better than BC yields
obtained with other conventional carbon sources under similar culture conditions
[11,34]. According to the comparison presented in Figure 11.2, BC yields in mannitol
and fructose media have almost the same yield curves as in an SFE medium during
the initial 7 days of culture. However, the BC yield rates in mannitol and fructose
media were slightly higher than in an SFE medium during the period from days 7 to
10. This further confirms that preferred carbon sources, fructose and glucose, were
at low concentrations during this time. BC yields in pure glucose and sucrose media
show consistently lower BC yield values during the entire culture period.
Figure 11.3 compares BC yields obtained using different carbon sources individu
ally after 10 days of culture. The BC yield of 255 mg in the SFE medium was almost
as high as those using fructose (270.3 mg) and mannitol (276.3 mg) media, which
have been regarded as two excellent carbon sources for BC production. Based on
these data, it may be concluded that SFE could be used as an excellent carbon source
for BC production and one of the least expensive ones. BC yield in the SFE medium
was also significantly higher than those obtained individually in raffinose (29.7 mg),
sucrose (72.8 mg), and glucose (128.2 mg) media. Glucose has also been reported as
an excellent carbon source for BC production [34]. However, in our trial carried out
350
300
BC yield (mg)
250
200
150
100
50
0
Raffinose Sucrose Glucose SFE Fructose Mannitol
Carbon sources
FIGURE 11.3 BC yields obtained for different carbon sources.

with a pure glucose medium, under the same conditions, the yield was lower because
the pH value of the medium was not regulated and gluconic acid generated by glu-
cose during the culture caused the pH value to change to less than 3.5, which was
not suitable for BC production. However, glucose in an SFE medium still could be
used as a good carbon source because the pH value in SFE medium did not change
significantly during the culture. The reason might be that the relatively small amount
of gluconic acid formed during culture was not able to change the pH of the medium.
Also, other sugars present in SFE created a buffer effect in the SFE medium. Since
sucrose partially hydrolyzed into glucose and fructose during autoclaving, as seen in
our results as well as by others [30–32], the BC yield obtained in a sucrose medium
was from the combined presence of sucrose, fructose, and glucose.
The BC yield in the SFE medium (255 mg) was also higher than previously
reported BC yields obtained using konjac powder hydrolyzate (212 mg, A. xylinum
ATCC 23770, 8 days) and processed rice bark (242 mg, A. xylinum ATCC 23769, 10
days) [11,20].
In addition, the cost of the carbon source, one of the major expenses for BC cul-
tures, is indeed reduced largely because the SFE is a by-product obtained from the
soy protein production process. The cost of SFE is almost nothing. Therefore, based
on rough calculation, the cost for BC production can be reduced by more than 30%.
11.3.4 Microstructure of BC–Soy Resin Composite

and Its Fabrication Mechanism
The photographs of BC–SF resin and BC–MSF resin composites are shown in
Figure 11.4a(i) and (ii), respectively. Both of these thin membrane-like composites
(a) (i) (ii)
200 nm 200 nm
(b) (i) (ii)
FIGURE 11.4 (a) Images of (i) BC–SF resin composite and (ii) BC–MSF resin composite.
(b) SEM micrographs of (i) freeze-dried BC and (ii) freeze-dried BC–MSF resin composite.
had fairly smooth surfaces. This indicates that both SF resin and MSF resin are fully
embedded into BC to fabricate corresponding composites.
Figure 11.4b shows SEM images of a freeze-dried BC pellicle and a BC–MSF
resin composite. Both specimens were not hot-pressed prior to SEM observations.
In Figure 11.4b(i), the BC network and porous structure can be observed clearly at
the surface of the membrane. Since these specimens were freeze-dried, the porous
structure has been maintained. The mean diameter of BC nanofibers is less than
100 nm, and the pore diameters range from several tens to several hundred nano-
meters. Figure 11.4b(ii) shows the structure of a BC–MSF resin composite. MSF
resin penetrated into the BC network structure and filled in most of the pores of
BC. Again, since the specimen was freeze-dried, the porous structure was partially
maintained. The spread of the MSF resin in between the BC nanofibers and their
embedment in the resin can be clearly seen. If this resin containing membrane is
hot-pressed, the gaps are fully filled as a result of the resin flow and consolidation,
forming a thin membrane-like composite. The SEM micrographs of freeze-dried BC
and BC–MSF resin composite specimens shown in Figure 11.4b(i) and (ii), respec-
tively, look identical.
MSF resin infiltration into the BC pellicle is facilitated by the ultrasonication
process. When the wet pellicle is immersed into the resin solution, there exists a
big concentration difference between water inside of the BC network structure and
MSF resin outside of BC. The mass transfer was driven by the concentration gradi-
ent, resulting in MSF resin penetrating into the BC network structure. Since BC is
hydrophilic and the resin is water based, the diffusion is easy, though assisted by the
ultrasonication. The equilibrium could be achieved until there is no concentration
gradient between the resin inside of BC and outside of BC. Once the MSF resin infil-
trates the BC pellicle, the wet BC–MSF resin composite is achieved. It is then dried
and hot-pressed to form a strong composite. This composite fabrication procedure
was also used for SF resin as well.
11.3.5 Tensile Properties of Resins and Green Composites

Table 11.2 presents the tensile test results for BC, SF resin, and MSF resin, as well
as BC–SF resin composites and BC–MSF resin composites. The soy resin content in
both composites was around 50% by weight.
The Young’s modulus value for BC with randomly organized and entangled
nanofibers (Figure 11.4b(i)) was 2493 MPa, while the value for the MSF resin was
104.3 MPa. The Young’s modulus value for the BC–MSF resin composite was
1231 MPa, which was between the modulus values of MSF resin and BC.
The tensile strength at a break of BC was 78.9 MPa, and for MSF resin, it was 8.1
MPa. The tensile strength value for BC–MSF resin composite was around 47.7 MPa,
which was between those of BC and MSF resin.
Both the modulus and tensile strength values of the BC–MSF resin composite
are close or even higher than those of many traditional plastic materials, including
polyethylene (modulus 800 MPa, tensile strength 15 MPa); polypropylene (modulus
1900 MPa, tensile strength 40 MPa); and nylon 6 (modulus 1800 MPa, tensile strength
70 MPa) [35]. The BC–MSF resin composites may be easily protected from water by
TABLE 11.2
Tensile Properties of BC, SF Resin, MSF Resin, and BC–Soy Resin Composites
Theoretical Theoretical
Young’s Young’s Tensile Tensile
Modulus Modulus Strength Strength Tensile Strain
(MPa) (MPa) (MPa) (MPa) (%)
BC 2493 (9.6)* 78.9 (13.7) 5.7 (17.5)
SF resin 62.3 (31.9) 7.5 (4.0) 110.7 (28.4)
MSF resin 104.3 (22.0) 8.1 (11.1) 23.4 (25.2)
BC–SF resin 1178 (2.4) 1183 43.3 (5.3) 40.4 3.4 (2.9)
composite
(50 wt% BC)
BC–MSF 1231 (2.6) 1206 47.7 (9.3) 40.8 3.1 (41.9)
resin
composite
(50 wt% BC)
*Values in parentheses are % coefficient of variation values.
applying a varnish or other water-resistant coatings to increase their durability. They

have the potential to replace traditional nonbiodegradable plastic materials in many
applications, including racket frames, ski poles, circuit boards, automobile insides,
and so forth.
It should be noted that the BC nanofiber Young’s modulus has been estimated to
be up to 78 GPa or even up to 114 GPa [36,37]. However, the nanofibers in the pel-
licles are not unidirectionally oriented but instead are highly entangled as a result of
the random path traveled by the A. xylinum (Figure 11.4b(i)). Their strength, like the
Young’s modulus, is reduced significantly. This is reflected in the composite strength
as well.
The tensile strain value for MSF resin was 23.4%, which was much higher than
the value of 5.7% obtained for BC. The tensile strain of the BC–MSF resin compos-
ite, however, was only 3.1%. Although the tensile strain is controlled by the BC, since
both BC and resin are hydrophilic in nature, the fiber/resin bonding is expected to be
good. Better bonding allows the resin to lock the fibers in place, reducing the tensile
strain of the composite.
The SF resin and BC–SF resin composites were fabricated and used for compari-
son. The results presented in Table 11.2 indicate that MSF resin and BC–MSF resin
composites have higher Young’s modulus and tensile strength than those of SF resin
and BC–SF composites, respectively. The enhanced MSF resin properties are due to
high protein content in the MSF resin compared to SF [28]. Also, the SF resin con-
tains low-molecular-weight sugars (up to 35% by weight), which plasticize the resin.
As mentioned earlier, these sugars were removed to obtain MSF. This is evident in
the higher tensile strain and lower Young’s modulus obtained for SF. The enhanced
properties of the MSF resin are reflected in the BC–MSF composite properties as
can be expected, though the difference in tensile properties between BC–MSF resin
composites and BC–SF resin composites here is not significant, only 5%–10%. This
is because the tensile values of MSF and SF resins are an order of magnitude lower
than BC and both composites have 50 wt% BC content; the tensile values of the com-
posites are dominated by BC properties. However, both experimental and theoreti-
cal values shown in Table 11.2 indicate that BC–MSF resin composites have better
tensile properties than BC–SF resin composites. This conclusion was further con-
firmed by composites with 10 wt% BC content for the two resins in our preliminary
experiments.
Table 11.2 also presents theoretically calculated values for Young’s modulus and
tensile strength values calculated using the rule of mixture [38]. The densities of cel-
lulose and soy resins were 1.52 and 1.30 g/mL, respectively [39], and the weight ratio
of BC to resin is 1:1 in the composites. It is clear that the theoretical and experimen-
tal Young’s modulus values are very close, but the experimental strength values are
slightly lower, which is controlled by the defects in the specimens.
11.3.6 Thermal Stability of BC–SF and BC–MSF Resin Composites

TGA studies were conducted to obtain information on the thermal decomposition
behavior of BC–SF and BC–MSF resin composites. Figure 11.5 shows typical TGA
thermograms obtained for the two composites. These curves confirm that the BC–
MSF resin composite has better thermal stability than the BC–SF resin composite, as
can be expected. The degradation temperature, Td, for BC–SF resin composites was
found to be around 205°C, compared to 220°C obtained for BC–MSF composites.
This was, again, due to the higher protein content in the MSF resin than that in the
SF resin after removing sugars. The sugars start to degrade at much lower tempera-
tures. Our preliminary experiments had shown that fructose, glucose, and sucrose
start to degrade at around 160°C, 170°C, and 200°C, respectively, much lower than
the degradation temperatures of proteins and cellulose. These results confirm that
the BC–MSF resin composite is thermally slightly more stable than the BC–SF resin
composite and hence may be used at slightly higher temperatures or would be more
durable at the same temperature.
120
100
(a) (b)
80
Weight (%)
60
40
20
0
0 100 200 300 400 500 600 700
Temperature (°C)
FIGURE 11.5 TGA thermograms of (a) BC–SF resin composite and (b) BC–MSF resin
composite.
11.4 CONCLUSIONS
The present study has shown for the first time that SFE can be an excellent car-
bon source for BC production. The BC yield obtained with the new carbon source
(SFE) was high and close to or even better than those obtained with other expensive
conventional carbon sources. The cost of the SFE carbon source can be very low
because it is a by-product of SF processing and is produced in abundance throughout
the world.
The study also showed that SFE contains at least five sugars, and three of them,
fructose, glucose and sucrose, can be utilized as an excellent carbon source for BC
production. In addition, the study confirmed that the autoclaving process can hydro-
lyze higher sugars in SFE.
The results further indicated that A. xylinum prefers to consume fructose and
glucose before sucrose and other higher sugars such as raffinose and stachyose dur-
ing the culture. Importantly, the results show that the rate of BC production is much
higher when the concentration of fructose and glucose combined is high.
The study successfully developed membrane-like “green” composites using BC
obtained by SFE and MSF resin produced by residual protein product after SFE
extraction. The BC–MSF resin composite showed better tensile and thermal prop-
erties than the BC–SF resin composite. These composites could provide a “green”
choice to replace many traditional plastic materials.
ACKNOWLEDGMENTS
This work was partly supported by the National Textile Center (NTC) and the Wallace
Foundation. The authors would like to thank Profs. Dan Luo, Antje Baeumner, and
Jocelyn Rose of Cornell University for allowing the use of their facilities. We would
also like to acknowledge Dr. Jun Tae Kim for his help in the SFE process and useful
discussions about the HPLC data.
REFERENCES
1. M. Iguchi, S. Yamanaka and A. Budhiono, J. Mater. Sci. 35, 261–270 (2000).
2. H. Baeckdahl, G. Helenius, A. Bodin, U. Nannmark, B. R. Johansson, B. Risberg and
P. Gatenholm, Biomaterials 27, 2141–2149 (2006).
3. D. Klemm, D. Schumann, U. Udhardt and S. Marsch, Prog. Polym. Sci. 26, 1561–1603
(2001).
4. D. Klemm, B. Heublein, H. P. Fink and A. Bohn, Angew. Chem. Int. Ed. 44, 3358–3393
(2005).
5. H. P. Fink, P. Weigel, H. J. Purz and J. Ganster, Prog. Polym. Sci. 26, 1473–1524 (2001).
6. Y. Wan, L. Hong, S. Jia, Y. Huang, Y. Zhu, Y. Wang and H. Jiang, Compos. Sci. Technol.
66, 1825–1832 (2006).
7. J. D. Fontana, A. M. De Souza, C. K. Fontana, I. L. Torriani, J. C. Moreschi, B. J.
Gallotti, S. J. De Souza, G. P. Narcisco, J. A. Bichara and L. F. X. Farah, Appl. Biochem.
Biotechnol. 24–25, 253–264 (1990).
8. H. Shibazaki, S. Kuga, F. Onabe and M. Usuda, J. Appl. Polym. Sci. 50, 965–969 (1993).
9. A. Svensson, E. Nicklasson, T. Harrah, B. Panilaitis, D. L. Kaplan, M. Brittberg and
P. Gatenholm, Biomaterials 26, 419–431 (2005).
10. J. A. Westland, S. R. Stephens, W. C. Johnston Jr. and H. J. Rosenkrans, US Patent,

5207826 (1993).
11. F. Hong and K. Qiu, Carbohydr. Polym. 72, 545–549 (2008).
12. J. N. Saddler and D. J. Gregg, In: Forest Products Biotechnology. A. Bruce and J. W.
Palfreyman (Eds.), Taylor & Francis, London, pp. 183–195 (1998).
13. K. V. Ramana, A. Tomar and L. Singh, World J. Microbiol. Biotechnol. 16, 245–248
(2000).
14. H. L. A. Tarr and H. Hibbert, Can. J. Res. 4, 372–388 (1931).
15. R. Kaushal and T. K. Walker, Biochem. J. 48, 618–621 (1951).
16. M. Ishihara, M. Matsunaga, N. Hayashi and V. Tišler, Enzyme Microb. Technol. 31,
986–991 (2002).
17. S. Hestrin and M. Schramm, Biochem. J. 58, 345–352 (1954).
18. S. Keshk and K. Sameshima, Appl. Microbiol. Biotechnol. 72, 291–296 (2006).
19. S. Keshk, T. Razek and K. Sameshima, Afr. J. Biotechnol. 5, 1519–1523 (2006).
20. F. D. E. Goelzer, P. C. S. Faria-Tischer, J. C. Vitorino, M.-R. Sierakowski and C. A.
Tischer, Mater. Sci. Eng. C 29, 546–551 (2009).
21. J. G. Endres, Soy Protein Products: Characteristics, Nutritional Aspects and Utilization
(Revised and Expanded Ed.). AOCS Press, Boulder, Urbana, IL, pp. 4–18 (2001).
22. H. Martin, H. Laswai and K. Kulwa, Afr. J. Food Agric. Nutr. Dev. 10, 2040–2049
(2010).
23. E. Giannoccaro, Y. J. Wang and P. Chen, Food Chem. 106, 324–330 (2008).
24. Y. K. Yang, S. H. Park, J. W. Hwang, Y. R. Pyun and Y. S. Kim, J. Ferment. Bioeng. 85,
312–317 (1998).
25. Y. C. Wang, R. C. Yu, H. Y. Yang and C. C. Chou, Food Microbiol. 20, 333–338 (2003).
26. A. N. Nakagaito, S. Iwamoto and H. Yano, Appl. Phys. A-Mater. 80, 93–97 (2005).
27. S. Ifuku, M. Nogi, K. Abe, K. Handa, F. Nakatsubo and H. Yano, Biomacromolecules 8,
1973–1978 (2007).
28. J. T. Kim and A. N. Netravali, J. Agric. Food Chem. 58, 5400–5407 (2010).
29. A. N. Netravali and K. Qiu, Intl. Pub. No.: WO 2010/135234 A2 (2010).
30. J. H. Dobbs and L. W. Roberts, Experiments in Plant Tissue Culture (3rd Ed.). Cambridge
University Press, Cambridge, U.K., p. 53 (1995).
31. E. Ball, Bull. Torrey Bot. Club 80, 409–411 (1953).
32. N. Schenk, K. C. Hsiao and C. H. Bornman, Plant Cell Rep. 10, 115–119 (1991).
33. J. De J Berrios, P. Morales, M. Cámara and M. C. Sánchez-Mata, Food Res. Intl. 43,
531–536 (2010).
34. S. Keshk and K. Sameshima, Afr. J. Biotechnol. 4, 478–482 (2005).
35. Tensile property testing of plastics-Mat Web. http://www.matweb.com/reference/
tensilestrength.aspx.
36. G. Guhados, W. Wan and J. L. Hutter, Langmuir 21, 6642–6646 (2005).
37. Y. C. Hsieh, H. Yano, M. Nogi and S. J. Eichhorn, Cellulose 15, 507–513 (2008).
38. S. B. Warner, Fiber Science. Prentice Hall, Upper Saddle River, NJ, pp. 205–206 (1995).
39. P. Lodha, Fundamental approaches to improving performance of soy protein isolate
based ‘green’ plastics and composites. PhD Thesis, Cornell University, Ithaca, NY, p. 47
(2004).
12 Preparing Medium
Density Fiberboard
Surfaces for Powder
Coating
Effects of Atmospheric
Plasma Treatment, Infrared
Irradiation, and Climatization
Christoph Jocham, Judith Sinic,
Thomas W. Schmidt, Uwe Müller,
and Andreas Kandelbauer*
CONTENTS
12.1 Introduction................................................................................................... 210
12.2.1 Medium Density Fiberboards............................................................ 211
12.2.2 Atmospheric Plasma Treatment of MDF Panels............................... 212
12.2.3 Contact Angle Measurements............................................................ 213
12.2.4 Measurement of Electrical Resistance and Moisture Content........... 214
12.2.5 Powder Coating of MDF Panels........................................................ 215
12.2.6 Quantifying the Electrostatic Powder Coatability of MDF
Panels via Coating Thickness Measurements................................... 215
12.2.7 Determination of Tendency toward Crack Formation in
MDF Panels....................................................................................... 216
12.3.1 Effect of Plasma Activation on Powder Coatability of MDF Panels..... 216
12.3.2 Effect of Infrared Irradiation on Powder Coatability of
MDF Panels....................................................................................... 219
12.3.3 Effect of Climatization on Powder Coatability of MDF Panels........ 222
* Prof. Andreas Kandelbauer was the lead author for this chapter and can be contacted at andreas.
kandelbauer@reutlingen-university.de.
209
12.4 Conclusion..................................................................................................... 225

References............................................................................................................... 225
The powder coating of wood products as an emerging environmentally sus-

tainable coating technology holds promise in terms of novel product qual-
ity features for engineered wood like medium-density fiberboards (MDFs).
However, one major limitation currently impeding widespread application of
powder coating technology is the availability of MDF panels that are suit-
able for this process. Typically, special-grade MDF panels are required that
are more costly than standard-grade MDF panels to provide reliable coat-
ing quality, which makes powder coating economically unattractive for many
users. Methods are needed that allow extending the range of available MDF
grades. In the present study, three surface pretreatment approaches for MDFs
were studied to increase the processability of standard-grade MDF in the
powder coating process: atmospheric plasma pretreatment, infrared irradia-
tion, and moisture equilibration in a climate chamber prior to electrostatic
powder application. While atmospheric plasma treatment had no beneficial
effect on the use of standard-grade MDF panels, both infrared preheating and
preconditioning of the panels under controlled temperature–humidity condi-
tions demonstrated that the range of MDF panels suitable for powder coat-
ing can be significantly extended by appropriate selection of the pretreatment
procedure.
12.1 INTRODUCTION
The powder coating of nonmetal substrates such as synthetic polymers; solid wood
and engineered wood-based products like particleboards, medium density fiber-
boards (MDFs), etc.; and polymers or natural fiber–based composites is an emerging
technology that holds promise in terms of economic and ecological sustainability of
the coating process [1–8].
By applying powder coatings onto such carrier materials, large amounts of
the currently used solvent-containing coating systems can be avoided. Thereby,
VOC emissions and energy consumption arising from solvent evaporation can be
reduced, and the coating process can be made more environmentally friendly. The
recyclability of unused powder and the material-efficient electrostatic application
method lead to improved economical sustainability of this process. Moreover,
three-dimensionally shaped pieces of complex geometry can easily be coated using
powders. With powder coating, surfaces possessing novel product property pro-
files such as increased hardness and scratch resistance or UV light stability can be
generated.
However, this technology is not yet widely used in the industrial coating of wood-
based panels [9,10] due to a number of reasons: Firstly, unlike metals, nonmetal
Preparing MDF Surfaces for Powder Coating 211
substrates are temperature sensitive, and hence, powder coatings are required that
melt, flow, and cross-link already at low reaction temperatures in the range of 130°C
to 140°C and heat treatments need to be carefully balanced regarding temperature
and duration [11]. Secondly, the surface texture of a natural fiber–based coating sub-
strate is very inhomogeneous and rough in comparison to that of a metal surface,
which leads to problems with respect to smooth film formation and appropriate sur-
face qualities [12,13]. Thirdly, since powder coatings are preferably applied by elec-
trostatic spraying, the electrical conductivity of the substrate is of major importance
and with nonmetal substrates is usually rather low.
Therefore, the industrial powder coating of wood-based products is currently
strongly limited to the surface finishing of MDFs, and typically, special grade MDF
panels are required that are specifically designed for powder application; such panels
are of higher than average density and display higher than average mechanical and
thermal stability, display a better than average homogeneity in density and roughness
profiles, and usually contain additives that increase the electrical conductivity of the
surface. All in all, such boards are much more expensive than standard grade MDF
panels and are manufactured by only a handful of MDF producers. This restricts the
use of powder coating to a very small fraction of the total number of manufactured
MDF panels, thus inhibiting the general industrial application of powder coating
technology. Strategies that enable the use of standard grade MDFs in powder coating
are currently highly sought after [14] and need to be developed in order to provide
widespread use of this innovative coating technology.
In the present contribution, three pretreatment methods for surface activation of
MDF panels (atmospheric plasma activation, infrared irradiation, and equilibration)
are evaluated for their potential to extend electrostatic powder application to standard
grade MDF panels. In the first set of experiments, the wettability of MDF surfaces
was modified by atmospheric plasma pretreatment, and the effect was investigated
in terms of achieved coating thickness and powder coating adhesion. In the second
set of experiments, infrared irradiation was applied to MDF panels prior to powder
application, and the achieved powder deposition was compared for standard grade
MDF panels and MDF panels containing conductivity enhancing additives. In the
third set of experiments, MDF equilibration climate conditions were systematically
varied to gain a deeper insight into the process requirements that need to be ful-
filled in order to be able to apply powder coating technology to standard grade MDF
panels.

12.2.1 Medium Density Fiberboards
Ten commercially available samples from different European MDF producers
(Glunz, Egger, and Finsa) were used for the experiments. The MDF samples were
either special grade MDF panels optimized for powder coating application and con-
tained an additive for conductivity enhancement or were standard grade MDF pan-
els without a conductivity enhancer. The thickness d of the MDF samples varied
from 19 to 28 mm. Other characteristics of the panels used are given in Table 12.1.
TABLE 12.1
MDF Panels Used in the Present Study
Additive for
Conductivity
Panel Grade d (mm) h (%) ρ (kg m−3) IB (N mm−2) s (%) Enhancement
A PG 19 6.5 831 1.2 9.5 +
B PG 19 5.7 854 1.1 7.5 +
C PG 25 5.8 825 0.8 9.1 +
D SG 19 6.2 819 0.9 7.1 –
E SG 25 5.5 753 0.7 17.2 –
F SG 28 7.0 674 1.7 5.6 –
G PG 19 5.9 817 1.6 10.6 +
H PG 25 5.9 799 1.2 19.2 +
I SG 19 5.6 803 0.7 13.5 –
J SG 22 5.4 798 0.7 15.6 –
Note: PG, powder grade; SG, standard grade; d, thickness; h, moisture content (as determined
before equilibration experiments); ρ, density; IB, internal bond strength; s, swelling in thick-
ness after immersion in water.
The dimensions of the sample panels (thickness × length × width [mm]) depended
on the type of experiment and were d × 80 × 80 for the panels used in the plasma
pretreatment experiments, d × 450 × 600 for the panels used in the infrared irradia-
tion pretreatment experiments, and d × 400 × 200 for the panels used for detailed
study of the dependence of the electrical conductivity on moisture content and tem-
perature of the MDF panel. All MDF panels were equilibrated in a climate chamber
(Binder MKF720, BINDER GmbH, Tuttlingen, Germany) under varying climate
conditions (Table 12.2). For each MDF type, each different climate condition, and
each preheating variation (preheated or not preheated), three sample specimens were
produced and used for the measurements.
12.2.2 Atmospheric Plasma Treatment of MDF Panels

For each MDF type, three sample panels of nominal dimensions of d × 80 ×
80 mm equilibrated at 20°C and 40% or 20°C and 60% relative humidity were
used for the measurements. The MDF samples were plasma pretreated with an
atmospheric plasma machine (Plasma generator FG3001 with nozzle type RD1004
from Plasmatreat GmbH, Steinhagen, Germany) using pressurized air. The dis-
tance between the plasma nozzle and substrate was set at 10 mm, whereas the mov-
ing speed of the sample relative to the plasma nozzle was either 50 or 100 mm s−1
to provide different plasma treatment intensities. For testing the effect of the plasma
treatment, contact angle measurements were performed immediately afterward
using the identical testing conditions as for the untreated samples. As the time
TABLE 12.2
Variations in Temperature and Relative Humidity Used for the
Equilibration of MDF Panel Samples in the Climate Chamber
Relative Humidity (%) at a Temperature of
20°C 30°C 40°C 50°C
30 a – – –
35 – – –
40 a,b 40 – –
45 a 45 45 45
50 50 50 50
55 a 55 55 55
60 b 60 60 60
– 65 65 65
– – 70 70
a Equilibration conditions applied to panels used in preliminary infrared irradiation

experiments.
b Equilibration conditions applied to panels used in preliminary plasma treatment
experiments.
between plasma treatment and contact angle measurement is an important factor

to consider, all contact angle measurements were performed within <1 h after the
plasma treatment.
12.2.3 Contact Angle Measurements

Contact angle measurements on untreated and plasma treated boards were performed
with the contact angle measurement device OCA 35 by DataPhysics Instruments
GmbH, (Filderstadt, Germany), using Software version SCA 20. Using the sessile
drop method, a drop of the test liquid (volume of 1.5 μL) was deposited on the board,
and the contact angle was studied versus time. As test liquids, water, ethylene glycol,
and diiodomethane were used. Using a high-speed camera, the contact angle was
automatically calculated over time (recording with 1 frame per second for a maxi-
mum of 7 min) according to the Young–Laplace equation. Automatic calculation of
the contact angle/time curves was started after stabilization of the deposited drop but
before drop spreading. The calculation ended either at the end of the maximum test-
ing time (about 7 min, maximum data memory capacity of the measurement device)
or at that time when the software was not able anymore to calculate the contact
angle correctly due to a poor contrast between drop and background. The magnifica-
tion of the high-speed camera was 0.7× to 1.25× depending on the spreading behav-
ior of the test liquid. For each sample specimen, five representative measurements
were performed at different spots of the panel. Figure 12.1 exemplifies the different
Contact angle on untreated MDF
Contact angle: 90° Contact angle: 55° Contact angle: 42°
Drop at time: 0 s Drop at time: 180 s Drop at time: 420 s

Drop dosing
Contact angle on plasma treated MDF
FIGURE 12.1 Example for the video analysis of the dynamic wetting behavior of untreated
and plasma treated MDF panels as the basis for calculation of contact angle/time curves (test
liquid: water).
wettabilities of untreated and plasma treated MDF panels as indicated by the dif-
ferent contact angles measured at three particular times during the video analysis.
For technological verification of these observations, additional varnishing experi-
ments with a water-based paint were performed. The water-based paint (SeidenAcryl
seidenmatt, supplied from ADLER Lackfabrik Johann Berghofer GmbH & Co KG,
Schwaz, Austria) was applied by means of a four-sided applicator frame with a layer
thickness of 120 μm and air-dried.
12.2.4 Measurement of Electrical Resistance and Moisture Content

Electrical resistance measurements were typically carried out after conditioning of
MDF panels according to DIN IEC 60093 [15] using a teraohmmeter (TO-3, H.-P.
FISCHER ELEKTRONIK GmbH & Co., Mittenwalde, Germany). The test voltage
was 250 V. The geometries of the electrodes were designed according to the recom-
mendations of DIN IEC 60093.
For more detailed studies on the influence of panel moisture content and panel
temperature on the surface electrical resistance and coatability of the MDF panel,
the electrical resistance measurements were carried out directly in the climate cham-
ber similar to an online procedure developed and described earlier [16] with the fol-
lowing modifications: two aluminum electrodes with a nominal length of 80 mm and
a diameter of 6 mm were fixed by pressing them into drilling holes (diameter 6 mm,
distance 30 mm) of the sample to measure the electrical resistance in the core layer
of the panels. The electrical resistance was recorded using a teraohmmeter (TO-3,
H.-P. FISCHER ELEKTRONIK GmbH & Co., Mittenwalde, Germany) with a test
voltage of 250 V. The measurements were carried out in the closed conditioning
chamber during panel equilibration at the different climate conditions.
The moisture content of the MDF panels was determined after conditioning in
accordance with the procedure described by DIN EN 322 [17].
12.2.5 Powder Coating of MDF Panels

Preliminary experiments with plasma treated and untreated reference MDF panels
were performed on a laboratory scale using a manual electrostatic powder coating
system (Prima Airfluid, J. Wagner GmbH, Markdorf, Germany). MDF samples of
nominal dimensions d × 150 × 200 mm were used, which were equilibrated at 20°C
and 40% relative humidity prior to powder application. For curing the powder coat-
ing, a laboratory gas catalytic infrared oven (Vulcan Catalytic Systems, Portsmouth,
U.S.A.) was used, and test specimens were cured for 5 min at a surface temperature
of 135°C.
For investigating the effect of infrared irradiation and climatization on the suit-
ability of standard grade MDF panels versus MDF panels especially designed for
powder coating, all MDF samples were powder coated after conditioning at an
industrial powder coating line (Sauter GmbH, Überlingen, Germany). The powder
coating process at this industrial line consists of the following steps: (1) fastening
the workpieces onto a handling device, (2) cleaning the surface with pressurized air,
(3) optional preheating of the MDF in an IR preheating oven, (4) electrostatic powder
application, and (5) curing of the powder coating in an IR oven for 5 min at 135°C
surface temperature. For the industrial trials to evaluate the primary coating quality,
each sample was subjected to this process once, and a coating layer consisting of
Drylac 512/10001 (TIGER Coatings GmbH & Co. KG, Wels, Austria) was applied
and cured.
To determine the influence of infrared preheating on the electrostatic powder
application, one half of the samples were either preheated or not preheated. Infrared
irradiation was performed such that an average surface temperature of at least 60°C
was achieved on the MDF panel in the preheating step. For the not preheated sam-
ples, the IR preheating oven was switched off before powder primer application. For
all samples (preheated and not preheated), the same electrostatic powder applica-
tion conditions and cure oven settings were used for comparison and to isolate the
effect of infrared irradiation and panel type. The dimensions of the panels used in
the industrial powder coating experiments were d × 450 × 600 mm. Adhesion of the
powder coating was determined by a cross-hatch test according to DIN EN ISO 2409
[18]. For this, a multiple line pattern was cut into the coating and delamination was
judged visually.
12.2.6 Quantifying the Electrostatic Powder Coatability of

MDF Panels via Coating Thickness Measurements
As an indicator of the ability to apply a powder coating onto the surface of an MDF
panel by using electrostatic coating technology, the thickness of the achieved powder
coating on the panel surface was selected. Since all parameters in the coating line
were kept constant except for the differences in pretreatment conditions (climate
chamber, IR irradiation) and the different types of MDF panels used, by measuring
the coating thickness, the suitability of a certain MDF panel type or pretreatment
method was quantified. The coating thickness was determined by the wedge cut
method according to ÖNORM EN ISO 2808 [19]. For each different sample, eight
coating thickness measurements were carried out on the front surfaces and on the
edges.
12.2.7 Determination of Tendency toward Crack Formation
in MDF Panels
The equilibrated MDF panel samples were subjected to a heat treatment: within
2 min, the surface temperature was raised to 150°C and kept at this temperature
for another 3 min. Immediately afterward, the edge of the coated MDF panel was
inspected using a magnifying glass, and the lengths Li of each observed crack i were
measured and added for all n observed cracks at the center of the edge over a length
of 100 mm along the edge. From these measurements, a crack value vc was calculated
according to the following equation representing the total crack length observed:
Contact angle on untreated MDF
Drop at time: 0 s Drop at time: 180 s Drop at time: 420 s
Drop dosing
(12.1)
Contact angle on plasma treated MDF

12.3.1 Effect of Plasma Activation on Powder Coatability
of MDF Panels
Standard MDF panels and MDF panels containing additives for increasing surface
conductivity from different suppliers were subjected to atmospheric plasma activa-
tion at different plasma intensities (as adjusted by the sample moving speed relative
to the plasma nozzle) prior to the electrostatic powder application. The changes in
surface polarity caused by plasma pretreatment were assessed by the determination
of contact angle/time curves using contact angle measurements and video analysis
(Figure 12.1).
Figures 12.2 through 12.4 show typical examples for contact angle/time curves
obtained with standard MDF panels containing no additive for conductivity enhance-
ment for the three test liquids water (Figure 12.2), ethylene glycol (Figure 12.3), and
diiodomethane (Figure 12.4).
The highest contact angles and thus poorest wetting are observed with water as
a test liquid (Figure 12.2) on the untreated MDF panels. Plasma treatment leads
to a significant improvement of wettability against water, with the treatment being
120
Water
100
Contact angle (˚) Untreated
80 Plasma treated with 100 mm/s
Plasma treated with 50 mm/s
60
40
20
0
0 30 60 90 120 150 180 210 240 270 300 330 360 390 420 450
Time (s)
FIGURE 12.2 Contact angle/time curves of untreated MDF and atmospheric plasma treated
MDF using a relative sample/nozzle speed of 100 mm s–1 and 50 mm s–1 based on contact
angle measurements using water as a test liquid. Typical results obtained with standard grade
MDF panels containing no conductivity enhancing additive.
more effective with higher plasma intensity (i.e., with the lower sample speed of
50 mm s−1).
With ethylene glycol, plasma treatment also leads to significant improvement of
MDF wettability (Figure 12.3). While the untreated panel shows a constant con-
tact angle after about 1 min after drop deposition, the plasma treated panels show
a continuous decrease in contact angle over time. Again, the more intensive plasma
120
Ethylene glycol
100
Contact angle (˚)
80
Untreated
60
40
20
0
0 30 60 90 120 150 180 210 240 270 300 330 360 390 420 450
Time (s)
angle measurements using ethylene glycol as a test liquid. Typical results obtained with stan-
dard grade MDF panels containing no conductivity enhancing additive.
120
Diiodomethane
100
Contact angle (˚) Untreated
80
60
40
20
0
0 5 10 15 20 25 30
Time (s)
angle measurements using diiodomethane as a test liquid. Typical results obtained with stan-
dard grade MDF panels containing no conductivity enhancing additive.
treatment at a sample speed of 50 mm s−1 was more effective as indicated by the

lower contact angles.
With diiodomethane (Figure 12.4), low contact angles and thus good wetting are
observed already without plasma treatment. Again, plasma treatment leads to an
improvement in droplet wetting; however, the differences are less distinctive than
with the more polar test liquids.
As Figures 12.2 through 12.4 exemplify, the atmospheric plasma treatment of
MDF panels generally resulted in an improvement in the wetting behavior of both
standard grade MDF and special grade MDF panels for powder coating contain-
ing conductivity enhancement additives toward polar and nonpolar test liquids.
Varnishing experiments on the MDF boards using a water-based paint that were
done for comparison of the coatability of the panels showed a significant improve-
ment in terms of increased drying speed of the applied paint film. Film thicknesses
of 120 μm were easily achieved with the liquid paint.
It was, therefore, hypothesized that a pretreatment could also have a beneficial
effect on the powder coating ability of the MDF panels. However, powder coating
experiments on a laboratory scale showed that there was no significant difference
between the plasma treated and the non-plasma treated panels in terms of achievable
powder coating thicknesses. The thickness of the powder coating layer is an indirect
measure for the efficiency of the electrostatic powder application process.
For both treated and untreated MDF panels, the thicknesses of the powder coating
layers did not increase upon plasma pretreatment. For instance, after conditioning
the standard grade MDF panels at 20°C and 40% relative humidity, the powder coat-
ing layers did not exceed 25 μm or the powder did not adhere at all. Conductivity-
enhanced panels displayed average thicknesses of about 60 to 80 μm on front and
edge surfaces (compare also Figures 12.5 through 12.7, not preheated samples).
Hence, no improvement in powder deposition by atmospheric plasma pretreatment

was noted.
Furthermore, also no improvement in terms of powder coating adhesion could
be found when cross-hatch tests were performed on plasma treated and nontreated
powder coated MDF samples (data not shown).
Measurements of surface electrical resistance on the atmospheric plasma treated
uncoated MDF panels without conductivity enhancing additive showed that there
was no significant increase in surface conductivity upon plasma treatment. With
rather dry equilibrated panels (20°C and 40% relative humidity), values for the
surface electrical resistance between 1011 and 1012 Ω were obtained independent of
whether they were plasma treated or not. MDF panels that were equilibrated at 20°C
and 60% relative humidity displayed values for the surface electrical resistance vary-
ing from 6 × 1010 to 7 × 1010 Ω. Upon plasma treatment, these values increased and
were between 7 × 1010 and 10 × 1010 Ω. A similar trend in surface electrical resistance
was observed with the MDF panels that were optimized for powder coating and were
supplemented with conductivity enhancing additive. With these, the average surface
electrical resistance increased from 2.7–3.6 × 109 to 4.6–10.0 × 109 Ω. Again, no
improvement in electrical conductivity but rather a slight deterioration was the result
of plasma treatment.
Hence, although the wetting experiments showed a significant improvement of
the surface for liquid coating application, the resistance measurements revealed
that the major key parameter surface conductivity was not positively affected by
the atmospheric plasma but on the contrary even decreased the susceptibility of the
MDF surface toward electrostatic powder application.
12.3.2 Effect of Infrared Irradiation on Powder Coatability

of MDF Panels
When MDF panels are heated prior to the electrostatic powder application, a reduc-
tion in electrical resistance is observed; preliminary experiments with a laboratory
powder coating device have shown earlier that this effect leads to an increase in
coating thickness or, in other words, an improvement in electrostatic powder applica-
tion efficiency [16]. In the present study, industrial powder coating experiments were
performed to verify the earlier finding on an industrial scale and to systematically
study if this effect is pronounced enough to allow powder application also onto the
surfaces of standard grade MDF panels.
In Figures 12.5 through 12.7, the influence of infrared irradiation on the achieved
coating thicknesses of a wide variety of MDF panels is illustrated.
Figures 12.5 and 12.6 depict the coating thicknesses obtained at the front edges
of two sets of MDF panels that were equilibrated at low (Figure 12.5) or high
(Figure 12.6) relative humidity at 20°C prior to the powder coating application. The
coating thickness on the edges is a crucial factor for judging the coating quality since
surface defects due to coating problems are especially manifest in this area.
The horizontal lines in Figures 12.5 and 12.6 represent the average coating thick-
nesses obtained from all tested boards. Average coating thickness increases when
the panels are subjected to higher initial moisture content during equilibration. The
140
130
120
Coating thickness–edges (µm)
110
100
90
80
70
60
50
40
30
20
10
0
A G B H C D F I J E A G B H C D F I J E
Not preheated Preheated
FIGURE 12.5 (See color insert.) Coatability of the edges of various types of MDF panels
in terms of film thickness. All panels were equilibrated at 20°C and 30% relative humidity.
One half of the panels (right side) were preheated to 80°C surface temperature using infrared
irradiation prior to the electrostatic powder coat application (one layer, primer). Columns A,
G, B, H, C: MDF panels containing conductivity enhancing additive, columns D, F, I, J, E:
standard grade MDF panels.
120
110
100
90
80
70
60
50
40
30
20
10
0
FIGURE 12.6 (See color insert.) Coatability of the edges of various types of MDF panels
in terms of film thickness. All panels were equilibrated at 20°C and 55% relative humidity.
One half of the panels (right side) were preheated to 80°C surface temperature using infrared
irradiation prior to the electrostatic powder coat application (one layer, primer). Columns A,
G, B, H, C: MDF panels containing conductivity enhancing additive, columns D, F, I, J, E:
standard grade MDF panels.
130
120
110
100
90
80
70
60
50
40
30
20
10
0
FIGURE 12.7 (See color insert.) Coatability of the flat front surface of various types of
MDF panels in terms of film thickness. All panels were equilibrated at 20°C and 30% rela-
tive humidity. One half of the panels (right side) were preheated to 80°C surface temperature
using infrared irradiation prior to the electrostatic powder coat application (one layer, primer).
Columns A, G, B, H, C: MDF panels containing conductivity enhancing additive, columns D,
F, I, J, E: standard grade MDF panels.
columns on the left of Figures 12.5 and 12.6 summarize the measured coating thick-
nesses obtained without infrared irradiation prior to powder application, while the
right-hand side columns show the results obtained when including such a preheating
step. With the MDF panels equilibrated at a low moisture level (20°C and 30% rela-
tive humidity), the positive effect of infrared irradiation is much more pronounced
than with the panels equilibrated at a high moisture level (20°C and 55% relative
humidity), although the latter also show significant improvement in powder coat-
ability. Since the samples equilibrated at higher moisture levels display a higher
electrical conductivity before entering the coating line, powder adhesion during elec-
trostatic coating application is much better than with the drier panels.
Especially notable are the pronounced differences in pretreatment effects when
standard grade MDF panels are compared with conductivity enhanced ones. It is evi-
dent from Figure 12.5 that the relative increase in coating thickness is much higher
with standard grade MDF when infrared preheating is included in the process (com-
pare columns D, F, I, J, and E in Figure 12.5) than it is with conductivity enhanced
panels. Panel thickness does not seem to play any significant role.
Figure 12.7 illustrates the effect of infrared irradiation on the thickness of the
powder coating layer on the flat front surfaces. Although here, too, a slight increase
in thickness upon preheating is observed, the effect is much less pronounced. This
is supported by the experience that flat surfaces generally are less prone to surface
defects in MDF powder coating than are edges and sharp profiles.
Summing up, infrared irradiation leads generally to a significant improvement

in the powder coating ability of MDF panels. Hence, this method is highly suit-
able to extend the range of MDF panels that can be used for powder coating. For
further understanding of the effects of heating in terms of moisture equilibration
and changes in surface electrical conductivity, a more systematic investigation of the
preconditioning of MDF panels under defined temperature and humidity conditions
was performed.
12.3.3 Effect of Climatization on Powder Coatability of MDF Panels

As shown in an earlier investigation [16], the panel moisture content is of great
importance for electrostatic powder application since the moisture distribution in
the panel governs the electrical resistance of the panel. In general, the conductivity
increases with increasing moisture content. If the MDF panel is too dry (high electri-
cal resistance), the required transport of electric charges during electrostatic powder
application is not at all or only insufficiently provided, leading to an insufficient film
formation and low coating thickness. The question now arises, if and how a suitable
process modification can be defined that allows processing of standard grade MDF
panels that per se display insufficient electrical conductivity for coat application. For
this, five different standard grade MDF panel types and five conductivity enhanced
MDF panel types were subjected to different climate conditions in an equilibrium
chamber, and their final moisture content and electrical resistivity were determined
directly in the chamber.
In industrial powder coating validation experiments, thicknesses of the cured
powder coatings were measured at both the flat fronts and at the edges of these
differently conditioned panels. Figure 12.8 shows a correlation between moisture
content and surface electrical resistance of panels equilibrated at 20°C, at a relative
humidity varying from 30% to 60%. As expected, resistance decreases with increas-
ing panel moisture content. Interestingly, panels without a conductivity enhancing
additive show similar electrical behavior when equilibrated at higher relative humid-
ity. Such standard grade MDF panels were found to show similar cured coating qual-
ity after being subjected to powder coating.
The influence of the electrical conductivity of MDF panels on coating thickness
at the edges is shown in Figure 12.9. It is evident that with increasing electrical resis-
tance, the coating thickness decreases steadily. At surface electrical resistance val-
ues >1011 Ω, the coatings formed were too thin, and the MDF substrate occasionally
was visible through the coating leading to an unsatisfactory coating quality. Again,
the edges were much more problematic than the plain front areas. Hence, a surface
electrical resistance value of about 1010 Ω can be regarded as a critical threshold
required for defect-free coating. While the plain surfaces were acceptable in many
cases even at such critical values for surface electrical resistance, the edges often
remained uncoated.
Figure 12.10 exemplifies the effects of the different pre-equilibration conditions
prior to the coating steps on the electrical resistance of the core layer of a typical
standard grade MDF panel. By applying appropriate combinations of temperature
1.00E+13
Surface electrical resistance (Ω)

1.00E+12
Standard grade MDF panels
1.00E+11
1.00E+10
Powder grade MDF panels

1.00E+09
5 5.5 6 6.5 7 7.5 8
Moisture content (%)
FIGURE 12.8 Surface electrical resistance in dependence of panel moisture content and
panel type. The shaded ellipse represents all tested standard grade MDF panels D, E, F, I, and J.
The symbols outside the shaded ellipse represent powder grade MDF panels (triangle: B;
circle A; star: C; square: G; and plus: H).
100
90
80
70
60
50
40
30
20
10
0
1.00E+09 1.00E+10 1.00E+11 1.00E+12 1.00E+13
Surface electrical resistance (Ω)
FIGURE 12.9 Thickness of cured powder coating at the edges in dependence of the surface
electrical resistance.
and relative humidity in a climatization chamber, the moisture content of the panel
can be adjusted within a wide range between 5% and 8%. This, in turn, leads to dra-
matic changes in the core electrical resistance, and upon decreasing the resistance
below <1011 Ω, the standard grade MDF panel is made suitable for electrostatic pow-
der coating application. Moreover, as seen from Figure 12.10, also a panel with 6%
moisture content displays a core electrical resistance of about 1012 Ω at 20°C, which
must be considered unsuitable for electrostatic powder application. Upon heat treat-
ment, the panel temperature may be raised to 60°C, which concomitantly reduces the
core resistance to about 1011 Ω and renders the panel suitable for electrostatic powder
1.00E+13
20°C
1.00E+12
Core electrical resistance (Ω)

30°C
1.00E+11 40°C
50°C
1.00E+10
60°C
1.00E+09
1.00E+08
1.00E+07
1.00E+06
5.0 5.5 6.0 6.5 7.0 7.5 8.0 8.5 9.0
Moisture content (%)
FIGURE 12.10 Core electrical resistance of a typical standard grade MDF panel in depen-
dence of panel temperature and moisture content prior to electrostatic powder application.
coating. This temperature-dependent effect explains the positive effect found with
the preliminary infrared irradiation experiments.
The target of the various temperature treatments (climate chamber, infrared irra-
diation) prior to powder application is hence the adjustment of surface temperatures
to ≧60°C in order to provide sufficient conductivity for powder adhesion during
electrostatic deposition. However, as shown in Figure 12.11, care must be taken that
the moisture content by such treatment does not reach too high levels. Although
favoring the surface resistance, the tendency of the coating toward formation of
cracks at the edges during the melting and curing processes in the oven is increased
at higher moisture levels. This may be caused by increased dimensional instabil-
ity of the engineered wood-based carrier material and degassing phenomenon.
Crack formation tendency is strongly dependent on the specific MDF panel and its
200
180
Panel B
160
Panel C
Crack probability (%)
140
120
100
80
60
40
20
0
5 5.5 6 6.5 7 7.5 8
Panel moisture content (%)
FIGURE 12.11 Dependence of crack formation on the moisture content of MDF panels.
mechanical performance, and a quantitative understanding of the interplay between

physical parameters of MDF panels and crack formation in the powder coating is
required. Hence, the pretreatment process conditions need to be carefully balanced
when defining a suitable MDF surface activation protocol.
12.4 CONCLUSION
The suitability of three surface pretreatment methods: atmospheric plasma treatment,
infrared irradiation, and climatization of MDF panels for improving the powder coat-
ability of standard grade MDFs, was studied. Although plasma treatment was found
to improve the wetting behavior of MDF surfaces for liquid coating formulations,
the atmospheric plasma had no positive effect on surface conductivity, which is the
major parameter governing the efficiency of the electrostatic powder coating process.
Infrared irradiation showed that by heating, a much higher susceptibility of the MDF
panel surfaces toward adhesion of the powder coating could be achieved.
It was shown that standard grade MDF panels, in principle, could be powder
coated in good quality not only at the flat front surfaces but especially also around
the edges, which are the areas usually prone to defect formation.
Detailed studies on the effect of climatization showed that the surface electrical
resistance of standard grade MDF panels can be changed much in favor of powder
coating adhesion; it is suggested that the main effect of infrared irradiation lies in
an increase in surface electrical conductivity and hence is related to the observed
improvement obtained by climatization. It was possible to obtain good powder adhe-
sion and defect-free coating formation with standard grade MDF. It is concluded
that by appropriate adjustment of pretreatment steps prior to the electrostatic powder
application step in the coating line, the surface characteristics of standard grade
MDF panels can be positively affected and adjusted so that they become suitable for
powder application. In the long run, surface conductivity control by climatization
and heat pretreatment will enable utilization of lower-price, standard grade MDF
panels for this developing and promising coating technology.
ACKNOWLEDGMENTS
The authors gratefully acknowledge funding by the European Union and the Kärntner
Wirtschaftsförderungsfond (KWF) within the project PPRSI, ID-4782, “Einsatz von
Plasma für eine bessere Dekorbehandlung bei Kunststoffen: der Schutzüberzug, der
Druck und die Klebeverbindung.” The authors would also like to thank the par
ticipating company partners for the good cooperation within a KWF-funded powder
coating project: Fritz Egger GmbH & Co, J. Wagner AG, Ramseier Woodcoat AG,
Sauter GmbH, TIGER Coatings GmbH and CoKG, and Vitra Logistik GmbH.
REFERENCES
1. M. C. Barbu, and T. W. Schmidt, Holztechnol. 50 (1), 32 (2009).
2. J. Prieto, and J. Kiene, Holzbeschichtung. Vincentz Network GmbH und Co, Hannover,
Germany (2007).
3. C. T. Lin, JCT Coatings Technol. 5 (3), 48 (2008).

4. E. Marelli, P. Dinning, S. Hobbs, O. Aixala, H. Bolm, C. Santamarta, and V. Rekowski,
Finish. Today 83 (12), 32 (2007).
5. R. Redshaw, Finish. Today 83 (2), 40 (2007).
6. S. Zaccaria, Industr. Paint Powder 82 (4), 8 (2006).
7. O. André, and K. Buysens, Surf. Coat. Intern. A Coatings J. 87 (8), 347 (2004).
8. M. Johansson, E. Malmström, A. Hult, and H. Claesson, Europ. Coatings J. 7, 26 (2002).
9. H. Bauch, R. Emmler, D. Krug, and I. Fuchs, Proceedings of the Interior Surface
Conference 2007, p. 159, Amsterdam, the Netherlands (2007).
10. D. Krug, and B. Lilie, Proceedings of the 6th International Symposium Holz Innovativ
2009, p. 34, Rosenheim, Germany.
11. G. Wuzella, A. Kandelbauer, A. R. Mahendran, and A. Teischinger, Prog. Org. Coatings
70, 186 (2011).
12. G. Wuzella, A. R. Mahendran, T. Bätge, C. Jocham, and A. Kandelbauer, Holztechnol.
51 (5), 39 (2010).
13. G. Wuzella, A. R. Mahendran, T. Bätge, C. Jocham, and A. Kandelbauer, Holztechnol.
51 (4), 34 (2010).
14. L. K. Friedrich, Holz- und Kunststoffverarbeitung 7 (1), 60 (2007).
15. DIN IEC 60093 (VDE 0303 Teil 30), Methods of test volume resistivity and surface
resistivity of solid electrical insulating materials (1993).
16. C. Jocham, T. W. Schmidt, G. Wuzella, A. Teischinger, and A. Kandelbauer, J. Adhesion
Sci. Technol. 25, 1937 (2011).
17. DIN EN 322, Wood-based panels—Determination of moisture content (1993).
18. DIN EN ISO 2409, Paints and varnishes—Cross-cut test (2010).
19. ÖNORM EN ISO 2808, Paints and varnishes—Determination of film thickness (2007).
13 Adhesion Aspects
in Dentistry
Bonding to Zirconia
Jukka P. Matinlinna* and Moustafa N. Aboushelib
CONTENTS
13.1 Introduction................................................................................................... 228
13.2 All-Ceramic Materials: Alumina and Zirconia............................................. 231
13.2.1 Alumina............................................................................................. 231
13.2.2 Zirconia.............................................................................................. 231
13.3 Resin–Zirconia Bonding................................................................................ 233
13.4 Conclusion..................................................................................................... 235
References............................................................................................................... 236
The introduction of zirconia (zirconium dioxide) frameworks for prosthetic

dentistry in the 1990s virtually opened the design and wide applications of all-
ceramic restorations with higher reliability. Today, these are used as prosthetic
implants for dental applications due to their outstanding biocompatibility and
chemical stability. However, there is a lack of long-term clinical reports regard-
ing this application. Nevertheless, zirconia frameworks have shown much suc-
cess in the field of all-ceramic restoration with high accuracy and success rate.
This review article covers basic properties of zirconia and focuses on new
trends used to enhance bonding to zirconia frameworks.
Establishing a reliable bond to zirconia-based materials has proven to be
difficult. The combination of airborne-particle abrasion and 10-methacryloyl
oxydecyldihydrogenphosphate monomer is the “gold standard” method for
bonding resin composites to zirconia. Selective infiltration etching (SIE) is a
novel surface treatment developed to transform the dense and nonbonding hard
surface of zirconia into a highly retentive and reactive surface. A new genera-
tion of specifically formulated primers may be required to optimize both the
wetting of the SIE-treated zirconia surfaces and chemical bonding to luting
* Dr. Jukka P. Matinlinna was the lead author for this chapter and can be contacted at jpmat@hku.hk.
227
resins. The novel zirconia primers based on bifunctional trialkoxysilane cou-

pling agents represent a promising step toward reliable resin–zirconia bonding.
13.1 INTRODUCTION
All-ceramic restorations have attracted rapidly increasing interest among patients and
clinicians, as demands and expectations for more natural looking material alternatives
have increased. In prosthetic dentistry, in many cases, restorations such as crowns
and bridges are based on a metallic framework that gives the restoration its required
strength, and an outer layer of veneer ceramic is used to provide the restoration with
aesthetically pleasing shade and color. The metallic alloys used must meet the demands
of biocompatibility and chemical stability [1]. On the other hand, metallic frameworks
offer high strength, high fracture toughness, and cheap and easy processing techniques
[2]. Mineral excavations containing zirconium have been discovered and utilized for
centuries, but it was not until 1789 that zirconia was incidentally identified by M. H.
Klaproth (1743–1817) from stones originating from Sri Lanka. Actually, zirconia was
first used as a pigment material in the manufacturing of ceramics [3].
Biomedical grade zirconia was invented and prepared for the first time over 25
years ago, but zirconia (ZrO2) as a new biomaterial was first introduced in the 1990s
as a femoral head material in orthopedics as a hip prosthesis [3]. Subsequently, its
properties as a dental biomaterial opened the design and wide applications of all-
ceramic restorations with higher success and reliability. Today, long span bridges
and complex all-ceramic restorations are feasible due to their unique biomechanical
properties. With the current computer-aided design/computer-aided manufacturing
(CAD/CAM) technology, the design, accuracy, and production of zirconia frame-
works are rapid, reliable, and economically favorable [4]. Surprisingly, only a few
cases of clinically observed fractures are reported [5]; however, there are several
reports about minor chip-off fracture of the o verlying veneer ceramic [6].
For biomedical applications, low-temperature degradation is a negative feature
that is related to the transformation ability of the tetragonal phase of zirconia under
the influence of water and temperature. It is also observed that aging proceeds even
more rapidly in a warm environment, that is, when the critical temperature is about
200°C–300°C [7,8]. This phenomenon was responsible for the fracture of zirconia
hip prostheses, which were sterilized by steam under pressure (autoclaving). Zirconia
also fulfills basic aesthetic demands, such as form, color, shade, glaze, and outlook.
Flexural strength of zirconia ranges between 800 and 1500 MPa [9], and its fracture
toughness, that is, the resistance to crack propagation from a flaw or a sharp notch,
ranges from 4 to 11.5 MN m3/2 [10].
Pure zirconia is a polymorphic inorganic material that is composed of micro-
scopic grains (Figure 13.1). With prolonged heating at high temperatures, the grains
fuse together and grow in size, which negatively influences the mechanical proper-
ties of the material (Figure 13.2). These grains appear in three main forms: cubic,
monoclinic, and tetragonal. Below about 1170°C, the monoclinic phase is stable.
However, the tetragonal phase could be preserved at room temperature by addi-
tion of a small concentration of what is known as the dopant phase. Controlling the
concentration of the dopant phase could result in full or partial stabilization of the
Adhesion Aspects in Dentistry 229
FIGURE 13.1 Scanning electron microscopic image of zirconia demonstrating typical

spherical rounded zirconia grains.
tetragonal phase depending on the intended use of the material. For dental applica-
tions, partial stabilization is an advantage because it increases fracture resistance of
the material and helps to arrest a propagating crack tip by a phenomenon known as
transformation toughening. The cubic form appears above 2370°C [3].
Its remarkable chemical stability has made it the optimal material, in particular,
in the presence of the hostile oral environment. In industrial applications, zirconia is
used under aggressive acidic and alkaline conditions where any ordinary metallic or
ceramic material would easily be deteriorated.
Zirconia is commonly machined in a weak stage called the green body, which
allows fast and easy cutting of the material. The green body needs to be sintered
Acc.V Spot Magn Det WD Exp 500 nm

30.0 kV 2.0 50000x SE 35.1 0 Cercon Base
FIGURE 13.2 Scanning electron microscopic image of zirconia demonstrating fusion of

grains with prolonged heating at higher temperatures resulting in larger grains.
Acc.V Spot Magn Det WD 1 mm

30.0 kV 5.0 32x SE 17.5 CAD/CAM trace lines
FIGURE 13.3 Scanning electron microscopic image of zirconia demonstrating characteris-

tic milling and grinding lines produced by the CAD/CAM fabrication technique.
in a special sintering furnace at about 1300°C–1600°C in order to achieve its final

mechanical properties. The sintering procedure is accompanied by approximately
18%–25% shrinkage, which has to be compensated for in the design phase of the
restoration [11]. This milling procedure (CAD/CAM) leaves characteristic markings
on the surface of the produced restoration (Figure 13.3). Because pure zirconia is
bright white in color and also more opaque than other ceramics used in dentistry,
new coloring pigments are available to modify its whitish color. These coloring
agents are based on some metallic oxides (such as the oxides of Fe, Co, Cu, Sn, Zn,
etc.), which are incorporated in the green body before the sintering procedure. As
Acc.V Spot Magn Det WD 2 µm

30.0 kV 1.0 15000x SE 35.6 Cercon Base Colored
FIGURE 13.4 Scanning electron microscopic image of zirconia demonstrating concentra-

tion of coloring pigments at grain boundary regions (longitudinal structures).
the milling process of the framework or coping is finished, some zirconia manu-
facturers recommend dyeing of zirconia prior to the sintering procedure in order
to improve the color match of the restoration. The dyeing agents contain selected
metallic oxides, and they were found to concentrate at the grain boundaries and thus
compete with some locations occupied by stabilizer oxides (Figure 13.4). Given this,
the process might change the surface quality of zirconia and thereby influence its
structural integrity and bonding with the veneering porcelain [12]. Extended dyeing
time has been reported to be detrimental to the hardness and strength of the copings
and frameworks. The conclusion was that coloring treatment should be performed
with extreme care [13].
13.2 ALL-CERAMIC MATERIALS: ALUMINA AND ZIRCONIA

Based on their chemical compositions, ceramic materials used in dentistry and den-
tal technology are usually divided into three groups [1]:
1.
Glass-ceramics (silica based), which enjoy high aesthetic qualities but are
generally weak and thus are used basically as veneering ceramics or for
fabrication of laminate veneers, inlays, and onlays.
2.
Glass infiltrated porous alumina, which is called a hybrid ceramic where
a specific infiltration glass material is used to fill the porosity associated
with the alumina core, in order to increase the mechanical properties of the
material.
3.
Polycrystalline ceramics such as alumina and zirconia, which are used
as frameworks for the construction of all-ceramic restorations and, when
combined with porcelain veneer, are known as core veneered all-ceramic
restorations.
13.2.1 Alumina
Aluminum trioxide, Al2O3 (alumina), sometimes also named aluminum oxide, can
be considered the predecessor of zirconium dioxide. Both are true inorganic crystal-
line ceramic materials. Alumina was first introduced as a reinforcing inclusion for
dental porcelain in the mid-1960s. However, the inherently low tensile strength of
porcelain does not allow it to be used in areas subjected to high stresses [14]. As
orthodontic brackets, however, alumina has found application in dentistry. Today
alumina is still used as a framework for the construction of crowns and small all-
ceramic fixed restorations.
13.2.2 Zirconia
Zirconia has a hard and dense surface that is ideal for resisting wear and contact
damage, which made it the material of choice where high functional demands are
required. Biomechanical properties, such as elastic modulus and hardness, are lower
for zirconia when compared to sintered dense alumina [3]. One of the main risks
associated with zirconia has been its potential radioactivity and was occasionally
mentioned to be a concern, but it has been shown to be negligible [1]. When used
as a femoral head material, it was observed that in 2001, about 400 femoral heads
failed in a very short period. This took place in two unsuccessful production batches
of Prozyr (St. Gobain Desmarquest, France) zirconia, and it was explained by its
sensitivity to low-temperature degradation [15].
Zirconia has multiple applications today: root canal posts used to reinforce
nonvital teeth since the mid-1990s [16], dental implant abutments [17], orthodon-
tic brackets [18], dental implants that offer adequate osseointegration [19], and as
frameworks for fixed all-ceramic dental prostheses where it is expected to resist
areas of stress concentration [20,21]. Zirconia structures are usually veneered with
porcelain to give the finish and tooth-like appearance. The bond strength between
zirconia [22] and the veneering porcelain was reported to be inferior compared to
other all-ceramic systems, and there are studies reporting delamination of the veneer
ceramic from intact zirconia frameworks [1,3,15,23]. Recently, a pressable veneering
porcelain was introduced for zirconia frameworks with claims of better performance
regarding easier fabrication technique and better marginal adaptation. The pressing
technology produces porcelain with minimum structural defects and flaws as air
bubbles and microcracks, which are known to weaken the zirconia–veneer interface,
especially when located at the interface between the two materials [24]. Pressing
veneer material was observed to enhance impact strength of the restoration with
reduced chances of delamination failure [12,25]. However, it should be mentioned
that pressing produces a monochromatic veneer in contrast to manually layered por-
celain where the restoration is provided with incorporation of different colors and
shades of the veneering material. Nowadays, other types of zirconia materials are
available in the market, and the biomechanical and other properties of two common
zirconia types, TZP (tetragonal zirconia, polycrystalline) and Mg–PSZ (Mg–
partially stabilized zirconia), are listed [22,26] in Table 13.1.
TABLE 13.1
Comparison of Some Properties of Two Different
Zirconia Types vis-à-vis Alumina Ceramics
Property Al2O3 TZP Mg–PSZ
Purity (%) >99.7 97 96.5
Y2O3 <0.3% 3 mol 3.4 wt%
Fracture toughness (MN m3/2) 5 7 8
Density (g/cm3) 3.98 6.05 5.72
Average grain size (μm) 3.6 0.2–0.4 0.42
Flexural strength (MPa) 595 1000 800
Compressive strength (MPa) 4250 2000 1850
Young’s modulus (GPa) 400 150 208
Hardness (HV) 2400 1200 1120
Sources: D. Liu et al., J. Adhesion Sci. Technol., 26, 1249, 2012. R. M.

Parker, Dentistry Today, 114, 26(3), 114, 2007.
13.3 RESIN–ZIRCONIA BONDING

Establishing a strong bond between all-ceramic restorations and the tooth structure
is a prerequisite for a successful restoration, as it will increase fracture resistance
of the restoration, prevent microleakage, and enhance the retention and stability of
the restoration [27]. However, establishing a strong and durable bond to zirconia has
been quite a challenge in the dental field due to the chemical inertness of the mate-
rial, and without a proper surface treatment or application of suitable priming agent,
the shear bond strength between zirconia and luting cements was reported to be only
about 1.5 MPa [28].
Numerous studies have focused on investigating the effect of different surface
treatments on the bond strength. Methods such as acid etching by hydrofluoric
acid, sand blasting using alumina particle powder (with various particle sizes, 50
to 110 µm), sand blasting with silica-coated alumina powder, and, more recently,
application of laser irradiation have been reported to significantly improve the bond
strength in vitro [29–33]. It may be summarized that these above-mentioned methods
produce mainly physical changes in the surface topography of zirconia and, after
application of the surface treatment of choice, the surface roughness. As a result, the
surface free energy is increased, which results in better wetting of the bonding agent.
In addition to these listed surface treatments, chemical treatment on the zirconia
surface with concentrated strong mineral acids and bases might be an alternative, but
zirconia remains to be more resistant to strong acids as hydrofluoric acid compared
to other glass ceramics [34]. Very recently, silica coating combined with etching
acids has been proposed as one approach for resin–zirconia adhesion promotion [35].
Approaches such as micromechanical retention, chemical bonding, or a combi-
nation of both have been described as prerequisites for achieving a reliable bond
between resin cements and zirconia restorations. Silane coupling agents (silanes) [36]
are well-known hybrid inorganic–organic compounds to form covalent chemical
bonds between dissimilar materials, that is, glassy ceramics (oxides) and organo-
functional monomers [37–40]. The silane monomer, in particular, with the meth-
acrylate functionality, 3-methacryloxypropyltrimethoxysilane, is widely used in
dentistry [36]. However, silanes cannot react directly with zirconia. Some other
reactive functional silanes than methacrylate silane have been recently evaluated
for different bonding resins and luting cements to silica coated zirconia [30,41]. In
combination with silica coating, silanes may enhance significantly resin–zirconia
bonding [42].
Application of a glassy layer on the fitting surface of zirconia restorations has
some drawbacks as it may interfere with the seating accuracy of these restorations,
and it is difficult to apply as an even layer, in particular, for small-sized restorations.
Today, the combination of acidic phosphate monomers such as 10-methacryloyl
oxydecyldihydrogenphosphate (MDP) monomer with airborne particle abrasion is
considered “the gold standard” for bonding to zirconia restorations. Bond strength
stability up to 150 days was reported for this bonding technique [43], but there is lack
of in-depth understanding of the mechanism of chemical reaction between the phos-
phate monomer and zirconia [44]. It should be mentioned that application of phos-
phate monomer agents on smooth as-sintered zirconia surface without application of
particle abrasion would result in quick deterioration of the established bond, indicat-
ing that micromechanical retention is an indispensable part of the bonding procedure
[13].
There are different forms of phosphate monomers supplied by different manu-
facturers available today that differ in structure and the number of active sites in
the molecular structure of the monomer. Three different phosphate monomer agents
were previously compared, and differences in bond strength values were observed
indicating different reactivity of these monomers [45]. Similar observation was
reported for the reactivity of different silane primers used to enhance wetting on
nanoporous zirconia surface as differences in the observed bond strength indicated
different reactivity of these silane primers. Different activation chemistries were
observed for these primers using Fourier transform infrared spectroscopy (FT-IR)
techniques [46].
Different laboratory approaches are used to evaluate bond strength values such
as shear, microshear, tensile, or microtensile bond strength tests. In each test setup,
the loading force delivered by the universal testing machine is transmitted through
the attachment unit to the specimen and then to the bonded interface, which has to
resist these forces until failure occurs. Microtensile bond strength test became one of
the most widely used methods to evaluate bond strength because it allows obtaining
numerous microbars from small-sized specimens and subjects the bonded interface
to a direct tensile force. However, these laboratory tests ignore the fact that stresses
in the oral cavity are not unidirectional, and thus, a new test, rotating fatigue, was
introduced, to subject the bonded interface to alternating compressive and tensile
stresses without the need to load the specimens to failure [47].
Establishing a strong and durable bond to zirconia is further complicated by the
problem of polymerization stresses of the adhesive resin [48] and the influence of
water absorption [49], which could significantly deteriorate the established bond.
For a stiff material like zirconia, any sort of deformation and compliance must be
compensated by the weaker adhesive resin. Additionally, the established bond must
resist the changing chemical environment of the oral cavity and hydrolytic enzymes
present in saliva.
Zirconia-based materials are dynamic materials on a microscopic level. Besides
the well-known tetragonal to monoclinic transformation of these materials [1,3], the
surface and bulk grains could grow in size with sufficient time and appropriate tem-
perature. Selective infiltration etching (SIE) technique utilizes this dynamic feature
to crease surface nanoporosity where the adhesive resin can infiltrate, interlock, and
polymerize. Through the application of a specific so-called infiltration glass powder
and then heating the specimen to the required temperature, the molten infiltration
glass powder forces the surface grains to split and slide over each other to create the
required surface porosity. After etching, rinsing, and washing off the infiltrated glass
layer, the new nanorough surface (Figure 13.5) is ready to establish a strong and
durable bond with the adhesive resin of choice [13]. Long-term bond stability was
reported when the SIE technique was combined with phosphate monomer contain-
ing bonding resin composite [46].
A major concern during performing the required surface treatment and bonding
procedure is the possibility of surface contamination during handling. Slight surface
Acc.V Spot Magn Det WD 20 μm

15.0 kV 4.0 1000x SE 11.8 MHT & SIE colored framework
FIGURE 13.5 Scanning electron microscopic image of zirconia demonstrating the rough
nanoporous surface produced by the SIE procedure used to enhance bonding with adhesive
resins.
contamination even in the form of adsorbed atmospheric gases could result in a

significant drop in bond strength. Given this, under clinical conditions, salivary con-
tamination presents the major risk of surface contamination [50]. Fortunately, with
proper surface cleaning and handling, a contaminated surface could be thoroughly
cleansed and prepared for a fresh bonding resin application [51]. Recently, a modi-
fied relatively rough surface was produced on zirconia by coating it, after having
been milled, with a slurry containing zirconia powder and a specific pore former
agent. The slurry was then sintered. This method produced significantly higher bond
strength than airborne-particle abrasion [52]. Some bifunctional silane monomers
have produced significantly higher resin–zirconia bond strength even after thermal
aging of the specimens [53], probably due to favorable postpolymerization of the
resin cement. Recently a comprehensive review [54] on bonding to alumina and zir-
conia was published that provides some additional information to this current review.
13.4 CONCLUSION
It has become more obvious that the bonding interface is a very complex region where
different materials meet. At this interface, chemical and micromechanical reten-
tion, surface roughness and porosity, chemical composition, polymerization stresses,
water absorption, and cyclic loading interact all together. The end result is the long-
term performance of the bond under the influence of fatigue and biodegradation.
ACKNOWLEDGMENTS
The authors would like to thank The University of Hong Kong (Hong Kong SAR,
PRC) and The University of Alexandria (Alexandria, Egypt) for financially support-
ing this study.
REFERENCES
1. J. P. Matinlinna and K. L. Mittal (Eds.), Adhesion Aspects in Dentistry. VSP/Brill,
Leiden (2010).
2. H. T. Shillingburg, In: Fundamentals of Fixed Prosthodontics, H. T. Schillingburg (Ed.),
pp. 455–483. Quintessence Publishing, Carol Stream, IL, USA (2002).
3. C. Piconi and G. Maccauro, Biomaterials 20, 1 (1999).
4. M. N. Aboushelib, N. de Jager, C. J. Kleverlaan and A. J. Feilzer, Dental Mater. 23, 952
(2007).
5. I. Sailer, B. E. Pjetursson, M. Zwahlen and C. H. F. Hammerle, Clin. Oral Implants Res.
18, 86 (2007).
6. I. Sailer, A. Feher and F. Filse, Quintessence Intl. 37, 685 (2006).
7. K. Kobayashi, H. Kuwajima and T. Masaki, Solid State Ionics 3–4, 489 (1981).
8. J. J. Swab, J. Mater. Sci. 26, 6706 (1991).
9. J. Tinschert, D. Zwez, R. Marx and K. J. Anusavice, J. Dentistry 28, 529 (2000).
10. H. Yilmaz, C. Aydin and B. E. Gul, J. Prosthet. Dentistry 98, 120 (2007).
11. J. Chevalier, S. Deville, E. Munch, R. Jullian and F. Lair, Biomaterials 25, 5539 (2004).
12. J. Hjerppe, T. Närhi, K. Froberg, P. K. Vallittu and L. V. Lassila, Acta Odontol. Scand.
66, 262 (2008).
13. M. N. Aboushelib, C. J. Kleverlaan and A. J. Feilzer, J. Prosth. Dentistry 98, 379 (2007).
14. J. W. McLean, J. Prosthet. Dentistry 85, 61 (2001).
15. J. Chevalier, Biomaterials 27, 535 (2006).
16. K. H. Meyenberg, H. Luthy and P. Scharer, J. Esthet. Dentistry 7, 73 (1995).
17. L. Canullo, Int. J. Prosthodont. 20, 489 (2007).
18. O. Keith, R. P. Kusy and J. Q. Whitley, Am. J. Orthod. Dentofacial. Orthop. 106, 605
(1994).
19. S. Tete, F. Mastrangelo, A. Bianchi, V. Zizzari and A. Scarano, Int. J. Oral Maxillofac.
Implants 24, 52 (2009).
20. M. Rosentritt, C. Kolbeck, S. Ries, M. Gross, M. Behr and G. Handel, Quintessence Intl.
39, 313 (2008).
21. P. Vult von Steyern, P. Carlson and K. Nilner, J. Oral. Rehab. 32, 180 (2005).
22. D. Liu, J. P. Matinlinna and E. H. N. Pow, J. Adhesion Sci. Technol. 26, 1249 (2012).
23. B. K. Choi, J. S. Han, J. H. Yang, J. B. Lee and S. H. Kim, J. Adv. Prosthodont. 1, 129
(2009).
24. M. V. López-Mollá, M. A. Martínez-González, J. F. Mañes-Ferrer, V. Amigó-Borrás and
K. Bouazza-Juanes, Med. Oral Patol. Oral Cir. Bucal. 15, 919 (2010).
25. M. N. Aboushelib, C. J. Kleverlaan and A. J. Feilzer, J. Prosthodont. 17, 401 (2008).
26. L. L. Hench and J. Wilson, An Introduction to Bioceramics. World Scientific Publishing,
Singapore (1999).
27. R. M. Parker, Dentistry Today 26(3), 114 (2007).
28. T. Derand, M. Molin and K. Kvam, Dental Mater. 21, 1158 (2005).
29. A. N. Cavalcanti, R. M. Foxton, T. F. Watson, M. T. Oliveira, M. Giannini and G. M.
Marchi, Operat. Dentistry 34, 280 (2009).
30. J. P. Matinlinna, T. Heikkinen, M. Özcan, L. V. J. Lassila and P. K. Vallittu, Dental
Mater. 22, 824 (2006).
31. S. S. Atsu, M. A. Kilicarslan, H. C. Kucukesmen and P. S. Aka, J. Prosthet. Dentistry 95,
430 (2006).
32. T. T. Heikkinen, J. P. Matinlinna, P. K. Vallittu and L. V. J. Lassila, Silicon 1, 199 (2009).
33. M. Kern, A. Arloi and A. B. Yang, J. Dental Res. 88, 817 (2009).
34. U. Lohbauer, M. Zipperle, K. Rischka, A. Petschelt and F. A. Müller, J. Biomed. Mater.
Res. Part B 87, 461 (2008).
35. C. Y. K. Lung, E. Kukk, T. Hägerth and J. P. Matinlinna, Appl. Surface Sci. 257, 1228
(2010).
36. C. Y. K. Lung and J. P. Matinlinna, Dental Mater. 28, 467 (2012).
37. K. L. Mittal (Ed.), Silanes and Other Coupling Agents, Vol. 5. VSP Brill, Leiden (2009).
38. K. L. Mittal (Ed.), Silanes and Other Coupling Agents, Vol. 4. VSP/Brill, Leiden (2007).
39. K. L. Mittal (Ed.), Silanes and Other Coupling Agents, Vol. 3. VSP, Utrecht (2004).
40. K. L. Mittal (Ed.), Silanes and Other Coupling Agents, Vol. 2. VSP, Utrecht (2000).
41. J. P. Matinlinna, L. V. Lassila and P. K. Vallittu, Acta Odontol. Scand. 65, 44 (2007).
42. J. P. Matinlinna and L. V. Lassila, Dental Mater. 27, 273 (2011).
43. M. Kern and S. M. Wegner, Dental Mater. 14, 64 (1998).
44. K. Yoshida, J. Biomed. Mater. Res. Part B 77, 28 (2006).
45. H. Mirmohammadi, M. N. Aboushelib, Z. Salameh, A. J. Feilzer and C. J. Kleverlaan,
Dental Mater. 26, 786 (2010).
46. M. N. Aboushelib, H. Mirmohamadi, J. P. Matinlinna, E. Kukk, H. F. Ounsi and Z.
Salameh, Dental Mater. 25, 989 (2009).
47. H. Mirmohammadi, M. N. Aboushelib, C. J. Kleverlaan, N. de Jager and A. J. Feilzer,
Dental Mater. 26, 627 (2010).
48. C. J. Kleverlaan and A. J. Feilzer, Dental Mater. 21, 1150 (2005).
49. A. J. Feilzer, A. I. Kakaboura, A. J. de Gee and C. L. Davidson, Dental Mater. 11, 186
(1995).
50. B. Yang, J. Biomed. Mater. Res. Part B 81, 283 (2007).
51. A. C. Quaas, B. Yang and M. Kern, Dental Mater. 23, 506 (2007).
52. J.-H. Phark, S. Duarte, M. Blatz and A. Sadan, J. Prosth. Dentistry 101, 29 (2009).
53. J. P. Matinlinna, A. H. Choi and J. K. H. Tsoi, Clin. Oral Implants Res. 24, 290 (2013).
54. M. Kern, J. Adhesion Sci. Technol. 23, 1097 (2009).
14 Escherichia coli O157:H7
Interactions and
Transport in Unsaturated
Porous Media
Yongjun Mao and Gang Chen*
CONTENTS
14.1 Introduction...................................................................................................240
14.2.1 E. coli O157:H7.................................................................................. 241
14.2.2 Surface Thermodynamic Measurements........................................... 241
14.2.3 Porous Medium.................................................................................. 242
14.2.4 Column Experiments......................................................................... 242
14.3 Results and Discussion..................................................................................244
14.3.1 E. coli O157:H7 Interactions in the Pore System..............................244
14.3.2 E. coli O157:H7 Transport Modeling................................................ 247
14.3.3 Effect of Air–Water Interface on Bacterial Transport....................... 249
14.4 Conclusions.................................................................................................... 250
Acknowledgment.................................................................................................... 250
References............................................................................................................... 250
During animal waste agricultural applications, the major concern is patho-

gen spreading, which may contaminate groundwater. Among the pathogenic
microorganisms found in animal waste, Escherichia coli O157:H7 is of par-
ticular concern. When transported in the vadose zone, E. coli O157:H7 may be
captured on the medium surface, at the air–water interface, or at the medium–
air–water three-phase interface depending on its predominant interactions
within the pore system. In this research, transport of E. coli O157:H7 in sil-
ica sand under water unsaturated conditions was investigated using column
* Dr. Gang Chen was the lead author for this chapter and can be contacted at gchen@eng.fsu.edu.
239
experiments.Its interactions within the system were characterized based on

E. coli O157:H7 and silica sand surface thermodynamic properties, which
were determined independently by means of contact angle measurements.
These calculated interactions were linked to E. coli O157:H7 transport param-
eters simulated from the column experiments.
14.1 INTRODUCTION
Appropriately managed animal waste material is a valuable nutrient source for sus-
tainable agriculture systems [1,2]. Animal waste can supply organic matter, nitrogen,
phosphorous, and potassium as well as other nutrients needed for plant growth when
it is used in the agricultural field. However, animal waste is rich in infectious agents
or pathogenic organisms, which can cause waterborne diseases [3,4]. Bacteria are
the most common pathogenic microorganisms found in animal waste. When animal
waste is used in an agricultural field, these pathogenic bacteria are usually retained
in the soil [5]. Infectious agents or pathogenic organisms in the soil are usually rep-
resented by Escherichia coli. Among the E. coli strains, E. coli O157:H7 is believed
to be the most notorious pathogen. E. coli O157:H7 mostly lives in the intestines of
cattle but has also been found in the intestines of chickens, deer, sheep, and pigs
[6,7]. E. coli O157:H7 does not cause illness in animals, but the animals can serve
as carriers for it. Infected animals, especially young ones, can shed the bacteria in
their feces.
Transport of the pathogenic bacteria in the soil is usually controlled by the pro-
pensity of these strains to adhere to the porous medium surfaces, which is deter-
mined by their interactions with the surrounding environment [8,9]. Bacterial
strains with different cell surface properties show different adhesion kinetics
and affinity for abiotic surfaces [10]. During transport in the subsurface agricul-
tural soil, E. coli O157:H7 may be retained in the system, which is controlled
by E. coli O157:H7 and porous medium surface properties and system mois-
ture content [11–14]. Specifically, under water-unsaturated conditions, E. coli
O157:H7 may be captured at the air–water–sediment three-phase interfaces due
to physical constraints [15]. Physical constraints are greatly affected by E. coli
O157:H7 interactions with the surrounding environment as well as water satura-
tion [16,17].
The aim of this study was to address the mechanisms of E. coli O157:H7 retention
in an unsaturated porous medium of silica sand. It was also the goal of this research
to validate that surface thermodynamic characterizations of E. coli O157:H7 and the
porous medium can help to elucidate E. coli O157:H7 transport mechanisms in the
pore system. We present here results of the transport of E. coli O157:H7 in a labora-
tory column packed with the model medium of silica sand. The transport was carried
out under unsaturated, steady-state flow conditions. Interactions of E. coli O157:H7
with silica sand as well as with the air–water interface were investigated based on
their independently determined surface thermodynamic properties. Traditional and
extended Derjaguin, Landau, Verwey and Overbeek (DLVO) theories were utilized
to calculate the interactions, which were then used in describing E. coli O157:H7
E. coli O157:H7 Interactions and Transport in Porous Media 241
adhesion in the porous medium [10,18,19]. The effect of the air–water interface on
E. coli O157:H7 retention was also discussed. The results of this research will lead
to a better understanding of E. coli O157:H7 transport and retention within the pore
system.

14.2.1 E. coli O157:H7
E. coli O157:H7 was obtained from the American Type Culture Collection (ATCC,
Manassas, VA; catalog no. 43895). After growth in Luria Bertani (LB) broth at
37°C until the late logarithmic growth phase, E. coli O157:H7 was collected by
centrifuging at 10,000 × g for 10 min. After washing twice with a sterilized buf-
fer solution (potassium phosphate monobasic/sodium hydroxide buffer, Fisher
Scientific, Pittsburgh, PA), it was resuspended in sterilized nanopure deionized water
(Barnstead, Dubuque, IA) to make a bacterial suspension. The number of cells per
milliliter of suspension was determined to be 1 × 108 cells/mL. For E. coli O157:H7
transport in the column, its growth was assumed to be minimal due to the short reten-
tion time. The size of E. coli O157:H7 was measured using a Malvern Zetasizer 3000
Hsa (Malvern Instruments Ltd., Malvern, Worcs, UK) as described by Meinders et
al. [20], and it was found to be 1.02 ± 0.11 μm. Since pure culture of E. coli O157:H7
was investigated, adenosine triphosphate (ATP) measurement was used for E. coli
O157:H7 quantification in this research. ATP measurement is a convenient bacte-
rial quantification technique, which provides a rapid and nonextractive technique for
estimation of bacterial concentration with advantages over other tedious and error-
prone methods of colony-forming unit (CFU) determination [10].
14.2.2 Surface Thermodynamic Measurements

Surface thermodynamic properties of E. coli O157:H7 were determined using con-
tact angle measurements [10,18]. An apolar liquid, diiodomethane, and two polar
liquids, formamide and water, were used as the measurement liquids. As weak
polar surfaces do not bind strongly with polar liquids, and strong polar surfaces
attract many types of impurities and are more soluble in polar liquids, formamide
and water were chosen as the polar liquids. Contact angle measurements were per-
formed using a contact angle meter (Tantec, Schaumburg, IL) following the method
described by Grasso et al. [18]. An E. coli O157:H7 suspension prepared as described
above was vacuum-filtered on silver metal membrane filters (0.45 μm, Osmonic, Inc.,
Livermore, CA) and air-dried to produce a lawn. The amount of E. coli O157:H7 on
the silver filter was approximately 13 mg to ensure a multilayer membrane coverage.
Each contact angle measurement was repeated 30 times, and the average results were
fitted in the van Oss–Chaudhury–Good equation to estimate E. coli O157:H7 surface
thermodynamic properties [21]:
(1 + cos θ) γ L = 2 ( γ SLW γ LW
L )
+ γ S+ γ −L + γ S− γ +L , (14.1)
where θ is the contact angle (degree); γL is the surface tension of the liquid that is used
for the measurement (J/m2); γLW is the Liftshitz–van der Waals component of surface
energy (subscript S denotes solid, i.e., E. coli O157:H7, and L denotes the measur-
ing liquid) (J/m2); γ+ is the electron-acceptor parameter; and γ– is the electron-donor
parameter of Lewis acid–base component of surface energy (J/m2). In addition, the
measuring liquid surface tension γL can be expressed in terms of Liftshitz–van der
+ −
Waals and Lewis acid–base components of surface energy γ LW L , γ L, and γ L:
γ L = γ LW − + . (14.2)
L + 2 γ Lγ L
14.2.3 Porous Medium

The porous medium used for this research was the same silica sand (8 mesh, Fisher
Scientific, Pittsburgh, PA) as reported by Chen and Zhu [22]. Silica sand was first
rinsed with deionized water and then treated with sodium acetate, hydrogen perox-
ide, sodium dithionate, and sodium citrate to remove organic matter. Treated silica
sand was then extensively flushed with sterilized nanopure deionized water until the
electrical conductivity was less than 1 dS/m. Surface thermodynamic properties of
silica sand were determined using the wicking method [23,24]. Similar to the con-
tact angle measurements, an apolar liquid, diiodomethane, and two polar liquids,
formamide and water, were used. The contact angles of silica sand were estimated
based on the Washburn equation [25]:
Ret γ L cos θ , (14.3)

h2 =
2η
where h is the height (m) of capillary rise of the wicking liquid at time t (s); Re is
the average interstitial pore radius (m), obtained by using low-surface-tension liquid
decane or hexadecane with cos θ = 1; and η is the viscosity of the measuring liquid
(Pa/s). The measurements were conducted using a Krüss K100 tensiometer (Krüss
GmbH, Hamburg, Germany) by dry-packing silica sand into a Krüss powder sample
holder in a closed chamber. Each measurement was repeated 30 times, and the aver-
age results were reported.
14.2.4 Column Experiments

The transport of E. coli O157:H7 through the porous medium of silica sand was
evaluated in column experiments. The experiments were conducted in a vertically
oriented custom-made column (5.0 cm internal diameter × 25.0 cm length). Silica
sand was packed in the column through CO2 solvation to eliminate air pockets. The
inflow was applied with a sprinkler from the top using a peristaltic pump (Masterflex,
Cole-Parmer, Vernon Hills, IL). During the column experiments, matric potential
(expressed as negative water height measurement due to fluid cohesion and surface
tension [Figure 14.1]) inside the column was monitored using three tensiometers
Pressure transducer
Soil sediments
Water
Ceramic porous cup
FIGURE 14.1 Matric potential measurement illustration.
mounted evenly along the length of the column and recorded using a Campbell
Scientific CR-7X data logger (Campbell Scientific, Inc., Logan, UT). Water content
within the column was predicted by the van Genuchten equation in terms of effective
water saturation, Se [26,27]:
Se = [1 + (αh)n](1/n−1), (14.4)
where α is the inverse of the air-entry potential (m–1; air-entry potential is expressed
as the water height measurement at which air passes through a water-saturated sedi-
ment); h is the matric potential (m-H2O); and n is the parameter related to pore size
distribution (–). α and n can be estimated by fitting the volumetric water content ver-
sus matric potential of the system using the van Genuchten fitting [28], which were
0.136 cm–1 and 4.776, respectively, for this research. During the transport experi-
ments, water saturation within the column was controlled by the flow rate and the
water suction by a hanging water column at the outlet. The hanging water column
was a saturated water tube that hung below that saturated bottom of the soil column
to provide suction for water to flow out of the column. The uniformity of water
saturation was achieved by adjusting the hanging water column until the three tensi-
ometers had consistent readings, which were then translated to the desired effective
water saturation according to the van Genuchten equation.
For each series of column experiments, a fresh column was packed, and prior to
starting each experiment, approximately 100 pore volumes of nanopure deionized
water were eluted through the column by the peristaltic pump to stabilize the col-
umn. A conservative pulse tracer (chloride) breakthrough curve (plot of eluted con-
centration versus time) was generated separately before the introduction of E. coli
O157:H7 suspension. For each column run, 1 pore volume of E. coli O157:H7 sus-
pension at a concentration of 1 × 108 cells/mL was introduced to the column, after
which the column was continuously flushed with nanopure deionized water until the
background signal was detected from the eluate collected by a fraction collector.
Collected eluate was measured for bacterial concentration using ATP assay [10]. For
each column run, three runs were performed, and the inconsistency of breakthrough
curves was within 5% (95% confidence interval).

14.3.1 E. coli O157:H7 Interactions in the Pore System
E. coli O157:H7 and silica sand surface thermodynamic properties were calculated
according to Equation 14.1 based on their contact angles measured with diiodometh-
ane, formamide, and water (Table 14.1). γLW was found to be 36.2 mJ/m2 for E. coli
O157:H7, which is consistent with van Oss’s prediction that γLW typically equals
40 mJ/m2 with minor variability for a considerable number of bacterial strains [21].
E. coli O157:H7 exhibited a monopolar surface; that is, the γ 1− was at least one order
of magnitude greater than γ 1+ (52.5 as compared to 0.8 mJ/m2) [21]. Silica sand had a
γLW value of 22.7 mJ/m2. The surface thermodynamic parameters of E. coli O157:H7
from this research were comparable to published data on other E. coli strains [29].
Similar to E. coli O157:H7, silica sand also exhibited a monopolar surface, with γ 1−
of 15.4 mJ/m2 and γ 1+ of 1.6 mJ/m2.
Assuming that E. coli O157:H7 behaves as inert particles and that its adhesion
can be understood by a physicochemical approach, traditional and extended DLVO
theories can be used in describing E. coli O157:H7 interactions in the pore sys-
tem [30–32]. The interactions that should be considered include the Lifshitz–van
der Waals, Lewis acid–base, and electrostatic interactions. The Lifshitz–van der
Waals interactions comprise Keesom, Debye, and London interactions [21]. Of these
three, Keesom and Debye interactions are only found among molecules that have
permanent dipole moments. The London interactions, however, are universal and are
of preponderate importance, especially in aqueous media that contain electrolytes.
Therefore, Lifshitz–van der Waals interactions are mainly contributed by London
interactions. Lifshitz–van der Waals interactions are initiated by the unevenness of
the electron cloud surrounding the molecules or particles. The electron-deficient part
of a molecule or particle is the positive pole, and the electron-excessive part is the
negative pole. Lifshitz–van der Waals interactions are thus electrostatic attractions
between oppositely charged poles of two or more dipoles. Owing to the induced
dipole potential, Lifshitz–van der Waals interactions occur between two particles
mediated in water. The positive induced pole face of one particle is directed to the
negative induced water pole face, and the positive induced water pole face is then
TABLE 14.1
E. coli O157:H7 and Silica Sand Contact Angles and Surface
Thermodynamic Properties
γLW γ+ γ–
θ (°)
a θ (°)
b θ (°)
c (mJ/ m2) (mJ/ m2) (mJ/ m2) ζ (mV)
E. coli O157:H7 47 ± 0.5 30 ± 0.4 26 ± 0.6 36.2 0.8 52.5 –7.7 ± 0.6
Silica sand 70 ± 1.2 60 ± 0.7 71 ± 0.4 22.7 1.6 15.4 –29.8 ± 0.8
a Contact angles measured with diiodomethane.

b Contact angles measured with formamide.
c Contact angles measured with water.
directed to the negative induced pole face of the other particle [21]. The Lifshitz–van
der Waals, Lewis acid–base, and electrostatic interaction free energies can be calcu-
lated as follows [33–35]:
∆G yLW
0 132
= −4 πRy0  ( γ 3LW − γ LW
2 )( )
γ 3LW − γ 1LW  (14.5)
∆G yAB
0 132
= 4 πRy0  ( γ 1+ − γ +2 )( γ 1− − γ −2 )

− ( +
γ − γ
1
+
3 )( −
γ − γ
1
−
3 )−( +
γ − γ
2
+
3 )( −
γ − γ
2
−
3 ) 
 (14.6)
  1 + e − κy  
∆G yEL
0 132
= πεε 0 R  2ψ 01ψ 02Ln 
 1− e − κy  ( )
+ ψ 201 + ψ 202 Ln(1 − e −2κy ) , (14.7)
 
where ∆G yLW0 132

, ∆G yAB
0 132
, and ∆G yEL
0 132
are the interaction free energies (superscript
LW denotes for Lifshitz-van der Waals, AB for Lewis acid-base, and EL for elec-
trostatic) (J) between E. coli O157:H7, 1 and silica sand or air–water interface,
2 immersed in the water, 3 evaluated at the equilibrium distance, y 0 of 1.57 ×
10 –10 m; R is the E. coli O157:H7 radius (m); ε and ε 0 are the relative dielectric
permittivities of water (78.55 for water at 25°C) and vacuum (8.854 × 10 –12 C/V
m), respectively; ψ01 and ψ02 are the potentials at the surfaces of E. coli O157:H7
and the silica sand or air–water interface; and 1/κ is the Debye–Hückel length and
also an estimation of the effective thickness of the electrical double layer [36]. ψ01
and ψ02 can be calculated based on the following equation:
ψ0 = ζ(1 + z/a) exp(κz), (14.8)
where ζ is the zeta potential measured at the slip plane (V); z is the distance from
the E. coli O157:H7 or silica sand surface to the slip plane (m), which is generally
on the order of 5 Å [21]; and a is the radius of E. coli O157:H7 or the silica sand (m).
E. coli O157:H7, silica sand, and the air–water interface were found to be nega-
tively charged, similar to prior observations [29,37,38]. Consequently, electrostatic
interactions between E. coli O157:H7 and silica sand and between E. coli O157:H7
and the air–water interface were positive, which served as a barrier to prevent E. coli
O157:H7 from getting close to silica sand and the air–water interface [39–41]. These
repulsive electrostatic interactions operated in the range of several tens of nano-
meters [19]. Once E. coli O157:H7 overcame the repulsive barrier and got close
to the silica sand surface or the air–water interface with the aid of hydrodynamic
forces, electrostatic interactions dropped dramatically, owing to the superimposition
of the double layers, and Lifshitz–van der Waals and Lewis acid–base interactions
became the actual driving forces for E. coli O157:H7 to interact with the surround-
ing environment [39]. Therefore, interactions evaluated at the equilibrium distance
where physical contact between E. coli O157:H7 and silica sand and between E. coli
O157:H7 and the air–water interface occurred were used to interpret E. coli O157:H7
adhesion in the pore system [20].
Interaction free energies between E. coli O157:H7 and silica sand and between
E. coli O157:H7 and the air–water interface evaluated at the equilibrium distance
were calculated according to Equations 14.5 through 14.7 (Table 14.2). E. coli
O157:H7 had attractive Lifshitz–van der Waals and repulsive Lewis acid–base inter-
actions with silica sand (–30.8 and 806.9 kT) and attractive Lifshitz–van der Waals
and Lewis acid–base interactions with the water–air interface (–6876.8 and –2921.8
kT). With silica sand, Lewis acid–base interaction free energy was one order of
magnitude greater than Lifshitz–van der Waals interaction free energy. However,
with the air–water interface, Lifshitz–van der Waals interaction free energy was
LW + AB+ EL
three times greater than Lewis acid–base interaction free energy. ∆G132 , the
sum of the interaction free energies of Lifshitz–van der Waals, Lewis acid–base,
and electrostatic interactions between E. coli O157:H7 and silica sand evaluated at
the equilibrium distance, was positive, that is, 785.2 kT, demonstrating repulsive
interactions of E. coli O157:H7 with silica sand. However, when interacted with the
air–water interface, this value was negative, that is, –9780.4 kT, indicating attractive
interactions. The above data were comparable to E. coli k12 interactions with silica
sand and air–water interface, for example, 1216.3 and –8695.8 kT, respectively [29].
Based on the analysis of E. coli O157:H7 interactions within the system, it could
be concluded that E. coli O157:H7 was retained at the medium–air–water three-phase
interface (Figure 14.2). E. coli O157:H7 diffused to the medium–air–water interface
and was retained there. Repulsive forces between E. coli O157:H7 and silica sand
aided E. coli O157:H7 attachment to the air–water interface. E. coli O157:H7 reten-
tion in these areas should correspond to the total “effective” interactions exerted on
TABLE 14.2
E. Coli O157:H7 Interaction Free Energy with Silica Sand and with
Air–Water Interface
With Silica Sand
LW AB EL LW + AB+EL
∆G 132 (kT )* ∆G132 (kT ) ∆G132 (kT ) ∆G132 (kT )
E. coli –30.8 806.9 9.1 785.2
With Air–Water Interface

LW + AB+EL
∆G LW
132 (kT )* AB
∆G132 (kT ) EL
∆G132 (kT ) ∆G132 (kT )
E. coli –6876.8 –2921.8 18.2 –9780.4
LW AB EL
Note: ∆G132 , ∆G132 , and ∆G132 were evaluated at the equilibrium distance where physical contact between
E. coli O157:H7 and the silica sand surface and between E. coli O157:H7 and the air–water inter-
face occurred.
* k is the Boltzmann constant (1.38 × 10–23 J/K), and T is absolute temperature (K). At 25°C, 1 kT =
4.11 × 10–21 J.
E. coli Water
O157:H7 Air
∆GTOT
132 (air−water interface)
E. coli
O157:H7
E. coli
O157:H7
E. coli
O157:H7 Silica sand surface
∆GTOT
132 (silica sand)
FIGURE 14.2 E. coli O157:H7 retention mechanism illustration.
E. coli O157:H7, that is, the sum of attractive interactions with the air–water inter-
face and repulsive interactions with silica sand (absolute values).
14.3.2 E. coli O157:H7 Transport Modeling

E. coli O157:H7 transport through homogeneously packed sand columns can be
mathematically described using the one-dimensional transport model, with the air–
water interface representing the sink for E. coli O157:H7 under water-unsaturated
conditions [42]. Based on the interaction analysis, E. coli O157:H7 should be retained
at the silica sand–air–water interface, as follows [43,44]:
∂ ∂  ∂C  ∂
[βC ] =  Dzβ − [qC ] = k1βC , (14.9)
∂t ∂z  ∂z  ∂z
where C is the E. coli O157:H7 concentration in the liquid phase (cells/m3); Dz is

the apparent diffusion coefficient (m2/s); β is the moisture content (moisture vol-
ume divided by the total volume of the porous medium; m3/m3); q is the specific
discharge (Darcian fluid flux; m/s); k1 is the deposition coefficient accounting for
E. coli O157:H7 retention at the medium–air–water interface (s–1); z is the axial coordi-
nate (m); and t is time (s). For each series of column experiments, a fresh column was
used. Owing to the low introduced E. coli O157:H7 concentration, the variation of k1
owing to E. coli O157:H7 attachment can be ignored.
E. coli O157:H7 breakthrough curves were characterized by a self-sharpening
front, which became broader and more diffused at the elution limb (Figure 14.3).
These E. coli O157:H7 transport observations were consistent with those of other
E. coli strains [45–48]. The long-lasting tails of the breakthrough curves indicated
kinetically controlled E. coli O157:H7 retention in the column [46]. The E. coli
O157:H7 breakthrough curves were simulated according to Equation 14.9 using
1.0
Effective saturation 0.8
0.8 Effective saturation 0.4
0.6
C/C0
0.4
0.2
0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
Pore volume (–)
FIGURE 14.3 E. coli O157:H7 breakthrough curves at variable water saturation.
an implicit, finite-difference scheme. All the parameters were optimized by mini-

mizing the sum of squared differences between observed and fitted concentrations
using the nonlinear least-squares method [49]. Based on the models, E. coli O157:H7
retention was assumed to occur at the medium–air–water interface [29]. With the
decrease of water saturation, E. coli O157:H7 retention increased accordingly, which
was evidenced by the corresponding peak values of the breakthrough curves (Figure
14.3). E. coli O157:H7 retention linearly decreased with the increase of water satura-
tion (Figure 14.4).
2.0
1.6
1.2
µ (min–1)
0.8
0.4
0.0
0.0 0.2 0.4 0.6 0.8 1.0
Effective saturation Se (–)
FIGURE 14.4 E. coli O157:H7 deposition coefficient as a function of effective water

saturation.
14.3.3 Effect of Air–Water Interface on Bacterial Transport

With the decrease of system saturation, the air–water interface area increased
accordingly, and more E. coli O157:H7 was retained in the system. The presence of
the air–water interface played an important role in controlling E. coli O157:H7 reten-
tion under water-unsaturated conditions [50,51]. The area of the air–water interface
can be estimated from pore size radii [52]:
1
β0  n n
ρg   β  1− n

S=
αγ ∫
β
 β  − 1 d β, (14.10)
 0 
where S is the air–water interfacial area per unit volume (cm 2/cm 3); ρ is the
water density (kg/m 3); g is the gravitational constant (9.8 m/s2); γ is the water
surface tension (72.69 mJ/m 2 at 20°C); and β 0 is the porous medium’s volume
fraction of pore space or porosity of the column. β, α, and n were defined previ-
ously. The air–water interfacial area increased with decreasing water saturation
(Figure 14.5). For this research, all the column experiments were conducted at
water saturation ranging from 0.2 to 0.8, within which the air–water interfacial
area displayed a linear relationship with water saturation. Consequently, the increase
of E. coli O157:H7 retention with decreasing water saturation should show the
same trend as with increasing air–water interfacial area. To reflect increased
E. coli O157:H7 retention with decreasing water saturation, the E. coli O157:H7
deposition coefficient was plotted against the air–water interfacial area (Figure
14.6). The linear relationship indicated that the increased E. coli O157:H7 reten-
tion was attributed to the increased air–water interfacial area with decreasing
water saturation.
400
Air−water interfacial area (cm2/cm3)
300
200
100
0
0.0 0.2 0.4 0.6 0.8 1.0
Effective saturation Se (–)
FIGURE 14.5 Air–water interfacial area as a function of effective water saturation.

2.0
1.6
1.2
µ (min–1)
0.8
0.4
0.0
80 100 120 140 160
Air−water interfacial area (cm2/cm3)
FIGURE 14.6 E. coli O157:H7 deposition coefficient as a function of air–water interfacial

area.
14.4 CONCLUSIONS
To investigate E. coli O157:H7 retention and transport in a pore system, interactions of
E. coli O157:H7 within the pore system were evaluated in this research. Based on the
analysis of E. coli O157:H7 interactions within the system, E. coli O157:H7 was found
to be retained at the medium–air–water three-phase interface. Specifically, E. coli
O157:H7 was attached to the air–water interface, which was aided by the repulsive
forces between E. coli O157:H7 and the medium. E. coli O157:H7 retention was deter-
mined by the total “effective” interactions exerted on E. coli O157:H7, that is, the sum
of attractive interactions with the air–water interface and repulsive interactions with
the medium surface. To account for increased bacterial retention with decreasing water
saturation, the air–water interfacial area was examined as a function of water satura-
tion. The linear relationship between E. coli O157:H7 retention and the air–water
interfacial area indicated that the increased E. coli O157:H7 retention was attributed
to the increased air–water interfacial area with decreasing water saturation.
ACKNOWLEDGMENT
The work was supported by the National Research Initiative of the USDA Cooperative
State Research, Education and Extension Service, grant no. 2007-35102-18111 to
Florida A&M University.
REFERENCES
1. G. Benckiser and T. Simarmata, Fertil Res 37, 1 (1994).
2. G.L. Meerdink, Vet Human Toxicol 42, 115 (2000).
3. M.L. Hutchison, L.D. Walters, T. Moore, D.J.I. Thomas and S.M. Avery, Appl Environ
Microbiol 71, 691 (2005).
4. M. Goberna, S.M. Podmirseg, S. Waldhuber, B.A. Knapp, C. García and H. Insam, Appl
Soil Ecol 49, 18 (2011).
5. L. Sahlstrom, Bioresource Technol 87, 161 (2003).
6. M.T. Jay, M. Cooley, D. Carychao, G.W. Wiscomb, R.A. Sweitzer, L. Crawford-Miksza,
J.A. Farrar, D.K. Lau, J. O’Connell, A. Millington, R.V. Asmundson, E.R. Atwill and
R.E. Mandrell, Emerg Infect Dis 13, 1908 (2007).
7. M. Muniesa, J. Jofre, C. Garcia-Aljaro and A.R. Blanch, Environ Sci Technol 40, 7141
(2006).
8. S.A. Bradford, J. Simunek and S.L. Walker, Water Resour Res 42, WR004904 (2006).
9. A. Jacobs, F. Lafolie, J.M. Herry and M. Debroux, Colloids Surf B 59, 35 (2007).
10. G. Chen and K.A. Strevett, Environ Microbiol 3, 237 (2001).
11. A.V. Semenov, L. van Overbeek and A.H. van Bruggen, Appl Environ Microbiol 75,
3206 (2009).
12. S.E. Walker, S. Mostaghimi, T.A. Dillaha and F.E. Woeste, Trans ASBE 33, 807 (1990).
13. S.B. Kim, S.J. Park, C.G. Lee and H.C. Kim, Hydrol Proc 22, 3856 (2008).
14. A.K. Guber, D.R. Shelton and Y.A. Pachepsky, Vadose Zone J 4, 828 (2005).
15. G. Gargiulo, S.A. Bradford, J. Šimùnek, P. Ustohal, H. Vereecken and E. Klumpp,
Vadose Zone J 7, 406 (2008).
16. H. Small, J Colloid Interface Sci 48, 147 (1974).
17. D.L. Liu, P.R. Johnson and M. Elimelech, Environ Sci Technol 29, 2963 (1995).
18. D. Grasso, B.F. Smets, K.A. Strevett, B.D. Machinist, C.J. van Oss, R.F. Giese and
W. Wu, Environ Sci Technol 30, 3604 (1996).
19. B.F. Smets, D. Grasso, M.A. Engwall and B.J. Machinist, Colloids Surf B 14, 121
(1999).
20. J.M. Meinders, H.C. van der Mei and H.J. Busscher, J Colloid Interface Sci 176, 329
(1995).
21. C.J. van Oss, Interfacial Forces in Aqueous Media. Marcel Dekker, New York (1994).
22. G. Chen and H.L. Zhu, Colloids Surf B 44, 41 (2005).
23. M.E.P. Walinder and D.J. Gardner, J Adhesion Sci Technol 13, 1363 (1999).
24. C.A. Ku, J.D. Henry, R. Siriwardane and L. Roberts, J Colloid Interface Sci 106, 377
(1985).
25. E.W. Washburn, Phys Rev 17, 273 (1921).
26. N. Toride and F.J. Leij, Soil Sci Soc Am J 60, 342 (1996).
27. N. Toride and F.J. Leij, Soil Sci Soc Am J 60, 352 (1996).
28. M.T. van Genuchten, Soil Sci Soc Am J 44, 892 (1980).
29. G. Chen, Colloids Surf B 67, 265 (2008).
30. P. Teixeira, J. Azeredo, R. Oliveira and E. Chibowski, Colloids Surf B 12, 69 (1998).
31. A. Vilinska and K.H. Rao, Miner Metall Proc 28, 151 (2011).
32. S. Bayoudh, A. Othmane, L. Mora and H. Ben Ouada, Colloids Surf B 73, 1 (2009).
33. S. Morisada, K. Muranishi, H. Shinto and K. Higashitani, Kagaku Kogaku Ronbun 31,
295 (2005).
34. R.F. Considine and C.J. Drummond, Langmuir 17, 7777 (2001).
35. J.P. Chapel, Langmuir 10, 4237 (1994).
36. K.C. Marshall, J.A. Breznak, G.B. Calleja, G.A. McFeters and P.R. Rutter, Microbial
Adhesion and Aggregation. Springer-Verlag, New York (1984).
37. M. Manciu and E. Ruckenstein, Colloids Surf A 400, 27 (2012).
38. H. Schwegmann, A.J. Feitz and F.H. Frimmel, J Colloid Interface Sci 347, 43 (2010).
39. M. Farahat, T. Hirajima, K. Sasaki and K. Doi, Colloids Surf B 74, 140 (2009).
40. J. Liu, W. Xu, Q. Zhao, M. Qiao, H. Zhang and G. Chen, Separ Sci Technol 42, 1031
(2007).
41. W. Zhang, B. Rittmann and Y.S. Chen, Environ Sci Technol 45, 2172 (2011).
42. A. Schafer, P. Ustohal, H. Harms, F. Stauffer, T. Dracos and A.J.B. Zehnder, J Contam
Hydrol 33, 149 (1998).
43. G.M. Hornberger, A.L. Mills and J.S. Herman, Water Resour Res 28, 915 (1992).
44. T.K. Sen, D. Das, K.C. Khilar and G.K. Suraishkumar, Colloids Surf A 260, 53 (2005).
45. M. Auset and A.A. Keller, Water Resource Res 42, W12S12 (2006).
46. G. Chen, Eur J Soil Sci 63, 239 (2012).
47. K. Garbrecht, G.A. Fox, J.A. Guzman and D. Alexander, Trans ASABE 52, 481 (2009).
48. E. Smith and A. Badawy, Water Sci Technol 57, 1123 (2008).
49. N. Toride, F.J. Leij and M. van Genuchten, The CXTFIT Code for Estimating Transport
Parameters from Laboratory or Field Experiments, Version 2.1. U.S. Salinity Laboratory,
Riverside, CA (1995).
50. M. Hermansson and B. Dahlback, Microbial Ecol 9, 317 (1983).
51. A. Schafer, H. Harms and A.J.B. Zehnder, Environ Sci Technol 32, 3704 (1998).
52. J.W. Cary, J Contam Hydrol 15, 243 (1994).
Section IV
Adhesives and Their Testing
15 Improved Approaches
to Analyze the Nonlinear
Behavior of Adhesives
in Bonded Assemblies
Jean Yves Cognard,* Romain Créac’hcadec,
Julien Maurice, Laurent Sohier, Bernard Gineste,
and Peter Davies
CONTENTS
15.1 Introduction................................................................................................... 256
15.2 Material Parameters....................................................................................... 257
15.3 Influence of Stress Singularities on the Analysis of the Behavior of the
Adhesive under Shear Loading...................................................................... 258
15.3.1 Presentation of the Modified TAST................................................... 258
15.3.2 Results of Asymptotic Analysis.........................................................260
15.3.3 Stresses throughout Thickness of Adhesive Joint............................. 261
15.3.4 Experimental Results......................................................................... 263
15.4 2D Modeling of Nonlinear Behavior of Adhesive.........................................264
15.4.1 Experimental Results under Tensile-Shear Loadings.......................264
15.4.2 Nonassociated Elastoplastic Model for Monotonic Loadings........... 265
15.5 Toward 3D Modeling of the Nonlinear Behavior of the Adhesive................ 269
15.5.1 Influence of Hydrostatic Stress on Adhesive Elastic Limit............... 269
15.5.2 Influence of Edge Effects on Industrial Assembly............................ 270
15.6 Conclusion..................................................................................................... 275
References............................................................................................................... 276
* Prof. Jean Yves Cognard was the lead author for this chapter and can be contacted at jean-yves.
cognard@ensta-bretagne.fr.
255
The objective of this study was to define a reliable tool for dimension-
ing adhesively bonded assemblies. Thus, numerical models that accurately
describe the complex nonlinear behavior of adhesives are required. Various
pressure-dependent constitutive models can be found in the literature, but
only a few experimental results are available. In addition, for ductile adhe-
sives, the strain evolution is strongly dependent on the tensile-shear loadings,
so a large experimental database is necessary. Experimental and numerical
analyses of the mechanical behavior of bonded joints are made particularly
difficult by the influence of edge effects, and stress singularities can initiate
cracks in the adhesive. Therefore, understanding the stress distribution in a
joint is essential to improve the performance of adhesively bonded assem-
blies. In order to obtain reliable experimental data to analyze the nonlinear
behavior of the adhesive in an assembly, there is an urgent need for tests that
limit the influence of stress singularities. Starting from experimental results
under various tensile-shear loadings obtained with a modified Arcan device,
which has been designed to strongly limit edge effects, an accurate 2D non-
associated elastoplastic model is presented. Moreover, experimental results
under quite large negative hydrostatic stress allow an accurate modeling of
the initial yield surface, which is an important point for the development of a
3D model. Such a model has been used here to analyze the influence of vari-
ous parameters on the transmitted load for an industrial assembly under com-
plex loading.
15.1 INTRODUCTION
This study is concerned with improving the performance of adhesively bonded
structures. Adhesively bonded joints offer many advantages [1,2], but a lack of con-
fidence limits the current use of this technology. The experimental and numerical
analyses of the mechanical behavior of bonded joints are made particularly diffi-
cult by edge effects associated with bimaterial interfaces [3,4]. Stress singularities
can contribute to the initiation and propagation of cracks in adhesive joints [5–7].
Therefore, understanding the stress distribution in adhesive joints is essential to
improving the performance of adhesively bonded assemblies. The development of
more reliable experimental tests to characterize the nonlinear behavior of the adhe-
sive in an assembly is also necessary. Several studies have described the influence
of the geometry on the stress singularities for bimaterial joints. However, precise
finite element computations are necessary to analyze stress singularities in order
to optimize the design of bonded assemblies and take into account the influence of
various parameters, such as the geometry of the assembly, the interactions between
the influences of the two interfaces in a thin joint, and the nonlinear behavior of the
adhesive [8].
The single lap joint is the most widely used test specimen configuration for
evaluating the behavior of adhesives in an assembly as the manufacturing of such
specimens is quite easy, and they require only a standard tensile testing machine.
Also various simplified models exist for single lap joints [9]. But such specimens
Nonlinear Behavior of Adhesives in Bonded Assemblies 257
are associated with complex loading of the adhesive, that is, nonuniform shear
stress along the overlap length, quite large peel stresses at the two ends of the
overlap, and large edge effects associated with geometrical and material param-
eters. Thus, peel and cleavage forces make the analysis of experimental results
quite difficult and can limit the transmitted load of the assembly despite various
techniques proposed to limit the influence of edge effects [10]. In the literature,
various geometries have been proposed to reduce the stress concentrations, such as
modifications to spew and chamfer size [11,12]; moreover, some work on the opti-
mization of such geometries has also been reported [13,14]. Simplified models give
only the average stress state in the adhesive and are not able to analyze the stress
concentrations close to the free edges of the adhesive [8–15]; thus, the optimiza-
tion of industrial bonded joints requires accurate numerical models. It is shown
that specific geometries, which depend on various parameters (material properties,
joint thickness, etc.), are required to limit the influence of edge effects significantly
[3,4,7,16]. Concerning pressure-dependent constitutive models, various proposals
can be found in the literature [6,17], but only a few experimental results are avail-
able to identify the material parameters. Thus, a large database of experimental
results under various tensile-shear loadings is necessary to develop representative
numerical models.
The first part of this paper presents the influence of stress singularities on the
mechanical behavior of an adhesive in an assembly under shear loadings, starting
from the widely used thick adherend shear test (TAST). Results of an asymptotic
analysis, assuming elastic behavior, are presented. These illustrate some of the
conditions necessary in order to design experimental devices that strongly limit
stress concentrations and show the strong influence of the local geometry at the free
ends of the adhesive. The influence of different parameters on the stress distribu-
tion through the adhesive joint thickness is presented using precise finite element
computations. In order to strongly limit the influence of edge effects, a modified
TAST fixture has been proposed [18]. Experimental results show that large stress
concentrations can lead to an incorrect analysis of the adhesive behavior. The sec-
ond part of the paper presents the development of a 2D model in order to describe
the nonlinear behavior of an adhesive starting from the results obtained with the
modified Arcan device developed to obtain reliable adhesive data under propor-
tional tensile/compression-shear loadings. The third part underlines the influence
of negative hydrostatic stress on the elastic limit using results from shear tests in a
pressure vessel. A pressure-dependent model is then used to analyze the influence
of various parameters on the transmitted load for an industrial structure under com-
plex loadings.
15.2 MATERIAL PARAMETERS
All the experimental results presented in this paper were obtained on assemblies
typical of those used in marine applications. These are based on relatively thick 0.4 mm
joints, Huntsman Araldite 420 A/B epoxy resin (Cambridge, UK), and aluminum
substrates. A simple surface preparation was used: abrasion with 120 grade paper
and acetone degreasing. Low temperature curing was used: the adhesive was left for
12 h at 20°C after assembly and then cured at 50°C for 4 h. A displacement rate of
the tensile testing machine of 0.5 mm/min was used throughout.
The numerical results presented in the paper for the study of the stress state in a
bonded joint were obtained using a 2D plane stress finite element analysis, with the
code CAST3M [19].
An accurate analysis of the mechanical behavior of bonded assemblies, that is,
the analysis of the stress distribution through the thickness of the adhesive joint,
requires refined meshes, especially for structures with edge effects. For the differ-
ent computations, 20 linear elements were used for a thickness of 0.1 mm [8]. The
elastic parameters used for aluminum substrates are Young’s modulus, Es = 75 GPa,
and Poisson’s ratio, νs = 0.3.
The behavior of the adhesive is assumed to be elastic and isotropic (Ea = 2.0 GPa,
νa = 0.3). For the adhesive used, the elastic limit was identified for 2D tensile-shear
loadings [20,21] and is proposed in the following form:
σ 2yy σ 2xy
+ = 1 or σ eq
2
= 1 (15.1)
Ra2 Sa2
where σyy and σxy are, respectively, the normal and shear stress components; Ra and Sa are
material parameters (Ra = 40 MPa and Sa = 25 MPa [21]). Thus, this equivalent stress, σeq,
is used to analyze the stress distribution in the joint; it is more appropriate than the usual
von Mises equivalent stress, as adhesives are pressure-dependent materials.
15.3 INFLUENCE OF STRESS SINGULARITIES ON THE ANALYSIS OF

THE BEHAVIOR OF THE ADHESIVE UNDER SHEAR LOADING
This section presents the influence of stress singularities on the mechanical behavior
of an adhesive in an assembly under shear loading. A methodology is proposed in
order to develop an experimental device that strongly reduces the influence of edge
effects in shear tests. As edge effects for bimaterial systems depend on the material
properties, on the local geometry of the assembly close to the substrate–adhesive
interface, and on the overall behavior of the assembly, all these issues have to be
analyzed in order to design the geometries of the bonded specimen, the support, and
the fixing system of the bonded specimen.
15.3.1
Presentation of the Modified TAST
The TAST (ISO 11003-2, ASTM D3983, Figure 15.1) is a logical extension of the
single lap shear (SLS) test method, which is widely used to evaluate adhesive sys-
tems. Using thick adherends, typically 10 mm thick (or 6 mm thick for the Althof test
[22]), and a short overlap enables the peel stresses, which complicate the SLS analy-
sis, to be significantly reduced. This fixture is often used to analyze the mechanical
behavior of adhesives in bonded assemblies [5,22]. Close observations of tests reveal
6.35 mm (0.25”) Adhesive tested

90.5 mm (3.56”) 9.53 mm (0.375”)
12.7 mm (0.5”) 9.53 mm (0.375”)

diameter pin hole
3.18 mm (0.125”) Adhesive tested
notch
(a) (b)
FIGURE 15.1 Representation of the TAST specimen (width: 25.4 mm). (a) Geometry and
(b) adhesive tested.
crack propagation close to the substrate–adhesive interfaces [5,7]; these cracks lead
to incorrect analysis of the behavior of the adhesive. Figure 15.2 presents the average
shear stress-displacement curve obtained with such a test. DT denotes the relative
displacements of the two ends of the adhesive in the tangential direction. It was
obtained with a noncontact measurement system based on image correlation. FT
represents the tangential component of the applied load in the tangential direction in
the midplane of the adhesive joint (the normal component is equal to zero for shear
loading). It can be seen that for such a test, cracks appear at the substrate–adhesive
interface even for quite low average shear stress. Thus, a correct analysis of such
results involving nonlinear material behavior and crack propagation can be complex.
The experimental and numerical analyses of the mechanical behavior of such
bonded joints are made particularly difficult by stress singularities due to edge
effects [3,23]. Therefore, understanding how stress is distributed in an adhesive
joint can lead to improvements in adhesively bonded assemblies. Since it is difficult
to take the effects of stress singularities into account when analyzing experimental
results, it is useful to design experimental fixtures that strongly limit edge effects
in order to obtain reliable data [7,24]. A modified TAST (Figure 15.3) has therefore
30 Modified TAST Substrate

Average shear stress (MPa)
Modified TAST
25 Adhesive
TAST Substrate
20
15
10
TAST Substrate
5
Adhesive
0
0 0.1 0.2 0.3 0.4 0.5 Substrate
DT (mm)
Crack Crack
(a) (b)
FIGURE 15.2 Experimental results from TAST and modified TAST for an adhesive thick-
ness of 0.4 mm. (a) Results in the load-displacement diagram (DT being relative displace-
ment of the two substrates in the loading direction) and (b) presentation of the adhesive
deformations.
x y
y z
z x
(1) Substrate
(2)
(1) Adhesive 2e
(3) Substrate
(c)
y d
(4)
x
z β
(1) Substrate h
1 - small bonded sample
2 - support Adhesive 2e
3 - fastening device ρ
4 - positioning system Substrate
(a) (b) (d)
FIGURE 15.3 Modified TAST fixture with a small bonded sample. (a) The experimental
device, (b) the fixture system for the small bonded sample (x being the loading direction),
(c) specimen without beak, and (d) specimen with beaks on two sides of the substrates and
clean free edges of the adhesive.
been proposed using the same bonded area as that of the TAST (adhesive section
Sc = 9.53 mm × 25.4 mm) [18]. In order to limit the bending of the device, which
increases the peel stress in the joint, a rigid support is proposed. A special system
was used (noted 4, Figure 15.3b) for precise positioning of the specimen in the
support. For the modified TAST, it can be seen in Figure 15.2 that homogeneous
deformation of the adhesive is observed (without crack propagation close to the
free edges). Moreover, the responses of the two tests are different, both in terms of
maximum average shear stress and deformation at failure.
15.3.2 Results of Asymptotic Analysis

Failure initiation in an adhesive joint between two metallic structures always starts at
stress concentration points, which are often associated with stress singularities [6,7].
For instance, using asymptotic analysis [3,25], for 2D problems with elasticity
assumptions, close to a corner the relevant parts of the displacement field u and the
stress tensor σ, using the polar coordinates (r: radius, θ: angle), are given by
u(r, θ) ≈ k r λ u(θ) (15.2)
σ(r, θ) ≈ k r λ−1 σ(θ) (15.3)
The singularity exponent λ is the solution of an eigenvalue problem defined by

the boundary close to the singular point; thus, this parameter depends on the geom-
etry close to the singular point. The intensity factor k of the problem depends on
(a) (b)
140
120
δs 100
λ < 1: edge effects
Substrate 80
δa (°)
Adhesive 60
δa 40
20
λ ≥ 1: regular solution
0
20 60 100 140 180
δs (°)
FIGURE 15.4 Results of asymptotic analysis for bimaterial systems (aluminum-adhesive)

under plane stress assumptions. (a) Definition of the local geometry close to the free edge
of a substrate–adhesive interface and (b) influence of the local geometry on the singularity
exponent λ.
the complete geometry of the structure and on the external loads. A regular stress
distribution is obtained when λ ≥ 1.
Close to a corner, two angles δs and δa can be used to define the local geometries
of the substrate and of the adhesive (Figure 15.4a). Figure 15.4b presents, in the case
of 2D plane stress, the possible choices that exist in order to design geometries for
bimaterial systems (aluminum-adhesive), which do not present stress concentrations
close to the interface [17]. For instance, for SLS specimens (Figure 15.1), two values
of the substrate angle δs exist: 90° and 180°. For δs = 90°, a regular stress is obtained
for δa < 60°; such a solution can be manufactured. But, for δs = 180°, a regular stress
is obtained for δa < 15°; such a solution is nearly impossible to manufacture. These
results underline that for constant thickness joints, it is necessary to use substrates
with quite sharp beaks (δs < 50°) in order to have a large choice for the design of
the free edge of the adhesive. It is important to note that these results depend on the
elastic properties of the two materials. Usual solutions generally reduce only the
stress singularities [5,11,12].
15.3.3 Stresses throughout Thickness of Adhesive Joint

Figure 15.5 presents the evolution of the shear and peel stresses in the adhesive for
a joint thickness of 2e = 0.4 mm with respect to the overlap length (the center of the
joint is associated with x = 0). Two geometries were used: straight substrates and
straight free edges of the adhesive (Figure 15.3c) and substrates with thin beaks
(h = 0.1 mm, d = 0.5 mm, β = 60°; Figure 15.3d) and cleaned adhesive free edges
(ρ = 1.5e; Figure 15.3d). The different curves are associated with different y posi-
tions in the adhesive; y = 0 represents the average plane of the adhesive, and y = e
is close to the adhesive–substrate interface. The average shear stress in the joint
was normalized to 1, in order to make analysis easier. It can be noted that for the
two geometries, the shear stress distributions are quite regular and similar. But for
(a) (b)
1.2 4.0
y=e
1.0 3.0
Shear stress (MPa)
2.0
Peel stress (MPa)

0.8
y=e 1.0 y = e/2
0.6 y=0
0.0
0.4 –6 –4 –2 0 2 4 6
–1.0
y = e/2
0.2 –2.0
y=0
0.0 –3.0
–6 –4 –2 0 2 4 6
–4.0
x (mm) x (mm)
(c) (d)
1.2 0.8
y = e/2
1.0 0.6
y=e
Shear stress (MPa)
0.4
Peel stress (MPa)
0.8
0.2
0.6 y=0
0.0
0.4 –6 –4 –2 0 2 4 6
–0.2
y=e
0.2 y = e/2 –0.4
y=0 –0.6
0.0
–6 –4 –2 0 2 4 6 –0.8
x (mm) x (mm)
FIGURE 15.5 Evolution of stress through the thickness of the adhesive with respect to
overlap length for two geometries, e being half of the adhesive thickness. (a) Shear stress for
straight substrates and straight edges of the joint, (b) peel stress for straight substrates and
straight edges of the joint, (c) shear stress for substrates with beaks and clean free edges of
the adhesive, and (d) peel stress for substrates with beaks and clean free edges of the adhesive
(note larger y-scale).
straight substrates and straight ends of the joint (Figure 15.5b), quite large peel stress
gradients are observed close to the free edges of the joint. Such results exist also for
the TAST, with a slightly larger shear stress value as the rigidities of the supports
are different [7]. The use of substrates with thin beaks and clean free edges of the
adhesive can significantly limit the peel stress state close to the free edges of the joint
(Figure 15.5d).
Figure 15.6 presents the evolution of the equivalent stress through the thickness
of the adhesive with respect to overlap length for the two previous geometries. It is
important to notice that the nonlinear behavior will appear at the point where this
equivalent stress is maximum. Figure 15.6 underlines the beneficial effect of the
beaks and of the clean free edges of the adhesive on the stress state close to the free
edges of the adhesive. Moreover, it can be noted that for straight edges of the adhe-
sive, the maximum value of the equivalent stress is obtained at the substrate–adhesive
interface, which is the weaker part of the assemblies. For clean free edges of the
adhesive and substrates with beaks, the maximum value of the equivalent stress is
obtained within the joint.
(a) (b)
0.06
0.10
Equivalent stress (MPa)
Equivalent stress (MPa)

y=e 0.05
0.08
0.04
0.06
0.03
y = e/2
0.04 0.02
y = e/2
0.02 y=e
0.01
y=0 y=0
0.00 0.00
–6 –4 –2 0 2 4 6 –6 –4 –2 0 2 4 6
x (mm) x (mm)
FIGURE 15.6 Evolution of the equivalent stress through the thickness of the adhesive
with respect to overlap length for the following geometries, e being the adhesive thickness.
(a) Straight substrates and straight joint and (b) substrates with beaks and clean free edges of
the adhesive.
Results from precise finite element computations and from asymptotic analysis
are in good agreement.
15.3.4 Experimental Results

Figure 15.7 presents the scatter in the experimental results using the modified TAST
for substrates with beaks and cleaning of the adhesive. This figure shows results
from tests on three specimens and indicates that test variability is quite low. For
specimens with straight substrates and straight edges of the adhesive, results similar
to those of TAST are obtained (Figure 15.2) [18].
These results provide some guidelines for manufacturing conditions in order to
optimize bonded joints. The lower the equivalent stress is (computed assuming lin-
ear elasticity), the lower the risk of first rupture near the edge of the joint will be. It
30
25
20
15
10
0
0 0.1 0.2 0.3 0.4 0.5
DT (mm)
FIGURE 15.7 Scatter in the experimental results for specimens with beaks and clean free
edges of the adhesive; results of three tests (DT denotes the relative displacement of the two
ends of the adhesive in the tangential direction).
is important to note that the nonlinear behavior of the adhesive joint does not seem
to limit the influence of the edge effects [7]. But, it has been shown that a coupling
between the nonlinear behavior of the adhesive and the stress concentration is
possible [17].
15.4 2D MODELING OF NONLINEAR BEHAVIOR OF ADHESIVE

This section presents the development of a 2D model in order to describe the nonlinear
behavior of an adhesive. An accurate description of the real response of the adhe-
sive in an assembly requires test results for various loadings. First, a representation
of the modified Arcan device developed to obtain reliable data under proportional
tensile/compression-shear loadings is proposed. Then, a model is proposed in order
to analyze the stress state in the joint under monotonic loadings for assemblies with
low stress concentrations.
15.4.1 Experimental Results under Tensile-Shear Loadings

The design of this modified Arcan fixture (use of beaks, clean free edges of the
adhesive, fixing system of the bonded specimen) allows to strongly limit the influ-
ences of the edge effects and to prevent preloading of the adhesive (Figure 15.8) [7].
For this test, a bonded specimen with a rectangular section (10 mm × 65 mm) was
proposed. The main parameters that define the geometry of the bonded specimen are
as follows: e = 0.2 mm, h = 0.2 mm, d = 0.5 mm, r0 = 0.8 mm, ρ = 1.5e, and α = 45°
(Figure 15.8c). This experimental fixture coupled with a noncontact extensometry
γ y
γ
(1)
(3) (2)
(3)
(2)
(4) u
(1)
x (4)
(a) (b)
d
r0 1 - support for the Arcan
Substrate
α h device
2 - bonded specimen
Adhesive 2e 3 - clamping system
ρ
4 - support for the
Substrate clamping system
(c) (d)
FIGURE 15.8 Modified Arcan device. (a) Test under shear loading, (b) test under tensile-
shear loading, (c) geometries of the substrates with beaks and of the bonded specimens, and
(d) clamping system.
30
Beaks

25
No beaks
20
15
10
0
0 0.1 0.2 0.3 0.4 0.5
DT (mm)
FIGURE 15.9 Experimental results from modified Arcan test under shear loadings for an
adhesive thickness of 0.4 mm; results of three tests for substrates with beaks and results of
two tests for substrates without beaks (DT denotes the relative displacements of the two ends
of the adhesive in the tangential direction).
system based on image correlation allows us to analyze, for proportional loadings,

the nonlinear behavior of an adhesive joint [20].
In the following, DN and DT denote the relative displacements of the two adhe-
sive–substrate interfaces, respectively, in the normal (y, Figure 15.8b) and tangential
(x, Figure 15.8b) directions of the midplane of the adhesive (measured using image
correlation techniques). FN and FT represent the components of the applied load in
the normal and tangential directions, respectively.
Figure 15.9 presents results from shear tests (γ = 90°, Figure 15.8b) for mono-
tonic loadings. Significant differences are obtained with beaks and without (i.e.,
with straight substrates); moreover, results are more scattered without beaks (three
results are presented for substrates with beaks and two for substrates without beaks).
Results are similar to those obtained with TAST type specimens. Other experimen-
tal results are presented in the next section; more results can be found in [20].
15.4.2 Nonassociated Elastoplastic Model for Monotonic Loadings

It has been shown that the stress distribution is not uniform in the adhesive; thus,
inverse identification techniques have to be used in order to analyze the experimental
results [21]. Using the proportional tensile-shear results (Figure 15.10a and b), for a
given loading rate, the elastic yield surface is shown in the peel stress (tn)–shear stress
(ts) diagram. Figure 15.10c also presents the fracture envelope. Moreover, it can be
seen in Figure 15.10a and b that the inelastic deformation is much larger in the tan-
gential direction than in the normal one; thus, nonassociated models have to be used.
For such specimens with low stress concentrations, the experimental response
can be accurately represented using the combination of an elastoplastic model and
interface elements [21]; for other assemblies, the use of various elements in the joint
thickness is necessary [7,10]. With interface elements, the bonded joint is modeled
by a surface Γ (with normal denoted by n) between the two substrates. This type
of model can be justified when the thickness of the joint is not too large and by the
(a) (c)
50
20
γ = 90˚
γ = 45˚ 40
15
γ = 37.5˚ 30
FT (kN)
10 γ = 30˚
20
5 10
0 0
Peel stress, tn (MPa)

0.00 0.10 0.20 0.30 0.40 0 10 20 30 40 50 60 70
DT (mm) –10
–20
(b)
γ = 0˚ –30
30
–40
γ = 30˚
20 –50 Yield surface
FN (kN)
γ = 37.5˚ (experimental)
γ = 45˚ Yield surface
–60
(tensile)
10 Yield surface
–70 (compression)
Failure
0 –80 (experimental)
0.00 0.01 0.02 0.03 0.04 0.05 Flow function
DN (mm) –90 (model)
Shear stress, ts (MPa)
FIGURE 15.10 Comparison between experimental and finite element analysis Arcan results
for different radial loadings (γ = 0°, 30°, 37.5°, 45°, 90°). (a) Behavior in the tangential direc-
tion, (b) behavior in the normal direction, and (c) yield surface and flow rules (Equation 15.7).
DN and DT denote the relative displacements of the two adhesive–substrate interfaces, respec-
tively, in the normal and tangential directions of the midplane of the adhesive. FN and FT rep-
resent the components of the applied load in the normal and tangential directions, respectively.
observation of a homogeneous deformation in the thickness of the adhesive joint. For

such elements (Figure 15.11), a constitutive law is written as a relation between the
interface tractions T and the displacement jump [u] across the joint (relative displace-
ment of Γ+ and Γ–):
[u] = u(Γ+) − u(Γ−) = un n + us s and T = tn n + ts s (15.4)
n n t ts Г+
Г+ n
us
un s s
Г– Г– ts tn
FIGURE 15.11 Variables for a 2D interface element.

Within the framework of continuum thermodynamics, a constitutive relation is

associated with the choice of a free energy and a dissipation potential [26]. The
additivity of elastic [u]e and irreversible [u]p relative displacements is assumed, that
is, [u] = [u]e + [u]p. Thus, for isotropic hardening, denoted by p, the plastic cumulative
displacement, the state equations are
 t   k 0  u e  ue
[T ] = [ K e ][u]e :  n  =  n   n  = [ K e ]  n  and R = Ri + Ap (15.5)
 ts   0 ks   use   use 
where Ke is the elastic matrix, and kn and ks, respectively, are the tensile modulus
and the shear modulus whose dimension is [F/L3] (F and L denote the force and the
length, respectively). A linear isotropic hardening is used (Ri and A are constants).
The yield surface (Figure 15.10c) was defined with two elliptic functions taking
into account the difference between traction and compression (or the influence of the
hydrostatic pressure), using material parameters, a+, a−, and b:
1/ 2
<tn > 2+ <tn > 2 ts2
f (T , R) = T R= + + 2 R. (15.6)
a+2 a2 b
<x>+ and <x>− are, respectively, the positive and negative parts of the real x. In
order to represent the large differences between the tensile and shear deformations
of the adhesive, a nonassociated model is proposed using a second elliptic function to
define the flow rules using two material parameters c and d (Figure 15.10c):
1/ 2
 t2 t2 
q(T , R) =  n2 + s2  − R. (15.7)
c d 
Thus, the evolution laws were defined as
∂q ∂q
[u ] p = λ and − p = λ = − λ . (15.8)
∂T ∂R
λ is the inelastic multiplier. It is important to note that the different parameters are
not independent, and thus, parameters b and d can be chosen equal to 1. Therefore,
this nonassociated elastoplastic model is defined using seven parameters (Table 15.1).
TABLE 15.1
Material Parameters of the Adhesive Nonassociated Elastoplastic
Model
kn (N/mm3) ks (N/mm3) a+ a− c Ri (MPa) A (N/mm3)
5575 1015 2.05 4 3.3 22 14
This model was implemented in the finite element code CAST3M using a return
mapping type algorithm [21].
Experimental results for γ = 0°, 30°, 37.5°, 45°, and 90° were used for the inverse
identification procedure, and Table 15.1 gives the values of material parameters.
Figure 15.10a and b presents, for the different radial loadings, comparison between
experimental data and results of finite element simulations. The results of the pro-
posed model are in good agreement with the experimental data. The flow rule func-
tion chosen allows the large ratios between the inelastic displacements in the normal
and tangential directions to be accurately represented. Figure 15.12 shows the results
for the tensile-shear test at γ = 45°. Figure 15.12a and b presents the experimental
and numerical load-displacement curves. Figure 15.12c and d represents the evolu-
tion of the adhesive shear and peel stresses along half of the overlap for increasing
stages of loading (steps: i, j, k, and l are defined in Figure 15.12a and b). After the
known nonuniform stress distribution in the elastic domain (i), the stresses became
nearly uniform for large inelastic displacements (before failure) (l). In the transition
part (j–k), large stress redistributions are observed associated with the development
of plasticity in the adhesive. The extension of this model has been proposed to take
(a) (b)
20 20
Experimental Experimental
k l k l
15 15
Numerical Numerical
j j
FN (kN)
FT (kN)
i i
10 10
5 5
0 0
0.00 0.05 0.10 0.15 0.20 0.25 0.00 0.01 0.02 0.03 0.04
DT (mm) DN (mm)
(c) (d)
25 l 25
l
k
Shear stress (MPa)
Peel stress (MPa)
20 20
k
15 15
j
10 10 j
i
5 5
i
0 0
0 10 20 30 0 10 20 30
x (mm) x (mm)
FIGURE 15.12 Comparison between experimental and numerical results: Arcan results for
tensile-shear loading (γ = 45°) at a displacement rate of 0.5 mm/min, and evolution of the
stress distribution in the adhesive along half of the overlap length. (a) Load-displacement
diagram in the tangential direction, (b) load-displacement diagram in the normal direction,
(c) evolution of the shear stress, and (d) evolution of the peel stress.
into account the influence of the strain rate on the mechanical behavior of the joint
(viscous effects) [21].
15.5 TOWARD 3D MODELING OF THE NONLINEAR

BEHAVIOR OF THE ADHESIVE
Various pressure-dependent constitutive models can be found in the literature [6,16],
but only a few experimental results are available to accurately represent the initial
yield surface. First, the influence of negative hydrostatic stress on the elastic limit is
presented using results from shear tests in a pressure vessel. In the second part, the
model is used to analyze the influence of geometric parameters and loadings on the
transmitted load for an industrial type structure.
15.5.1 Influence of Hydrostatic Stress on Adhesive Elastic Limit

The CHEM system [27] was designed at the IFREMER Brest Centre in order to
study the influence of hydrostatic stress on the mechanical behavior of materials.
This pressure vessel allows pressures up to 100 MPa to be applied during mechani-
cal testing (Figure 15.13). A view-glass can be used to see inside the pressure vessel
up to a pressure of 60 MPa, which allows the use of a noncontact measurement
system based on image correlation (Figure 15.13a). After positioning the specimen
in the tensile testing machine, the vessel is closed and is filled with tap water. The
required pressure is applied using a special pump; while the pressure increases, a
specially designed two-chamber equilibrium system ensures no other preloading
of the specimen. Then, a mechanical test, imposed by the tensile testing machine,
can be performed at a given pressure in the vessel. Thus, this system allows the
mechanical behavior of adhesives to be studied under quite large negative hydro-
static stresses.
(a) (b)
Pressure
vessel
View-glass
Modified
Digital TAST
camera
FIGURE 15.13 CHEM system description. (a) CHEM system with a digital camera and
(b) positioning of the modified TAST inside the pressure vessel.
60
Modified
TAST
von Mises stress (MPa)

Arcan
45
30
Model
15
0
–90 –75 –60 –45 –30 –15 0 15 30
Hydrostatic stress (MPa)
FIGURE 15.14 Elastic limit in the von Mises equivalent stress–hydrostatic stress diagram
for different tensile-shear tests and the proposed identified yield function (Equation 15.9).
TABLE 15.2
Material Parameters for the Initial Yield
Surface (Defined by Equation 15.9)
m (SI Units) l pt0 (MPa)
1. E-6 4.87 31.7
Figure 15.14 shows the experimental elastic limit in the von Mises equivalent
stress (σvm)–hydrostatic stress (ph) diagram. Results were obtained under mechani-
cal shear loadings with the modified TAST specimen for four hydrostatic pressures
(0, 30, 60, and 90 MPa). Moreover, the results obtained with the modified Arcan
device are also plotted in this diagram. An exponential Drucker–Prager yield func-
tion (F0 = 0) allows a good representation of the experimental data for the initial
so-called elastic limit [28]:
F0 = m(σvm)l + ph – pt0 = 0 (15.9)
m, l, and pt0 are material parameters. The results of the identification are presented
in Table 15.2, and the initial yield surface is drawn in Figure 15.14.
15.5.2 Influence of Edge Effects on Industrial Assembly

The aim of this section is to analyze the influence of the local geometry close to
the free edges of the adhesive on the performance of an industrial bonded assem-
bly. Including the yield function presented previously in a finite element simulation
allows the maximum load transmitted by the assembly to be determined. This is a
first approach, as simplified geometries and elastic material behavior of the different
parts are assumed.
(a) (b) (c)
Load
Load
(4)
(2) (3)
(4)
(3)
(2’)
(1) (1)
1 - support; 2 - mast; 3 - rail; 4 - trolley
FIGURE 15.15 Representation of the problem studied. (a) Industrial application: mainsail
and rail of a racing boat, (b) the experimental device with a part of the composite mast (2), and
(c) experimental device using an intermediate aluminum support (2’) to replace the composite
mast part.
The starting point of this study is the characterization of the bonding of a rail
(used to guide the mainsail) to a composite mast on a racing yacht (Figure 15.15a).
For this application, the loads imposed by the sail lead to a quite complex loading
of the adhesive (bending and shear). This application can be experimentally studied
using a specially designed fixture presented in Figure 15.15b. In order to analyze the
influence of geometric effects, it is possible, in a first step, to replace the composite
mast part (2, Figure 15.15b) by an intermediate aluminum support (2’, Figure 15.15c).
A complementary problem is to optimize the design of the composite based in par-
ticular on resistance to delamination under out-of-plane loadings [27].
Four different geometries have been numerically analyzed (Figure 15.16) with
a joint thickness of 0.4 mm. The two main geometric parameters are h1 = 30 mm
and h2 = 28 mm; the angle α defines the direction of the load with respect to the
α Loading
h2
(b) (c)
h1 “Rail”
Adhesive
Support
(a) (d) (e)
FIGURE 15.16 Model used. (a) 2D representation of the problem, (b) geometry A: large
support, straight “rail” and straight adhesive edges, (c) geometry B: large support, straight
“rail” and clean free edges of the adhesive, (d) geometry C: straight support, straight “rail”
and clean free edges of the adhesive, and (e) geometry D: straight support, “rail” with beaks
and clean free edges of the adhesive.
normal direction of the bonded joint. For this application, the load is characterized
by tension-shear and bending; the influences of edge effects are different from those
for Arcan tests with only tension-shear loadings. Therefore, such assemblies require
a specific study.
Figure 15.17a presents results for the four geometries for a loading associated
with an angle α = 45° and for a normalized load, that is, F/S = 1 MPa, where S is
the area of the joint, and the von Mises equivalent stress–hydrostatic stress diagram
showing the stress state in the adhesive. These simplified computations were made in
2D without taking into account the exact complex geometries of the rail and trolley;
thus, an overestimation of the transmitted load is obtained, but results give important
information about the behavior of this bonded assembly. The hydrostatic stress and
the von Mises stress are computed for each integration point. In Figure 15.17, only
the envelope of these points is drawn to simplify the presentation (all the points are
therefore inside the envelope). Using the linear property of elastic problems, we can
determine the maximum transmitted load by the bonded assembly Fmaxi for a rail
length of 100 mm (Table 15.3). Figure 15.17b presents the von Mises equivalent
stress–hydrostatic stress diagram with the stress in the adhesive for the maximum
transmitted load, under elastic behavior, and for the four geometries. It can be noted
that the envelopes are similar for the different geometries. For the load used, the
more highly stressed part of the adhesive is loaded under tensile shear. It can be
noted that a reduction of the edge effects (Figure 15.4) leads to an increase in the
transmitted load.
As presented in Figure 15.17b, for the maximum load transmitted and for a given
assembly, only some of the characteristic points reach the elastic limit in the von
Mises equivalent stress–hydrostatic stress diagram; for such points, F0 = 0, and
thus, the distance between this point and the elastic limit is equal to 0. Therefore,
for the other points, the distance from the elastic limit in the von Mises equivalent
(a) (b)
Geometry A 9 40
Geometry B A
8
Geometry C 35
Geometry D 7 Geometry A 30
6 Geometry B
Geometry C 25
5 Geometry D
4
B Elastic limit 20
C 15
3
D 10
2
1 5
0 0
–4 –3 –2 –1 0 1 2 3 4 5 –40 –30 –20 –10 0 10 20 30 40
Hydrostatic stress (MPa) Hydrostatic stress (MPa)
FIGURE 15.17 Envelope of the stress state in the adhesive for the four geometries (see
Figure 15.16) in the von Mises equivalent stress–hydrostatic stress diagram for a loading
associated with an angle α = 45°. (a) Stress state for an average tensile stress of 1 MPa and
(b) stress state for the maximum transmitted load.
TABLE 15.3
Maximum Load Transmitted for a Rail Length of
100 mm by Different Geometries (see Figure 15.16)
for a Loading Associated with Angle α = 45°
Geometry A B C D
Fmaxi (kN) 11.8 19.8 27.0 37.4
stress–hydrostatic stress can be computed using the properties of proportional load-

ings: it is determined in the direction of the stress path associated with an increase
in the load. Figure 15.18a presents for the four geometries and for the maximum
transmitted load the evolution of the minimum values (throughout the joint thick-
ness) of the distance of a characteristic point from the elastic surface in the adhesive
thickness, plotted on the von Mises equivalent stress–hydrostatic stress diagram,
with respect to the overlap length. It can be noted that in the middle of the overlap
(x = 0 mm), the minimum distance from the yield surface for the four geometries can
be classified in a decreasing way as A, B, C, and D. In fact, naturally, a decrease in
the stress singularity leads to an increase in the maximum transmitted load by the
assembly, which can be seen as a decrease in this so-called average distance from the
yield surface over the overlap length (the optimum value is the value 0, which cannot
be obtained for the whole bonded area). Moreover, it can be noticed that the elastic
limit is not reached at one free edge of the joint only for geometry D, in contrast to
the other geometries (associated with quite large edge effects). The bending load
leads to a maximum stress state in the left part of the adhesive. An increase in the
stress concentration leads to an increase in the stress close to the adhesive edges and
thus limits the transmitted load. Figure 15.18b presents the minimum and maximum
(a) (b)
Geometry A 35 40
Geometry B Maximum
Geometry C 30 35
Minimum
Geometry D 30
25
Distance (MPa)
Distance (MPa)
25
20
20
15
15
10
10
5 5
0 0
–20 –10 0 10 20 –20 –10 0 10 20
x (mm) x (mm)
FIGURE 15.18 Distance from the elastic surface in the von Mises equivalent stress–
hydrostatic stress diagram with respect to the overlap joint. (a) Minimum values of the dis-
tance for the four geometries and (b) minimum and maximum values of the distance for
geometry D.
values of the distance for geometry D; it can be seen that the stress state in the joint
thickness is nearly constant for a given position x in the overlap length. Close to the
adhesive free edges, an evolution of the stress state occurs in the joint thickness.
Figure 15.19 presents the influence of the loading angle α on the maximum trans-
mitted load (Fmaxi) for a rail length of 100 mm. It can be noted that an increase
in the angle α leads to a decrease in the transmitted load for all four geometries
(Figure 15.19a). In fact, an increase in bending leads to an increase in the normal
stress at the left end of the joint. Figure 15.19b shows the envelope of the stress state
in the joint for three values of the angle α. In order to make analysis easier, this figure
is drawn for a normalized load, that is, F/S = 1 MPa, where S is the area of the joint.
For α = 0°, nearly a pure tensile loading state is observed (positive hydrostatic stress),
and for an angle α = 45°, both compression loading (negative hydrostatic stress) and
tension loading (positive hydrostatic stress) associated with the bending loading are
observed combined with shear loading. Thus, as the stress state in the adhesive can
be complex, the use of a simplified method, which generally cannot analyze large
stress concentrations, for dimensioning industrial type bonded joints can overesti-
mate the maximum load transmitted by the bonded assembly [10].
Figure 15.20 presents the transmitted load (at failure) for the first experimental
tests for three different geometries with an intermediate aluminum support (B, C,
and D). For each geometry, four tests have been performed in order to analyze the
results’ variability. Significant differences are obtained for the different configura-
tions. The important point is that experimental results are less scattered for geometry
D, associated with low edge effects. Moreover, it is important to notice that the rail
and trolley geometries are quite complex and can lead to more complex stress states
than the simplified model used for the computations. A detailed analysis of this
test requires comparison between experimental and numerical results taking into
(a) (b)
120 Geometry D 6
Maximum transmitted load Fmaxi (kN)
Geometry C α = 45˚
100 Geometry B 5
Geometry A
80 4
60 3
α = 25˚
40 2
1 α = 0˚
20
0 0
0 10 20 30 40 50 –2 –1 0 1 2 3 4
α (˚) Hydrostatic stress (MPa)
FIGURE 15.19 Influence of the loading angle α on the maximum transmitted load.
(a) Maximum transmitted load for a rail length of 100 mm and (b) envelope of the stress state
in the adhesive in the von Mises equivalent stress–hydrostatic stress diagram for a normal-
ized load.
35
Transmitted load F (kN)

30
25
20
15
10
5
0
B C D
Geometry
FIGURE 15.20 Experimental results for three geometries (see Figure 15.16); results are
presented for four specimens per geometry and α = 30°.
account the nonlinear behavior of the adhesive; it is an interesting example to vali-

date the numerical tools that are under development in order to optimize adhesively
bonded structures.
15.6 CONCLUSION
In order to analyze the behavior of an adhesive in an assembly and to obtain reliable
data, experimental fixtures that strongly limit edge effects must be devised. The
properties of such a test for shear loading have been presented. Moreover, a decrease
in the influence of edge effects leads to a decrease in the influence of defects on
the experimental results. On the contrary, the use of experimental tests character-
ized by quite large stress concentrations can lead to inappropriate analyses of the
experimental results due to crack initiation. Stress singularities in adhesively bonded
joints, which depend mainly on the relative elastic properties of the two materials
and the geometry of the substrates, can be analyzed using either asymptotic analy-
sis or precise finite element models. The numerical determination of how stresses
evolve throughout the thickness of the adhesive is essential for optimizing adhe-
sively bonded assemblies. Moreover, it has been shown that using a 2D nonasso-
ciated elastoplastic model, it is possible to accurately represent the behavior of a
ductile adhesive in an assembly. But, as the strain evolution is strongly dependent on
the loading path, a large database of experimental results under various tensile-shear
loadings is required in order to accurately represent the adhesive behavior. In addi-
tion, to develop precise 3D pressure-dependent constitutive models, it is important
to have experimental results under quite large negative hydrostatic stress, especially
to describe the initial yield surface. Such a model has been used to analyze the influ-
ence of various parameters on the transmitted load in an industrial assembly. The
results underline the possibility of such a model for optimizing complex bonded
structures, especially in order to evaluate the influence of edge effects. It is important
to emphasize that the simplified methods frequently used for dimensioning bonded
joints, which generally cannot analyze large stress concentrations, can overestimate
the maximum transmitted load in the assembly.
To fully represent the experimental results, the proposed model must be further
enhanced to describe the various viscous effects observed under cyclic, creep, and
relaxation loadings.
REFERENCES
1. R. D. Adams (Ed.), Adhesive Bonding: Science, Technology and Applications. Woodhead
Publishing Limited, England (2005).
2. L. F. M. da Silva and A. Öchsner (Eds.), Modeling of Adhesive Bonded Joints. Springer,
Heidelberg (2008).
3. D. Leguillon and E. Sanchez-Palancia, Computation of Singular Solutions in Elliptic
Problems and Elasticity. Editions Masson, Paris (1987).
4. C. H. Wang and L. R. F. Rose, Int. J. Adhesion Adhesives 20, 145–154 (2000).
5. F. Kadioglu, L. F. Vaughn, F. J. Guild and R. D. Adams, J. Adhesion 78, 355–381 (2002).
6. G. Dean, L. Crocker, B. Read and L. Wright, Int. J. Adhesion Adhesives 24, 295–306
(2004).
7. J. Y. Cognard, R. Créac’hcadec, L. Sohier and P. Davies, Int. J. Adhesion Adhesives 28,
393–404 (2008).
8. J. Y. Cognard, Comput. Struct. 86, 1704–1717 (2008).
9. L. F. M. da Silva, P. J. C. das Neves, R. D. Adams and J. K. Spelt, Int. J. Adhesion
Adhesives 29, 319–330 (2009).
10. J. Y. Cognard, H. Devaux and L. Sohier, Int. J. Adhesion Adhesives 30, 706–719 (2010).
11. R. D. Adams and J. A. Harris, Int. J. Adhesion Adhesives 7, 69–80 (1987).
12. G. Belingardi, L. Goglio and A. Tarditi, Int. J. Adhesion Adhesives 22, 273–282 (2002).
13. A. R. Rispler, L. Tong, G. P. Steven and M. R. Wisnom, Int. J. Adhesion Adhesives 20,
221–231 (2000).
14. J. Wang, A. N. Rider, M. Heller and R. Kaye, Int. J. Adhesion Adhesives 25, 410–426
(2005).
15. J. Wang and C. Zhang, Int. J. Adhesion Adhesives 29, 495–502 (2009).
16. J. Y. Cognard, R. Créac’hcadec, L. Sohier and D. Leguillon, Int. J. Adhesion Adhesives
30, 257–266 (2010).
17. R. S. Raghava and R. M. Cadell, J. Mater. Sci. 8, 225–232 (1973).
18. J. Y. Cognard and R. Créac’hcadec, J. Adhesion Sci. Technol. 23, 1333–1355 (2009).
19. CAST3M documentation, www.cast3m.cea.fr/cast3m.
20. J. Y. Cognard, P. Davies, L. Sohier and R. Créac’hcadec, Compos. Struct. 76, 34–46
(2006).
21. R. Créac’hcadec and J. Y. Cognard, J. Adhesion 85, 239–260 (2009).
22. W. Althof, J. Reinf. Plast. Compos. 1, 29–38 (1982).
23. Z. Q. Qian, Comput. Struct. 79, 53–64 (2001).
24. Z. Wu, Compos. Struct. 65, 339–345 (2004).
25. E. D. Reedy and T. R. Guess, Int. J. Solids Struct. 30, 2929–2936 (1993).
26. J. Lemaitre and J. L. Chaboche, Mechanics of Solid Materials. Cambridge University
Press, Cambridge, (1994). ISBN: 0521477581.
27. J. Y. Cognard, P. Davies and L. Sohier, Advances in Structural Adhesive Bonding, Ed.
D. A. Dillard. Woodhead Publishing, Cambridge, pp. 437–465 (2010).
28. J. Y. Cognard, R. Créac’hcadec, J. Maurice, P. Davies, M. Peleau and L. F. M. da Silva,
J. Adhesion Sci. Technol. 24, 1977–1994 (2010).
16 Soy Proteins as
Wood Adhesives
Charles R. Frihart,* Christopher G. Hunt,
and Michael J. Birkeland
CONTENTS
16.1 Introduction................................................................................................... 278
16.2 Demands of an Industrial-Scale Wood Adhesive.......................................... 278
16.3 Protein Fundamentals....................................................................................280
16.3.1 Changes in Protein Adhesive Technology.........................................280
16.3.2 Protein Structure................................................................................280
16.3.3 Protein Dispersion and Modification................................................. 282
16.3.4 Current Soy Protein Adhesives.......................................................... 282
16.3.5 Protein Adhesive Cross-Linking Agents...........................................284
16.4 Properties of Soy Adhesives.......................................................................... 285
16.4.1 Rheology............................................................................................ 285
16.4.2 Bond Strength.................................................................................... 286
16.4.3 Environmental Impact....................................................................... 287
16.4.4 Opportunities..................................................................................... 288
16.5 Conclusions.................................................................................................... 288
References............................................................................................................... 289
Protein adhesives allowed the development of bonded wood products such as

plywood and glulam in the early 20th century. Petrochemical-based adhesives
replaced proteins in most wood bonding applications because of lower cost,
improved production efficiencies, and enhanced durability. However, several
technological and environmental factors have led to a resurgence of proteins,
especially soy flour, as an important adhesive for interior nonstructural wood
products. This paper discusses important aspects of protein structure and
recent successful advances in higher performance soy flour adhesives for wood
bonding. Despite these advances, we believe there is even greater potential for
protein adhesives if the reactive groups can be better utilized for bonding and
more efficiently cross linked. Protein wood adhesives have recently displaced
fossil fuel–based adhesives in some markets and have the potential to replace a
significant percentage of fossil fuel–based wood adhesives worldwide.
* Dr. Charles R. Frihart was the lead author for this chapter and can be contacted at cfrihart@fs.fed.us.
277
16.1 INTRODUCTION
The use of protein-based adhesives has a long history that includes examples of
adhesive systems based on several protein sources, such as animal protein, casein
from milk, soy flour, and blood [1]. Animal protein–based hot melt glues were used
in many applications, including furniture manufacturing, because they set rapidly
[2]. Caseins precipitated from skim milk under acidic conditions were used in the
development of the glulam industry [3]. Soy flour adhesives were the dominant glue
in plywood well into the 20th century [1]. Blood was used mainly with other pro-
tein glues to provide improved moisture resistance [4]. The ability of a wide variety
of proteins to serve as wood-bonding adhesives indicates a good natural affinity
between proteins and wood. The main limitation of protein adhesives is that syn-
thetic, fossil fuel–based adhesives tend to be more durable than protein adhesives.
For a widespread resurgence in the use of protein adhesives, new technologies
needed to be developed.
Because of the low cost and wide availability of soy flour, soy adhesives have
the greatest potential for widespread use in the wood products industry. The origi-
nal soy flour adhesives used in plywood were made in plywood plants under highly
alkaline conditions. They typically had a high viscosity, a low solids content, and
a usable pot life of only a few hours [1]. This plywood was adequate for interior
applications. However, fossil fuel–based adhesives, such as urea-formaldehyde
(UF), replaced soy-based adhesives for interior applications because UF adhesives
were cheaper, easier to use, and more water resistant. In recent years, develop-
ments in protein adhesive technology [5,6], volatility in petroleum prices, and more
stringent regulatory limits on formaldehyde emissions from composite wood pan-
els [7,8] have created a renewed commercial interest in soy adhesive technology.
Additionally, these developments have closed the gap in both performance and
pricing differences between fossil fuel–based adhesives and soy-based adhesives.
Despite these factors, there are still many obstacles in the path to greater usage
of soy-based adhesives in composite wood panels. The focus of this chapter is to
review these challenges and discuss some of the fundamentals of protein-based
adhesives, focusing mainly on soy-based adhesives, to provide an understanding of
current technology and a background for future improvements to protein adhesive
technologies.
16.2 DEMANDS OF AN INDUSTRIAL-SCALE WOOD ADHESIVE

For a new adhesive to be successful on a large scale, it must meet certain minimum
requirements. First, it must be relatively inexpensive and abundant and have consis-
tent properties. Second, given the great expense associated with changing process
designs and equipment capabilities in the wood products industry, a new adhesive
should preferably work with current process equipment and conditions.
One very important aspect of these demands is the adhesive’s fluid properties.
Adhesives are typically pumped through pipes and applied to wood by a variety of
processes, including atomization, shear blending, and roll or curtain coating. Any
Soy Proteins as Wood Adhesives 279
new adhesive technologies must be produced with a suitable viscosity needed for
them to flow through existing equipment for a specific process.
Another highly important process parameter in wood product manufacturing is
moisture content. Adhesives must not greatly raise the moisture content of the wood
product during processing or in the final product. Current composite products are
heat-cured, and excess water in the processing of a product results in problems in
production speed or quality. In end products, changes to moisture content can alter
the working properties of the panel, such as machining, or lead to problems with
dimensional stability.
The final requirement for a new adhesive is to provide finished products with
sufficient strength and stiffness properties as defined by the appropriate industry
standards. In the wood products industry, there are three main classes of materials:
interior nonstructural, exterior nonstructural, and structural. Each class has its own
set of standards and testing requirements, but in general, all three classes require
some level of both dry and wet strengths. Interior products are judged mostly by dry
strength and stiffness properties; however, wet strength and dimensional stability are
also used to gauge product quality. Adhesives for exterior products must have great
water resistance, while structural adhesives must be resistant to water, extreme heat
(fire), and creep. An adhesive that promotes mildew, mold, or insect infestations is
not acceptable, so these issues must also be considered.
In recent years, low formaldehyde emissions have become a major perfor-
mance criterion for interior composite wood products in North America, Europe,
and Japan. With the change in the classification of formaldehyde from a suspected
to a known carcinogen, the California Air Resources Board (CARB) was required
to consider setting new standards for formaldehyde emissions [7,9]. However,
CARB could not act unless technology was available that could meet a new stan-
dard. A soy-based adhesive system was one way to meet the new standard and had
been demonstrated to be commercially viable for decorative (interior) plywood.
CARB ultimately set regulations that apply to decorative plywood, particleboard,
fiberboard, and products made using these materials [9], and recently, a national
standard was passed based on the CARB limits for formaldehyde emission in
wood products [8].
The low formaldehyde emissions requirement has spurred commercial interest in
protein-based adhesive technology, especially as a replacement for UF, which may
have difficulty achieving consistent compliance with CARB formaldehyde emis-
sions limits in all composite wood panels. Although protein-based adhesives have
continued to be used in several niche applications for wood products, now soy-based
adhesives are again being used at increasing levels in the interior plywood industry,
even though the processes and properties of wood products have been designed and
defined for use with synthetic adhesives. Synthetic adhesives have high productiv-
ity in composite or engineered wood product manufacturing processes because of
their design; thus, natural adhesives need to perform equally well. For the protein
adhesives to effectively displace the synthetic adhesives, they need to be of low cost
and in large enough supply, to fit with the current manufacturing processes, and to
produce products that meet all the relevant specifications.
16.3 PROTEIN FUNDAMENTALS

Proteins have long been used for bonding wood because of their strong adhesion to
wood surfaces. Proteins have many of the characteristics of good adhesives, such as
having surfactant character, both polar and nonpolar domains, and good cohesive
strength. Surfactant character means that proteins can wet surfaces well. Once the
surface has been effectively wetted, the polar and nonpolar domains can interact
with the polysaccharide and lignin domains, respectively, in wood. The long history
of proteins used as adhesives, as well as their ability to form gels, demonstrates that
proteins also adhere well to themselves.
16.3.1 Changes in Protein Adhesive Technology

Compared to the early 1900s, when proteins were used widely in wood products,
the knowledge about protein structure and properties has advanced tremendously.
The driving force has been to better understand the biological function of proteins
(enzymatic processes, pharmaceutical interactions, cellular adhesion) and how they
behave in food products. For food applications, many properties of proteins have
been studied extensively, including the following [10,11]:
• Primary, secondary, tertiary, and quaternary structures

• Gel properties, especially under heating
• Foaming ability and water retention
• Interaction with other proteins, carbohydrates, and salts
• Thermal transitions for characterizing protein structure
Despite the tremendous advances and sophisticated analytical methods, many

aspects of proteins are still not well understood. For instance, the effect of the com-
mon denaturant urea is still debated [12]. One reason for the lack of progress in
research is that crude protein products are mixtures of proteins with different com-
positions, structures, and properties. Another reason is that the energy difference
between one folded conformation and another is often so small that changes in the
environment can affect the protein’s tertiary conformation [13]. Given the heteroge-
neity of most protein sources, understanding structure–property relationships on a
fundamental level as they relate to adhesion is a daunting task made more difficult
by the myriad of complex interactions and influences.
16.3.2 Protein Structure
Proteins are linear polypeptides made from α-amino acids, whose sequence is deter-
mined by the genetic code. Amino acids are distinguished by their side chains,
which, in addition to steric differences, can be categorized as hydrophobic or hydro-
philic. Depending on the pH of the medium as well as the immediate environment of
a given side chain, the hydrophilic residues can also exhibit various states of charge.
This backbone sequence is referred to as the primary structure (see Figure 16.1). As
the polypeptide is synthesized, certain sections have the proper sequence for forming
Secondary
Primary β-sheet
α-helix Polar and covalent bond

formation
Crystallite formation
Original polymer chain
Quaternary
Tertiary
Hydrophobic collapse
Subunit association
FIGURE 16.1 The formation of secondary and tertiary structures where each chain folds
and coils in specific ways into a relatively compact structure with a characteristic three-
dimensional shape. These folded chains can interact to form a quaternary structure.
α-helices or β-sheets. These sections constitute the secondary structure resulting

from hydrogen bonds joining the atoms of the peptides along the polypeptide chain
[10]. The α-helix involves a single chain of adjacent amino acids that require certain
sequences in order to form a stable helix. The β-sheets contain a series of adjacent
amino acids as well as segments from other parts of the same chain. The fraction
of a protein in helices and sheets can be determined by several methods, including
infrared spectroscopy and optical rotation [11].
Whereas most polymers depend primarily on interchain interactions to provide
structure, proteins use intrachain interactions extensively to form a tertiary structure.
Therefore, understanding the properties of protein begins with the intrachain proper-
ties and how the molten globular structure is influenced by the external environment.
The term “molten globule” has been used by protein scientists to reflect the ease
with which the tertiary structure rearranges as a result of changes in the surrounding
environment and the small energy differences between different folded conforma-
tions. In a normal aqueous environment, nonpolar amino acid side chains want to
minimize the amount of surface area exposed to water. This aversion results in a
“hydrophobic collapse,” with more of the nonpolar groups ending up on the inside
of the molten globule and more of the polar groups residing on the outside [13].
Intrachain stabilization is provided by van der Waals interactions of the hydrophobic
groups, hydrogen bonding of polar groups, ionic bonding between acids and bases,
and disulfide bonds between thiols.
Protein interchain interactions result in a quaternary structure, made up of two or
more kinds of protein subunits, and offer a significant level of additional structure
and stability. Interchain associations involve the same kinds of hydrophobic, polar,
ionic, and disulfide interactions as the secondary and tertiary structures. Although
it is energetically favorable to have all nonpolar side chains buried inside the glob-
ule and all polar groups on the outside, the backbone sequence does not allow this.
This is especially true for proteins like soy with a high proportion of hydrophobic
residues. As a result, proteins often have nonpolar groups on the surface. This leads
to interchain associations to reduce the interaction of surface hydrophobic groups
with water.
Most proteins serve a structural, enzymatic, or storage role in their native biologi-
cal system. Storage proteins (e.g., soy, casein, wheat gluten) [14] supply the amino
acids that are needed to make the structural or enzymatic proteins of the growing
organism and can also be used as adhesives with little modification. In contrast,
structural proteins (e.g., bone, hide, blood, skin) often need to be treated to make
them useable as adhesives [2,4,15]. Interestingly, there is no clear difference in wood
adhesion properties between storage and structural proteins, with both having been
used for wood bonding. However, there are substantial differences in their structures
and other properties [14]. In addition, protein adhesives are not equal in their wood
adhesion durability [1–4], the general order being animal, soy, casein, and blood,
from least to greatest water resistance.
16.3.3 Protein Dispersion and Modification

Proteins as they exist in living organisms are referred to as being in their native state.
In their native state, many proteins are water dispersible so that they can migrate
from their synthesis site to where they are used or stored. Either intentionally or
unintentionally, processing of the proteins can denature (change from native state)
them in a reversible or nonreversible fashion. Proteins may be intentionally altered to
make them more digestible; destroy enzymes that can cause food degradation; pre-
cipitate them (e.g., casein); or alter water retention properties, dispersibility, swelling
characteristics, or accessibility of functional groups. Common ways to denature pro-
teins are to change the pH, apply heat, and add denaturants such as urea.
Proteins can also be modified by chemical reactions, especially if cross-linking
is desired [16,17]. Often, primary amino groups, such as the side chain of a lysine
amino acid, are used as reactive sites. Amino groups can be alkylated, acylated,
hydroxymethylated, and so forth. These reactions can be done either to reduce water
solubility (such as a reaction with acetic anhydride) or increase water solubility (such
as a reaction with succinic anhydride). Reactions with aldehydes, such as formal-
dehyde, usually decrease the water solubility of proteins by causing gelation. Other
groups that are relatively easily modified include carboxylic acids, thiols, hydroxyls,
and a variety of nitrogen-containing groups.
16.3.4 Current Soy Protein Adhesives

Soybeans are a major crop in the United States and are noted for having high pro-
tein content. For adhesive applications, the most common raw material is soy flour
ground from defatted meal. Defatted meal is the residue after the oil has been pressed
or extracted from the dehulled soybean [18]. If mild conditions are used to process
the flour, it has a high protein dispersibility index (PDI). PDI is the ratio of protein
content in the supernatant to the total protein in the flour after dispersion of the
flour and centrifugation. Greater heat treatment gives a lower PDI, which is desirable
for a number of applications [19]. As shown in Figure 16.2, soy flour has slightly
more protein than carbohydrates. The carbohydrate contains a variety of insoluble
and soluble components [20]. Extraction with aqueous ethanol removes many of the
soluble carbohydrates to produce soy concentrate [18]. The purest commercial soy
protein product is the soy protein isolate (SPI), which has almost all of the insoluble
and soluble carbohydrates removed [18].
With over 90% protein content, SPI is the commercial soy product used in many
soy reaction studies. SPIs have been altered using a variety of denaturants, includ-
ing surfactants, amino-containing agents, alkali, and enzymes, in efforts to improve
adhesive properties. For example, addition of 1% of the anionic surfactant sodium
dodecyl sulfate (SDS) resulted in the maximum improvement of SPI adhesion to ply-
wood [21]; cationic surfactants had a similar effect [22]. Urea at 1 and 3 molar con-
centrations provided improved bond strength and water resistance for SPI over 0.5
and 8 molar concentrations [23]. Studies have also investigated the effect of alkali
and enzyme treatments [24–26]. Much of the recent work on modifying the perfor-
mance of SPI and soy protein fractions without curing chemicals has been discussed
in terms of mechanical and adhesive properties [27–29].
While soy contains a number of proteins, the two that account for most of the
mass are glycinin (11S) and conglycinin (7S) [30]. S is the Svedberg number, an
expression of the sedimentation rate, and larger numbers signify higher molecular
weights. These two proteins have distinctive differences in tertiary and quaternary
structures. The subunits in 11S are either acidic or basic and are held in their quater-
nary structure by acid–base interactions, disulfide bonds, and hydrophobic interac-
tions [30]. On the other hand, the 7S subunits are more neutral and are held together
mainly by hydrophobic interactions. Although the 7S and 11S proteins are very dif-
ferent structurally, this difference has not resulted in observation of any large differ-
ences in their wood-bonding properties [27].
Although most of the recent adhesive research has been with soy proteins, other
proteins have also been studied. The concentration on soy research in the United
• Whole soybeans
≈ 16–17¢/lb, 36% protein, 18% oil, 36% carbohydrates, 10% moisture
• Defatted meal
≈ 15¢/lb, 48% protein, 0% oil, 44% carbohydrates, <10% moisture
• Soy flour
≈ 18–25¢/lb, 50% protein, 0% oil, 40% carbohydrates, <10% moisture high (90%) to low
(20%) PDI (dispersible protein)
• Soy protein concentrate
≈ 50–90¢/lb, 65+% protein, 0% oil, up to 35% carbohydrates
• Soy protein isolate
≈ $1.50–2.00+/lb, 90+% protein, 0% oil, up to 10% carbohydrates
FIGURE 16.2 Soybean products with their 2010 prices and composition, with the flour and
isolate highlighted because they are the most commonly investigated products for adhesives.
States has been due to the availability of soy flour in large quantities at reason-
able prices, along with financial support of the United Soybean Board and similar
producer-based organizations. However, in other countries, especially in Europe,
soybeans are not a major crop. Thus, research in Europe tends to use wheat gluten
and other protein sources [31,32].
16.3.5 Protein Adhesive Cross-Linking Agents

Traditional methods of making protein adhesives have not provided as much water
resistance as their competing synthetic adhesives. Un–cross-linked soy proteins are
good adhesives for wood under dry conditions, but their water resistance is typically
poor. Older soy or casein formulations, made under alkaline conditions and using cal-
cium compounds, had good water resistance [1,3]. The problem is that none of these
methods makes products that can fully meet current performance standards, such as
ANSI/HPVA HP1-2009 4.6 three-cycle soak test for decorative plywood. Thus, in
the past dozen years, more emphasis has been placed upon covalent cross-linking.
Some recent research has emphasized using soy as a part of a phenol-formaldehyde
(PF) adhesive matrix in contrast to earlier work that treated soy as a filler. Several
programs attempted to keep the soy–PF viscosity at the same low level as typical PF
adhesives. This limited the level of soy incorporation [33–36]. However, viscosity
limits are more flexible for soy adhesives because soy is shear thinning, whereas PF
adhesives are Newtonian. Thus, a soy adhesive of higher apparent viscosity can per-
form just as well as a PF with a much lower viscosity with current process equipment
without significantly altering adhesive distribution on the wood. This has allowed
a higher incorporation of soy into soy–PF adhesives [36,37]. An interesting aspect
of these alkaline soy–PF adhesives is that they can be acidified to provide stable
dispersions [38,39]. Both the alkaline-dispersion and acid-dispersion soy–PFs have
shown good performance as adhesives for oriented strand board (OSB) face layers.
The acid-dispersion adhesives are lighter in color and do not give the caustic burn
spots in OSB that occur with normal PF formulations, while still giving good bond
durability. The stability of these dispersions is surprisingly good, as it is usually
difficult to neutralize alkaline PF products without precipitation of the PF unless
surfactants are used [40].
Formaldehyde is well known for cross-linking proteins [16,17], but there are con-
cerns about the safety of formaldehyde. This has led to the examination of other
types of aldehydes. Among the other aldehydes are glyoxal, glutaraldehyde, and
starch dialdehyde. Improved properties have been reported when using glutaralde-
hyde and glyoxal with soy flour or SPI [41–43].
The biggest commercial success in water-resistant soy adhesives has been achieved
with the addition of polyamidoamine-epichlorohydrin (PAE) resin. Kaichang Li
[5,6] developed the PAE cross-linked soy technology, and others have studied PAE–
soy interactions [44]. This technology has been licensed and is currently used by
several major North American plywood manufacturers. Soy–PAE resins have been
demonstrated to make products that meet the standards for interior (decorative) ply-
wood, engineered wood flooring, particleboard, and fiberboard [45–47]. It has also
been tested commercially on all these products. A main advantage of these products
is that, because they contain no added formaldehyde (NAF), the formaldehyde emis-
sions of these products are well below all of the standards [46–48]. The reaction
mechanism of PAE with soy proteins has been investigated [44], but many of the
details are still unknown.
Other technologies have also been developed to make soy-based adhesives that
have good water resistance. Liu and Li [49] have grafted dopamine groups onto SPIs
so that they could cure in a manner similar to that used by mussels. The modified soy
provided much better water resistance than the unmodified soy for bonding maple
plywood. SPI can be reacted with polyethylenimine (PEI) and maleic anhydride
(MA) to make an adhesive with suitable water durability for interior plywood bond-
ing [50]. The PEI and MA combinations can also be used with the more economical
soy flour for interior plywood bonding [51]. Cysteamine has been grafted onto SPI
by amide linkages to increase the mercapto content; this provides improved strength
and water resistance of plywood [52].
16.4 PROPERTIES OF SOY ADHESIVES

16.4.1 Rheology
Although proteins have good adhesion to wood, their rheological properties cause
complications. Most proteins make water dispersions of high apparent viscosities,
with both the words dispersions and apparent being important. Proteins, under most
conditions, do not make true solutions. Globular proteins have enough hydrophobic
domains to prevent them from forming a true solution. Even though many of the
hydrophobic groups are located inside the globule because of the hydrophobic col-
lapse, not all these groups can be buried [19]. Thus, there are hydrophobic patches
on the surface of folded protein chains (tertiary structure). To minimize the energy
involved, the individual proteins associate with each other in the hydrophobic regions
to form quaternary structures. The hydrophobic attraction of the chains is balanced
by the electrostatic repulsion because each of the chains is negatively charged above
a pH of about 4.5. Thus, understanding the colloidal properties of the protein is
important to understanding how these dispersions behave. The native tertiary and
quaternary structures of the protein can be disturbed by a variety of conditions such
as heat, pH changes, or added chemicals. The tendency of proteins to fold into glob-
ules, and the tendency of these globules to associate with one another, cannot be
avoided because of the hydrophobic side chains [19]. These disruptions often result
in permanent changes. Thus, heating soy flour causes the dispersibility of the protein
to decrease. For example, heating the 90 PDI flour results in a 20 PDI flour. It has
been determined that upon heating, the 11S and 7S subunits can separate and recom-
bine to form combinatorial versions of both 11S and 7S subunits [53].
The associations of proteins also result in the need to refer to apparent viscosi-
ties, because the shear conditions and history of the dispersions greatly influence
the measured viscosity. Protein dispersions are shear thinning (Figure 16.3); thus,
the apparent viscosity measured using a Brookfield viscometer with a low shear
rate does not represent the fluidity under the high-shear-rate environment found in
most mixing, pumping, and application equipment. Consequently, incorporating low
1.E+07
pH
1.E+06
Viscosity (Pa/s) 5.5
1.E+05
7.0
1.E+04
9.0
1.E+03
1.E+02
1.E+01
1.E–02 1.E–01 1.E+00 1.E+01 1.E+02 1.E+03
Frequency
FIGURE 16.3 Viscosity (Pa/s) vs. frequency for soy flour measured on a Paar Physica UDS
200 with a 25 mm serrated parallel plate [54], with the TS5525 using Prolia 100-90 flour from
Cargill at different pH values.
amounts of soy in a PF adhesive or the enzymatic degradation of the protein to obtain

a low Brookfield viscosity is unnecessary and often counterproductive. Considerable
education of wood adhesive users is necessary, because they are used to Newtonian-
type adhesives, in which high viscosity means difficulty in pumping and applying
the adhesive. Often, protein adhesives can be in the range of 2 to 75 Pa/s apparent
viscosity and still work in typical application equipment, with the lower end for
spray application and the higher end for roll coaters.
In spite of the shear-thinning nature of protein adhesives, workable viscosities
are usually found at lower solids content than with synthetic adhesives. This causes
a problem in many wood applications because of excessive steam pressure during
the making of composite panels or in having a bonded product at a low enough
moisture content to meet its standards. Most wood composite panels are produced by
hot pressing the wood–adhesive combination under heat. In products such as OSB,
particleboard, and fiberboard, this heating generates an internal steam pressure that
the adhesive bonds need to withstand to prevent blows in the composite when the
pressure is released. More water in the adhesive produces greater steam pressure and
requires high adhesion under wet conditions to withstand the internal forces.
16.4.2 Bond Strength
Protein adhesives have been used with wood because they give good bond strength
under dry conditions, but wet strength has been an issue. Some proteins, like casein
and blood, have been able to provide more water-resistant adhesive bonds than soy
proteins [1,4], but they have tended not to be as resistant as most synthetic adhesives.
Good synthetic adhesives have a curing mechanism that generates covalent bond
formations between the polymer chains. The newer soy adhesive technologies do
incorporate some type of covalent bond formations. Even though not all cured soy
products provide good water resistance, the soy–PF products [36–39] and soy–PAE
products [45–47,54] discussed above have been shown to produce wood products
using standard production techniques with sufficient water resistance to meet normal
product standards for the specific application.
Heat resistance has also been important in structural applications and has become
more of an issue in some of these areas. The ability to withstand load at elevated tem-
peratures has been important for many years for engineered wood products because
of the concern about creep, especially for products used in floor and roof supports.
In recent years, the ability of adhesives to hold up under fire conditions has become
important [55]. In the past, protein adhesives have shown good heat resistance at
higher temperatures; however, the ability of soy flour-based adhesives to withstand
more extreme heat treatment was unclear. Recent work has shown that properly for-
mulated soy flour adhesives maintain strength at elevated temperatures [56].
16.4.3 Environmental Impact
Although ultra-low-emitting urea-formaldehyde (ULE-UF) adhesives can meet the
current regulations, the soy–PAE products are much lower in formaldehyde emis-
sions using current tests [48]. Furthermore, unlike UF and ULE-UF formulations,
soy–PAE formulations are not susceptible to long-term emissions by hydrolysis
under conditions of increased heat and humidity (Figure 16.4a and b) [54,57].
(a) NAF ULEF-UF

60
CH2O emissions (µg/g)
50
40
30
20
10
0
0 20 40 60 80 100 120
% Relative humidity at 25°C
(b) NAF ULEF-UF

200
CH2O emissions (µg/g)
180
160
140
120
100
80
60
40
20
0
0 20 40 60 80 100 120
% Relative humidity at 35°C
FIGURE 16.4 (a) Formaldehyde (CH2O) emissions at various relative humidity levels
at 25°C (4-day samples) using modified EN-317 with commercially produced hardwood
plywood with NAF (soy–PAE) and ULE-UF (ultra-low-emitting urea-formaldehyde).
(b) Formaldehyde emissions at various relative humidity levels at 35°C (4-day samples) using
the same testing and hardwood plywood as in Figure 14.4a. Note different scale on y-axis.
Additional reasons for using soy proteins as wood adhesives are their strong wood-
bonding ability, their excellent heat resistance, a desire to be more environmentally
friendly, and the increasing price of fossil fuel. All of the proteins are derived from
natural, renewable resources. Most of the proteins are available in large quantities, and
even though the wood adhesive market is large, many of the proteins are available in
much larger quantities, making them different from tannins, which are in limited sup-
ply. Thus, soy proteins are environmentally friendly and offer long-term sustainability.
For a long time, the price of fossil fuels was decreasing compared with agricultural
feedstocks. Increasing competition for and limited supplies of fossil fuels coupled with
improvements in the agricultural yields have led to a shift in petroleum versus agricul-
tural economics. These trends are expected to continue in the future. The advantage of
soy protein is the availability of a very large supply of soy flour that can readily meet
the needs of even a large market, such as wood adhesives.
16.4.4 Opportunities
Although performance of soy adhesives has vastly improved, there is still a need for
better products to make them more competitive with synthetic adhesives. One area of
focus is the properties of soy itself. Generally, proteins in water at high concentrations
generate high viscosities and are very shear thinning. Making higher percent solids soy
formulations at equivalent or lower viscosities is an important goal in order to compete
more effectively with fossil fuel–based adhesives. This involves understanding what
contributes to the high viscosity and designing processes to reduce this effect. Perhaps
alterations of processes are available to control the viscosity through avoiding denatur-
ation in some cases or by selectively denaturing the proteins in others.
In soy flour, about half the material is carbohydrates. These are likely to be a
detriment to adhesive performance due to their hygroscopic nature. This portion is
relatively unstudied in adhesives and is simply assumed to contribute little to bond-
ing and/or rheological properties. Understanding the role of these carbohydrates may
provide further insight into the use of soy flour as a protein-based adhesive. Other
questions for consideration include the following: Can the carbohydrates be removed
from soy flour at low cost so that the product contains more of the higher-bond-
strength proteins? Can the carbohydrates in the flour be reacted so that they contrib-
ute to the strength of the adhesive network?
In general, proteins need to be cross-linked to provide good wet strength. Are
there types of chemicals, in addition to those already investigated, that can be incor-
porated to provide greater wet strength? Are there ways to denature the protein to
expose additional functional groups for increased cross-linking? Can soy protein
adhesive be made less water sensitive like the naturally water-resistant hair, colla-
gen, or mussel protein adhesives?
16.5 CONCLUSIONS
The bonded wood products were originally based on proteins from various sources,
but fossil fuel–based adhesives have generally displaced these protein adhesives over
the past 60 years. The need to make wood composite panels with lower formaldehyde
emissions for interior uses and the desire for bio-based adhesives have opened an
opportunity for a resurgence in protein adhesives. New technology for cross-linking
soy adhesives has shown the ability to make composite products that meet current
product specifications. Proteins differ from most other polymers in that intrachain
interactions play a very important role in influencing interchain interactions and
adhesion. To develop better protein-based adhesives, more of the fundamental struc-
ture–property relationships, as they relate to adhesion, must be understood. Although
the push toward green technology and some environmental factors have provided a
renewed level of commercial interest in protein adhesives, it is clear that to gain more
widespread use and acceptance, advances are still needed. This review hopefully has
provided some of the basic background information to spark the research necessary
to achieve the goal of using a renewable adhesive source with the lowest long-term
environmental impact.
ACKNOWLEDGMENTS
We thank the United Soybean Board (grant 0458) and Ashland Water Technologies
for their support of this work.
REFERENCES
1. A. L. Lambuth, in: Handbook of Adhesive Technology, A. Pizzi and K. L. Mittal (Eds.),
2nd Ed., pp. 457–478, Marcel Dekker, New York (2003).
2. J. R. Hubbard, in: Handbook of Adhesives, I. Skeist (Ed.), 2nd Ed., pp. 139–151, Van
Nostrand Reinhold Co., New York (1977).
3. H. K. Salzberg, in: Handbook of Adhesives, I. Skeist (Ed.), 2nd Ed., pp. 158–171, Van
4. A. L. Lambuth, in: Handbook of Adhesives, I. Skeist (Ed.), 2nd Ed., pp. 181–191, Van
5. K. Li, U.S. Patent 7,252,735 (2007).
6. K. Li, S. Peshkova and X. Gen, J. Am. Oil Chem. Soc. 81, 487 (2004).
7. ATCM, Airborne toxic control measure to reduce formaldehyde emissions from com-
posite wood products. Health and Safety Code: Title 17 California Code of Regulations,
Section 93120-93120.12. (2009).
8. U. S. Public Law No: 111-199, http://www.govtrack.us/congress/bill.xpd?bill=s111-1660.
9. J. R. Williams, in: Wood Adhesives 2009, C. R. Frihart, C. G. Hunt and R. J. Moon
(Eds.), pp. 12–16, Forest Products Society, South Lake Tahoe, CA (2010).
10. S. Damodaran, in: Food Proteins and Their Applications, S. Damodaran and A. Paraf
(Eds.), pp. 1–24, Marcel Dekker, New York (1997).
11. J. I. Boye, C.-Y. Ma and V. R. Harwalkar, in: Food Proteins and Their Applications, S.
Damodaran and A. Paraf (Eds.), pp. 25–56, Marcel Dekker, New York (1997).
12. D. R. Canchi, D. Paschek and A. E. Garcıa, J. Amer. Chem. Soc. 132, 2338 (2010).
13. R. H. Pain, Mechanisms of Protein Folding, 2nd Ed., Oxford University Press USA,
New York (2000).
14. S. Damodaran and A. Paraf (Eds.), Food Proteins and Their Applications, Marcel
Dekker, New York (1997).
15. R. E. Norland, in: Handbook of Adhesives, I. Skeist (Ed.), 2nd Ed., pp. 152–157, Van
16. G. E. Means and R. E. Feeney, J. Food Biochem. 22, 399 (1998).
17. J. Bjorksten, in: Advances in Protein Chemistry, M. L. Anson (Ed.), Vol. 6, pp. 343–381,
Academic Press, New York (1951).
18. X. S. Sun, in: Bio-Based Polymers and Composites, R. P. Wool and X. S. Sun (Eds.),
pp. 33–55, Elsevier-Academic Press, Burlington, MA (2005).
19. J. E. Kinsella, S. Damodaran and B. German, in: New Protein Foods, A. M. Altschul and
H. L. Wilcke (Eds.), Vol. 5, pp. 108–179, Academic Press, New York (1985).
20. E. M. Bainy, S. M. Tosh, M. Corredig, V. Poysa and L. Woodrow, Carbohydr. Polym. 72,
664 (2008).
21. W. Huang and X. Sun, J. Amer. Oil Chem. Soc. 77, 705 (2000).
22. Y. Wang, D. Wang and X. S. Sun, J. Amer. Oil Chem. Soc. 82, 357 (2005).
23. W. Huang and X. Sun, J. Amer. Oil Chem. Soc. 77, 101 (2000).
24. N. S. Hettiarachhy, U. Kalapathy and D. J. Myers, J. Amer. Oil Chem. Soc. 72, 1461
(1995).
25. M. A. Hanna, J. Amer. Oil Chem. Soc. 72, 507 (1995).
26. U. Kalapathy, N. S. Hettiarachhy, D. J. Myers and K. C. Rhee, J. Amer. Oil Chem. Soc.
73, 1063 (1996).
27. X. Mo, X. Sun and D. Wang, J. Amer. Oil Chem. Soc. 81, 395 (2004).
pp. 292–326 Elsevier-Academic Press, Burlington, MA (2005).
pp. 327–368, Elsevier-Academic Press, Burlington, MA (2005).
30. S. Utsumi, Y. Matsumura and T. Mori, in: Food Proteins and Their Applications, S.
Damodaran and A. Paraf (Eds.), pp. 257–291, Marcel Dekker, New York (1997).
31. S. Khosravi, F. Khabbaz, P. Nordqvist and M. Johansson, Ind. Crop. Prod. 32, 275
(2010).
32. P. Nordqvist, F. Khabbaz, E. Malmström, Int. J. Adhesion Adhesives 30, 72 (2009).
33. J. D. Clay, B. Vijayendran and J. Moon, in: Abstracts of Papers, pp. 1298–1301, SPE
ANTEC 1999 (1999).
34. C.-Y. Hse, F. Fu and B. S. Bryant, in: Wood Adhesives 2000, pp. 13–20, Forest Products
Society, Madison, WI (2001).
35. M. Kuo, D. J. Myers, H. Heemstra, D. Curry, D. O. Adams and D. D. Stokke, U.S. Patent
6,306,997 (2001).
36. J. M. Wescott, C. R. Frihart and A. E. Traska, J. Adhesion Sci. Technol. 20, 859 (2006).
37. L. Lorenz, C. Frihart and J. Wescott, J. Amer. Oil Chem. 84, 769 (2007).
38. J. M. Wescott, C. R. Frihart, A. E. Traska and L. Lorenz, in: Wood Adhesives 2005,
pp. 263–269, Forest Products Society, Madison, WI (2006).
39. J. M. Wescott and C. R. Frihart, U.S. Patent 7,345,136 (2008).
40. B. Broline, R. Adams and R. Schmidt, in: Proc. 36th International Wood Composites
Symposium, pp. 23–29, Washington State University, Pullman, WA (2002).
41. Y. Wang, X. Mo, X. S. Sun and D. Wang, J. Appl. Polym. Sci. 104, 130 (2007).
42. C. Marquie, J. Agric. Food Chem. 49, 4676 (2001).
43. G. A. Amaral-Labat, A. Pizzi, A. R. Gonçalves, A. Celzard, S. Rigolet and G. J. M.
Rocha, J. Appl. Polym. Sci. 108, 624 (2008).
44. Z. Zhong, X. S. Sun and D. Wang, J. Appl. Polym. Sci. 103, 2261 (2007).
45. M. Fong, J. M. Wescott and J. A. Yavorsky, in: Proceedings for the 9th Pacific Rim Bio-
Based Composites Symposium, pp. 233–237, Roturua, New Zealand (2008).
46. A. J. Allen, B. K. Spraul and J. M. Wescott, in: Wood Adhesives 2009, C. R. Frihart,
C. G. Hunt and R. J. Moon (Eds.), pp. 176–184, Forest Products Society, Madison, WI
(2010).
47. J. M. Wescott, M. J. Birkeland, J. Yarvoski and R. Brady, in: Wood Adhesives 2009,
C. R. Frihart, C. G. Hunt and R. J. Moon (Eds.), pp. 136–141, Forest Products Society,
Madison, WI (2010).
48. M. J. Birkeland, L. Lorenz, J. M. Wescott and C. R. Frihart, Holzforschung 64, 42

(2010).
49. Y. Liu and K. Li, Macromol. Rapid Commun. 23, 739 (2002).
50. Y. Liu and K. Li, Int. J. Adhesion Adhesives 27, 59 (2007).
51. J. Huang and K. Li, J. Am. Oil Chem. Soc. 85, 63 (2008).
52. Y. Liu and K. Li, Macromol. Rapid Commun. 25, 1835 (2004).
53. S. Damodaran and J. E. Kinsella, J. Agric. Food Chem. 30, 812 (1982).
54. C. R. Frihart, M. J. Birkeland, A. J. Allen and J. M. Wescott, Society of Wood Science
and Technology Proceedings, WS-23, Geneva, Switzerland (2010).
55. ASTM International, D 5456-09a Standard Specification for Evaluation of Structural
Composite Lumber Products, Vol. 04.10.
56. C. G. Hunt, J. O’Dell and C. R. Frihart, in: Wood Adhesives 2009, South Lake Tahoe,
CA, C. R. Frihart, C. G. Hunt and R. J. Moon (Eds.), pp. 214–219, Forest Products
Society, Madison, WI (2010).
57. C. R. Frihart, J. M. Wescott, M. J. Birkeland and K. M. Gonner, Society of Wood Science
and Technology Proceedings, WS-24, Geneva, Switzerland (2010).
Section V
Nanomaterials and
Nanocomposites
17 Carbon Nanotube-
Based Interphase Sensor
for Structural Health
Monitoring in Continuous
Glass Fiber-Reinforced
Polypropylene
Julius Rausch and Edith Mäder*
CONTENTS
17.1 Introduction................................................................................................... 296
17.2 Experimental................................................................................................. 298
17.3 Results and Discussion..................................................................................300
17.3.1 CNT-Modified GF Yarns...................................................................300
17.4 Embedding CNT-Modified GF Yarns into the PP Matrix............................304
17.5 Mechanical Testing of Model Composites with Interphase Sensors.............307
17.5.1 Quasi-Static Tensile Tests..................................................................307
17.5.2 Cyclic Loading and Quantification of Interphase Damage............... 311
17.6 Conclusions.................................................................................................... 316
References............................................................................................................... 317
In this study, we present a novel approach for structural health monitoring

in continuous glass fiber (GF) reinforced polypropylene (PP). The concept is
based on a carbon nanotube (CNT) filled coating, which is applied on the
GF yarns in order to create conductive paths on the fiber surface. This allows
monitoring the interphase strain and the occurrence of first interphase defects
associated with mechanical loading in the composites. We report on issues rel-
evant for the manufacturing and application of the interphase sensor in GF/PP
composites. This involves the modification of the coating system by CNTs, the
application of the coating onto the GF yarn, the embedding of the coated GF
* Prof. Edith Mäder was the lead author for this chapter and can be contacted at emaeder@ipfdd.de.
295
into the thermoplastic matrix, and finally the mechanical testing and quantifi-
cation of the interphase damage related to tensile loading of model composites.
The results show that the sensor can be used to detect local interphase failure
caused by microcracks as well as the breakage of GF. Moreover, by adjusting
the amount of the coating as well as the CNT content of the coating, the resis-
tivity and the sensitivity of the sensor, respectively, can be adjusted within a
large range.
17.1 INTRODUCTION
Over the past several years, there has been an increasing effort in developing strat-
egies for structural health monitoring of fibrous composites [1–6]. Among the vari-
ety of different sensor concepts, the application of carbon nanotubes (CNTs) shows
great potential for this purpose in composite materials. When the amount of CNTs
exceeds the system’s percolation threshold, a CNT network within the matrix is
formed, and any applied mechanical strain causes a dimensional change on the
nanoscale. Thus, the electrical resistance is a function of the applied strain and
allows us to in situ monitor the strain and, thereby, possibly the associated dam-
age in the composite. As the size of many defects in composite materials is in the
submillimeter range, CNTs are well suited for strain mapping on this scale due to
their small dimensions.
Schulte and coworkers [7] were the first to explore the suitability of percolated
CNT networks for structural health monitoring of continuous glass fiber (GF)/
epoxy composites, and in recent years, numerous studies have followed dealing
with different aspects of in situ strain sensing by recording the change in electrical
resistance upon variation of mechanical strain [8,9]. The sensitivity of CNT net-
works with regard to the onset of matrix dominated failure and the failure propaga-
tion could be demonstrated [10,11] as well as the possibility to monitor the fatigue
life of GF-reinforced epoxy composites [12]. The current state of research related
to sensing with CNTs has been comprehensively reviewed by Chou and coworkers
[13,14]. Most of the literature using CNTs for sensing mechanical strain involves
CNTs distributed throughout the bulk matrix. Consequently, any change in resis-
tance is related to the response of the whole CNT network in the bulk. However,
as the mechanical performance of fiber-reinforced composites is closely related to
the fiber/matrix interface and its ability to transfer the applied load from the bulk
into the fibers, it is desirable to assess the interfacial strain separately from that
of the bulk. In this context, interface strain mapping of GF/epoxy composites has
already been performed by Raman spectroscopy relating certain band shifts to the
externally applied strain [15–17].
In this study, we present a new concept for structural health monitoring in
GF-reinforced composites by depositing a multiwalled CNT containing coating onto
the GF surface. Hence, the formation of the CNT network is restricted solely to
the interphase, which forms upon consolidation of the composite. Thus, any change
in electrical resistance can be regarded as a true interphase response rather than a
bulk-related phenomenon. We see the potential of this approach in its versatility:
CNT Sensor for Structural Health Monitoring in GF-Reinforced PP 297
(1) as only the coating recipe needs to be modified, it can be tailored for improved
adhesion strength and adopted to both thermoplastic and thermoset matrices, with-
out causing internal stress concentrations; (2) only very little volume of the compos-
ite contains CNTs, that is, the absolute weight fraction of CNTs in the composite is
very low compared to bulk nanocomposites; (3) the GF yarns can be coated on a
large scale employing industrial processing routes; and (4) single GF yarns can be
introduced into macroscopic textile composites for monitoring of highly stressed
areas. Generally, CNT-modified GF yarns can be used in any kind of textile struc-
ture; however, it has to be assured that the textile processing does not result in a
complete peeling off of the electrically conductive coating. Figure 17.1 illustrates the
interphase sensor in different length scales in a composite structure. The incorpora-
tion of the GF yarn into the textile structure relates to the macroscopic level, whereas
the coating homogeneity on single filaments refers to the microscopic length scale.
Effective dispersion and distribution of single CNTs within the coating has to be on
the nanolevel.
In this work, different issues associated with the interphase sensor are discussed.
This includes the manufacturing of CNT-coated yarns and their characterization
with regard to their electrical resistance. Moreover, the effect of processing tempera-
ture on the formation of the initial resistance of the composites and the mechanical
testing of composites with simultaneous resistance measurements are discussed.
The latter includes quasi-static and cyclic tensile loading as well as the determina-
tion of parameters allowing for an assessment of the accumulated interphase damage
during cyclic loading.
CNT network on GF
CNT modified GF surface
200 nm
2 µm
FIGURE 17.1 Different length scales related to an interphase sensor based on CNT-coated
GF integrated into a woven GF structure. The insets show GF surfaces being covered by the
coating system containing CNTs. CNTs form percolated networks within the coating system
providing for an electrical conductivity on the electrically insulating GF surface.
17.2 EXPERIMENTAL
Multiwalled CNTs (NC3150, Nanocyl S.A., Belgium) were added to Brij 76 (nonionic
surfactant, molecular weight = 711 g/mol, Sigma Aldrich) surfactant solution and treated
with a horn sonicator for 120 min at constant output power of 140 W. The aqueous sur-
factant solution was prepared by dissolving surfactant in deionized water. The surfactant-
to-CNT ratio was adjusted to 1.5/1, using 0.3 wt% CNT in 100 ml deionized water.
In the next step, different quantities of predispersed CNTs were added to a polypro-
pylene (PP) film former. In order to obtain a homogeneous system, the predispersed
CNTs and the film former were magnetically stirred for several minutes. Afterward,
the system was freeze-dried (Alpha 1–2 LD plus, Christ GmbH, Germany), and the
obtained powder was used for compression molding to form thin films (K207, Rucks
GmbH, Germany). In detail, the powder was heated from 25°C to 200°C at 15°C/
min, including a degassing step at 180°C. After having reached 200°C, the films
were compression-molded at 30 bar for 2 min, before cooling down to 25°C. As
described elsewhere [18], volume resistivity measurements of the 80-μm-thick films
were carried out using a four-point test fixture in combination with the electrometers
6517a and DMM2000 (Keithley Instruments Inc.). The films were cut into pieces of
about 30 × 25 mm and the through-thickness volume resistivity measured separately.
The values presented are mean values of at least four measurements.
For preparation of the CNT-coated GF yarns, E-glass fibers with an average diam-
eter of 15 μm were continuously spun and sized with an aqueous 1 wt% γ-amino
propyltriethoxysilane (APS) solution at the Leibniz Institute of Polymer Research
Dresden. The yarn count of the GF filament yarn was 111 tex, containing 204 filaments.
In a subsequent step, the GF yarns were coated with a CNT/PP film former system con-
taining 0.5 wt% CNTs relative to the solid content of the film former Permanol 602
(Clariant, Germany) on a horizontal vertical padder type (HVF) in combination with
a continuous coating system type (KTF) (both from Werner Mathis AG, Switzerland).
The solid content of the CNT coating on the GF yarns was determined by thermogravi-
metric analysis (TGA; Q500, TA Instruments). For obtaining images of the CNT-coated
GF surfaces, a scanning electron microscope (SEM), Ultra 55 (Carl Zeiss SMT AG,
Germany), was employed. Except for the charge contrast image, all samples were sput-
tered with a 5-nm-thick platinum layer before SEM characterization.
Prior to measuring the electrical resistance using a Keithley 6514 programmable
electrometer, the yarns were annealed in an oven at 200°C for 15 min. Depending on
the measured yarn length of 5, 10, 20, and 30 cm (cf. Figure 17.6), for each measure-
ment length, at least five GF yarns were used.
In Figure 17.2, the experimental setup for resistance change measurements during
the embedding of the CNT-coated GF yarn into PP is shown. The GF yarn was placed
on the two supports it was contacted using electrically conductive silver paint and cop-
per plates. Between the supports, the GF yarns were surrounded by PP granules, which
started impregnating the fiber bundle when the melting temperature of the matrix was
reached. The heating and cooling was performed at 1ºC/min using dynamic thermal anal-
ysis (DMA; Q800, TA Instruments). The thermal properties of the film former and PP
were determined by differential scanning analysis (DSC; Q2000, TA Instruments). Simi
larly to the conditions in the DMA, a heating and cooling rate of 1 K/min was chosen.
R PEEK
support Copper
plate
T < Tm matrix T > Tm matrix
2 mm
8 mm 40 mm GF yarn 10 mm
contacted with
conductive silver paint
PP granules PP melt
FIGURE 17.2 (See color insert.) Experimental setup for resistance change measurements dur-
ing the embedding of the CNT-coated GF yarn into PP. The GF yarn is placed on the two poly-
etheretherketone (PEEK) supports, where it is contacted using electrically conductive silver paint
and copper plates. Between the supports, the GF yarns are surrounded by PP granules, which
impregnate the fiber bundle when the melting temperature of the matrix is reached.
For the preparation of the tensile specimens, the CNT-coated GF yarns were
embedded in polypropylene (Borealis HG455 FB) and compression-molded (K207,
Rucks GmbH, Germany). The computer-controlled cycle under vacuum comprised
a heating step from 25°C to 225°C. Five minutes after reaching this temperature, a
pressure of 30 MPa was applied for 1 min, before cooling down to 25°C. The heat-
ing and cooling were performed at a rate of 15°C/min. Tensile specimens with the
dimensions of 60 × 15 × 4 mm were cut out of the compression-molded PP plate
CNT coated F
GF yarn
4 mm
15 mm
R 60 mm
PP
matrix
Conductive
silver paint F
FIGURE 17.3 (See color insert.) Schematic of a specimen used for mechanical charac-
terization with simultaneous resistance change measurement. The CNT-coated GF yarn is
embedded in the center of the specimen pointing in the direction of the applied tensile load.
containing the embedded GF yarns. As the yarn length was chosen to be longer than
the specimen length of 60 mm, the overlapping GF yarn could be contacted using
electrically conductive silver paint. Figure 17.3 shows a schematic of the specimens
used for mechanical characterization with simultaneous resistance change measure-
ments. The tensile testing was conducted on a UPM Zwicki 2.5 (Zwick GmbH & Co
KG, Germany) with a crosshead velocity of 1 mm/min. No additional extensometer
was used due to the small gauge length of 20 mm. The change in resistance of the
embedded GF yarn during tensile testing was recorded simultaneously with the ten-
sile loading using a Keithley 6514 programmable electrometer.

17.3.1 CNT-Modified GF Yarns
The functionalization of a composite interphase can be achieved by incorporating
CNTs into an aqueous film former system. In a subsequent step, the CNT-modified
sizing/coating system can be applied onto the GF yarn, where it forms an electrically
conductive layer on the fiber’s surface. During consolidation of the composite, the
fiber bundle is impregnated by the matrix, turning the electrically conductive layer
into a functional interphase; that is, an electrical conductivity of the interphase is
added due to the concentrated CNTs within the interphase of the composite.
In order to realize electrically conductive interphases, it has to be assured that the
percolation threshold of the coating system is exceeded. Adding different amounts of
dispersed CNTs to the aqueous film former system, the specific resistivity of the sys-
tem can be adjusted. Figure 17.4 shows the volume resistivity as a function of the CNT
1011
1010 –1.5
109 –2.0
R2 = 0.98
Volume resistivity (Ω*cm)
108 –2.5
log σ
107 –3.0
106 –3.5
105 φperc = 0.09 wt%
–4.0
104 –5 –4 –3 –2 –1 0 1 2
103 log (φ – φperc)
102
101
100
0.0 0.5 1.0 1.5 2.0 2.5 3.0
CNT weight fraction (%)
FIGURE 17.4 Volume resistivity as a function of CNT weight fraction of nanocomposites

prepared by adding predispersed CNTs to an aqueous film former system. The inset shows
that the linear fit of a log–log plot of the conductivity σ versus the CNT content of the system
for different percolation thresholds, ϕperc, can be used for determination of the system’s per-
colation threshold.
weight fraction. The system shows a typical percolation behavior, as known for CNT
nanocomposites. However, the percolation threshold depends on different parameters,
for example, the CNTs used [19,20], the matrix system, and the processing conditions
[18,21]. For CNT contents above 0.5 wt%, the volume resistivity can only be reduced
to a minor extent. However, compared with melt-processed PP nanocomposites, the
comparably low percolation threshold as well as the low volume resistivity prove that
the chosen processing route is highly efficient with respect to dispersion of the CNTs.
A log–log plot of the conductivity versus the CNT content of the system for different
percolation thresholds, ϕperc, can be used for the determination of the system’s percola-
tion threshold (cf. inset of Figure 17.4). Here, ϕperc varied between 0.07 and 0.15 wt%,
with an increment of 0.01. The best linear fit was obtained for ϕperc = 0.09 wt% CNTs,
indicating the system’s percolation threshold.
Figure 17.4 shows that a film former system commonly used for GF sizings can
be modified in its electrical properties by adding predispersed CNTs. Therefore, this
system was applied as a coating onto the GF yarns. In detail, 0.5 wt% CNTs with
regard to the solid content of the film former were used for the coating system. The
coating content of the yarns was measured by TGA and was found to be 5.5, 10.4, 14.5,
21.6, and 29.4 wt%. Figure 17.5 shows SEM micrographs of the coated GF yarns with
5.5 wt% solid content. From Figure 17.5a, it can be inferred that relatively large areas
of the GF appear to be uncovered by the film former, while others are covered by a thin
layer or show cluster-like structures. This illustrates the challenges of this approach:
Theoretically, a very thin and homogeneous layer of the film former/CNT system on
the GF surface would be ideal, as this allows the formation of continuous conductive
paths. However, when GF yarns are sized or coated, the surface can rarely be assumed
to be homogeneously and completely covered by the sizing/coating. Hence, the amount
of film former on the GF has to be increased until continuous and interconnected areas,
rather than an island-in-the-sea topography, are formed and one can take advantage of
the electrical properties of the CNTs forming a network on the GF surface. This means
that higher film former contents on the fibers are necessary, compared with GF used as
a reinforcing fiber, where the solid content on the fibers is about 1 wt%.
In spite of the initial film former distribution after the coating process, the topogra-
phy and homogeneity of the film former/CNT layer can be affected by annealing the
GF yarns. Figure 17.5b shows a charge contrast image [22–25] of surface of a GF yarn
with 5.5 wt% solid content, which had been annealed before taking the SEM image.
The PP particles of the film former are in molten state and have wetted the GF yarn
surface. Thus, the topography is smooth, and the film former/CNT system covers the
GF yarn surface in a continuous manner. This latter effect is most important as this
supports the formation of continuous conductive paths within the film former/CNT
system on the GF surface even for low solid contents. For the prepared GF yarns with
higher solid contents of coating, the strand integrity of the yarn is improved, and the
individual filaments are held together by the coating forming fiber bundles. Figure
17.5b is a visualization of the CNT network within the PP film former on an annealed
GF yarn employing the charge contrast imaging technique. It can be observed that the
CNTs are both well dispersed and well distributed within the PP film former. A similar
distribution and state of dispersion of CNTs is expected for all areas on the GF yarn
covered by the film former. Hence, in the case of interphase sensors based on CNT
(a)
CNTs
400 nm
(b)
1 µm
FIGURE 17.5 SEM micrographs of GF with CNT coating. The solid content on the GF yarn
is 5.5 wt%. (a) GF with patchy coverage of GF surface by CNT coating showing dispersed
CNTs as coated within the spherically shaped PP particles of the film former. (b) Charge
contrast SEM image of CNT coating on a GF after annealing at 200°C for 15 min.
networks on the GF yarn surface, any change in electrical resistance as a function of

applied stress provides information on all continuously connected areas on the GF
yarn covered with the CNT coating as well as on the conductive paths between adja-
cent fibers. Whenever a single GF fails or the interfacial shear stress exceeds a critical
limit and locally causes the interphase to fail, this will be reflected in a permanent
increase of resistance, as certain conductive paths cease to exist.
In Figure 17.6, the dependence of the GF yarn resistance on its length is illustrated
for varying amounts of CNT coating on the GF yarn. Firstly, one can observe that for
an identical yarn length, the resistance decreases with increasing solid content on the
fiber. This is expected since the resistance of the yarn is given by
l
Ryarn = ρcoating ⋅ . (17.1)
Acoat
1.2 × 107
1.00 × 107 5.5 wt%
m = 2,924,510 10.4 wt%
7.50 × 107 14.5 wt%
1.0 × 107 21.6 wt%
5.00 × 107 29.4 wt% m = 358,192
8.0 × 106 2.50 × 107

Resistance (Ω)
0.00
0 5 10 15 20 25 30
6.0 × 106 Length (cm)
m = 183,047
4.0 × 106 m = 120,872
2.0 × 106
m = 76,422
0.0
0 5 10 15 20 25 30
Length (cm)
FIGURE 17.6 Dependence of resistance of CNT-coated GF yarn on the yarn length for
different coating contents. Linear lines show the least-squares fits through the origin of the
resistance data and the corresponding slopes. The inset figure shows a coating content of
5.5% on CNT-coated GF yarn.
Taking the volume resistance, ρcoating, as well as the yarn length, l, as constant,
an increasing amount of CNT coating is related to an increase of the cross-sectional
area, Acoat, of the CNT coating on the GF yarn. However, an exact value for Acoat can-
not be provided since it does not have a geometrically defined shape. Moreover, the
thickness of the coating layer varies locally; thus, Acoat is not the same for the whole
measurement length. For the different amounts of CNT coating on the GF yarns, dis-
tinct resistances are observed. For all GF yarns with coating contents >10 wt%, the
resistance ranges between several hundreds of kΩ for 5 cm yarn length and 10 MΩ
for 30 cm yarn length. The GF yarn with 5.5 wt% CNT coating lies well above this
range, between 15 and 86 MΩ. Moreover, it is noteworthy that the standard deviation
becomes bigger for the yarns with lower coating content, which is related to inhomo-
geneities of the coating layers on the fibers. For higher coating contents on the GF
yarn, it is more likely that in spite of inhomogeneities in thickness, a continuous layer
is formed, providing continuous conductive paths on the fiber surface, whereas for
lower coating contents, a growing percentage of the film former is not connected to
the continuous structure of the film former and therefore does not contribute to the
charge transfer. Similar to this, the increase in the standard deviation with increas-
ing measurement length is related to statistically distributed inhomogeneities in the
conductive layer along the measured length of the yarn (cf. Figure 17.6). These inho-
mogeneities can be regarded as a bottleneck for the charge transfer. Consequently,
a reduced measurement length results in less scattered and lower resistance values,
similar to the effect of gauge length on the tensile strength of brittle fibers as a
consequence of the statistical distribution of the critical flaws [26]. Furthermore, a
pronounced linear dependence of the resistance on the yarn length can be observed
as predicted by Equation 17.1. A linear least-squares fit through the origin was per-
formed in order to derive the different slopes of the data sets (Figure 17.6). For all
fits performed, R2 was found to be >0.99. As expected, the smallest slope was found
for the yarn with the highest coating content, and a clear trend toward steeper slopes
with decreasing amount of CNT coating on the yarn is observed. The differences
between the yarns with coating contents of 14.5, 21.6, and 29.4 wt% are relatively
modest, but the yarns with lower coating contents show a comparably pronounced
increase in the slope.
17.4 EMBEDDING CNT-MODIFIED GF
YARNS INTO THE PP MATRIX
As the concept of interphase sensing is based on CNT-coated yarns being embed-
ded in a polymeric matrix, it is of special interest to investigate how the consolida-
tion of the composites affects the electrical resistance of the sensor. In Figure 17.7,
a schematic of a GF yarn cross section below and above the melting temperature
of the matrix is shown. Moreover, Figure 17.7 shows three phenomena, which are
expected to affect the resistance of the interphase sensor when embedding a CNT-
coated yarn into the PP matrix. All phenomena are independent of each other and
can be assumed to occur after the melting temperatures of the thermoplastic coating
and the PP matrix are exceeded.
When the melting temperature of the thermoplastic coating system is exceeded,
additional surface area of the GF yarn is wetted by the CNT-filled coating, resulting
in a more homogeneous layer providing additional conductive paths. At the same
time, secondary agglomeration of the CNTs within the molten coating takes place,
resulting in a decrease of resistance as it has been observed previously for CNT-filled
thermoplastic systems [18,21,27]. The third phenomenon is the interdiffusion of the
molten coating layer with the PP matrix. Here, one could expect that the process of
CNT-based
PP granules coating PP melt
Dynamic percolation
of CNTs within the
coating
Additional wetting of the

GF surface by the coating
GF Interdiffusion of the coating

with the PP matrix
FIGURE 17.7 (See color insert.) Schematic of a CNT-coated GF cross section below and
above the melting temperature of the matrix resin. Above the melting temperatures of the PP
and the coating, different phenomena affect the electrical resistance of the CNT-coated GF,
as mentioned on the right-hand side.
interdiffusion causes locally a depletion of the CNTs in the region where the CNT-
rich coating is in direct contact with the matrix resin.
In the following, it is shown how and to what extent the initial resistance of a
CNT-coated GF yarn changes upon embedding into the PP matrix. The resistance
of GF yarn coated with a PP film former containing CNTs is measured during the
first heating, an isothermal and a cooling step, representing the typical regimes for
consolidation of thermoplastic composites. However, for this measurement, the heat-
ing and cooling rates were chosen to be 1°C/min, and the isothermal step lasted 180
min, as this allows a better assessment of the different effects mentioned in Figure
17.7. In Figure 17.8a, the resistance change during the heating from –20ºC to 200°C
is shown. Additionally, the melting region and melting temperatures as determined
by DSC measurements are indicated by the hatched areas. At temperatures below
these areas, only a modest decrease of resistance is observed, which is associated
with the negative temperature coefficient as it is typically found for CNT-filled nano
composites [28]. However, a pronounced and rapid change in resistance occurs
around 165°C, which coincides with the melting temperature of the PP film for-
mer system. This change is associated with the melting of the CNT coating and the
additional wetting of the GF yarn surface, resulting in a higher homogeneity of the
coating layer and the formation of new conductive paths. From the stepwise decrease
of the resistance value, it can be concluded that the process of additional wetting of
the GF yarn surface by the coating takes place on a comparably short time scale and
reaches its steady state after a few minutes, which is evidenced by the plateau region
at around 170°C. As soon as the melting temperature of the CNT-filled coating is
exceeded, dynamic percolation of the CNTs within the melt can take place. Strictly
speaking, this process starts at the same time as the additional wetting of the fibers.
However, its effect can be seen only at temperatures above 175°C—and thereafter,
a monotonous decrease in resistance up to 200°C occurs—and where the additional
wetting of the GF surface by the coating can be assumed not to play an important
role anymore. The decrease in resistance due to dynamic percolation continues in
the isothermal step shown in Figure 17.8b, although the curve starts leveling off
after longer periods of time, indicating that the system is approaching an equilibrium
state. The dynamic percolation and secondary agglomeration have been observed
for both CNT-filled thermoset and thermoplastic matrices, respectively [18,21,27].
The driving force for this process is assumed to depend on the specific CNT/CNT
and CNT/polymer interactions. The process of dynamic percolation is reversible and
highly dependent on the processing temperature, as this affects the viscosity of the
resin. In Figure 17.8c, the cooling step results in an increase of resistance due to the
negative temperature coefficient of the resistance in the CNT-filled system. However,
a stepwise increase in resistance is observed at temperatures below the crystalliza-
tion regions of the coating and PP matrix. While the thermal analysis of the PP film
former and PP matrix was carried out by DSC measurements with sample weights of
a few milligrams and correspondingly little thermal inertia, the experimental setup
for the resistance change measurements of the embedded GF yarn (cf. Figure 17.2)
used a significantly higher sample weight. Although the heating and cooling rates
were chosen to be only 1°C/min in order to minimize the effect of thermal inertia, it
is not surprising to find an offset toward lower temperatures compared with the DSC
(a)
PP matrix
106 Melting temp.
169°C
Melting range
Resistance (Ω) 125°C–174°C
105 Film former

Melting temp.
164°C
Melting range
114°C–169°C
First heating
104
0 50 100 150 200
Temperature (°C)
(b)
6.5 × 104
6.0 × 104
Resistance (Ω)
5.5 × 104
5.0 × 104
Isothermal step, 200°C, 180 min

4.5 × 104
0 50 100 150 200
Time (min)
(c)
6.5 × 104
6.0 × 104
Film former
Resistance (Ω)
Crystallization temp.
121°C
5.5 × 104 Crystallization range
112°C–132°C
PP matrix
5.0 × 104
121°C
Crystallization range
109°C–135°C
4.5 × 104
200 150 100 50 0 –50
Temperature (°C)
FIGURE 17.8 (See color insert.) Resistance change of the interphase sensor upon embed-
ding into a PP matrix. (a) Heating from –20°C to 200°C, (b) isothermal step at 200°C for
180 min, (c) cooling from 200°C to –20°C. Heating and cooling rate is 1°C/min. The hatched
areas refer to the thermal analysis of the film former and PP matrix by DSC measurements,
indicating the temperature ranges for melting and crystallization.
data for any effects related to the phase transition of the polymer. Therefore, the
increase of resistance between approximately 110°C and 75°C can be related to the
crystallization of the thermoplastic coating with CNT networks. After the process of
crystallization has come to an end within the whole sample, the resistance continues
to increase similarly as observed at temperatures below 200°C. The increase in
resistance between 75°C and –20°C is found to be linear, with a temperature coef-
ficient of –59 ohm/°C.
Considering the resistance change in the different regimes (heating, isothermal,
cooling) quantitatively, it can be concluded that the highest resistance change of
more than one order of magnitude occurs around the melting temperature of the
CNT-filled coating system, which is related to the additional wetting of the GF yarn
surface by the coating. Moreover, the dynamic percolation of the CNTs within the
molten polymer causes a further decrease in resistance. However, even for 180 min
at 200°C, this change is small compared to the one taking place during the first heat-
ing. A significant effect of the interdiffusion between the CNT-rich coating and the
polymer matrix on the resistance cannot be found, as this would have resulted in an
increase of resistance during the isothermal treatment. Although this mechanism is
likely to occur, its effect on the resistance is outweighed by the dynamic percolation
of the CNTs.
17.5 MECHANICAL TESTING OF MODEL COMPOSITES

WITH INTERPHASE SENSORS
17.5.1 Quasi-Static Tensile Tests
In order to verify the feasibility of the CNT coating to be used as an interphase sen-
sor on GF yarn, single rovings of the CNT-coated GF yarns were embedded into a
PP matrix (cf. Figure 17.3) and subjected to tensile loading. The concept of the inter-
phase sensor is based on the resistance change of the percolated CNT network in
the fiber/matrix interphase due to mechanical loading. Figure 17.9 shows schemati-
cally the cross section of a GF with a CNT-modified interphase serving as a sensor
PP matrix
1 2 3
GF
CNT-modified
interphase
FIGURE 17.9 (See color insert.) Schematic of a GF cross section with CNT-modified coat-
ing serving as interphase sensor. On the right-hand side, different types of interphase-related
defects are shown: (1) GF failure with associated interphase failure, (2) microcracks in the
interphase without GF failure, (3) local debonding at the fiber/matrix interface.
element. Any externally applied load is transferred via the interphase region to the
GF, resulting in a mechanical strain in this region, which can be assessed indirectly
by the resistance change. If the mechanical loading causes the interphase to fail, as
schematically shown in Figure 17.9, certain conductive paths are destroyed, and a
sudden resistance change is measured.
In Figure 17.10, two representative quasi-static tensile tests with simultaneous resis-
tance change measurements are shown. Figure 17.10a shows the stress–displacement
curve of a GF yarn with 29.4% coating content and 0.5 wt% CNTs. Although the CNT
affected volume is very low compared to the specimen volume and its extension is
restricted to the interphase only, a pronounced resistance change with increasing dis-
placement can be observed. This change in resistance can be regarded as an indirect
measure of the interphase strain. With increasing displacement, the slope becomes
gradually steeper until the resistance changes stepwise at a certain stress level. This
point coincides with the appearance of first discontinuities in the stress–displacement
curve as can be inferred from the inset in Figure 17.10a. The first local maximum in
the stress curve corresponds to the onset of GF fiber breakage, which is still below the
maximum stress of the sample and demonstrates the sensitivity of the approach with
regard to the detection of the initial failure of the reinforcing fibers.
While in Figure 17.10a, the resistance change follows an exponential behavior,
as has also been found for CNT/epoxy nanocomposites [12], the resistance change
curve shown in Figure 17.10b shows a different behavior. The reason for this dif-
ferent behavior is related to the comparatively small amount of coating on the yarn
(5.5 wt%) resulting in a higher sensitivity of the system for detecting microcracks
in the interphase region. From the stepwise changes in the resistance change data,
it can be inferred that the interphase region has been subjected to a critical load,
while the stress–displacement curve does not show similar changes. Distinct regions
can be identified in Figure 17.10b: (1) The resistance change curve shows a steady
growth until a displacement of 2.15 mm, corresponding to approximately 80% of the
maximum stress of the composite. Up to this value, the interphase is strained homo-
geneously, and no failure has occurred yet. (2) Exceeding this stress level, a first
local interphase failure occurs, which is characterized by a sudden change in slope
of the resistance change curve. However, after this event, the resistance change con-
tinues to grow steadily until a displacement of 2.62 mm. (3) From here on, multiple
interphase failures occur shortly one after another, causing a fivefold increase of the
resistance change. However, these events are not reflected in the stress–displacement
curve, which does not show any irregularities up to a displacement of 3.24 mm.
(4) After reaching approximately 97.5% of the maximum stress, the GFs start to fail,
and resistance change increases beyond the measurement limits. This demonstrates
that the onset of interphase failure is well before the onset of GF failure as the two
events occur at 80% and 97.5% of the maximum stress, respectively. Based on this
information, one can conclude that for the given system and loading case, stresses
of around 21 MPa are critical for the interphase. This corresponds well with the
transverse tensile strength of unidirectional GF yarn/PP composites based on hybrid
yarns, mentioned in our previous work [29,30], which was found to be about 25 MPa.
The CNT-filled coating can be regarded as the sensor element of the system.
Thus, a varying coating thickness as well as different CNT contents of the coating
(a) 30 30
25 25
25 24 25
Resistance change ΔR/ΔR0 (%)

20
23
20 20
Stress (MPa)
22 15
3.0 3.5 4.0
15 15
10 10
5 5
0 R0 = 373.25 kΩ 0
0 1 2 3 4
Displacement (mm)
(b) 30 350
26.0 325
300
25 300
Resistance change ΔR/ΔR0 (%)

25.5 275
(4)
250 250
20
Stress (MPa)
25.0 225
3.00 3.25 3.50 200
15
150
10 (3)
100
5 50
(1)
(2) R0 = 36.41 kΩ
0 0
0 1 2 3 4
Displacement (mm)
FIGURE 17.10 Stress–displacement curve and simultaneous recording of the resistance

change for GF yarns embedded in PP matrix. (a) GF yarn with 29.4 wt% coating content and
0.5 wt% CNTs, (b) GF yarn with 5.5 wt% coating content and 0.5 wt% CNTs. For the yarn
with 5.5 wt% coating content, multiple interphase failures before the first failure of the GF
yarn can be identified by the resistance change measurement. (The insets show higher mag-
nifications of first kinks in the stress–displacement curves. For details, see text.)
will affect the resistance of the interphase sensor and its sensitivity with respect to
mechanical loading.
First, the volume resistivity of the sensor is investigated based on the initial resis-
tance of the compression-molded model composites (cf. Figure 17.3). Figure 17.11
depicts the dependence of the volume resistivity of the interphase sensor on the yarn
105
Volume resistivity (Ω cm)

Volume resistivity
104 (Ω cm):
2.0E+04
103
102 7.5E+02
101
30
100
10–1 10
Yar
10
n coa
20
tin
1.00
0.75
g con
0.50
(wt%)
30 0.25
of the coating
ten
ontent
CN T c
t
(wt
%)
FIGURE 17.11 (See color insert.) Dependence of the volume resistivity of the interphase
sensor on the yarn coating content and the CNT content of the coating. The volume resistivity
values are based on the initial resistances of the compression-molded GF yarn/PP composites
and represent averaged values of at least five individual measurements.
coating content and the CNT content of the coating. A pronounced change in resis-
tivity for variation in yarn coating content can be seen, resulting in comparatively
high resistivity values for lower yarn coating contents. This is related to the amount
of interconnected area of CNT coating on the GF yarn surface, which increases with
higher coating content, as discussed elsewhere [31]. Similarly to the effect of the yarn
coating content, the resistivity can be affected by a variation of the CNT content of
the coating. Not surprisingly, a decrease in resistivity for higher CNT loadings can
be found, similar to that shown in Figure 17.4. Combining these two effects, one can
adjust the volume resistivity of the interphase sensor within a large range by choos-
ing different yarn coating contents and CNT contents of coating. Generally, the low-
est resistivities can be found for systems with a high yarn coating content as well as
a high level of CNT loading.
Similarly to the resistivity of the interphase sensor, its sensitivity can be affected
by varying yarn coating content and CNT content of the coating. The sensitivity,
expressed in terms of the change in resistance, depends on the interphase strain, that
is, displacement during the tensile loading. Therefore, the data displayed in Figure
17.12 show the resistance change value during tensile loading at a displacement of
3 mm. It can be seen that these two parameters have significant influence on the sen-
sitivity. For the systems with low yarn coating content as well as low CNT content in
the coating, the resistance change is found to be the highest, whereas the lowest sen-
sitivities are found for thick coating layers with high CNT contents. Generally, above
the system’s percolation threshold, an increasing CNT content results in additional
conducting paths as well as in an increased number of contact points between the
CNTs. Thus, the charge transfer can be distributed over all conductive paths. Upon
40
Resistance change (%) 30

Resistance change (%):
0.000
20
8.000
16.00
10 24.00
32.00
40.00
0 45.00
Yar
10
n coa
20
ting
0.5
0.4
con
0.3 (wt%)
30
of the coating
ten
0.2
ontent
CN T c
t (w
t%)
FIGURE 17.12 (See color insert.) Dependence of the sensitivity of the interphase sensor
during tensile loading on the yarn coating content and the CNT content of the coating. The
resistance change values represent averaged values of at least five individual measurements
on the compression-molded GF yarn/PP composites. All resistance change values were taken
at 3 mm displacement during tensile loading of the model composites.
mechanical loading of such a system, any local failure of a conductive path does
not result in significant changes in the resistance value, since the electrical current
is bypassed via the nearby existing and undisturbed conductive paths. In the case of
systems close to the percolation threshold, significantly fewer conductive paths exist,
and the failure of any of these connections is related to a correspondingly higher
relative change in resistance than is found for systems with CNT contents well above
the percolation threshold. When designing interphase sensors, the interplay of the
yarn coating content and the CNT weight fraction of the coating has to be consid-
ered. Aiming at low initial resistances, the coating content of the yarn as well as the
amount of CNTs should be as high as possible. However, the low initial resistance
would be achieved at the cost of reduced sensitivity and will cause misleading infor-
mation on the interphase strain due to a heavily bundled structure of the GF yarn. In
terms of sensitivity, a low coating content as well as CNT weight fractions slightly
above the system’s percolation threshold are favorable, although this can only be
achieved for systems with comparatively high initial resistances.
17.5.2 Cyclic Loading and Quantification of Interphase Damage

In the following, interphase structural health monitoring is investigated by perform-
ing cyclic loading in order to assess the formation and growth of interphase defects
at loading levels below GF yarn failure. Additionally, new parameters are introduced
for assessing the overall damage of the interphase throughout the mechanical testing.
1.0 7.5 0.5

R0 = 188.48 kΩ
0.4
Resistance change ΔR/R0 (%)

Displacement (mm)
5.0
Stress (MPa) 0.3
0.5
0.2
2.5
0.1
0.0 0.0 0.0

0 50 100 150 200
Time (s)
FIGURE 17.13 First load cycles of a specimen with interphase sensor during stress-
controlled cyclic loading between 0 and 5 MPa.
Figure 17.13 shows the three signals captured during stress-controlled cyclic load-
ing. Combining mechanical testing with simultaneous resistance measurements, a
set of data is obtained consisting of force and displacement as well as information on
resistance change. For CNT-filled nanocomposites, it has been shown that resistance
change follows the displacement of the system and allows monitoring of the system’s
deformation as well as the occurrence of microcracks [8,11,12,28,32].
In Figure 17.13, it can be seen that the resistance change is in agreement with
the stress and displacement, which becomes evident from the fact that the change
in slope of the resistance curve occurs at the same time as the displacement and
stress signal. As a stress-controlled test regime was chosen, the specimen is elon-
gated slightly more with every load cycle due to viscoplastic and viscoelastic defor-
mations. Although this gradual elongation occurs very slowly, it is reflected in the
resistance change curve by a steady increase of the value after specimen unloading
and demonstrates the sensitivity of the interphase sensor with respect to the sample
deformation and, thereby, the associated interphase strain.
Figure 17.14 shows the behavior of the resistance change during the cyclic loading
between 0 and 20 MPa. After the first load cycle, the viscoelastic and/or viscoplastic
deformation results in an increase in resistance change of 6%. Any further load cycle
causes an additional increase in resistance change due to sample deformation but
only to a minor extent. For up to 22 h, the amplitude of the resistance change remains
stable at low levels, indicating that no severe damage of the interphase has occurred.
However, after longer periods of time, the resistance change increases stepwise,
related to the initiation of first larger defects, which continue to grow with additional
load cycles. This is indicative of a severe interphase failure, cutting off the majority
of the conductive paths within the interphase. For further information on the concept
of conductive paths in conductive interphases, the reader is referred to Figure 4 by
15,000
15
Resistance change ΔR/R0 (%)

10
10,000
5000
0
0 4 8 12 16 20 24
0 R0 = 342.13 kΩ
0 4 8 12 16 20 24 28 32
Time (h)
FIGURE 17.14 Resistance change of interphase sensor during stress-controlled cyclic load-
ing between 0 and 20 MPa. The inset shows the amplitude of the resistance change before
the occurrence of severe interphase damage, which causes a distinct change in the amplitude
pattern.
Rausch and Mäder [31]. Although the resistance change curve indicates clearly when
the failure of the interphase takes place, it remains unclear whether the preexisting
microcracks have reached a critical size, causing unstable crack growth, or whether
the sudden initiation of a larger single crack in the interphase is the reason for the
observed resistance change.
In general, the behavior of the resistance change up to 22 h resembles the dis-
placement or strain during fatigue loading. Within the first 4 h, cyclic loading the
resistance changes faster as is the case between 4 and 21 h, where a fairly linear
increase is observed. This is due to the higher extent of plastic deformation at the
beginning of the loading. In the linear part between 4 and 21 h, the additional plastic
deformation in each load cycle is comparably small. As the resistance change cor-
responds to the integrity of the conductive paths, one can assume that its increase
is due to a deformation of the whole CNT network structure as well as stable and
slow growth of all existing microcracks. As can be inferred from the inset in Figure
17.14, at approximately 21 h, right before the resistance change increases drastically
for the first time, the resistance change no longer follows the linear relationship as
before. Rather, it shows a distinct increase in resistance change with every load cycle,
indicating the onset of a faster crack growth in the interphase region, resulting in
significantly increased amplitudes of resistance change.
By recording a resistance change curve of a specimen with an interphase sensor
during cyclic loading, one can define characteristic points allowing us to describe the
severity of the interphase defects and their evolution during testing quantitatively. In
this context, we introduce new parameters for the quantification of the accumulated
overall damage of the interphase during cyclic loading. Figure 17.15 shows sche-
matically the resistance change curve of three successive load cycles. Two features
c3max
c2max
c1max
a3
ΔR/R0 a2 c4min
Amplitude
of resistance c3min
change Δρ3
a1 c2min
Δρ2
c1min Difference in resistance
Δρ1
change after unloading
Load cycle
FIGURE 17.15 Definition of characteristic points of successive load cycles in a schematic

resistance change curve during cyclic loading. In order to quantify the accumulated inter-
phase damage upon loading (Equation 17.2), ci max and ci min denote the maximum and the
minimum of the resistance change for each load cycle, and the ratio of Δρi /ai–1 instead of
Δρi /ai has to be used, since upon damage initiation, ai as well as Δρi increase simultaneously.
characterize the behavior of the resistance change. First is the amplitude of the resis-
tance change during loading, ai, and second is the difference between the resistance
change of successive load cycles after unloading, Δρi. The amplitude of the resistance
change, ai, can be regarded as the sensitivity of the sensor with regard to any damage
occurring within the interphase during the mechanical loading of the system. Unless
no severe defects are present, for the systems investigated, ai is found to be below 3%.
Δρi refers to the resistance change due to the viscoelastic and/or viscoplastic flow
within the interphase. Even when ai remains constant and no microcracks are intro-
duced in the interphase region, the resistance change increases steadily. Although this
is not necessarily linked with interphase defects at the early stage of cyclic loading, it
represents indirectly the overall strain the interphase region has suffered.
Using the ratio of successive resistance change amplitudes, we introduce the
parameter ψloaded for quantification of the accumulated interphase damage upon
loading of the sample.
n n
 − ci min 
ψ loaded = ∑
i=2
ai
ai −1
−1 = ∑  c c
i=2
i max
i −1 max − ci −1 min
 − 1 (17.2)
n 
ci max and ci min denote the maximum and the minimum of the resistance change for
each load cycle, respectively (cf. Figure 17.14). As long as the structural integrity of
the interphase is maintained and no cracks result in a cutoff of conductive paths, the
magnitude of the amplitude remains constant for every load cycle, yielding a value
of 1 for the ratio of ai/ai–1. Summing up the differences from the value of 1, ψloaded
expresses the extent of the accumulated damage during interphase loading. However,
the parameter ψloaded is not sufficient for describing the effect of cyclic loading on the
interphase as it does not consider the permanent resistance change after unloading
the sample due to viscoelastic and/or viscoplastic flow. Hence, a second parameter,
ψunloaded, is introduced, relating the resistance change due to the viscoelastic and/or
viscoplastic flow to the amplitude of resistance change. ψunloaded can be written as
n n
 − ci min 
ψ unloaded = ∑
i=2
∆ρi
ai −1
= ∑  cc
i=2
i +1 min
 . (17.3)
i −1 max − ci −1 min 
If the permanent resistance change after unloading of each load cycle does not
increase, for example, in a purely elastic loading regime, Δρi then ψunloaded would
be zero. Hence, ψunloaded expresses the degree of plastic deformation within the
interphase region as well as the impact of possible defects on the resistance in the
unloaded state of the sample. In order to identify the occurrence of defects using
ψunloaded, the ratio of Δρi/ai–1 instead of Δρi/ai has to be used, since upon damage
initiation, ai as well as Δρi increase simultaneously and might result in a similar ratio
as for the load cycles before the damage has occurred.
Combining Equations 17.2 and 17.3,
ψint Dam = ψloaded + ψunloaded, (17.4)
a parameter is obtained describing the accumulated interphase damage, ψint Dam,

based on the effect of interphase damage on the resistance change in both loaded
and unloaded conditions.
Figure 17.16 shows the introduced parameters calculated from the resistance
change data for the cyclic loading between 0 and 20 MPa (cf. Figure 17.14). Up to
the initiation of the first interphase failure after approximately 550 load cycles, all
30
25
Interphase damage (–)
20 ψloaded
ψunloaded
15
ψint Dam
10
0 100 200 300 400 500 600 700 800

Load cycle
FIGURE 17.16 Assessment and quantification of the interphase damage during cyclic load-
ing between 0 and 20 MPa by using the parameters ψloaded, ψunloaded, and ψint Dam, from the
resistance change data based on amplitudes of resistance change (Equations 17.2 and 17.3).
parameters show only a very modest increase without any sudden change in slope.
However, right after this point, they show a stepwise increase and continue to grow
faster as compared to the beginning of the cyclic loading. However, with further
cyclic loading, ψloaded starts to increase steadily, reflecting a certain amount of inter-
phase damage. Interestingly, after the sudden increase of ψloaded at approximately 550
load cycles, two regions with distinct slopes can be observed. Obviously, the crack
continues to grow relatively fast, before crack growth slows down again, resulting
in a change in slope after approximately 620 load cycles. The behavior of ψunloaded
clearly shows the effect of the viscoelastic and/or viscoplastic deformation of the
interphase at the very beginning of the cyclic loading. Here, ψunloaded increases rela-
tively fast before leveling off and increasing linearly in the course of cyclic loading.
However, right before its sudden increase at approximately 550 load cycles, ψunloaded
starts to grow faster again, indicating the accumulation of new interphase damage.
17.6 CONCLUSIONS
By coating GF yarns with a CNT-modified film former system, electrically conduc-
tive GF fiber surfaces can be generated, allowing for interphase structural health
monitoring in continuous GF–reinforced composites.
The dispersion of CNTs in an aqueous medium and their addition to the film
former system results in a coating system with adjustable resistivity value. Besides
the CNT content of the coating, it was shown that the electrical resistance of CNT-
coated GF yarns depends also on the yarn coating content, that is, the thickness
of the coating layer. The highest resistance values were found for yarns with low
coating contents, as these result in an increased amount of discontinuous coating
structures, which cannot provide continuous conductive paths on the surface of the
GF yarn. Upon embedding the CNT-coated yarns into the PP matrix, it could be
shown that the electrical resistance of the GF yarn is affected, to a large extent, by
the chosen temperature regime. The biggest change in resistance is found during the
first heating, when the melting temperature of the CNT-filled coating is exceeded
and the coating starts to wet an additional area of the GF yarn surface, resulting in
new conductive paths. Moreover, the dynamic percolation also results in lowered
resistance values. However, this effect is comparatively small.
The mechanical testing of model composites with interphase sensor reveals that
the system is capable of detecting microcracks in the interphase as well as the break-
age of the GF. Whereas the failure of the GF is also visible in the stress–displacement
curve of the tensile test, the occurrence of the interphase failure is only detectable
by the resistance change of the interphase sensor. Both the resistivity of the sensor as
well as the sensitivity with regard to the mechanical loading can be adjusted by the
yarn coating content and the CNT content of the coating. The highest resistivities as
well as the highest sensitivities are found for the systems with both low yarn coat-
ing content and low CNT content of the coating. Besides a quasi-static tensile test,
cyclic tensile loading was performed in order to assess the integrity of the interphase
region. At the beginning of the cyclic testing, the amplitude of the resistance change
remained small, indicating that no severe interphase defects had occurred. The first
interphase failure occurs after approximately 22 h of cyclic loading and can result in
a significant increase of resistance change. By defining certain characteristic points

in the resistance change curve during cyclic loading, we introduce parameters allow-
ing for a quantification of the accumulated interphase damage.
ACKNOWLEDGMENTS
The funding of this work within the framework of the collaborative research cluster
SFB639, subproject A1, by the Deutsche Forschungsgemeinschaft (DFG) is highly
appreciated. Moreover, the authors wish to thank Dr. Rong-Chuan Zhuang for help-
ful discussions. The experimental assistance of Dr. Wolfgang Jenschke, Dr. Petra
Pötschke, and Holger Scheibner is acknowledged.
REFERENCES
1. R.M.M. Measures, Comp. Eng. 2, 597 (1992).
2. Y.J. Rao, Opt. Laser Eng. 31, 297 (1999).
3. Y. Zou, L. Tong and G.P. Steven, J. Sound Vib. 230, 357 (2000).
4. Y. Okabe, T. Mizutani, S. Yashiro and N. Takeda, Compos. Sci. Technol. 62, 951
(2002).
5. G. Zhou and L.M. Sim, Smart Mater. Struct. 11, 925 (2002).
6. S. Zhongqing, Y. Lin and L. Ye, J. Sound Vib. 295, 753 (2006).
7. K. Schulte, B. Fiedler, F.H. Gojny, M.H.G. Wichmann and W. Bauhofer, Ann. Chim.-
Sci. Mater. 29, 81 (2004).
8. J.M. Park, D.S. Kim, S.J. Kim, P.G. Kim, D.J. Yoon and K.L. deVries, J. Compos.: Part B
38, 847 (2007).
9. M.H.G. Wichmann, S.T. Buschhorn, L. Böger, R. Adelung and K. Schulte, Nanotechnology
19, 475503 (2008).
10. E.T. Thostenson and T.W. Chou, Adv. Mater. 18, 2837 (2006).
11. E.T. Thostenson and T.W. Chou, Nanotechnology 19, 215713 (2008).
12. L. Böger, M.H.G. Wichmann, L.O. Meyer and K. Schulte, Compos. Sci. Technol. 68,
1886 (2008).
13. T.W. Chou, C. Li and E.T. Thostenson, Compos. Sci. Technol. 68, 1227 (2008).
14. T.W. Chou, L. Gao, E.T. Thostenson, Z. Zhang and J.H. Byun, Compos. Sci. Technol. 70,
1 (2010).
15. Q. Zhao, J.R. Wood and H.D. Wagner, Appl. Phys. Lett. 78, 1748 (2001).
16. P. Sureeyatanapas and R.J. Young, Compos. Sci. Technol. 69, 1547 (2009).
17. P. Sureeyatanapas, M. Hejda, S.J. Eichhorn and R.J. Young, Compos. Sci. Technol. 70,
88 (2010).
18. G. Kasaliwal, A. Göldel and P. Pötschke, J. Appl. Polym. Sci. 112, 3494 (2009).
19. H.P. Wu, X.J. Wu, M.Y. Ge, G.Q. Zhang, Y.W. Wang and J.Z. Jiang, Compos. Sci.
Technol. 67, 1116 (2007).
20. F.H. Gojny, M.H.G. Wichmann, B. Fiedler, I. Kinloch, I.A. Kinloch, W. Bauhofer, A.H.
Windler and K. Schulte, Polymer 47, 2036 (2006).
21. W. Bauhofer and J.Z. Kovacs, Compos. Sci. Technol. 69, 1486 (2009).
22. C.W. Oatley and T.E. Everhart, J. Electron 2, 568 (1957).
23. K.T. Chung, J.H. Reisner and E.R. Campbell, J. Appl. Phys. 54, 6099 (1984).
24. J. Loos, A. Alexeev, N. Grossiord, C.E. Koning and O. Regev, Ultramicroscopy 104, 160
(2005).
25. J.Z. Kovacs, K. Andresen, J.R. Pauls, C.P. Garcia, M. Schossig, K. Schulte and W.
Bauhofer, Carbon 45, 1279 (2007).
26. P. Zinck, M.F. Pays, R. Rezakhanlou and J.F. Gerard, J. Mater. Sci. 34, 2121 (1999).
27. I. Alig, D. Lellinger, M. Engel, T. Skipa and P. Pötschke, Polymer 49, 1902 (2008).
28. S.L. Gao, R.C. Zhuang, J. Zhang, J.W. Liu and E. Mäder, Adv. Funct. Mater. 20, 1885
(2010).
29. J. Rausch, R.C. Zhuang and E. Mäder, Compos.: Part A 41, 1038 (2010).
30. J. Rausch, R.C. Zhuang and E. Mäder, Mater. Technol. 24, 29 (2009).
31. J. Rausch and E. Mäder, Compos. Sci. Technol. 70, 1589 (2010).
32. J. Zhang, R.C. Zhuang, J.W. Liu, E. Mäder, G. Heinrich and S.L. Gao, Carbon 48, 2273
(2010).
18 Reaction of Multiwalled
Carbon Nanotubes
with Gaseous Oxygen
and Chlorine Atoms
L. Oliveira, T. Debies, and G. A. Takacs*
CONTENTS
18.1 Introduction................................................................................................... 320
18.2 Experimental................................................................................................. 320
18.2.1 MWNT Paper.................................................................................... 320
18.2.2 Microwave Plasma Discharge of Ar–O2 Mixture.............................. 321
18.2.3 UV Photochlorination........................................................................ 321
18.2.4 X-Ray Photoelectron Spectroscopy................................................... 322
18.3 Results............................................................................................................ 322
18.3.1 Untreated MWNT Paper................................................................... 322
18.3.2 Reaction of O Atoms with MWNT Paper......................................... 322
18.3.3 Reaction of Cl2 with MWNT Paper................................................... 324
18.3.4 Photochlorination of MWNT Paper.................................................. 324
18.4 Discussion...................................................................................................... 326
18.5 Conclusions.................................................................................................... 327
References............................................................................................................... 328
Multiwalled carbon nanotube (MWNT) paper was surface-modified with gas-

eous oxygen and chlorine atoms and analyzed by X-ray photoelectron spectros-
copy (XPS). The oxygen atoms were produced by a low-pressure microwave
(MW) plasma discharge of an Ar–O2 mixture, and the results were compared
to a previous study using VUV photooxidation. Exposure to VUV photooxida-
tion (λ = 104.8 and 106.7 nm) showed approximately the same level of oxida-
tion (about 7.4 at% O) as the MW results (about 6.3 at% O) that were conducted
* Prof. G. A. Takacs was the lead author for this chapter and can be contacted at gatsch@rit.edu.
319
in the absence of radiation from the plasma. Oxygen atoms reacted with the
π-conjugation of the MWNTs to primarily produce the epoxide/ether groups.
Chlorine atoms, which were generated by UV photolysis of Cl2, reacted more
readily than chlorine molecules with the nanotubes to yield a saturation level
of about 13 at% Cl.
18.1 INTRODUCTION
Carbon nanotubes (CNTs) have many desirable bulk properties, such as very good
mechanical strength, high thermal stability, and excellent electrical conductance
for potential use in a variety of applications. However, their surfaces often require
modification in order to achieve functionality. To help control electronic properties,
electron-withdrawing oxygen and halogen atoms may be covalently bonded to the
surface to assist in the conversion of metallic-like conductivity of CNTs to semicon-
ducting properties.
Oxygen-containing functional groups on single-walled carbon nanotubes
(SWNTs) have been shown to increase electrical resistance when treated with ozone
[1], demonstrate semiconductor properties when oxidized with hydrogen peroxide
[2], display persistently photoconductive properties when carboxylated [3], and result
in photoinduced charge separation and generation of photocurrent upon illumination
with the attachment of n-pentyl ester moieties [4].
Theoretical calculations predict that sidewall fluorination can yield either metal-
lic or semiconducting nanotubes depending on the type of fluorination and composi-
tion [5,6].
The fluorinated CNTs are highly electrically resistive and can survive tempera-
tures up to 325°C [7]. Functionalization with bromine leads to an enhancement in
the Fermi level absorptions due to metallic intraband transitions occurring with the
valence band of the semiconducting CNTs [8]. The reaction of diatomic molecular
chlorine with CNTs has been achieved using mechanochemical ball milling [9,10],
electrochemical gas sensing [11], and sonicating with chlorinated solvents [12].
There were no reports of the reaction of Cl atoms with CNTs.
In this study, a MWNT paper was reacted with gaseous oxygen atoms, produced
using a microwave discharge (MW) of an Ar–O2 mixture, and chlorine atoms formed
by UV photolysis of Cl2 gas. The resulting surface modification was investigated
with x-ray photoelectron spectroscopy (XPS).
18.2 EXPERIMENTAL
18.2.1 MWNT Paper
The MWNT paper was synthesized using an injection chemical vapor deposition
(CVD) reactor having a coaxial injection design [13]. The injection tip was fabricated
from quartz capillary tubing having an o.d. of 3 mm and an i.d. of 1 mm, which
was encased in a stainless steel jacket. A precursor solution of cyclopentadienyl iron
dicarbonyl dimer in xylene (0.08 M) was injected into the hot zone of the furnace
Reaction of MWNTs with Gaseous Oxygen and Chlorine Atoms 321
by a mechanical pump for consistent delivery rates. The material was synthesized at
725°C, with an Ar flow rate of 0.75 L/min and a precursor delivery rate of 3.5 mL/h.
The MWNTs were dispersed in acetone, sonicated for 30 min, and filtered through
a Teflon filter with 1 μm pores to form the paper. The MWNT paper produced was
an ~1.5 cm diameter circle with a thickness of ~10 μm. Characterization of the high
purity MWNTs was carried out using intensity ratios of Raman spectroscopy peaks,
which were consistent with scanning electron microscopy (SEM) and thermal gravi-
metric analyses (TGA) [14]. Scanning electron micrographs showed a range of tube
diameters from 11 to 139 nm and an average diameter of 58 ± 4 nm at a 95% confi-
dence level [14,15].
The bottom side of the MWNT paper was attached, using a double-sided adhesive
tape, to a microscope coverslip and the top side was reacted with either oxygen or
chlorine atoms.
18.2.2 Microwave Plasma Discharge of Ar–O2 Mixture

Atomic oxygen was generated from a MW discharge of an Ar–O2 mixture, as shown
in Figure 18.1, having flow rates of 50 and 10 sccm, respectively, with the pressure
in the chamber at 1.3–4.0 Pa. The MW plasma had a frequency of 2.45 GHz and
absorbed power of 27–29 W. The vacuum system was designed so that the discharge
was located ~41 cm upstream from the MWNT, and the radiation from the discharge
was not directed at the sample [16].
18.2.3 UV Photochlorination
A 25.4 cm i.d. Rayonet photochemical reactor (RPR) (manufactured by Southern
New England Ultraviolet Co., Inc., Branford, CT) was equipped with 16 lamps
(RPR-300 nm, Sunlight Phosphor), which emitted a broad band of wavelengths from
Plasma cavity
Mass flow Mass flow

controller control system Microwave guide
Microwave
MKS power source Reactor chamber
Mass flow
controller Samples
Pressure Throttle valve

gauge
Substrate and
holder
Ar tank Oxygen tank Mechanical vacuum pump
FIGURE 18.1 Schematic diagram showing microwave discharge of Ar–O2 mixture.

~250 to 400 nm with a maximum intensity at ~300 nm. The emission spectrum over-
lapped with the photoabsorption spectrum of gaseous chlorine, which has a weak con-
tinuum from 250 to 450 nm with a maximum photoabsorption cross section of ~2.4 ×
10 –19 cm2 at a wavelength of ~330 nm [17]. A cylindrical photochemical cell (2.54 cm
in diameter, 17.8 cm long) was constructed of Suprasil quartz to allow transmission
of UV radiation into the cell and fitted with a Cajon removable high vacuum stainless
steel fitting for placement and removal of the sample from the cell [18].
The photochemical cell was attached to a Pyrex vacuum line and filled with 1.1–
1.6 × 104 Pa of ultrahigh purity Cl2 (Matheson Tri-Gas, Parsippany, NJ) as measured
with a Hg manometer that was topped off with ~2 cm of dibutyl phthalate to prevent
reaction of chlorine with the Hg. The chlorine was degassed of air using at least two
freeze–thaw cycles at 77 K. Unreacted chlorine was neutralized by condensing the
gas into an evacuated Pyrex flask containing a 1 M NaOH solution frozen at 77 K
and then warmed up to room temperature.
18.2.4 X-Ray Photoelectron Spectroscopy

The samples were analyzed with a Physical Electronics Model 5800 XPS that exam-
ines the top 2 to 5 nm of a sample’s surface using a take-off angle of 45° between
the sample and the analyzer. A region of about 800 μm in diameter was analyzed.
The sample attached to the microscope coverslip was mounted directly on the XPS
sample holder. The monochromatic Al Kα (1486 eV) x-ray beam was focused to
accept only photoelectrons emitted from the nanotubes. The quantitative analyses
are precise to within 5% for major constituents and 10% for minor constituents.
The samples were charge-neutralized with a flood of low energy electrons from a
BaO field emission charge neutralizer in order to minimize radiation damage to the
samples.
18.3 RESULTS
18.3.1 Untreated MWNT Paper
The MWNT paper contained only carbon and oxygen, indicating the lack of other
elemental contamination. The average amount of oxygen for eight untreated samples
was 1.9 ± 0.8 at%, which is in good agreement with previous measurements [15,19].
18.3.2 Reaction of O Atoms with MWNT Paper

The MWNT paper exposed to O atoms in the absence of radiation showed increas-
ing amounts of oxygen with increasing time of exposure up to 60–90 min where a
plateau value is reached at ~6.3 at%. Figure 18.2 compares these data to samples of
the MWNT paper that were treated in the presence of radiation using vacuum UV
(VUV) photooxidation where an O atom saturation level of ~7.4 at% was obtained
[15]. The scatter in the data at short treatment times is probably due to the variability
in the O at% observed for the untreated samples.
6
Oxygen (at%)
5
0
0 20 40 60 80 100 120
Treatment time (min)
FIGURE 18.2 Plot of atomic percent of oxygen as a function of treatment time for MWNT
paper samples treated with the microwave discharge method for producing oxygen atoms (◾)
(this work) and VUV photooxidation (⬥). (From M. Krysak et al., J. Adhesion Sci. Technol.
21, 999–1007, 2007.)
As shown in Figure 18.3, the overlapped C 1s peaks exhibit a steady growth of

the shoulder (~286 eV) of the sp3-hybridized C 1s peak for treatment times ranging
from 0 to 90 min. The C 1s spectra curve fitting results are given in Table 18.1 for the
control and 90 min exposure time. Binding energy values reported in the literature
[20,21] were utilized to assign the following peaks with increasing binding energy:
1
0.9
0.8
Normalized intensity
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
296 294 292 290 288 286 284 282
Binding energy (eV)
FIGURE 18.3 (See color insert.) Overlapped C 1s XPS spectra for MWNT paper as a
function of treatment time with O atoms from 0 to 90 min. A steady increase on the shoulder
(~286 eV) of the sp3-hybridized C 1s peak is observed with treatment time as shown in the
direction of the arrow.
TABLE 18.1
XPS C 1s Curve Fitting Results for Untreated and Treated MWNT Paper
Sample for 90 min with O Atoms
Binding Control Control 90 min 90 min
Energy (eV) Assignment Peak Area% Absolute% Peak Area% Absolute%
284.8 C–C sp2 66.4 65.3 61.4 57.5
285.2 C–C sp3 18.2 17.9 13.7 12.8
286.2 O 2.1 2.1 7.2 6.7
—
—
C–O–C, C—C
286.5 Energy loss 2.0 2.0 2.0 1.9
287.2 C=O 1.0 1.0 3.6 3.4
287.7 Energy loss 1.8 1.8 1.5 1.4
288.6 O–C=O 1.5 1.5 2.5 2.3
289.3 O=C–O–C=O, 2.5 2.5 3.4 3.2
O–(C=O)–O
290.8 Energy loss 2.1 2.1 3.0 2.8
292.2 Energy loss 2.4 2.4 2.6 2.4
Total = 98.3 Total = 93.7
C–C sp2, C–C sp3, C–O–C as epoxy and/or ether, C=O, O–C=O, O=C–O–C=O,
and/or O–(C=O)–O. As a result of the variety of species and similar contributions,
the high binding energy region of the spectra has a broad undulating appearance.
The peaks were modeled with different full width at half maximum. The C 1s peaks
due to carbon–carbon bonding were fitted with peaks whose full width at half maxi-
mum was about 0.7 eV, while the peak due to energy loss was fitted with a peak hav-
ing a full width at half maximum of 1.8 eV. The percentage of carbon species was
estimated by curve-fitting the minimum number of peaks necessary to achieve chi-
squared values of 2.0 or less. The absolute percentages of carbon were calculated by
multiplying the C at% from the quantitative analyses with the percent contribution
from the species obtained from the curve fitting. Therefore, the sum of the carbon
concentrations in Table 18.1 equals the concentration of carbon obtained from the
quantitative analyses and thus is not 100%.
18.3.3 Reaction of Cl2 with MWNT Paper

Control experiments were conducted with Cl2 gas in the presence of MWNT paper
for 60 min treatment time. The amount of Cl detected for the two control experi-
ments was 0.84 and 0.85 at%, and the O at% was within experimental error of the
untreated MWNT samples.
18.3.4 Photochlorination of MWNT Paper

Figure 18.4 shows the Cl and O at% on the surfaces of MWNT paper as a function
of treatment time with UV photochlorination. There is a steady increase in the Cl
16
14
12
10
Atomic %
8 at% Cl
at% O
6
0
0 20 40 60 80 100
Treatment time (min)
FIGURE 18.4 Atomic percent of chlorine (⬥) and oxygen (⦁) as a function of treatment time
for MWNT paper samples exposed to UV photochlorination.
TABLE 18.2
XPS C 1s Curve Fitting Results for MWNT Paper Treated 45 min with
UV Photochlorinationa
Binding Energy Peak Area% Assignment Absolute%
284.8 60.2 C–C sp 2 49.8
285.6 11.2 C–C sp , –C =C–C –C
3 * * 9.3
|
Cl
286.2 6.2 C*–CCl2, –C=C*–C–C, ether, glycol 5.1
|
Cl
286.5 5.5 Energy loss 4.6
287.2 7.8 –C=C*–Cl 6.5
288.6 1.6 C–C*Cl2, O–C=O 1.3
Total = 82.9
a Quantitative analysis: 82.9 at% C, 15.1 at% Cl, and 2.0 at% O.
×104
1.5
Counts/s
0.5
298 296 294 292 290 288 286 284 282 280 278
Binding energy (eV)
FIGURE 18.5 (See color insert.) XPS C 1s curve fitting for MWNT paper treated 45 min
with UV photochlorination.
concentration up to a saturation level of ~13 at%. For the initial ~20 min of treatment,
the O at% is within the experimental error of the untreated samples, while at longer
treatment times, the amount increases to ~5 at%.
Table 18.2 reports the curve fitting results of the C 1s spectrum (Figure 18.5)
for the treated sample that showed the highest level of chlorination with the O at%
within the error limits of the untreated MWNTs (45 min). The Cl 2p and O 1s spectra
(not shown here) are difficult to curve-fit because the two peaks that appear in the Cl
2p spectrum are not due to different chemical states but arise from two spin orbital
components, 2P3/2 (200.5 eV [20]) and 2P1/2 (202.1 eV [20]), while the O 1s spectrum
shows a single peak without any resolvable structure.
18.4 DISCUSSION
The MWNT paper treated downstream from the MW plasma discharge of Ar–O2 mix-
ture in the absence of photons contains saturation levels of oxidized species (~6.3 at%
O) about the same for the nanotubes exposed to VUV photooxidation (~7.4 at% O).
These detected levels of oxidation are in good agreement with the liquid phase acidic
oxidation of MWNT powder (5–6 at% [22], ~6 at% [23]). The VUV photons emitted
from excited Ar atoms have sufficient energy (11.6 and 11.8 eV) to break carbon–car-
bon bonds and form excited oxygen atoms from the photodissociation of O2 [15], thus
producing more oxygenated functional groups. For SWNT powder, the saturation level
achieved from the MW Ar–O2 plasma discharge without photons was found to be ~15
at% O compared to ~24 at% O for VUV photooxidation [16]. The larger diameter outer
shells of the MWNTs probably account for their decreased reactivity relative to the
smaller diameter SWNTs, which possess greater curvature (i.e., strain).
Oxygen atoms are well known to undergo addition to unsaturated sp2 C bonds
to form a triplet biradical, which may then follow one of three pathways: (1) ring
closure to an excited epoxide, (2) rearrangement to an excited carbonyl compound,

or (3) decomposition into free radicals [24]. The epoxide/ether groups are predomi-
nantly observed in this study (Table 18.1), VUV photooxidation of MWNTs [15] and
SWNTs [16], and the oxidation of graphite [25,26]. The reaction of oxygen atoms
with SWNTs in the absence of VUV radiation showed more carbonyl moiety than
the epoxide/ether groups [16].
The functionalization of MWNTs by ball milling in reactive atmospheres, such
as Cl2, H2S, NH3, etc., is an effective method for the large-scale production of func-
tionalized nanotubes [9,10], while the photochemical procedure used in this study
may be applicable for nanoelectronic devices where control is needed of the metallic
and semiconductive properties of the functionalized CNTs. The rate of chlorina-
tion of MWNT paper with Cl2 in the absence of photons was shown to be slow
relative to photochlorination where Cl atoms are photochemically produced. Like
oxygen atoms, chlorine atoms readily add across a π-conjugated system like CNTs
to form chlorine-containing radicals that are then able to react with either Cl2 to
form a dichloro-moiety or O2 to generate ultimately chlorine-containing oxygenated
products [27]. The curve fitting results show a drop in sp2 hybridization (Table 18.2,
49.8 Absolute%) compared to the control sample (Table 18.1, 65.3 Absolute%). The
observed binding energies of Cl 2P3/2 and Cl 2P1/2 at 200.5 and 202.1 eV, respectively,
indicate that chlorine atoms are covalently bonded to sp2 carbon [9,28]. In addition,
the weak energy loss structure due to π → π* around 291 eV [9,10] in the C 1s spec-
trum (Figure 18.5) shows only little change in sp2-bonding after the reaction with
chlorine atoms. Photochlorination produces a steady increase in covalently bonded
Cl with treatment time up to a saturation level of ~13 and ~5 at% O (Figure 18.4). The
oxygen may come from impurities or reaction of free radical sites with oxygen in air
when the sample is removed from the photochemical cell. Table 18.2 shows some of
the chlorinated species that may account for the observed C 1s XPS spectrum after
UV photochlorination.
Since this work was presented at the “Special Symposium on Recent Advances
in Adhesion Science and Technology in Honor of Dr. Kash Mittal” at the 240th
American Chemical Society National Meeting, Boston, MA, August 22–26, 2010,
theoretical density function calculations were performed on the structural and elec-
tronic properties of chlorinated SWNTs [29]. The chemisorption of Cl atoms to
SWNTs was found to be an exothermic process that showed strong aromaticity with
a significant increase in the highest occupied to lowest unoccupied molecular orbital
(HOMO–LUMO) energy gap, hence making the chlorinated nanotubes promising
candidates for use in nanoelectronics [29].
18.5 CONCLUSIONS
Two electron-withdrawing groups, oxygen and chlorine atoms, were reacted with
MWNT paper and studied by XPS. The oxygen atom reaction yielded approximately
the same level of oxidation (~6.3 at%) as VUV photooxidation, which was conducted
in the presence of radiation (~7.4 at% O). In contrast to Cl2 molecules, chlorine atoms
produced by UV photolysis readily reacted with the π-conjugation of the nanotubes
to yield saturation levels of ~13 at% Cl.
ACKNOWLEDGMENTS
The authors gratefully acknowledge help from R. A. DiLeo and Dr. B. J. Landi
for the supply of MWNT paper and their support and encouragement during the
research.
REFERENCES
1. L. Cai, J. L. Bahr, Y. Yao and J. M. Tour, Chem. Mater. 14, 4235–4241 (2002).
2. M. Yudasaka, M. Zhang and S. Iijima, Chem. Phys. Lett. 374, 132–136 (2003).
3. R. F. Khairoutdinov, L. V. Doubova, R. C. Haddon and L. Saraf, J. Phys. Chem. B 108,
11976–11981 (2004).
4. M. Alvaro, P. Atienzar, P. de la Cruz, J. L. Delgado, H. Garcia and F. Langa, Chem. Phys.
Lett. 386, 342–345 (2004).
5. K. N. Kudin, H. F. Bettinger and G. E. Scuseria, Phys. Rev. B 63, 045413/1–045413/8 (2001).
6. K. A. Park, Y. S. Choi, Y. H. Lee and C. Kim, Phys. Rev. B 68, 045429/1–045429/9 (2003).
7. E. T. Mickelson, C. B. Huffman, A. G. Rinzler, R. E. Smalley, R. H. Hauge and J. L.
Margrave, Chem. Phys. Lett. 296, 188–194 (1998).
8. H. Hu, B. Zhao, M. A. Harmon, K. Kamaras, M. E. Itkis and R. C. Haddon, J. Am.
Chem. Soc. 125, 14893–14900 (2003).
9. R. Barthos, D. Mehn, A. Demortier, N. Morciaux, G. Demortier, A. Fonseca and J. B.
Nagy, Carbon 43, 321–325 (2005).
10. Z. Konya, I. Vesselenyi, K. Niesz, A. Kukovecz, A. Demortier, A. Fonseca, J. Delhalle,
Z. Mekhalif, J. B. Nagy, A. A. Koos, Z. Osvath, A. Kocsonya, L. P. Biro and I. Kiricsi,
Chem. Phys. Lett. 360, 429–435 (2002).
11. G. Sun, S. Liu, K. Hua, X. Lv, L. Huang and Y. Wang, Electrochem. Commun. 9, 2436–
2440 (2007).
12. K. R. Moonoosawmy and P. Kruse, J. Am. Chem. Soc. 130, 13417–13424 (2008).
13. J. D. Harris, R. P. Raffaelle, T. Gennett, B. J. Landi and A. P. Hepp, Mater. Sci. Eng. B
116, 369–374 (2005).
14. R. A. DiLeo, B. J. Landi and R. P. Raffaelle, J. Appl. Phys. 101, 064307/32–064307/36
(2007).
15. M. Krysak, B. Parekh, T. Debies, R. A. DiLeo, B. J. Landi, R. P. Raffaelle and G. A.
Takacs, J. Adhesion Sci. Technol. 21, 999–1007 (2007).
16. L. Oliveira, T. Debies and G. A. Takacs, Mater. Res. Soc. Symp. Proc. 1142, JJ10/01–
JJ10/06, Warrendale, PA (2009).
17. H. Okabe, Photochemistry of Small Molecules. John Wiley & Sons, New York (1978).
18. U. Sener, A. Entenberg, B. Kahn, F. D. Egitto, L. J. Matienzo, T. Debies and G. A. Takacs,
in: Polyimides and Other High Temperature Polymers: Synthesis, Characterization and
Applications, K. L. Mittal (Ed.), Vol. 3. VSP/Brill, Leiden, pp. 535–552 (2005).
19. M. Krysak, A. Jayasekar, B. Parekh, L. Oliveira, T. Debies, K. S. V. Santhanam, R. A.
DiLeo, B. J. Landi, R. P. Raffaelle and G. A. Takacs, in: Polymer Surface Modification:
Relevance to Adhesion, K. L. Mittal (Ed.), Vol. 5. VSP/Brill, Leiden, pp. 125–137
(2009).
20. G. Beamson and D. Briggs, High Resolution XPS of Organic Polymers: The Scienta
ESCA300 Database. Wiley, Chichester, UK (1992).
21. J. F. Moulder, W. F. Stickle, P. E. Sobol and K. D. Bomben, in: Handbook of X-ray
Photoelectron Spectroscopy, J. Chastin and R. C. King, Jr. (Eds.). Physical Electronics,
Eden Prairie, MN, p. 216 (1995).
22. S. Banerjee and S. S. Wong, J. Am. Chem. Soc. 125, 10342–10350 (2003).
23. Y. Xing, L. Liang, C. C. Chusuei and R. V. Hull, Langmuir 21, 4185–4190 (2005).
24. B. J. Finlayson-Pitts and J. N. Pitts, Atmospheric Chemistry. Wiley & Sons, New York
(1986).
25. N. A. Asrian, G. N. Bondarenko, G. I. Yemelianova, L. Y. Gorlenko, O. I. Adrov, R.
Marassi, V. A. Nalimova and D. E. Sklovsky, Mol. Cryst. Liquid Cryst. Sci. Technol.
A340, 331–336 (2000).
26. J.-L. Li, K. N. Kudin, M. J. McAllister, R. K. Prud’homme, I. A. Aksay and R. Car,
Phys. Rev. Lett. 96, 176101/1–176101/4 (2006).
27. B. J. Finlayson-Pitts and J. N. Pitts, Jr., Chemistry of the Upper and Lower Atmosphere.
Academic Press, London, UK (1999).
28. A. F. Perez-Cadenas, F. J. Maldonado-Hodar and C. Moneno-Castilla, Carbon 41, 473–
478 (2003).
29. D.-L. Wang, H.-L. Xu, Z.-M. Su, S. Muhammad and D.-Y. Hou, ChemPhysChem 13,
1232–1239 (2012).
19 Nanoparticle Deposition
on Packaging Materials
by Liquid Flame Spray
Generation of
Superhydrophilic and
Superhydrophobic Coatings
Hannu Teisala,* Mikko Tuominen, Mikko Aromaa,
Milena Stepien, Jyrki M. Mäkelä, Jarkko J. Saarinen,
Martti Toivakka, and Jurkka Kuusipalo
CONTENTS
19.1 Introduction................................................................................................... 332
19.2 Experimental................................................................................................. 333
19.2.1 Materials............................................................................................ 333
19.2.2 PPS Roughness.................................................................................. 334
19.2.3 Water Contact Angle Measurements................................................. 334
19.2.4 FEG-SEM Imaging........................................................................... 334
19.2.5 Coating Amount................................................................................ 334
19.2.6 Yield................................................................................................... 335
19.2.7 Surface Chemistry............................................................................. 335
19.3.1 Characterization of Substrate Materials............................................ 336
19.3.2 Coating of Paperboard Substrate....................................................... 336
19.3.3 Coating of Paper Substrate................................................................ 337
19.3.4 Coating of LDPE Substrate............................................................... 338
19.3.5 Coating Amount and Yield................................................................ 339
19.3.6 Line Speed......................................................................................... 339
19.3.7 Surface Chemistry.............................................................................340
* Dr. Hannu Teisala was the lead author for this chapter and can be contacted at hannu.teisala@tut.fi.
331
19.4 Conclusions.................................................................................................... 342

References............................................................................................................... 343
Superhydrophilic and superhydrophobic nanoparticle coatings were generated

on fiber-based packaging materials using liquid flame spray. The coating was
deposited in a continuous roll-to-roll process on paperboard, paper, and poly-
ethylene at line speeds of up to 150 m/min. The coatings were studied by water
contact angle measurements, field emission gun scanning electron microscopy,
inductively coupled plasma mass spectrometry, and X-ray photoelectron spec-
troscopy. The mechanisms of superhydrophilicity and superhydrophobicity on
the surfaces as well as the processing parameters required for the coating are
discussed. The amount of the coating on the substrates was no higher than 45
mg/m2, and the yield in the process was estimated to be as high as 22.5%.
19.1 INTRODUCTION
Wettability of surfaces is closely related to many converting processes of packaging
materials, including coating, laminating, gluing, printing, etc. Proper wettability of
the surface maximizes the liquid–solid contact area and thus creates a framework
for strong mechanical and chemical bonding at the interface between the liquid and
the solid, and is therefore essential for good adhesion in, for example, extrusion
coating [1]. From the point of view of printability, appropriate wettability of the sub-
strate is needed for sharp impression and adequate ink adhesion. Furthermore, the
surface properties of the substrate should be adjustable depending on the ink proper-
ties, for example, viscosity and surface tension. This is important especially in more
demanding applications, such as printed electronics [2,3].
Sometimes low wettability is a desired property. For example, on nonsticky sur-
faces or waterproof wrappings, water repellency is a benefit. In addition, fiber-based
materials such as paper and paperboard easily lose their stiffness and strength prop-
erties after wetting. A hydrophobic coating would thus contribute to their maintain-
ing the mechanical properties also in wet conditions.
Flame and corona discharge treatments are traditional methods for improving
surface wettability of various materials in roll-to-roll processes. They are widely
used, for example, in extrusion coating to improve adhesion between polymer and
paper or paperboard. These well-established methods affect only the topmost layer
of the surface, and the improved wettability is mainly based on the oxidation of the
surface. In addition, some microroughening of the surface may also occur during the
flame and corona treatments [1,4,5].
A shortcoming of the flame and corona treatments is their limited capability to
affect the surface chemistry. Therefore, they are mainly used for improving the wet-
tability of substrates. Moreover, the effect of these traditional surface treatments
is sometimes not enough. With special polymers, such as biopolymers, adequate
wettability and adhesion are always an issue because the target is toward thinner
coatings and higher line speeds. Therefore, more efficient roll-to-roll techniques for
Nanoparticle Deposition on Packaging Materials by LFS 333
surface modification, for example, atmospheric plasma [4] and nanoparticle deposi-
tion [6–8], have been developed in recent years. These novel techniques enable both
hydrophilicity and hydrophobicity improving surface treatments.
In general, surface wettability can be affected by changing the chemistry of
the surface by bringing either hydrophilic or hydrophobic groups onto the surface.
Depending on the surface chemistry, greater roughness of the surface can further
increase either hydrophilicity or hydrophobicity [9–12]. Chemical modification of
the rough surface can dramatically change its wettability from perfect hydrophilicity
to extreme hydrophobicity and vice versa [13–15]. On rough hydrophilic surfaces,
water is able to penetrate into cavities and pores of the substrate, and water spreading
is enhanced, whereas on rough hydrophobic surfaces, air can get entrapped between
a water droplet and a rough solid, leading to a formation of a triple phase contact line.
Surface treatments or coating processes generating extremely hydrophilic or
hydrophobic surfaces are quite often multistep and batch-like processes, which lim-
its their utilization in online processing of materials. If a process takes place at low-
pressure conditions, the investment costs grow significantly. On the other hand, the
substrate itself may also set limits for the treatment method. For example, paper and
paperboard are flexible and versatile materials, but processing of these materials is
challenging because of their rough and porous surface structure, complex surface
chemistry, and limited thermal resistance.
Recently, we introduced a method called liquid flame spray (LFS) for generating
nanoparticle coatings in an atmospheric roll-to-roll process [6–8] and demonstrated
creation of transparent and highly hydrophobic titanium dioxide nanocoating onto
a paperboard surface. In the present study, silicon dioxide (SiO2), titanium dioxide
(TiO2), aluminum oxide, and zirconium oxide nanoparticle coatings were gener-
ated using LFS technique onto various types of packaging materials. Wettability of
the coated substrates varied from extreme hydrophilicity to extreme hydrophobic-
ity. Both the chemistry and the topography required for the highly hydrophilic and
hydrophobic coatings were created in a one-step process.
19.2 EXPERIMENTAL
19.2.1 Materials
The LFS coating was performed at Tampere University of Technology (TUT) on
the Paper Converting and Packaging Technology (PCPT) pilot line. A detailed
description of the LFS coating procedure can be found in our previous studies [6–8].
The combustion gases were hydrogen and oxygen with flow rates of 50 and 15 L/
min, respectively. Precursors for the SiO2, TiO2, aluminum oxide, and zirconium
oxide coatings were tetraethyl orthosilicate (98% pure, Aldrich), titanium tetraiso-
propoxide (97% pure, Aldrich), aluminum sec-butoxide (97% pure, Alfa Aesar), and
zirconium isopropoxide (97% pure, Alfa Aesar), respectively. The precursor for the alu-
minum oxide coating was dissolved in xylene, while the solvent for the other precursors
was isopropanol. The concentration of all the precursor solutions was 50 mg (atomic
metal)/mL. The substrate materials were pigment coated paperboard (200 g/m2, Stora
Enso, Skoghall mill, Sweden), machine glossed paper (83 g/m2, UPM, Valkeakoski
TABLE 19.1
LFS Processing Parameters
Concentration Feed Rate Burner Distance Line Speed
Notation (mg atomic metal/ml) (ml/min) (cm) (m/min)
1 50 32 15 50
2 50 12 15 50
3 50 12 6 50
mill, Finland), and low-density polyethylene (LDPE) coated paper (15 g/m2 of LDPE
on the machine glossed paper). The LFS processing parameters are presented in
Table 19.1, and the related notations for the different sets of parameters (1–3) are used
throughout the article. Unless mentioned otherwise, the line speed was 50 m/min.
19.2.2 PPS Roughness
Parker Print-Surf (PPS) roughness was measured from five different points of each
substrate using H.E. Messmer Ltd. (London, England), model MK 2, equipment.
The measurements were carried out according to the ISO 8791-4:1992 and SCAN-P
21:67 standards.
19.2.3 Water Contact Angle Measurements

The contact angle (CA) measurements were done with distilled water using KSV
CAM200 equipment. Static CAs were measured 3 s after droplet deposition (volume
of 2 μL). In 3 s, the vibration of the droplet had ceased, while water evaporation and
penetration into the substrate had not yet affected the droplet volume or the CA. At
least five droplets were used for determining the static CA for each surface.
19.2.4 FEG-SEM Imaging

A field emission gun scanning electron microscope (FEG-SEM), Zeiss ULTRAplus,
was used for imaging the nanoparticle coatings. Due to their resistive nature, the
samples were sputter coated with thin carbon or gold films prior to the FEG-SEM
imaging.
19.2.5 Coating Amount

Inductively coupled plasma mass spectrometry (ICP-MS) was used for chemical
quantification of the deposited TiO2 mass on the substrates. Samples (2 × 30 cm) were
cut from the centerline of the coating stripe, dissolved in nitric acid, and injected into
a plasma, from which the mass spectrum of Ti-ions was quantified (SFS-EN ISO
17294-1:2006 and SFS-EN ISO 17294-2:2005 [TL25]). The amount of TiO2 coating
per square meter was then obtained by normalizing the number of detected Ti-atoms
with the surface area of the samples [8].
The width of the coating stripe was estimated by CA measurements on the SiO2
coated paperboard surfaces because even a small number of SiO2 particles on the
substrate noticeably increase hydrophilicity of the surface (wettability of SiO2 dif-
fers more from paperboard surface than that of TiO2; see Results and Discussion).
Because the size of the SiO2 and TiO2 particles was of the same order of magnitude,
particle spread and coating width were assumed to be similar for both particle mate-
rials. CAs were measured in a cross direction toward the coating stripe with 1 cm
gaps between the droplets. The width of the SiO2 coating stripe, where hydrophilic-
ity had clearly increased, was approximately 15 cm. On this area, there were a large
number of particles. A slighter increase in hydrophilicity was observed on the edges
of the stripe, indicating that the particles were distributed at least in a 25 cm wide
area [5]. The width of the coating stripe was similar regardless of the processing
parameters used (parameter sets 1 or 3; see Table 19.1).
19.2.6 Yield
The yield of nanoparticles deposited on the substrates, that is, the number of particles
deposited compared with the total number of particles generated, was evaluated on
the basis of the ICP-MS results and the estimated coating width. A very rough model,
where the amount of coating was assumed to be uniform at 15 cm wide stripe, was
used to quantify the total amount of deposited coating material. It is likely that the
amount of coating decreased toward the edges of the stripe, but, on the other hand,
because there were particles also outside the 15 cm stripe, the model is thought to be
valid for a rough estimation of the yield.
19.2.7 Surface Chemistry
Surface chemistry was studied by x-ray photoelectron spectroscopy (XPS), using a
PHI Quantum 2000 instrument equipped with a monochromatic Al Kα x-ray source
operated at 25 W. Survey spectra were recorded with a pass energy of 184 eV from
three different points. The take-off angle was 45° in relation to the sample surface,
and the charge compensation was enhanced by combination of electron flood and
ion bombarding.

In general, metal oxides are of high surface energy and should therefore attract mate-
rials of lower surface energy such as water. This also applies to TiO2 and SiO2,
which are known to be hydrophilic materials. Static water CAs on smooth and clean
TiO2 and SiO2 surfaces are generally reported to be between 0° and 40°. However,
surfaces exposed to ambient conditions are hardly ever completely clean, because
adsorption of molecules from the surrounding atmosphere begins instantly after
cleaning the surface. Because of organic molecules that have adsorbed onto the sur-
face and decreased its energy, CA on smooth TiO2 surface is usually reported to
be around 80°. In comparison, on smooth SiO2 surface, CA remains much lower,
only around 30°−40° [16,17]. Hierarchical roughness of the LFS-made nanoparticle

coatings [6] further enhances the hydrophobicity and hydrophilicity of the TiO2 and
SiO2 surfaces, respectively; thus CAs of over 160° and below 10° can be achieved on
several LFS coated substrates.
Various types of substrates can be coated using the LFS method, including
paperboard, paper, and polymers with varying thicknesses and surface rough-
nesses. The thickness of LFS-generated coatings is typically less than 1 μm [6].
In order to obtain extreme hydrophilicity or hydrophobicity, different processing
parameters (Table 19.1) are required depending on the properties of the substrate.
By controlling the processing parameters and the number of nanoparticles on the
substrate [6,8], surface wettability can be adjusted precisely to any level between
the extremes.
19.3.1 Characterization of Substrate Materials

Before the LFS coating, water CAs on the pigment coated paperboard, machine
glossed paper, and LDPE were 77°, 120°, and 98°, and PPS roughnesses of the sub-
strates were 1.3, 4.7, and 0.9 μm, respectively. In comparison with other substrates,
the paper surface had significantly higher roughness on the micrometric scale
because of its fibrous surface structure. This surface also had the highest initial CA,
and it is likely that some air pockets remained trapped between the rough substrate
and the water droplets.
19.3.2 Coating of Paperboard Substrate

The physical structures of the SiO2 and TiO2 nanoparticle coatings on the paper-
board substrate resembled each other (Figure 19.1). Both the SiO2 and TiO2
nanoparticles were highly agglomerated creating a hierarchically structured coat-
ing that fully covered the substrate. CA was measured to be as high as 156° on
the TiO2 coated substrate, while it was only 12° on the SiO2 coating prepared with
the same parameters (parameter set 1; see Table 19.1). The opposite wettabilities
of the two surfaces are related to their different surface chemistries as discussed
earlier in this section.
(a) (b) (c)
200 nm 200 nm 200 nm
2 µm 2 µm 2 µm
FIGURE 19.1 FEG-SEM images of the pigment coated paperboard before (a) and after the
SiO2 (b) and TiO2 (c) nanoparticle coating (parameter set 1).
19.3.3 Coating of Paper Substrate

The high roughness of the paper substrate led to uneven distribution of nanoparticles
on the LFS coated surface. Particles were prone to attach onto the highest areas of
the surface, that is, on the top of the fibers. Figure 19.2e clearly illustrates that on
the SiO2 coated (parameter set 2; Table 19.1) paper surface, only a few particles have
deposited onto the areas between the fibers. Although the cavities of the surface are
not fully coated, the hydrophilic SiO2 particles significantly improved the wettability
of the paper surface (CA = 7°) by enabling water to spread over the fibers and fill the
cavities between them.
Contrary to the SiO2 particles, the TiO2 particles did not have any significant
effect on the wettability of the paper surface when only the top areas of fibers were
well coated. That is, when the TiO2 coating of paper was carried out with the pro-
cessing parameter set 1 (Table 19.1), which turned the relatively smooth pigment
coated paperboard surface highly hydrophobic (Figure 19.1c; CA = 156°), the areas
between the fibers were not sufficiently coated and water was able to fill some of the
cavities. Therefore, the CA of 127° on the TiO2 coated (parameter set 1; Table 19.1)
paper surface was almost the same as it was before the LFS coating (120°). However,
also the cavities of the paper surface can be coated sufficiently by utilizing a high
velocity of the gas flow. The flow velocity of the gases is on the order of hundreds of
meters per second immediately after the burner face, but slows down quickly with
the increasing distance from the torch head [8]. Therefore, with the shorter burner
distance of 6 cm (parameter set 3; Table 19.1) instead of 15 cm (parameter sets 1 and
2), the forced gas flow transported the particles more efficiently into the cavities of
the paper surface and enabled particle deposition onto the areas between the fibers
(Figure 19.2f). As the grooves and gaps of the surface were sufficiently coated with
the TiO2 particles, the water spreading and penetration into the micrometric cavities
(a) (b) (c)
100 µm 100 µm 100 µm

(d) (e) (f )
200 nm 200 nm 200 nm
2 µm 2 µm 2 µm
FIGURE 19.2 FEG-SEM images of the paper surface before (a and d) and after the SiO2 (b
and e) and TiO2 (c and f) nanoparticle coating (parameter set 2 for the SiO2 and parameter set
3 for the TiO2 coating).
was reduced, and hence the hydrophobicity of the surface increased significantly
(CA = 164°).
19.3.4 Coating of LDPE Substrate

The LDPE surface partly melted during LFS processing, and therefore, some of the
particles penetrated into the polymer. The level of penetration seemed to vary across
the surface (Figure 19.3), so that there were areas with a large number of particles
on the top of the surface, but, on the other hand, there were also areas where all the
particles had penetrated into the polymer. In addition, wrinkled structures can be
observed on the nanoparticle coated LDPE surface. Several factors play a role in the
particle penetration into the polymer and deformation of the surface. For example,
the existence of crystalline and amorphous areas on the polymer surface might have
affected the particle penetration and surface deformation, but also the first roll con-
tacts of the partly melted LDPE surface might have pushed the particles deeper into
the polymer matrix or removed them from the surface.
Although there were some areas on the LDPE surface where all the particles had
penetrated into the polymer, it did not prevent high hydrophobicity (Figure 19.3)
because LDPE was initially a hydrophobic material. CA of 151° was measured on
the TiO2 coated LDPE surface (parameter set 1; Table 19.1). On the other hand,
when the aim is hydrophilicity, it is important that the hydrophobic polymer surface
is covered with SiO2 particles as fully as possible. Because the temperature of the
substrate can be controlled by the processing parameters [6,8], so the melting of
the LDPE surface and the penetration of the particles into the polymer can also be
controlled.
The SiO2 nanoparticles on the LDPE surface, produced with the small feed rate
of the precursor (parameter set 2; Table 19.1), were quite uniformly distributed and
gave a good coverage of the polymer substrate, although some cracking of the sur-
face, presumably caused by the temperature changes and postcrystallization of the
polymer, can be observed (Figure 19.3). The CAs on the SiO2 coated LDPE surfaces,
produced with the low (parameter set 2; Table 19.1) and the high (parameter set 1)
200 nm 1 µm
200 nm
1 µm 10 µm
FIGURE 19.3 FEG-SEM images of the TiO2 (left, parameter set 1) and SiO2 (right, param-
eter set 2) nanoparticle coated LDPE surfaces.
precursor feed rates, were 33° and 48°, respectively. The better wettability of the
former surface can be attributed to better coverage of the substrate by SiO2 particles.
However, with TiO2 particles, the improved coverage of the LDPE surface does not
directly mean higher water repellency. The TiO2 nanoparticle coating generated with
the low precursor feed rate (parameter set 2; Table 19.1) did not increase the hydro-
phobicity of the LDPE surface at all (CA = 91°), which can be explained by the lack
of submicrometer and micrometer scale structures on the surface [6,18].
19.3.5 Coating Amount and Yield

It is obvious that the LFS processing parameters affected the amount of coating
on the substrates, but also the substrate material itself had an effect on the particle
deposition. The amounts of TiO2 coating (parameter set 1; Table 19.1) on the pigment
coated paperboard and LDPE surfaces were 38 and 40 mg/m2 (estimated yields of
7.2% and 7.5%), respectively. Despite the many differences between the two sub-
strates, for example, the melting of the polymer and particle penetration, the yield of
the deposited particles seems to be similar on both substrates.
When the LFS coating was carried out using the shorter burner distance together
with the smaller feed rate (parameter set 3; Table 19.1), the amount of TiO2 on the
pigment coated paperboard was 45 mg/m2, with the estimated yield being as high as
22.5%. The yield of TiO2 on the pigment coated paperboard substrate increased over
200% when the burner distance was decreased from 15 to 6 cm. The result confirms
the conclusion [6] that both the yield and efficiency of the LFS coating improve sig-
nificantly when the burner distance is reduced.
The amount of TiO2 (parameter set 3; Table 19.1) on machine glossed paper sub-
strate was only 30 mg/m2 (estimated yield of 15.0%), indicating clearly lower coat-
ing amount in comparison with the pigment coated paperboard substrate. Substrate
material and its porosity are known to affect the deposition efficiency of nanopar-
ticles also in other contexts. For example, in impactors, the collection efficiency of
particles is higher with rough substrate materials in comparison with smooth ones
[19]. Here, both of the present substrates were actually very rough, which might have
benefited particle deposition. The lower yield of particles on the paper substrate can
be attributed to the fact that the particles had difficulties in getting into the spaces
between the fibers, and therefore, the surface area on which particle deposition
occurred most effectively actually decreased on the fibrous substrate in comparison
with the pigment coated substrate.
19.3.6 Line Speed
The wettability of the nanoparticle coated surfaces as a function of the increasing
line speed is illustrated in Figure 19.4. The line speed was the only varied parameter
during the coating. Although the coating process was not optimized for the high line
speeds, both high hydrophilicity and hydrophobicity of the surfaces were obtained
with the line speed of 150 m/min. The only exception was the SiO2 coated LDPE sur-
face, on which the hydrophilicity clearly decreased with the increasing line speed.
It is likely that the decrement in the number of SiO2 particles on top of the LDPE
90 180
80 170
70 160
60 150
CA (°)
CA (°)
50 140
40 130
30 120
20 110
10 100
0 90
50 100 150 50 100 150
Line speed (m/min) Line speed (m/min)
1 Pigment coated paperboard 1 Pigment coated paperboard
2 Fibrous paper 3 Pigment coated paperboard
2 LDPE coated paper 3 Fibrous paper
1 LDPE coated paper
FIGURE 19.4 CAs on the LFS coated substrates as a function of the line speed. (SiO2
coatings on the left, TiO2 coatings on the right.) Indexes refer to the processing parameters
(Table 19.1).
substrate was more significant than it was on the other substrates because (1) small
feed rate of the precursor was used (parameter set 2; Table 19.1) and (2) particles
were able to penetrate into the polymer.
Especially, the hydrophilicity on the paperboard and paper substrates remained
well with the increasing line speed. This is hardly surprising, because the hydro-
philicity of a substrate can be increased by fewer particles than hydrophobicity, as
was seen on the paper substrate (Figure 19.2). Surface chemistry of SiO2 is more
hydrophilic than that of paperboard, paper, and LDPE substrates, and therefore, even
a monolayer of SiO2 particles improves hydrophilicity of the substrates noticeably.
With the TiO2 coating, the hydrophobicity improvement is based more on the changed
physical structure of a surface than on the changed chemistry, and thus more than
a monolayer of particles are needed to create the particle aggregates responsible for
the hydrophobicity improvement [6,18]. Here, the benefit of the short burner distance
usage (parameter set 3; Table 19.1) became relevant, because despite the low precur-
sor feeding, CAs of over 150° were achieved on the TiO2 coated paperboard substrate
even at the line speed of 150 m/min.
19.3.7 Surface Chemistry
The XPS analysis of the SiO2 and TiO2 coated paperboard surfaces (Figure 19.5)
revealed important chemical differences, which were responsible for the different
wettabilities of the two coatings. First of all, the XPS spectra peak positions for Si2p
at the binding energy of 103.0 eV and for Ti2p a double peak at 457.9 and 463.6 eV
indicate that both silicon and titanium on the surfaces were in dioxide form [20,21].
The relative concentrations of the most common chemical elements on the surfaces,
that is, oxygen (O), carbon (C), silicon (Si), and titanium (Ti), were obtained from
the spectra. Here, a significant difference between the two coatings is observed in
the relative number of carbon atoms, which was clearly higher on the surface of TiO2
(45.9%) in comparison with the SiO2 (28.2%). When the amount of carbon is com-
pared with the amounts of titanium and silicon, the difference between the surfaces
was even greater: C/Si ratio on the SiO2 coating was 1.5 and C/Ti ratio on the TiO2
coating was 3.9.
× 104
5
-O1s
SiO2 SiO2
4.5 C1s = 28.2%
TiO2
O1s = 52.3%
4 Paperboard Si2p = 19.0%
3.5 TiO2
Signal intensity (c/s)
C1s = 45.9%
3 O1s = 39.6%
-Ti2p3
Ti2p = 11.9%
-O KLL
2.5
-O KLL
-Ti2s
2
-C1s
-Na KLL
1.5
-Ca2p3
-Si2p
-Si2s
1
-Ti3s
-Al2p
-Al2s
0.5
0
1100 1000 900 800 700 600 500 400 300 200 100 0
Binding energy (eV)
FIGURE 19.5 XPS survey spectra of the SiO2 and TiO2 nanoparticle coatings on the pig-
ment coated paperboard (parameter set 1) and relative concentrations of the most common
elements on the surfaces.
The XPS and CA measurements confirmed that the surface of SiO2 nanoparticles
remained moderately free of carbonaceous substances through the LFS coating pro-
cess, and thus the SiO2 nanocoatings showed hydrophilic properties. In contrast, the
high number of carbon atoms on the TiO2 surface indicated that a carbonaceous layer
from, for example, hydrocarbons, was formed on the TiO2 particles during the coat-
ing process, making the TiO2 nanoparticle coatings eventually hydrophobic. That is,
both the hydrophilic and hydrophobic coatings were created in a one-step process,
and no further alteration of the surface chemistry was needed.
The different amounts of carbonaceous material on the TiO2 and SiO2 nanopar-
ticle coatings can be explained with the dissimilar surface chemistries of the two
metal oxides, which seem to be related especially to hydroxyl groups on the surfaces.
It is known that hydroxyl groups existing on metal oxides are responsible for many
phenomena on the surfaces, including wettability [16,17,22]. Takeda et al. [16] found
a correlation between the hydroxyl group density, the amount of organic contamina-
tion, and the CA on surfaces of various metal oxides, and pointed out that hydroxyl
groups might serve as good adsorption sites for organic molecules. Later, Kanta et al.
[17] studied contamination and wettability of hydroxylated and dehydroxylated SiO2
and TiO2 surfaces, and the results gave strong support for the conclusions of Takeda
et al. The aforementioned studies strongly suggest that the higher hydroxyl group
density on TiO2 in comparison with SiO2 [16,17,22] was the reason for the higher
adsorption of carbonaceous material on TiO2 surfaces.
Our recent studies [5,23] confirmed that both LFS-made SiO2 and TiO2 coat-
ings were able to maintain their high hydrophilicity and hydrophobicity with time.
Changes in the surface wettability of the coatings were not observed during a storage
period of 6 months. The unchanged wettability during the storage can be explained
by the fact that neither of the freshly prepared SiO2 or TiO2 nanoparticle coatings
were clean metal oxides (Figure 19.5); their chemistries were rather comparable to
contaminated surfaces.
The LFS-made superhydrophobic TiO2 coating can be converted to superhy-
drophilic by modifying the surface chemistry, for example, by removing carbona-
ceous material from the photocatalyst surface using ultraviolet (UV) illumination.
However, the photoinduced hydrophilicity of the TiO2 coating is not a permanent
state because a new layer of carbonaceous material will adsorb on TiO2 after clean-
ing the surface by UV. More information on the photoinduced wettability conversion
on the LFS-made TiO2 coating can be found elsewhere [23,24].
Similar to TiO2, also aluminum oxide and zirconium oxide have high surface
hydroxyl group density [16,22], which indicates, according to the previous discus-
sion, existence of a carbonaceous layer on the particles and thus the hydrophobicity
of the coatings. LFS-made TiO2, aluminum oxide, and zirconium oxide coatings
(parameter set 1; Table 19.1) on the paperboard substrate were all highly hydropho-
bic with CAs of 156°, 159°, and 156°, respectively. On the SiO2 coating, prepared
with the similar parameters as the hydrophobic coatings, the CA was only 12°. This
result is in good agreement with the discussion above and supports the conclusions
of Takeda et al. [16].
19.4 CONCLUSIONS
Superhydrophilic and superhydrophobic nanoparticle coatings were generated on
paperboard, paper, and LDPE using LFS. Different types of materials were able to
be LFS coated simply by adjusting the processing parameters. The coating was car-
ried out in a continuous roll-to-roll process at high line speeds up to 150 m/min. TiO2
coated surfaces were extremely hydrophobic with water CAs of over 160°, whereas
SiO2 coating showed high hydrophilicity with CAs of less than 10°. Different surface
chemistries of the two coatings were the main reason for their dissimilar wettabil-
ity behaviors, and the high surface roughness of the coatings further enhanced this
difference.
The amount of TiO2 coating on the substrates was 45 mg/m2 or less, depending on
the processing parameters and the substrate material. The particle deposition and the
efficiency of the coating process could be controlled by controlling the processing
parameters. When the burner distance was decreased from 15 to 6 cm, the particle
deposition increased over 200%, leading to a surprisingly high yield of 22.5% of
TiO2 nanoparticles on the paperboard substrate.
The LFS is an efficient method for controlling wettability behavior of various
types of fiber based packaging materials. Therefore, the LFS coatings may offer
completely new possibilities in further processing of these materials. The LFS coat-
ing combines the excellent properties of fiber based materials, for example, flexibil-
ity, recyclability, and renewability, with the enormous prospects of nanomaterials.
ACKNOWLEDGMENTS
Tekes (Finnish Funding Agency for Technology and Innovation) is acknowledged for
the financial support of this study. The work was done in the Functional Materials
2007−2013 programme, under the project called Liquid Flame Spray Nanocoating
for Flexible Roll-to-Roll Web Materials. The authors would also like to thank Mrs.
Mari Honkanen (TUT, Department of Materials Science) for the FEG-SEM images.
REFERENCES
1. M. Tuominen, M. Vähä-Nissi, and J. Kuusipalo, in: Paper and Paperboard Converting,
J. Kuusipalo (Ed.), 2nd edition. Paperi ja Puu Oy, Jyväskylä, p. 13 (2008).
2. H. Sirringhaus, T. Kawase, R.H. Friend, T. Shimoda, M. Inbasekaran, W. Wu, and E.P.
Woo, Science 290, 2123 (2000).
3. U. Caglar, M. Keskinen, J. Valkama, E. Kunnari, K. Kaija, and M. Mäntysalo, in:
Proceedings of the IMAPS Advanced Technology Workshop and Tabletop Exhibition
on Printed Devices and Applications, Orlando, Florida, 4 p. (2009).
4. M. Tuominen, J. Lahti, J. Lavonen, T. Penttinen, J.P. Räsänen, and J. Kuusipalo,
J. Adhesion Sci. Technol. 24, 471 (2010).
5. M. Tuominen, H. Teisala, M. Aromaa, M. Stepien, J.M. Mäkelä, J.J. Saarinen, M. Toivakka,
and J. Kuusipalo, J. Adhesion Sci. Technol. doi: 10.1080/01694243.2012.697744 (2012).
6. H. Teisala, M. Tuominen, M. Aromaa, J.M. Mäkelä, M. Stepien, J.J. Saarinen, M.
Toivakka, and J. Kuusipalo, Surf. Coat. Technol. 205, 436 (2010).
7. M. Stepien, J.J. Saarinen, H. Teisala, M. Tuominen, M. Aromaa, J. Kuusipalo, J.M.
Mäkelä, and M. Toivakka, Appl. Surf. Sci. 257, 1911 (2011).
8. J.M. Mäkelä, M. Aromaa, H. Teisala, M. Tuominen, M. Stepien, J.J. Saarinen, M.
Toivakka, and J. Kuusipalo, Aerosol Sci. Technol. 45, 827 (2011).
9. M. Ma, and R.M. Hill, Curr. Opin. Colloid Interface Sci. 11, 193 (2006).
10. P. Roach, N.J. Shirtcliffe, and M.I. Newton, Soft Matter 4, 224 (2008).
11. D. Quéré, Annu. Rev. Mater. Res. 38, 71 (2008).
12. J. Bico, U. Thiele, and D. Quéré, Colloids Surfaces A 206, 41 (2002).
13. C. Dorrer, and J. Rühe, Adv. Mater. 20, 159 (2008).
14. A. Nakajima, K. Hashimoto, and T. Watanabe, Langmuir 16, 7044 (2000).
15. X. Zhang, M. Jin, Z. Liu, D.A. Tryk, S. Nishimoto, T. Murakami, and A. Fujishima,
J. Phys. Chem. C 111, 14521 (2007).
16. S. Takeda, M. Fukawa, Y. Hayashi, and K. Matsumoto, Thin Solid Films 339, 220 (1999).
17. A. Kanta, R. Sedev, and J. Ralston, Langmuir 21, 2400 (2005).
18. H. Teisala, M. Tuominen, M. Aromaa, M. Stepien, J.M. Mäkelä, J.J. Saarinen, M.
Toivakka, and J. Kuusipalo, Langmuir 28, 3138 (2012).
19. M. Marjamäki, and J. Keskinen, J. Aerosol Sci. 35, 301 (2004).
20. C.D. Wagner, W.M. Riggs, L.E. Davis, J.F. Moulder, and G.E. Muilenberg, Handbook of
X-Ray Photoelectron Spectroscopy. Perkin-Elmer, Eden Prairie, Minnesota (1979).
21. NIST X-ray Photoelectron Spectroscopy Database, Version 3.5 (National Institute of
Standards and Technology, Gaithersburg, Maryland, 2003), http://srdata.nist.gov/xps/.
22. E. McCafferty, and J.P. Wightman, Surf. Interface Anal. 26, 549 (1998).
23. H. Teisala, M. Tuominen, M. Stepien, J. Haapanen, J.M. Mäkelä, J.J. Saarinen, M.
Toivakka, and J. Kuusipalo, Cellulose. doi: 10.1007/s10570-012-9825-y (2012).
24. M. Stepien, J.J. Saarinen, H. Teisala, M. Tuominen, M. Aromaa, J. Kuusipalo, J.M. Mäkelä,
and M. Toivakka, Surf. Coat. Technol. doi: 10.1016/j.surfcoat.2012.08.008 (2012).
(a)
3.5 (iv)
ωc (pendant)
(iii)
3
(iv)
Angular velocity, ω, (Rad/s)

ωc (sessile)
2.5 (iii)
2 trest (pendant)
trest (sessile)
1.5 tactive
(pendant)
1 tactive
(sessile)
0.5 tstill
(i) (ii)
0
0 2 4 6 8 10 12 14
Time (min)
(b) (i)–(ii) (iii) (iv)
Sessile
Pendant
FIGURE 2.2 A single-force datum measurement using a hexadecane drop on a Teflon-

coated silicon surface. The Teflon used was EGC 1700 from 3M. See also the work of Priest
et al. [9]. (a) The applied angular velocity, ω, measured during the experiment as a function
of the measurement time. For a prescribed period, tstill, the drop is left to rest in the station-
ary CAB, after which ω is steadily increased until, at a critical angular velocity, ω c, the drop
begins to slide on the surface. The drop is pinned to the surface during tstill and when ω is
increased until just before ω c is reached; this entire time is termed trest. tactive is the time from
when the CAB is started (i.e., at the end of tstill) to when the drop begins to move. During tactive,
the applied angular velocity of CAB is gradually increased until the drop begins to move. We
consider pendant and sessile drops. (b) Pictures of the drop taken at different stages of the
experiment. From (i) to (ii), no lateral force is applied, and the drop is symmetric and pinned
to the surface; during the active stage (iii), it is deformed but is still pinned to the surface.
Once ω c is reached, the drop slides, and hence, in (iv), we see only part of it in the frame. We
see that the lateral force required to slide the drop is higher when the normal force is lower.
1.2
-CF2
1.0
-C-CFn
-CF-CFn
-CF3
0.8
Intensity (a.u.)
-C-F
0.7
-C-C
0.5
0.3
0.2
296 292 288 284

Binding energy (eV)
FIGURE 5.6 C1s XPS spectrum of an rf-sputtered Teflon film. (From Sarkar, D.K. et al.,
Mater. Lett., 62, 1226, 2008.)
(a)
0.7 2.0
C1s O1s
Intensity (a.u.)
1.6
0.6 1.2
0.8
-CF2
0.4
Intensity (a.u.)
0.5 538 536 534 532 530 528 526
-CH2-CF2
Binding energy (eV)
-C-O
-CF3
0.4
-C-C
-C-Metals
-C-Si
0.3
0.2
295 290 285 280
Binding energy (eV)
(b)
2.5
O1s C1s
Intensity (a.u.)
2.0
0.6
1.5
-C-C
1.0
0.5 0.5
Intensity (a.u.)
538 536 534 532 530 528 526

Binding energy (eV)
-C-O
-CH2-CF2
0.4
-CF2
-CF3
-C-Metals
-C-Si
0.3
0.2
295 290 285 280
Binding energy (eV)
FIGURE 5.15 High-resolution C1s core-level spectra of aluminum surfaces treated with
FAS-17/NaOH ratio of (a) 0.4 and (b) 0.1. Insets show the corresponding high-resolution O1s
spectra. (From Saleema, N. et al., ACS Appl. Mater. Interfaces, 3, 4775, 2011).
(a) (b)
100 100
Hydrolysis temp. 70°C Hydrolysis temp. 80°C

Hydrolysis time 96 h Hydrolysis time 96 h

72 h
50 48 h 50 72 h
24 h
48 h
24 h
0 0
7.5 15 23 46 70 7.5 15 23 46 70
NaOH conc. (%) NaOH conc. (%)
FIGURE 6.5 (a) Cross-cut test results for electroless Cu-metallized ABS hydrolyzed at 70°C
as functions of the NaOH concentration and hydrolysis time. (b) Cross-cut test results for
electroless Cu-metallized ABS hydrolyzed at 80°C as functions of the NaOH concentration
and hydrolysis time.
140
130
120

110
100
90
80
70
60
50
40
30
20
10
0
FIGURE 12.5 Coatability of the edges of various types of MDF panels in terms of film thick-
ness. All panels were equilibrated at 20°C and 30% relative humidity. One half of the panels (right
side) were preheated to 80°C surface temperature using infrared irradiation prior to the electro-
static powder coat application (one layer, primer). Columns A, G, B, H, C: MDF panels containing
conductivity enhancing additive, columns D, F, I, J, E: standard grade MDF panels.
120
110
100
90
80
70
60
50
40
30
20
10
0
FIGURE 12.6 Coatability of the edges of various types of MDF panels in terms of film thick-
ness. All panels were equilibrated at 20°C and 55% relative humidity. One half of the panels (right
side) were preheated to 80°C surface temperature using infrared irradiation prior to the electro-
static powder coat application (one layer, primer). Columns A, G, B, H, C: MDF panels containing
conductivity enhancing additive, columns D, F, I, J, E: standard grade MDF panels.
130
120
110
100
90
80
70
60
50
40
30
20
10
0
FIGURE 12.7 Coatability of the flat front surface of various types of MDF panels in terms of
film thickness. All panels were equilibrated at 20°C and 30% relative humidity. One half of the
panels (right side) were preheated to 80°C surface temperature using infrared irradiation prior to
the electrostatic powder coat application (one layer, primer). Columns A, G, B, H, C: MDF panels
containing conductivity enhancing additive, columns D, F, I, J, E: standard grade MDF panels.
R PEEK
support Copper
plate
2 mm
8 mm 40 mm GF yarn 10 mm
contacted with
conductive silver paint
PP granules PP melt
FIGURE 17.2 Experimental setup for resistance change measurements during the embed-
ding of the CNT-coated GF yarn into PP. The GF yarn is placed on the two polyetherether-
ketone (PEEK) supports, where it is contacted using electrically conductive silver paint and
copper plates. Between the supports, the GF yarns are surrounded by PP granules, which
impregnate the fiber bundle when the melting temperature of the matrix is reached.
CNT coated F
GF yarn
4 mm
15 mm
R 60 mm
PP
matrix
Conductive
silver paint F
FIGURE 17.3 Schematic of a specimen used for mechanical characterization with simulta-
neous resistance change measurement. The CNT-coated GF yarn is embedded in the center
of the specimen pointing in the direction of the applied tensile load.
CNT-based
PP granules coating PP melt
Dynamic percolation
of CNTs within the
coating
Additional wetting of the

GF surface by the coating
GF Interdiffusion of the coating

with the PP matrix
FIGURE 17.7 Schematic of a CNT-coated GF cross section below and above the melting
temperature of the matrix resin. Above the melting temperatures of the PP and the coating,
different phenomena affect the electrical resistance of the CNT-coated GF, as mentioned on
the right-hand side.
(a)
PP matrix
106 Melting temp.
169°C
Melting range
Resistance (Ω) 125°C–174°C
105 Film former

Melting temp.
164°C
Melting range
114°C–169°C
First heating
104
0 50 100 150 200
Temperature (°C)
(b)
6.5 × 104
6.0 × 104
Resistance (Ω)
5.5 × 104
5.0 × 104
Isothermal step, 200°C, 180 min

4.5 × 104
0 50 100 150 200
Time (min)
(c)
6.5 × 104
6.0 × 104
Film former
Resistance (Ω)
121°C
5.5 × 104 Crystallization range
112°C–132°C
PP matrix
5.0 × 104
121°C
Crystallization range
109°C–135°C
4.5 × 104
200 150 100 50 0 –50
Temperature (°C)
FIGURE 17.8 Resistance change of the interphase sensor upon embedding into a PP matrix.
(a) Heating from –20°C to 200°C, (b) isothermal step at 200°C for 180 min, (c) cooling from
200°C to –20°C. Heating and cooling rate is 1°C/min. The hatched areas refer to the thermal
analysis of the film former and PP matrix by DSC measurements, indicating the temperature
ranges for melting and crystallization.
PP matrix
1 2 3
GF
CNT-modified
interphase
FIGURE 17.9 Schematic of a GF cross section with CNT-modified coating serving as inter-
phase sensor. On the right-hand side, different types of interphase-related defects are shown:
(1) GF failure with associated interphase failure, (2) microcracks in the interphase without
GF failure, (3) local debonding at the fiber/matrix interface.
105
Volume resistivity (Ω cm)
Volume resistivity
104 (Ω cm):
2.0E+04
103
102 7.5E+02
101
30
100
10–1 10
Yar
10
nc
oat
20
1.00
ing
0.75
con
0.50
30 g (wt%)
0.25 e coatin
ten
t of th
CN T conten
t (w
t%)
FIGURE 17.11 Dependence of the volume resistivity of the interphase sensor on the yarn
coating content and the CNT content of the coating. The volume resistivity values are based
on the initial resistances of the compression-molded GF yarn/PP composites and represent
averaged values of at least five individual measurements.
40
Resistance change (%)

30
Resistance change (%):
0.000
20
8.000
16.00
10 24.00
32.00
40.00
0 45.00
Yar
10
n coa
20
ting
0.5
0.4
con
0.3 t%)
30 oating (w
of the c
ten
0.2 nt
nte
N T co
t (w
C
t%)
FIGURE 17.12 Dependence of the sensitivity of the interphase sensor during tensile load-
ing on the yarn coating content and the CNT content of the coating. The resistance change
values represent averaged values of at least five individual measurements on the compression-
molded GF yarn/PP composites. All resistance change values were taken at 3 mm displace-
ment during tensile loading of the model composites.
1
0.9
0.8
Normalized intensity
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
296 294 292 290 288 286 284 282
Binding energy (eV)
FIGURE 18.3 Overlapped C1s XPS spectra for MWNT paper as a function of treatment
time with O atoms from 0 to 90 min. A steady increase on the shoulder (~286 eV) of the sp3-
hybridized C1s peak is observed with treatment time as shown in the direction of the arrow.
×104
1.5
Counts/s
0.5
298 296 294 292 290 288 286 284 282 280 278
Binding energy (eV)
FIGURE 18.5 XPS C1s curve fitting for MWNT paper treated 45 min with UV
photochlorination.
M at erials S ci ence
Recent Advances in Adhesion

Science and Technology
The surface of an object is the first thing we see or touch. Nearly
every article or object we encounter at home, in industry, land
transportation, aerospace, or the medical field in some way uses
an adhesive, a sealant, or a decorative coating. Adhesion science
provides the technology and the know-how behind these applica-
tions. Recent Advances in Adhesion Science and Technology in
Honor of Dr. Kash Mittal is dedicated to Dr. Mittal’s outstanding
contributions to the global adhesion community and his achievements
in disseminating the science of adhesion. This Festschrift volume
contains selected papers from the Special Symposium on Recent
Advances in Adhesion Science and Technology held in honor of Dr.
Mittal to commemorate the publication of his 100th edited book.
Written by world-renowned researchers, the papers have been

updated for inclusion in this volume. They offer insight into recent
developments and the significant ramifications to adhesion science
and adhesive technology. Nineteen articles are divided into five
sections: Interfaces, Wettability and Adhesion; Surface Modification of
Polymers; Adhesion Aspects of Bio-Based Materials and
Bio-Adhesion; Adhesives and Their Testing; and Nanomaterials
and Nanocomposites.
Reflecting the multidisciplinary nature of adhesion science, the topics

covered include metal–polymer interfaces and ways to improve
adhesion, lateral force at liquid–solid interfaces, particle adhesion in
pharmaceutical sciences, wood joints formed without use of adhe-
sives, reinforced polymer composites using different fillers, “green”
composites, medium density fiber board surfaces for powder coating,
adhesion aspects in dentistry, E. coli interactions in porous media,
analysis of adhesive behavior in bonded assemblies, soy proteins as
wood adhesives, carbon nanotube–based interphase sensors, and
reaction of multiwalled carbon nanotubes with gaseous atoms.
K15549

Recent Advances in Adhesive Science and Technology - Wojciech Gutowski PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Recent Advances in Adhesive Science and Technology - Wojciech Gutowski PDF

Uploaded by

Copyright:

Available Formats

o f l

Wojciech (Voytek) Gutowski

Boca Raton London New York

CRC Press is an imprint of the

No claim to original U.S. Government works

International Standard Book Number-13: 978-90-04-20174-3 (eBook - PDF)

Section I Interfaces, Wettability and Adhesion

Chapter 1 Study of Metal–Polymer Interfaces: Relevance to Adhesion................ 3

Chapter 2 Lateral Adhesion Force Measurements at the Drop–Surface

Chapter 3 Particle Adhesion in the Pharmaceutical Sciences............................. 41

Chapter 4 The Effects of Nanostructure and Composition of Solid

Chapter 5 A Review on the Fabrication of Nanostructured

Section II Surface Modification of Polymers

Chapter 6 Surface Modification of ABS for Electroless Copper Plating

Chapter 7 Achieving Nanoscale Surface Structure on Polyester Film by

Chapter 8 Aerosol-Based DBD: A Simple Way to Provide Polymers with

Section III Adhesion Aspects of Bio-Based

Chapter 9 Wood Joints Adhesion and Performance in Mechanical Friction

Chapter 10 Effect of Pretreatment of Rice Straw Used as a Bio-Filler in

Chapter 11 “Green” Composites Based on Bacterial Cellulose Produced

Chapter 12 Preparing Medium Density Fiberboard Surfaces for Powder

Chapter 13 Adhesion Aspects in Dentistry: Bonding to Zirconia....................... 227

Chapter 14 Escherichia coli O157:H7 Interactions and Transport in

Section IV Adhesives and Their Testing

Chapter 15 Improved Approaches to Analyze the Nonlinear Behavior of

Chapter 16 Soy Proteins as Wood Adhesives...................................................... 277

Section V Nanomaterials and Nanocomposites

Chapter 17 Carbon Nanotube-Based Interphase Sensor for Structural

Chapter 18 Reaction of Multiwalled Carbon Nanotubes with Gaseous

Chapter 19 Nanoparticle Deposition on Packaging Materials by

• Title of Doctor honoris causa awarded by the Maria Curie-Sklodowska

Dr. Mittal’s outstanding contributions to the global adhesion community and

• Special issue of Colloids and Surfaces A: Physicochemical and Engineering

• Special Symposium on “Surfactant Science and Technology: Retrospects

He is listed in numerous biographical references: American Men and Women

He is also a member of the editorial advisory board of the Encyclopedia of Surface

LIST OF BOOKS EDITED BY DR. KASH MITTAL

47. Particles on Surfaces: Detection, Adhesion and Removal 1995

91. Contact Angle, Wettability and Adhesion, Vol. 5 2008

Section I: Interfaces, Wettability and Adhesion

Section III: Adhesion Aspects of Bio-Based Materials and Bio-Adhesion

Prof. Dr.-Ing. W. (Voytek) Gutowski

Prof. Hanna Dodiuk

Prof. Dr.-Ing. W. (Voytek) Gutowski

Prof. Hanna Dodiuk

Professor Hanna Dodiuk earned her PhD with honors

Mikko Aromaa Romain Créac’hcadec

Frank M. Etzler Christoph Jocham

M. Masudul Hassan S. Kenig

Aisha Leh Uwe Müller

Edith Mäder Anil N. Netravali

Jarkko J. Saarinen Rafael Tadmor

Judith Sinic Martti Toivakka

Mikko Tuominen C. W. M. Yuen

* Feature report dedicated to K. L. Mittal on his 65th birthday.

To enhance earlier concepts, a new architecture of interfaces between metals

Ni(CO)4 + plasma → Ni↓ + 4 CO↑

Amino groups introduced onto CF surfaces by exposing them to the ammonia

CF−NH2 + epoxy−R → CF−NH−CH2−CHOH−R

1.1.2 Interactions at Metal–Polymer Interfaces

1.2 FACTORS INFLUENCING ADHESION

FIGURE 1.1 Dependence of tensile shear strength of polyurethane (PUR)–steel laminates

Transmittance (%) 100

Section III Adhesion Aspects of Bio-Based

1.2 FACTORS INFLUENCING ADHESION

1.2.4 Dynamics at Plasma-Treated Polymer

1.2.8 Influence of Metal–Polymer Interactions on

1.3 NEW PROCEDURES FOR ENHANCING

1.3.4 Introduction of Flexibilizing Spacers or