You are on page 1of 81

The metallurgy of zinc-coated steel

A.R. Marder
1
Department of Materials Science and Engineering, Lehigh University, Bethlehem, PA, 18015-3195, USA
Abstract
The generation of zinc and zinc alloy coatings on steel is one of the commercially most
important processing techniques used to protect steel components exposed to corrosive
environments. From a technological standpoint, the principles of galvanizing have remained
unchanged since this coating came into use over 200 years ago. However, because of new
applications in the automotive and construction industry, a considerable amount of research
has recently occurred on all aspects of the galvanizing process and on new types of Zn coatings.
This review will discuss the metallurgy of zinc-coated steel from a scientic standpoint to
develop relationships to practical applications. Hot-dip zinc coating methods, i.e. batch and
continuous processes, will rst be reviewed along with FeZn phase equilibria and kinetics.
Commercially, the addition of aluminum to the zinc bath results in three important types of
coatings, galvanized, galfan and galvalume, and produces complex reactions at the coating/
substrate interface. FeZnAl equilibrium will be reviewed in the light of recent studies of
solubility and inhibition layer formation and breakdown. The eect of steel substrate
composition on these reactions will also be critically analyzed. The overlay coating formation,
or the coating alloy, is specically chosen for its desired properties. The morphology of the
galvanize, galfan and galvalume coating overlays will be reviewed, as well as the eect of heat
treatment to produce a galvanneal coating. Finally, the eect of the microstructures of these
coatings on the important properties of corrosion, formability, weldability and paintability will
be discussed. #2000 Elsevier Science Ltd. All rights reserved.
Contents
1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193
2. Hot-dip zinc coating methods. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193
Progress in Materials Science 45 (2000) 191271
0079-6425/99/$19.00 # 2000 Elsevier Science Ltd. All rights reserved.
PII: S0079- 6425( 98) 00006- 1
PERGAMON
1
Tel.: +1-610-758-4197; fax: +1-610-758-4244; e-mail: arm@lehigh.edu
2.1. Batch galvanizing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193
2.2. Continuous processing. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 195
3. FeZn phase equilibria and kinetics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 200
3.1. Phase diagram. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 200
3.1.1. Zeta (z) phase . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 202
3.1.2. Delta (d) phase . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 202
3.1.3. Gamma (G
1
) phase . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 202
3.1.4. Gamma (G) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 202
3.2. FeZn phase formation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 202
3.3. FeZn reaction kinetics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 205
4. Interface reactions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 209
4.1. FeZnAl equilibrium . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 209
4.2. Low Al additions (galvanized <1%). . . . . . . . . . . . . . . . . . . . . . . . . . . . 212
4.2.1. Inhibition of FeZn reactions by Al bath additions . . . . . . . . . . . 213
4.2.2. Morphology of interface reactions . . . . . . . . . . . . . . . . . . . . . . . 217
4.2.3. Inhibition layer break-down. . . . . . . . . . . . . . . . . . . . . . . . . . . . 218
4.2.4. Zn(Al) ``outburst'' reactions . . . . . . . . . . . . . . . . . . . . . . . . . . . . 223
4.2.5. Substrate grain size . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 223
4.2.6. Steel solute additions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 225
4.3. High Al additions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 227
4.3.1. 5 wt% Al (galfan) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 227
4.3.2. 55 wt% Al (galvalume) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 229
5. Overlay coating formation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 232
5.1. ZnAl equilibrium. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 233
5.2. Zinc coating types . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 235
5.2.1. Galvanized (<1 wt% Al). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 235
5.2.2. Galfan (5 wt% AlZn) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 238
5.2.3. Galvalume (55 wt% AlZn). . . . . . . . . . . . . . . . . . . . . . . . . . . . 240
6. Heat treatmentgalvannealing. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 244
6.1. Galvanneal microstructure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 245
6.2. Process variables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 250
6.3. Alloy additions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 253
7. Coating properties. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 255
7.1. Corrosion resistance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 256
7.2. Formability. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 258
7.3. Weldability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 261
7.4. Paintability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 262
Acknowledgements. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 263
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 264
A.R. Marder / Progress in Materials Science 45 (2000) 191271 192
1. Introduction
Zinc coatings are predominantly used to improve the aqueous corrosion of steel
by two methods, barrier protection and galvanic protection. In barrier protection,
the zinc coating, which separates the steel from the corrosion environment, will
rst corrode before the corrosive environment reaches the steel. In galvanic
protection, zinc is less noble or anodic to iron at ambient conditions, and will
sacricially corrode to protect the substrate steel, even if some of the steel is
exposed as cut edges or scratches in the coating. Typical processing methods used
in producing zinc coatings include hot-dip galvanizing, thermal spraying and
electrodeposition. This review will be limited to hot-dip galvanizing, i.e. the
immersion of a steel article in a liquid bath of zinc or a zinc alloy, by batch or
continuous processing. The continuous process is more advantageous for coiled
products such as sheet, wire and tube, whereas the batch process is normally used
for bulk products.
In general, prior to immersion in the liquid zinc bath, the steel article to be
galvanized is rst cleaned to eliminate any surface oxide that may react in the zinc
bath. After hot-dipping, in which the steel reacts with the bath forming the
coating, the article is withdrawn, cooled and sometimes subsequently heat treated.
The anatomy of a zinc coated steel part consists of (1) the overlay or coating
alloy, (2) an interfacial layer between the overlay and the substrate steel
containing a series of intermetallic compounds, and (3) the substrate steel. Each of
these regions can be aected by the bath time and temperature, as well as the
chemistry of both the bath and the substrate steel.
It is the purpose of the present review to supplement and update the excellent
reviews of the 1970 s [1, 2]. In recent decades there has been a dramatic increase in
zinc coatings research worldwide led by new applications in the automotive and
building industries and indicated by the number of international conference series,
particularly Galvatech [36] and the TMS [79]. The object of this review is to
provide metallurgists with fundamental knowledge of a complex coating process as
well as to aid those scientists and engineers who are directly involved in the
production and application of zinc coatings. After briey providing a background of
the coating production processes, FeZn phase equilibria will be discussed, followed
by a critical analysis of substrate/alloy bath interface reactions and overlay coating
formation. The post coating heat treatment process will next be reviewed along with
a description of the various commercial coating types and their properties.
2. Hot-dip zinc coating methods
2.1. Batch galvanizing
In the batch galvanizing process the steel article to be galvanized is cleaned,
pickled and uxed prior to immersion. The part is rst degreased usually in an
A.R. Marder / Progress in Materials Science 45 (2000) 191271 193
alkali solution, followed by rinsing in water, pickling in hydrochloric or sulfuric
acid, and rinsing in running water [10]. The two types of conventional practices
used at the present time are the wet process and the dry process. The wet process
involves passing the article through a blanket of molten ux salts on top of the
molten zinc bath to remove impurities from the surface of the steel and to also
keep that portion of the surface of the zinc bath, through which the steel is
immersed, free from oxides. The wet process requires less plant equipment and
space and, because of the strong cleansing action of the ux blanket, it is less
liable to give badly galvanized patches. In general, because of the wiping action,
the wet process tends to produce thinner coatings when the work is withdrawn
through the ux. Any adherent ux is removed by subsequent water quenching.
Alternatively, the ux blanket may be held back so that the article may be
withdrawn through clear zinc. The ux blanket has the following main
functions [10]:
. To clean the surface of the article and the molten zinc so that the zinc and steel
can react.
. To reduce the danger of splattering when wet articles are dipped.
. To wipe the article during withdrawal, producing thinner coatings.
. To reduce oxidation of the surface of the molten zinc and thus reduce ash
formation.
. To prevent burning or overheating when immersing large objects or during
``double dipping''.
. To reduce distortion by preheating thin articles.
The ux blanket can be made of ammonium chloride with additions of cryolite,
a sodium aluminum uoride for higher Al contents, or zinc ammonium chloride.
The ux blanket must ow freely over the bath and remain in contact with the
article as it is immersed. The activity of the ux blanket depends initially on the
ammonium chloride content and with time will become more viscous due to the
formation of high melting point Zn compounds. However, the uidity of the
ux can be maintained for some time by periodic additions of ammonium
chloride.
In the dry process after cleaning, the article is preuxed in an aqueous solution,
dried and then dipped in the molten zinc bath. The ux used is aqueous zinc
ammonium chloride with a small quantity of wetting agent added. The solution
can be mixed in several ways, including special proprietary uxes that are used for
the higher aluminum content zinc coatings. The temperature of the preux
solution may range from room temperature up to 808C. It is essential that the
article be thoroughly dried before immersion in the molten bath. The optimum
condition for drying is about 1208C for up to 5 min and should not exceed 1508C,
when decomposition of the ux occurs. Wet patches on the piece can cause spatter
and splashing that can result in splash marks and bare patches on the nished
part. The explosion hazard is also more acute in dry galvanizing than in wet
galvanizing. Articles that have been preuxed and dried should be galvanized
A.R. Marder / Progress in Materials Science 45 (2000) 191271 194
immediately, as the ux coating picks up moisture from the air and also tends to
oxidize [10].
In both processes, the molten zinc galvanizing bath is maintained at
temperatures between 445 and 4558C and immersion times are in the range of 3
6 min depending upon the thickness of the workpiece. The predominant coating
made by the batch process is galvanized, although some small quantities of galfan
(5% Al) and galvalume (55% Zn) are also made. The time of immersion can be
varied to control the thickness of the coating that consists of ironzinc alloy
phases at the interface along with a top coat of pure zinc. Good cooling control is
also necessary since the zinc can continue to react with the substrate to produce
further alloying and detrimentally aect the properties of the coating.
2.2. Continuous processing
In continuous hot-dip processing, welded coils of steel are coated at speeds of
up to 200 m/min. The ux or CookNorteman line is similar to the batch process
in that the sheet is cleaned and uxed in line prior to immersion and is generally
termed a cold line method. This process relies on a pre-dip cleaning method using
aqueous alkaline degreasing, acid pickling and a nal uxing treatment to prevent
oxidation. In the cold line method, most strip material must be fully annealed
prior to dipping since the process does not provide for a heating and annealing
step.
Most sheet product is made on a Sendzimir-type hot-process continuous coating
line [11]. Initially, the sheet undergoes a pre-cleaning treatment in which rolling
oils, loose soils, surface carbon and iron nes are removed from the cold rolled
strip surface to enhance coating adherence and prevent contaminants such as iron
nes from entering the metal bath. The newer modern lines utilize an alkali brush
system and electrolytic cleaning stage, together with the advantages of a direct
red reducing treatment. An alkali spray and brush section uses sodium hydroxide
with concentrations varying from 1.5 to 2.5% to remove loose soil contaminants
by washing and brushing the steel surface. Afterwards, an electrolytic pre-cleaning
step removes contaminants tightly adhered to the strip surface. By hydrolyzing
water, hydrogen and oxygen molecules are released at the strip surface creating a
bubbling action that eectively blasts the remaining contaminants from the strip
surface. Next, the clean strip passes through a high volume low pressure air
blower to remove moisture from the strip and prevent oxidation.
The steel sheet then enters the Sendzimir Mill cleaning section, Fig. 1, at about
5007608C in a N
2
/H
2
reducing atmosphere that will further reduce residual
organic contaminants and surface oxides. In the process of cold rolling prior to
hot-dip galvanizing, the newly created surface of the sheet is spontaneously
oxidized to Fe-oxide and Fe-hydroxide, and upon heating is completely reduced.
However, selective surface oxidation of the strong oxide forming alloying elements
in the substrate (e.g. Mn, Si, Al, V and Ti) will occur [12]. The total oxide
coverage as a function of the sum of the oxide forming elements is given in Fig. 2.
A.R. Marder / Progress in Materials Science 45 (2000) 191271 195
The oxides do not form a continuous layer on the surface but are present as
islands. The nonuniform distribution is to be expected since the amount available
for the diusion of the elements is so limited that easy diusion paths, such as
grain and subgrain boundaries, must clearly be favored sites for oxidation of the
Fig. 1. Schematic of a typical continuous hot-dip galvanizing line.
Fig. 2. Total surface coverage of the oxides versus the sum of the bulk composition from [12].
A.R. Marder / Progress in Materials Science 45 (2000) 191271 196
reactive elements. It is of interest to note that, despite the formation of discrete
particles at the grain boundaries, there are still other particle free grain boundary
areas that can react with the Zn bath. Fig. 3 shows the equilibrium diagram for
the reactions of the type:
Fig. 3. Equilibrium partial pressure of water pH
2
O for reactions of the type:
x/y M+H
2
O \y
1
M
x
O
y
+H
2
from Ref. [12].
Fig. 4. Typical recrystallization curves for IF steels from Ref. [13].
A.R. Marder / Progress in Materials Science 45 (2000) 191271 197
x=y M H
2
O == y
1
M
x
O
y
H
2
(1)
where M is the cation forming element. The shaded area in the gure represents
the range of temperatures and dew points investigated [12]. From these results it
can be seen that Fe
3
O
4
will be reduced above 1508C at a dew point of 20 to
308C, while the other alloy elements will be oxidized over the entire temperature
range. It should be noted that the equilibrium curves for oxidation of Ti, Si and
Al are beneath the bottom of the gure.
After cleaning, the sheet enters the heating and holding zones of the furnace
where it is annealed above the recrystallization temperature in excess of 7008C.
The specic recrystallization temperature used will depend on the type of steel
being coated. For example, Fig. 4 shows that for a given set of cold rolling,
coiling and annealing temperature (6508C) conditions, the kinetics of
recrystallization for an Al killed steel is faster than steels containing solute
additions of Ti, Ti/Nb or Nb [13]. At faster line speeds the annealing temperatures
would have to be higher to produce the same amount of recrystallization. As a
result of this step in the process the sheet is suciently heated to enter the bath
without aecting the bath temperature.
Fig. 5. Schematic of the pot region in a typical continuous hot-dip galvanizing line.
A.R. Marder / Progress in Materials Science 45 (2000) 191271 198
On exiting the furnace, the strip passes through a gas jet section capable of
cooling the strip temperature at 508C/s to 4608C prior to coating by immersion in
the zinc molten bath. As the strip exits the bath, Fig. 5, the thickness of the
molten metal lm is controlled by gas wiping dies that remove excess coating
metal. After coating, the sheet is either cooled by forced air or subjected to an in-
line heat treatment, called galvannealing, before being rewound into coil or
sheared into cut lengths at the exit of the line.
The bath temperature will depend upon the composition of the coating the bath
alloy melting point and the strip temperature prior to dipping. A galvanized Zn
(<0.3% Al) bath is kept between 445 and 4558C. Because the galfan (Zn5% Al)
composition is a eutectic, a lower bath temperature of 4258C is usually used and
for a galvalume (Zn55% Al) bath, the temperature of the bath is near 6008C.
Bath temperature and dipping time, a result of line speed, are specic operating
parameters that can be optimized to control ecient surface cleaning, strip
heating and minimization of alloy layer growth during the immersion step.
Typical dipping times range between 48 seconds for line speeds in excess of
175 m/min.
Dross in the Zn pot can be classied as oxide types (Zn and/or Al) and
intermetallic compound types (ZnFe and FeAl). The latter type tends to cause
dross problems and form in the Zn pot when Al and Fe are present in the bath in
concentrations above the solubility limits [14]. Specically, the problematic
intermetallic compounds have been identied as Fe
2
Al
5
Zn
x
``top'' or ``oating''
dross and aluminum saturated d (FeZn
7
) phase ``bottom'' dross [15]. Even if
perfect Zn bath chemistry management is maintained, dross crystallization is
unavoidable due to the Al additions to the bath, Fe dissolution from the steel
strip and the insucient temperature uniformity and chemical homogeneity
obtained with conventional galvanizing technology. However, specic bath
management techniques utilizing computer software to measure eective Al have
been suggested to minimize dross formation [16]. Recently, the thermochemistry of
dross formation in the continuous galvanizing process has been reviewed [17].
On exiting from the liquid metal pot, the excess liquid is forced back into the
bath by gas wiping dies that blow either air or nitrogen onto the surface of the
steel, Fig. 2. This step is designed to precisely control and maintain a uniform
coating thickness and is usually monitored by on-line X-ray coating thickness
measuring equipment. As the sheet moves further up the cooling tower, Fig. 1, the
coating can be additionally cooled at this point or, on some lines, nucleation can
be induced by impingement of Zn particles. In the galvanneal process, the cooling
tower is transformed into an annealing furnace either by introducing an induction
furnace above the gas wiping jets and using a good part of the tower height as a
holding furnace, or introducing a gas heating furnace above the gas knives as
depicted schematically in Fig. 1. In either case, producing an optimized galvanneal
product is dependent upon controlling processing parameters such as heating rate,
peak or hold temperature, and cooling rate, as they all aect the amount of
alloying that will occur in the coating prior to reaching the tower roll. At this
point, all of the coating must be solidied to avoid sticking on the tower roll.
A.R. Marder / Progress in Materials Science 45 (2000) 191271 199
Subsequent processing and nishing steps such as roller leveling, temper rolling,
chromating and/or oiling, cutting and slitting, complete the manufacturing
process.
3. FeZn phase equilibria and kinetics
When the substrate steel is immersed in the liquid zinc bath, a number of
reactions occur depending upon the bath composition and the solutes found in the
steel. Before discussing compositional eects, it will be helpful to develop an
understanding of the reactions between iron and zinc.
3.1. Phase diagram
The FeZn system has been the subject of a number of review papers, e.g. [1, 2].
The phase diagram has been modied several times, especially the zinc-rich
section, since it was presented in 1938 [1823]. The most widely accepted ironzinc
Fig. 6. FeZn phase diagram From Massalski TB. Binary Alloy Phase Diagram. ASM International
1986. # 1986 ASM International. [24].
A.R. Marder / Progress in Materials Science 45 (2000) 191271 200
equilibrium phase diagram (Fig. 6) is that of Kubachewski [23]; the zinc portion
of the diagram is seen in Fig. 7. The phases found in the current diagram are
outlined in Table 1. The primary phases formed during long-time immersion
galvanizing (or post-dip annealing) are zeta (z), delta (d), gamma
1
(G
1
) and
gamma (G). Although not represented in the equilibrium diagram (Figs. 6 and 7),
eta (Z) phase is a solid solution of Fe in Zn with an iron solubility of 0.03 wt%.
Although older ironzinc equilibrium phase diagrams show the existence of both a
delta (d) phase and a delta
1
(d
1
) phase, the X-ray analysis of Bastin et al. [21]
indicates that only one delta phase exists up to 6708C.
Table 1
FeZn phase characteristics.
Phases Formula Crystal structure VHN (25 mg) [20] VHN (25 g) [27]
aFe Fe(Zn) BCC 104 86
G Fe
3
Zn10 BCC 326
G
1
Fe
5
Zn
21
FCC 505
d FeZn
10
Hexagonal 358 273
z FeZn
13
Monoclinic 208 118
ZZn Zn(Fe) HCP 52 41
Fig. 7. Zinc rich corner of the FeZn binary phase diagram.
A.R. Marder / Progress in Materials Science 45 (2000) 191271 201
A brief description of each ironzinc intermetallic phase found in hot-dip
galvanized coatings appears below in order of increasing iron content.
3.1.1. Zeta (z) phase
The zeta (z) phase, FeZn
13
, has an iron content of approximately 56 wt% [24].
It is formed from the peritectic reaction between the delta (d) phase and liquid
zinc at 5302108C. During controlled diusion studies in the absence of
aluminum, the zeta (z) phase was found to form between the free zinc eta (Z)
phase and the delta (d) phase. The zeta (z) phase is isomorphous with a
monoclinic unit cell and an atomic structure that contains an iron atom and a
zinc atom surrounded by 12 zinc atoms at the vertices of a slightly distorted
icosahedron. The icosahedra link together to form chains and the linked chains
pack together in a hexagonal array [25].
3.1.2. Delta (d) phase
The delta (d) phase, FeZn
10
, has an iron composition range of 7.011.5 wt%
and a hexagonal unit cell. It is formed from another peritectic reaction, gamma
(G) and liquid, at 6658C. In the past the delta (d) phase was separated out into
two morphologies, delta
1P
, palisade morphology found on the zinc rich side, and
delta
1 K
, compact morphology found on the iron rich side, which was found for
long-term (4 h) and high-temperature (5538C) immersions [2]. Both morphologies
were found to have the same crystallographic structure [26] and are now referred
to as delta (d) phase. For the short-term immersions found in galvanizing, only
one delta (d) phase morphology was reported [27].
3.1.3. Gamma
1
(G
1
) phase
The gamma
1
(G
1
) phase, Fe
5
Zn
21
, has a face centered cubic lattice structure
with an iron composition of 1719.5 wt% at 4508C. It forms as a result of a
peritectoid reaction between the gamma (G) phase and delta (d) phase at
5502108C. The gamma
1
(G
1
) phase appears as an uninterrupted layer between
the gamma (G) and delta (d) layers, and can be produced upon heating at low
temperatures over long periods of time [21]. The Gamma
1
(G
1
) phase has the
highest reported microhardness values, Table 1.
3.1.4. Gamma (G) phase
The gamma (G) phase, Fe
3
Zn
10
, is body centered cubic, and has an iron
composition range of 23.528.0 wt% at 4508C. It forms as a result of a peritectic
reaction at 7828C between a iron and liquid zinc, and exhibits a maximum
solubility of Fe in Zn at the delta (d) phase peritectic temperature of 6658C [28].
3.2. FeZn phase formation
When iron is immersed in molten zinc at the typical galvanizing temperatures
(4504908C), according to the FeZn phase diagram, Horstmann [1, 29] proposed
A.R. Marder / Progress in Materials Science 45 (2000) 191271 202
that the following layers should form: zinc saturated a-iron, gamma (G) phase
layer, gamma
1
(G
1
) phase layer, delta (d) phase layer, zeta (z) phase layer and an
eta (Z) phase layer. However, the sequential nucleation of the FeZn phases
occurs at the interface beginning with (1) the zeta (z) phase layer, followed by (2)
the delta (d) phase layer, and after some incubation time, (3) the gamma (G) phase
layer, e.g. Fig. 8. In light optical microscopy, the gamma (G) phase layer is so
small that the layer is considered to contain both the G and G
1
phases. The FeZn
phase layer development is also shown schematically in Fig. 9, where the reaction
sequence is represented chronologically. Zero time is represented by t
0
, and the
phase development occurs according to time so that t
0
<t
1
<t
2
<t
3
<t
4
. In
studies on iron substrates (0.003 wt% C, 0.258 wt% Mn), it was found that zeta
(z) phase nucleation is immediately followed by delta (d) phase formation (t
2
) at
the a-iron/zeta (z) interface [27]. There was no apparent delay in the formation of
zeta (z) or delta (d) phases, as both were found to form a continuous layer after a
5 s immersion time. On the other hand, the gamma (G) phase was found to form
(t
3
) after an incubation time of 30 s.
A typical morphology found for a pure Zn hot-dipped coating is shown in
Fig. 8. Each of these phases have been conrmed by EPMA measurements of the
iron content [27]. The gamma (G+ G
1
) phases appear as a thin layer with a
planar interface between the steel substrate and the delta (d) phase layer. The
delta (d) phase has a columnar morphology as a result of a preferred growth
perpendicular to the interface in a direction along the (0001) basal plan of the
Fig. 8. Microstructure of Zn coating formed after 300 s immersion in a 4508C, 0.00 wt% Al bath on a
ULC steel substrate. (1) gamma (G) phase, (2) delta (d) phase (3) zeta (z) phase.
A.R. Marder / Progress in Materials Science 45 (2000) 191271 203
hexagonal structure [30]. After time, cracks form along this basal plane of the
delta (d) phase layer, that can extend into the zeta (z) phase layer above and the
gamma (G+ G
1
) phase layer below. The zeta (z) phase has two layers depending
upon the supersaturation of Fe in the melt. Adjacent to the delta (d) phase layer,
the zeta (z) phase grows in a columnar morphology that is supersaturated in Fe.
Continued growth of these crystals occurs rather than the formation of new zeta
(z) phase crystals. However, if the zinc melt is supersaturated with iron, and there
is sucient nucleation of new crystals, numerous tiny zeta (z) crystals can form in
the melt, that are separated from each other by the solidied zinc Z phase [29],
Fig. 8.
Fig. 9. A schematic representation of FeZn phase layer formation in a 0.00 wt% AlZn galvanizing
bath. t
0
corresponds to zero time, and development occurs according to time such that t
1
<t
2
<t
3
<t
4
from Ref. [27].
A.R. Marder / Progress in Materials Science 45 (2000) 191271 204
3.3. FeZn phase reaction kinetics
Each phase layer in the zinc coating exhibits dierent growth kinetics,
depending upon the immersion temperature, that together aect the total layer
kinetics. For example, in short time immersions up to 300 s at 4508C, Fig. 10, the
zeta (z) phase layer grows rapidly at rst, then slows down, while the delta (d)
phase layer grows slowly and after time its thickness increases more rapidly. The
gamma (G+ G
1
) phase layer forms only after long periods of time and seems to
reach a maximum thickness of about 1 mm [27]. Similar eects were reported at
4578C for times up to 6 h [2]. Horstmann [1, 29] reported that there is an overall
inward movement of the gamma (G+ G
1
) phase layer towards the iron, whereas
the zeta (z) phase layer is displaced towards the zinc melt. The delta (d) phase
layer expands in both directions, but generally towards the zinc melt. Thus, as the
gamma (G+ G
1
) phase layer grows into the iron, it is also consumed by the
growing delta (d) phase layer. Similarly, the delta (d) phase layer expands into the
growing zeta (z) phase layer that is advancing into the zinc melt. All of these
transformations are governed by the diusion of Zn into the iron substrate based
on inert marker experiments in both liquid zinc [31] and solid-state diusion
studies [32]. However, Fe probably also diuses outward through the alloy phases
into the zinc melt, but at a much slower rate [31].
To evaluate the kinetics of FeZn alloy layer growth, a power-law growth
equation is generally used to interpret the growth rate data [1], as follows:
Y = Kt
n
(2)
where: Y=growth layer thickness, K=growth rate constant, t =reaction time,
and n=growth-rate time constant.
The growth-rate time constant, n, is an indication of the type of kinetics
controlling the growth of the layer under study. An n value of 0.5 is indicative of
parabolic diusion controlled growth, while an n value of 1.0 is representative of
Fig. 10. Individual FeZn gamma (G) phase, delta (d) phase, and zeta (z) phase layer growth for a
ULC steel substrate hot dipped at 4508C in a 0.00 wt% AlZn bath. From Ref. [27].
A.R. Marder / Progress in Materials Science 45 (2000) 191271 205
linear kinetics in which growth is interface controlled. Table 2 is taken from a
compilation by Mackowiak and Short [2], of the many determinations made in the
past. Unfortunately, many of the studies were made on steel substrates with
signicant additions of carbon and other alloying elements. Most of the studies
were made for long immersion times in excess of 1 h, which is not indicative of
the continuous hot dip process. Nevertheless, it was concluded from these results
that in the lower parabolic range, values of n tend to be around 0.5 for delta (d)
phase layer growth and total layer growth. Values for the zeta (z) phase layer and
gamma (G+ G
1
) phase layer growth were lower, around 0.35 and 0.25,
respectively. In a more recent experiment [27] with short time immersions of up to
300 s at a temperature of 4508C, Table 3, the short time immersion results are
remarkably close to the previous work. The only signicant dierence is found
with the total layer n value. However, the total layer should reect the dominant
alloy layer phase in the total coating. For short times, less the 300 s, the zeta (z)
phase layer dominates the coating morphology [27], whereas for long times, the
delta (d) phase layer dominates the coating structure [33].
The eect of temperature on the kinetics of the reaction between liquid zinc
and pure iron, based on iron loss experiments [34], shows that up to 4958C the
total layer kinetics for long immersion times is parabolic (lower parabolic
region), and it is again parabolic above 5208C (upper parabolic). Linear attack
Table 2
Values of n for growth of alloy layers in the lower parabolic range [2]
Source G d z Total layer
Allen 0.25 0.65 0.35 0.55
Rowland 0.13 0.53 0.31
Blickwede 0.10 0.60 0.16
Horstmann and Peters 0.50 0.50
Sjoukes 0.23 0.58 0.26
Onishi et al. 0.23 0.49 0.36 0.43
Table 3
Individual phase layer growth at 4508C
up to 300 s on a iron substrate (0.003
wt% C) in a pure zinc iron saturated
(0.03 wt% Fe) bath [27]
Alloy layer n value
Gamma (G) 0.2420.06
Delta (d) 0.5120.11
Zeta (z) 0.3220.03
Total 0.3520.02
A.R. Marder / Progress in Materials Science 45 (2000) 191271 206
(linear region) occurs between these two regions, but short time experiments up
to 30 min also show parabolic attack in this region [34]. With increasing
temperature, the zeta (z) phase layer decreases until 4958C, when it is no longer
possible to form a continuous zeta (z) phase layer. At about 5008C, the
nucleation of the zeta (z) phase layer is so slow that this compound no longer
appears in the iron-zinc coating, even though the phase diagram, Figs. 6 and 7,
shows that it is stable up to 5308C [29]. In the lower parabolic range, the delta
(d) phase layer continues to increase parabolically with temperature, until the
range of linear attack, where the delta (d) phase layer only grows to a certain
thickness as the outer portion continues to break away [34] up to the
temperature at which the delta (d) phase layer is no longer stable, 6658C.
Breakaway attack is commonly indicative of linear kinetics. According to the
phase diagram, Figs. 6 and 7, the double layer gamma (G+ G
1
) phase can form
at the interface between the iron substrate and the delta (d) phase layer.
Between 550 and 6658C, only the gamma (G) phase layer is stable along with
the constant thickness delta (d) phase layer. Above 6658C, the delta (d) phase
layer is no longer stable and only a gamma (G) phase layer forms. Since there is
no evidence for aking of the gamma (G) phase layer, this layer probably grows
parabolically in the linear region just as it grows in both the lower parabolic and
upper parabolic regions [29].
Substrate solute additions have been reported to aect the rate of attack of zinc
on steel during galvanizing [2]. However, most of the research has been conducted
on steel substrates containing a multiplicity of solute additions that can confound
the determination of any systematic solute eect. Few studies have been concerned
with pure Zn baths without the addition of Al. In general, small additions of
carbon are found to segregate to grain boundaries, impeding the reaction rate [35],
whereas Mn additions of up to 12 wt% have very little eect on weight loss and
coating microstructure [36]. Interstitial-free (IF) steel [37], produced by the
addition of titanium and/or niobium to an extra low-carbon grade to precipitate
interstitial carbon and nitrogen atoms, is being utilized as a galvanized coated
sheet product [38]. Phosphorous additions are also sometimes added to these steels
to improve mechanical properties. The eect these solute additions on the
formation and kinetics of galvanizing has recently been studied in Al containing
Zn baths [3941].
The eect of FeZn alloy formation in interstitial-free steels galvanized in a
pure Zn (Al free) bath has been investigated [27]. Table 4 shows the total alloy
layer growth-rate time constant, n, for a series of IF steels hot-dip galvanized at
4508C for up to 300 s and Table 5 shows the growth-rate time constant, n, for
the individual FeZn phases formed. It was reported [27] that phosphorous
solute additions retard the kinetics of gamma (G) phase layer growth, but did
not aect the growth kinetics of any other FeZn phases. On the other hand,
titanium and titanium+niobium solute additions had no eect on the growth
kinetics on any of the FeZn phase layers found in the coating. As in the 0.003
wt% C substrate, the IF substrates exhibited a parabolic growth-rate time
constant, n =0.5, for the delta (d) phase layer, indicative of bulk diusion and
A.R. Marder / Progress in Materials Science 45 (2000) 191271 207
the zeta (z) phase layer had a value of n =0.33, that may be related to the
solidliquid reaction at the interface. The gamma (G) phase layer for the 0.003
wt% C substrate and the other IF steels (except for the phosphorous containing
alloys) had a growth-rate time constant of n=0.25, indicative of grain
boundary diusion controlled growth. The phosphorous added alloys reduced
the gamma (G) phase layer n to 0.03520.13, most likely by blocking Zn
diusion down solute grain boundaries and preventing gamma (G) phase
formation [42].
Table 5
Individual FeZn phase layer growth-rate time-constant, n, values for steels hot-dip galvanized in a
0.00 wt% AlZn bath [27]
Sample/layer Growth-rate time constant, n
Gamma-phase layer
ULC 0.2420.06
ULCP 0.1020.08
Ti IF 0.2220.11
TiP IF 0.03520.080
TiNb IF 0.2320.05
TiNbP IF 0.1320.28
Delta-phase layer
ULC 0.5120.11
ULCP 0.4420.12
Ti IF 0.3420.12
TIP IF 0.3720.13
TiNb IF 0.4420.15
TiNbP IF 0.3920.15
Zeta-phase layer
ULC 0.3220.03
ULCP 0.3220.03
Ti IF 0.2820.05
TiP IF 0.2920.03
TiNb IF 0.3320.05
TiNbP IF 0.3320.03
Table 4
Total FeZn alloy layer growth-rate time-constant, n,
values for steels hot-dip galvanized in a 0.00 wt% Al
Zn bath [27]
Sample Growth-rate time constant, n
ULC 0.3520.02
ULC-P 0.3520.03
TiIF 0.3120.02
TiP IF 0.3320.02
TiNbIF 0.3720.03
TiNbP IF 0.3420.03
A.R. Marder / Progress in Materials Science 45 (2000) 191271 208
4. Interface reactions
An understanding of the phase transformations that take place at the liquid
zinc/steel substrate interface is necessary in order to predict and control the
microstructure of galvanized coatings. This understanding is further compounded
by the use of use of zinc baths with additions of aluminum and substrates that
contain deliberately added solute elements to improve substrate steel properties.
These complexities in understanding and the subsequent underlying mechanisms
have been related to three main factors [43]:
. Several reactions are occurring at the same time, including (1) wetting of the
solid substrate by liquid zinc, (2) dissolution of the steel by the zinc, (3)
isothermal solidication of FeAlZn intermetallic compounds, (4) solid state
diusional phase transformations, and (5) solidication of the liquid Zn alloy.
. The speed of the reactions are very fast and in some cases take place in less
than a second.
. The transformation front often becomes unstable and therefore is not governed
by simple equilibrium thermodynamics.
4.1. FeZnAl equilibrium
The rst systematic investigation covering the entire FeZnAl system was
conducted by Ko ster and Go decke [44] by means of metallography observations
and thermal- and X-ray analysis on alloys. Urednicek and Kirkaldy [45] conrmed
these ndings in a study at 4508C by means of microprobe analysis and
metallographic observations on alloys and solidliquid diusion couples,
Fig. 11 [46]. As a result of the increased attention to galvanized steel, more recent
results have focused on the zinc-rich corner of the ternary diagram; in particular,
the solubility of (1) Zn in the various FeAl intermetallic compound, and (2) Fe
in molten Zn.
Perrot et al. [47] studied the metastable equilibrium conditions that arise during
short time (<30 min) immersion in Zn(Al) baths. They found that for these
relatively short times, an FeAlZn ternary isotherm at 4508C generated from
EPMA data, Fig. 12, showed extended solubility ranges when compared to the
previously accepted ternary equilibrium diagram of Urednicek and Kirkaldy [45].
For comparison, a long-term equilibrium phase diagram is also included in Fig. 12.
In particular, Fe
2
Al
5
was found to have a solubility for Zn of up to 22.87 wt%
and 13.92 wt% Zn in FeAl
3
. Similarly, Chen [48] found extended solubility for Zn
in Fe
2
Al
5
to be 21.0 wt%. Long-term equilibrium immersion studies reported that
the Zn solubility in Fe
2
Al
5
was 18.71 wt% [47] or 14 wt% [45]. Perrot et al. [47]
also observed a transient phase for dipping times less than 2 min that had an
average composition of ZnFeAl
3
, that afterwards formed Fe
2
Al
5
crystals saturated
in Zn. The phase corresponded to a previously reported transient phase [49], that
A.R. Marder / Progress in Materials Science 45 (2000) 191271 209
F
i
g
.
1
1
.
I
s
o
t
h
e
r
m
a
l
s
e
c
t
i
o
n
o
f
t
h
e
F
e

A
l

Z
n
p
h
a
s
e
d
i
a
g
r
a
m
a
t
4
5
0
8
C
,
(
a
)
o
v
e
r
a
l
l
s
e
c
t
i
o
n
,
(
b
)
z
i
n
c
r
i
c
h
c
o
r
n
e
r
.
F
r
o
m
R
e
f
.
[
4
6
]
.
A.R. Marder / Progress in Materials Science 45 (2000) 191271 210
F
i
g
.
1
2
.
E
v
o
l
u
t
i
o
n
o
f
F
e

Z
n

A
l
p
h
a
s
e
d
i
a
g
r
a
m
w
i
t
h
t
i
m
e
a
t
4
5
0
8
C
,
(
a
)
m
e
t
a
s
t
a
b
l
e
:
<
3
0
m
i
n
,
(
b
)
e
q
u
i
l
i
b
r
i
u
m
:
>
1
0
0
0
h
.
F
r
o
m
R
e
f
.
[
4
7
]
.
A.R. Marder / Progress in Materials Science 45 (2000) 191271 211
was observed during the rst step of galvanizing steel in ZnAl baths. Chen's [48]
short term immersion results (30 min) also showed that the maximum solubility of
Zn in alpha iron was 2.0 wt% in a Zn bath containing 0 wt% Al. Perrot et al. [47],
also found that the delta (d) phase exhibited an extended solubility range of 3.71
wt% Al and an extended Fe solubility range of up to 16 wt% in a Zn bath
containing 0 wt% Al (normally 11.5 wt%. Fe). Similarly, the eta (Z) phase was
also found to be oversaturated in Fe for short-term immersions.
The determination of bath iron saturation has been studied by a number of
investigators in ZnAl baths [45, 47, 50, 51]. However, in a more recent critical
analysis, an improved zinc rich ZnFeAl ternary equilibrium phase diagram was
developed [52, 53]. Additional studies by Tang et al. [54, 55], on the iron solubility
limits in molten zinc as a function of aluminum content, were conducted and the
intermetallic compounds in thermal equilibrium with the liquid were identied. It
was previously established [23, 56] that the iron solubility limit ([Fe] in wt%) in
pure zinc is a function of temperature ([T] in Kelvin):
ln[Fe] = 17:78 15388=[T] (3)
For ZnAl melts, the following was established at 4608C for the zinc corner of the
ZnFeAl ternary equilibrium phase diagram, Fig. 13 [53, 54]:
. When the bath aluminum content is less than 0.10 wt%, the phase in
equilibrium with the liquid is the zeta (z) phase.
. When the bath aluminum content is between 0.10 and 0.14 wt%, the phase in
equilibrium with the liquid is the delta (d) phase.
. When the bath aluminum content is greater than 0.14 wt%, the phase in
equilibrium with the liquid is the Fe
2
Al
5
Zn
x
(Z) phase.
These results conrm previous data [45, 47, 50, 51], that the solubility limit of Fe
decreases continuously with increasing Al content at a given temperature
isotherm. The eect of temperature on the solubility product, [Fe]
2
[Al]
3
in wt%, in
the Fe
2
Al
5
Zn
x
(Z) phase dominant region satises the following equation:
ln[Fe]
2
[Al]
3
= 32:3 36133=T (4)
4.2. Low Al additions (galvanized <1%)
Low aluminum additions to the Zn bath are deliberately added to form Zn
coated galvanized steel. These additions have been made to (1) improve the luster
or reectivity of the coating, (2) reduce oxidation of the zinc bath, and (3) to
obtain a ductile coating by suppressing the formation of brittle FeZn phases. In
practice, 0.10.3 wt% Al is added to the Zn galvanizing bath to ``inhibit'' FeZn
intermetallic compounds from forming. Inhibition is transitory, and results in an
incubation period that can increase (1) with increasing Al content in the bath, (2)
using low bath temperatures, (3) having low bath iron content, (4) increasing
A.R. Marder / Progress in Materials Science 45 (2000) 191271 212
agitation, (5) increasing the presence of Si in the steel, and (6) decreasing surface
roughness [2]. The complex solidication and solid state diusional reactions
during galvanizing has been reviewed recently [57].
4.2.1. Inhibition of FeZn reactions by Al bath additions
Temporary inhibition of the formation of FeZn compounds can be a result of
the development of a continuous layer of the compound that is in equilibrium
with the Zn bath. Depending upon the Al content of the Zn bath, this can lead to
the formation of a zeta (z), delta (d) or Fe
2
Al
5
phase layer, Fig. 14 [58]. The zeta
(z) phase layer is capable of growing epitaxially on the substrate at an extremely
high rate, but zeta (z) phase will be out of equilibrium with the substrate alpha (a)
iron, according to the ZnFeAl ternary equilibrium phase diagram, Fig. 11 [46].
Therefore, equilibrium of the interfaces will only be reached once all the FeZn
intermetallics have appeared. Delta (d) phase, according to Fig. 11 [46], will be in
thermodynamic equilibrium with the alpha (a) iron substrate, as will an Fe
2
Al
5
phase layer. The origin of the potential inhibition layer compound will depend
upon the Al concentration in the bath, thus it can be seen from Fig. 13 that the
minimum Al content necessary for a the full inhibition eect by Fe
2
Al
5
Zn
x
(Z) is
approximately 0.15 wt% Al at 4508C, that is slightly higher than the
concentration corresponding to the change-over from delta (d) phase to Fe
2
Al
5
Zn
x
Fig. 13. Zinc-rich corner of the 4508C isothermal section of the FeZnAl phase diagram. From
Ref. [54].
A.R. Marder / Progress in Materials Science 45 (2000) 191271 213
(Z) being the thermodynamically stable phase [52]. However, in continous
galvanizing, the transient chemistry in the vicinity of the Zn/substrate steel sheet
interface will deviate signicantly from the bulk bath chemistry and a
thermodynamic equilibrium state is not always established there due to the
extremely short reaction times involved [59]. Table 6 summarizes the coating
microstructures found in continuous galvanizing based on aluminum content [59],
and indicates that only Fe
2
Al
5
Zn
x
(Z) phase is capable of inhibiting the formation
of FeZn intermetallic compounds.
Aluminum control in the Zn bath is complicated by the fact that aluminum
exists in two forms in the bath. Some Al is dissolved in the liquid Zn phase and
the rest of the Al is present in the intermetallic particles entrapped in the bath. It
is the aluminum in liquid solution, commonly referred to as ``active'' or
``eective'' aluminum, that can perform the function of inhibiting the FeZn
reaction during galvanizing [54]. Eective aluminum is also strongly dependent on
Fig. 14. Mechanism of formation of the inhibiting layers. From Ref. [58].
Table 6
Summary of coating microstructure in continous galvanizing [59]
Al content
(wt%)
Equilibrium
compound
Intermetallics
in coating
Alloy layer
characteristics
Nucleation rate (s) Growth rate (mm/s)
<0.100 z z/d/G
/
/G Continuous z110
18
Up to 110
0.1000.135 d z/d/G
/
/G Gaps exist z110
18
11
0.1350.140 Z Mostly z Discontinuous z110
15
z10.5, Z10.05
0.1400.145 Z z plus Z z dissolution z, Z110
12
z10.1, Z10.05
0.1450.150 Z Mostly Z z dissolution Z110
12
Z10.05
>0.150 Z Z Full inhibition Z110
12
Z10.05
A.R. Marder / Progress in Materials Science 45 (2000) 191271 214
the amount of dissolved Fe in the bath, since supersaturated Fe can combine with
Zn and Al to form Al-containing zeta (z), delta (d), and Fe
2
Al
5
Zn
x
(Z) particles,
further reducing the amount of Al in the liquid Zn. Avoiding the entrapment of
intermetallic particles in bath samples can be very dicult, thus determining the
eective aluminum from the bath assay is problematic [57]. Based on the revised
ZnFeAl ternary equilibrium phase diagram [53, 54], Fig. 13, new technologies
for determining eective bath Al, such as the DEAL
TM
computer program along
with Al sensors [59, 60] have been developed. The program calculates the eective
Al content from bath assays of total Al and Fe in Zn. The program also
calculates the dissolved Fe content and predicts the types and amounts of
intermetallic dross particles entrapped in the bath assay sample [54].
The rate of Al uptake in Zn coatings was found to be dependent on strip-entry
temperature and to change rapidly with time [61, 62]. Isobe [61] showed that in the
time required for steel to pass through a bath on a continuous galvanizing line
(<3 s), a signicant amount of the total Al uptake occurs, Fig. 15. The amount of
Al in the AlFe alloy layer increased continuously with Al bath content [63],
Fig. 16. Thus, the incubation period (i.e. the time for FeZn phases to form)
increases with an increase in Al content in the bath and decreasing bath
temperature [64], Fig. 17. Faderl et al. [62] found that for a bath containing 0.18
wt% Al, the Al content in the interface increased approximately linearly with the
dierence between the strip-entry temperature and the bath temperature.
However, long immersion time studies [65, 66], resulted in the growth kinetics of
the inhibition layer following a parabolic time law and that bath Fe contents
inuenced the overall growth kinetics. Tang [55] showed that the formation of the
Fig. 15. Growth kinetics of Fe
2
Al
5
layers at 4708C for various bath Al contents. From Isobe M. Initial
alloying behavior in galvannealing process. CAMP-ISIJ 1992;5:1629. # 1992 The Iron and Steel
Institute of Japan. [61].
A.R. Marder / Progress in Materials Science 45 (2000) 191271 215
inhibition layer is a two-stage process. The rst stage, associated with a high rate
of Al uptake at the coating/substrate interface, is controlled by continuous
nucleation of Fe
2
Al
5
, followed by a second stage diusion-controlled growth
process. He proposed a model in that Al uptake was shown to increase with
increasing strip-entry temperature and thickness, since both work to increase the
eective temperature for nucleation and growth processes. Al enrichment was
Fig. 16. Eect of bath aluminum content on amount of aluminum reaction (bath temperature4608C;
immersion time1 s) from. Ref. [63].
Fig. 17. Eects of bath aluminum content and alloying temperature on incubation period. From
Yamaguchi H, Hisamatsu Y. Reaction mechanism of the sheet galvanizing. Trans ISIJ 1979:19;649. #
1979 The Iron and Steel Institute of Japan. [64].
A.R. Marder / Progress in Materials Science 45 (2000) 191271 216
shown as a function of bath Al content, bath temperature, strip-entry
temperature, strip thickness, immersion time and coating weight.
4.2.2. Morphology of interface reactions
There now appears to exist a transition point at approximately 0.15 wt% Al in
the revised isothermal FeAlZn phase diagram, Figs. 13 and 14, at which point
the morphology of the interfacial layers and particularly the inhibition layer type
changes [54, 58]. For Al contents below this point (low aluminum baths), the
thermodynamically stable phases are zeta (z) phase and delta (d) phase without
any detectable Fe
2
Al
5
. If the Al content exceeds 0.15 wt% Al (high aluminum
baths), Fe
2
Al
5
will be the thermodynamically stable phase. Thus, dierent
morphologies can be found at the interface. In low aluminum baths (<0.10 wt%
Al) at very short dipping times, only a small amount of FeZn phases form at the
interface, Fig. 18 [67], depending upon the substrate composition and bath
temperature. Ohtsubo et al. [68] reported a zeta (z) phase layer saw tooth
morphology, with a specic crystallographic orientation between the zeta (z) phase
crystals and the substrate ferrite:
[001]
z
[[[111]
a
(100)
z
[[(101)
a
(010)
z
[[(121)
a
As the aluminum content is increased to between 0.12 and 0.14 wt%, FeAlZn
ternary phases begin to appear along with zeta (z) phase and/or delta (d)
phase [6971]. The Fe
2
Al
5
layer had a ne granular structure, whereas the
morphology for the zeta (z) phase was ``pillar-like'' and the delta (d) phase was
columnar. In most studies, it is unclear whether the ZnFe phases are nucleated
on top of the Fe
2
Al
5
phase or if the Fe
2
Al
5
phase and ZnFe crystals nucleated
Fig. 18. Microstructure of hot-dip galvanized DQSK steel in a 0.10 wt% AlZn bath # 1993 Elsevier
Science Ltd. [67].
A.R. Marder / Progress in Materials Science 45 (2000) 191271 217
side-by-side at the steel interface surface. In more recent results, an in-situ HVEM
study showed the existence of both z phase and d phase at the interface between
the FeAl inhibition layer and liquid zinc containing 0.2 wt% Al [72]. The relative
amounts of zeta (z), delta (d) and Fe
2
Al
5
phases, are dependent upon the substrate
steel reactivity, bath temperature and immersion time.
Increasing the Al content or decreasing the bath temperature, increases the
stability of the FeAl inhibition layer. In baths containing in excess of 0.15 wt%
eective Al, a continuous FeAl inhibition layer forms upon immersion. The Fe
Al layer was found to be essentially the orthorhombic Fe
2
Al
5
Zn
x
(Z) containing
approximately 23 wt% Zn [43]. Guttmann et al. [43], found that the layer had an
overall thickness of 250 nm and was made up of two sublayers. The lower layer in
contact with the substrate was a compact layer of very small, roughly equiaxed,
closed packed crystals, having a diameter of the order of 60 nm. The crystals are
arranged in colonies of specic crystallographic orientation whose boundaries
delineate the underlying steel grain boundary. Several variants can exist within the
area of one substrate grain. The orientation relationship between the substrate
steel and the Fe
2
Al
5
layer is reported to be [57]:
(311)
Fe
2
Al
5
[[(110)
a
or
(221)
Fe
2
Al
5
[[(110)
a
Tang and Adams [52] found that the Fe
2
Al
5
Zn
x
compound had a growth
direction of [100]. The thicker upper layer on the coating side of coarser,
pancaked shaped crystals, approximately 300600 nm in diameter and
approximately 200 nm in thickness, exhibits random crystalline orientations and
morphologies. These ndings can help to explain the two-stage growth kinetics
observed previously by Isobe [61], Fig. 15 and Al uptake by Tang [55].
. During the initial growth period (<1 second), the lower layer colonies grow to
meet each other forming a compact layer. This very rapid interface controlled
reaction probably exhibits linear kinetics.
. The second growth period (>2 seconds) corresponds to the formation of the
upper, coarse crystal layer. The slower kinetics may reect control by solid state
diusion of Fe through the already formed lower layer.
4.2.3. Inhibition layer breakdown
The inhibition of FeZn reactions is always transient. Al delays the FeZn
reaction rather than suppressing it completely and eventually FeZn ``outbursts''
form, Fig. 19 [73]. Outbursts appear to occur in association with the inhibition
layer [45], since it has been shown that normal linear interface growth occurs in
the absence of Al in the Zn bath [27]. For example, outbursts were not seen for a
low carbon steel (0.006 wt% C) dipped in a 0.12 wt% Al bath at 4608C after
A.R. Marder / Progress in Materials Science 45 (2000) 191271 218
immersion times of up to 10 s, however, the appearance of outbursts were
observed after 3 s in a 0.16 wt% Al bath [71]. Even thick layers of Fe
2
Al
5
are
destroyed after suciently long holding times in high aluminum baths [74, 75].
According to the ternary equilibrium phase diagram, liquid Zn with more than
0.12 wt% Al is in principle saturated with respect to Fe
2
Al
5
and this phase should
not redissolve in the liquid after it is formed. Thus the transient character of
inhibition cannot be explained by thermodynamic equilibrium between Fe
2
Al
5
and
the liquid. In fact, lumps of Fe
2
Al
5
inhibition layer have been found scattered
throughout an FeZn outburst as if the outburst had caused the Fe
2
Al
5
to break
apart by nucleating beneath the inhibition layer [74, 75]. In addition, outbursts
have been shown to nucleate at substrate grain boundaries [74, 76], Figs. 20 and
21.
Numerous models have been proposed to explain the breakdown of the
inhibition layer:
1. Enrichment of the inhibition layer [49, 65]Al or Zn enriches the Fe
2
Al
5
inhibition layer until the solid solubility limit is reached, initiating the
transformation of the saturated inhibition layer into other phases. This
mechanism is unable to account for the mechanical breakup of the inhibition
layer [43].
2. Depletion of aluminum in the liquidlower Al adjacent to the inhibition layer
results in a dissolution of the Fe
2
Al
5
inhibition layer and the subsequent
Fig. 19. Typical outburst formation in a 0.20 wt% AlZn bath immersed at 4508C for 3600 s. From
Ref. [73].
A.R. Marder / Progress in Materials Science 45 (2000) 191271 219
formation of more stable FeZn phases [45, 66, 76]. Guttmann [57] also showed
by diusion path analysis that this model was unable to predict the FeZn
phase growth rate and morphologies that would result from Al depletion.
Nishimoto et al. [76] proposed that rapid precipitation of Fe
2
Al
5
at the
substrate grain boundaries would produce a greater Al depletion at the
boundaries than in front of the grain interiors, Fig. 21. Model measurements
show that Al depletion would be minimum and would be replenished by the
bath Al content [43]. The reduction of surface oxides by Al dissolved in the
bath has also been proposed as mechanism for Al depletion [77], however,
oxide reduction by Al was experimentally shown not to take place [43, 78].
3. Diusion of zinc through the inhibition layerZn atoms reach the substrate and
cause nucleation and growth of FeZn intermetallic compounds at the Fe
2
Al
5
/
substrate interface. Guttmann [43, 57] has proposed that inhibition layer break
down occurs by Zn diusion down short circuit paths in the Fe
2
Al
5
inhibition
layer, as original suggested by Hisamatsu [74]. In reported studies on thick
Fig. 20. Schematic diagram showing the FeZn outburst growth behavior. From Hisamatsu Y. Science
and technology of zinc and zinc-alloy coated sheet steel. GALVATECH '89. Tokyo: The Iron and Steel
Institute of Japan. 1989. p. 3. # 1989 The Iron and Steel Institute of Japan. [74].
A.R. Marder / Progress in Materials Science 45 (2000) 191271 220
Fe
2
Al
5
, Guttmann [43] estimated that the average Zn diusion coecient
would be 510
11
cm
2
/s
1
. This implies that Zn could diuse across a 70 nm
thick Fe
2
Al
5
layer in 1 s and a 250 nm layer in 3 s at 4608C. It was found that
appreciable intergranular diusion took place, thus real inhibition layers in
galvanized steel can be traversed under usual galvanizing conditions. These
results support the hypothesis that Fe
2
Al
5
destabilization results from Zn
diusion through the Fe
2
Al
5
layer along short circuit paths [57]. When Zn
reaches the substrate, it reacts with Fe, nucleating FeZn intermetallic phases
at the Fe
2
Al
5
inhibition layer/substrate interface that bursts the layer apart into
the surrounding bath. Short circuit paths through the Fe
2
Al
5
are expected to be
the grain boundaries of the pancake structure layer or interfaces between the
surface particles and the layer, Fig. 22.
In a recent study [78] iron oxide was used as a simulation inhibition layer to test
the Guttmann hypothesis [43, 57]. In these experiments, a pure Zn bath (0.00 wt%
Al) was used in order to prevent the confounding eect of the formation of an
additional Fe
2
Al
5
inhibiting layer. The iron oxide initially acted as a physical
Fig. 21. Schematic diagram of the formation of FeZn phase growth at substrate grain boundaries.
From Nishimoto A, Inagaki J, Nakaoka K. Trans ISIJ 1986;26:807. # 1986 The Iron and Steel
Institute of Japan. [76].
A.R. Marder / Progress in Materials Science 45 (2000) 191271 221
barrier for Zn diusion, but eventually outbursts form beneath the oxide layer
broke through the layer, causing sections to break away and become incorporated
into the coating, Fig. 23. Increased immersion time for a constant oxide layer
thickness led to an increase in the number of outbursts. Cracks and other
microdefects in the oxide acted as fast diusion paths for Zn, simulating the short
circuit diusion mechanisms for Fe
2
Al
5
inhibition layer break down in Al-
containing Zn baths [43, 57].
Fig. 22. Short circuit Zn diusion path through Fe
2
Al
5
. From Ref. [43].
Fig. 23. Outburst formation in a 0.00 wt% Al bath with a simulated iron oxide inhibition layer. From
Ref. [78].
A.R. Marder / Progress in Materials Science 45 (2000) 191271 222
4.2.4. Zn(Al) ``outburst'' reactions
The diusion of Zn through the inhibition layer can account for the emergence
of outbursts at steel grain boundaries [74]. Guttmann proposed that the Fe
2
Al
5
diusion short circuits would coincide with emerging substrate grain
boundaries [43]. Because of the Fe
2
Al
5
/substrate steel orientation relationship, the
high angle boundaries between colonies of the inhibition layer could be coincident
with random substrate grain boundaries. Oxide/Fe
2
Al
5
interface short circuits may
also be preferentially situated at substrate steel grain boundaries. Thus, both the
eect of steel solute additions on ferrite grain texture and grain size can aect the
rate of inhibition layer breakdown or outburst formation. The fast diusion path
mechanism could also depend on cracking of the inhibition layer, as was
demonstrated for the iron oxide simulated inhibition study [78]. Growth stresses
that result, for example, from epitaxial constraints due to mismatch between the
inhibition layer and the substrate, may cause cracking. Although relief of stresses
in the inhibition layer may be accommodated at high temperature as a result of
grain boundary sliding, creep or dislocation climb, at lower temperatures the relief
of the growth stresses could occur by fracture. This process would enable liquid
zinc to attack the substrate at the grain boundary faster than Zn atoms diusing
down a grain boundary.
4.2.5. Substrate grain size
Hisamatsu [74] suggested that a ner grain size substrate will be more reactive,
since more grain boundary area is available for reaction with the liquid zinc bath
on a ne grain size surface resulting in more rapid FeZn phase growth. In a
recent study [79], the substrate grain size was systematically changed and the steel
sheet was galvanized at 4508C for varying times in Zn baths with 0.00 and 0.20
wt% Al. Uniform attack of the substrate occurred in the Zn0.00 wt% Al bath,
since the Fe
2
Al
5
inhibition layer did not form and thus no barrier to FeZn
nucleation existed. The results showed that grain size did not aect the kinetics
(growth rate time constant, n) of FeZn phase formation for the 0.00 wt% Al
Table 7
Total alloy layer and individual FeZn phase layer growth rate time constant (n) values for the 15- and
85-mm grain size ULC steel galvanized in a 0.00 wt% AlZn bath [79]
Sample/layer Growth rate time constant, n
15-mm ULC/total FeZn alloy layer 0.3520.02
85-mm ULC/total FeZn alloy layer 0.3320.21
15-mm ULC/gamma layer 0.2420.06
85-mm ULC/gamma layer 0.2620.02
15-mm ULC/delta layer 0.5120.11
85-mm ULC/delta layer 0.4320.18
15-mm ULC/zeta layer 0.3220.03
85-mm ULC/zeta layer 0.4020.08
A.R. Marder / Progress in Materials Science 45 (2000) 191271 223
bath, Table 7. In the 0.20 wt% Al baths, outburst formation readily occurred due
to the initial formation of the Fe
2
Al
5
inhibition layer; however, increased grain
size signicantly retarded incubation time and promoted Fe
2
Al
5
inhibition layer
stability, Fig. 24. An SEM analysis of an 85 mm grain size substrate with outbursts
at grain boundaries [79, 80], Fig. 25, showed that the Fe
2
Al
5
inhibition layer was
faceted and cracked, conrming the Zn short circuit diusion mechanism through
the Fe
2
Al
5
inhibition layer [43, 57].
Fig. 24. Incubation time for outburst formation as a function of substrate steel grain size (log scale).
Ref. [79].
Fig. 25. Outbursts in an 85 mm grain size ULC steel immersed in a 0.20 wt% AlZn bath for 1800 s.
From Ref. [80].
A.R. Marder / Progress in Materials Science 45 (2000) 191271 224
4.2.6. Steel solute additions
According to Hisamatsu [74], FeZn reactivity at grain boundaries of the
substrate will depend upon the ability of the solute element to segregate to these
sites. These alloying additions can be separated into elements that segregate to the
grain boundary, (e.g. C, P), and those that will form compounds (e.g. Ti, Nb) that
will precipitate throughout the grain, leaving the grain boundary ``pure'' or clean.
Clean grain boundaries will have no barrier to FeZn compound formation, while
segregated boundaries will reduce the thermodynamic activity at these sites,
decreasing outburst and Fe
2
Al
5
inhibition layer break down.
. Carbonit is well known that outburst formation decreases with increased
carbon content, since carbon segregates to the grain boundary [75].
. Phosphorousthe growth of FeZn phases is retarded by phosphorous
additions [81]. Phosphorous was found to segregate to ferrite grain boundaries
in rephosphorized low carbon steels, blocking the diusion of Zn along grain
boundaries and lowering the thermodynamic activity [82, 83]. It has also been
proposed that P segregates to the steel surface during recrystallization
annealing, stabilizing the inhibition layer and retarding the rate of FeZn phase
growth reactions during galvanizing [84, 85]. A recent study ion implanted P on
one surface of a large grain (1020 mm) low-carbon steel sheet galvanized in a
Zn0.02 wt% Al bath and showed that FeZn phase growth only occurred
Fig. 26. Outbursts at the same grain boundary in a phosphorous ion-implanted (P) side and a
nonimplanted side (NP). From Ref. [86].
A.R. Marder / Progress in Materials Science 45 (2000) 191271 225
after extended reaction times on both the P ion implanted and nonimplanted
surfaces [86]. The FeZn outbursts were found to directly correspond to the
location of a single grain boundary site, Fig. 26, and that a greater FeZn
reaction occurred on the P-free side (NP), since there was no P segregation in
the grain boundary. It was shown that substrate grain boundary cleanliness is
the dominant steel substrate structural feature that controls the kinetics of Fe
Zn alloy phase formation in aluminum containing zinc baths. Recently, these
results were conrmed with a study of phosphorous additions in IF-P steels.
Outbursts were rst found after 100 h in the 0.105 P and was related to a
Sandelin eect [87].
. Interstitial free (IF) alloy additionsIF steels can contain additions of titanium
and/or niobium at extremely low carbon levels. It is well known that TiIF
steels are more sensitive to outburst formation than Al-killed or even NbTi
steels [74]. Because these steels are more likely to form carbide, nitride, sulde
and phosphide precipitates in the grains, preventing segregation to the grain
boundaries, the concept of excess titanium, Ti**, has been introduced [39].
Ti
++
= Total Ti 3:99C 1:49S 3:42N 1:55P (5)
A positive Ti** indicates excess titanium and therefore a clean, carbide free and
reactive grain boundary. A negative value of Ti** would indicate that not all of
the solute carbon is tied up and zinc diusion down the boundaries would be
blocked.
Fig. 27. Comparison of iron losses for dierent immersion times and dierent Si contents in the steel.
From Ref. [89].
A.R. Marder / Progress in Materials Science 45 (2000) 191271 226
. Silicon (Sandelin eect)the presence of small concentrations of Si (about 0.1
wt%) leads to linear growth in which the usual uniform attack in a pure zinc
bath characterized by FeZn layers is replaced by a mass of zeta (z) phase
crystallites surrounded by liquid Zn (see [1]). At short times (23 min) the
layered structure of the coating is observed, independent of the Si content of the
steel [88] and the growth kinetics are parabolic during this period. At the end of
this incubation period, the morphology, growth kinetics and iron loss, Fig. 27,
change for Sandelin steels with 0.070.12 wt% Si. Recently, a novel approach
to this long misunderstood phenomenon has been proposed [89] in that
thermodynamics and ternary diusion are used to explain the eect.
The kinetics of FeZn phase formation in Zn coatings containing an Fe
2
Al
5
inhibition layer (Zn0.20 wt% Al bath) has recently been studied [90] for a series
of IF steels and contrasted to coatings without an inhibition layer (Zn0.00 wt%
Al bath [27]). FeZn phase formation followed the sequence outlined in Fig. 28.
Delta (d) phase formed rst, followed by gamma (G) phase, however, zeta (z)
phase did not form with this bath composition. Ti and Ti +Nb solute additions,
which enhance grain boundary reactivity, resulted in more rapid growth kinetics
of the gamma (G) and delta (d) phases than an ultra-low carbon steel. P additions,
which decreased grain boundary activity, increased incubation time and retarded
the growth rate of the gamma (G) phase, Table 8. These results along with the
0.00 wt% Al bath data [27], were explained by the use of the excess Ti**
calculation proposed by Toki et al. [39], as plotted in Fig. 29 [80]. Grain boundary
reactivity is not important in the formation of both delta (d) and zeta (z) phases.
However, gamma (G) phase growth is retarded by P additions, conrming the
concept of segregation of P at grain boundaries that prevent the FeZn
reaction [82].
4.3. High Al additions
4.3.1. 5 wt% Al (galfan)
The galfan composition corresponds to the zinc aluminum eutectic (5 wt% Al)
with additions of up to 0.05 wt% mischmetal. A modied composition, superzinc,
substitutes 0.1 wt% Mg for the mischmetal. One of the objects in producing a
galfan coating was to develop a zinc-rich coating without intermetallics at the
coating/steel interface for improved formability. Surprisingly, little is known of the
inhibiting layer composition, but many investigators indicate the presence of an
Fe
2
Al
5
Zn
x
after short dipping times, which suggests a mechanism of inhibition
layer break down similar to galvanized coatings. At short dipping times at low
bath temperatures, Ghuman and Goldstein [49] showed no reaction occurred
between iron and zinc baths containing 510 wt% Al. The inhibiting layer was
found to be rich in Al and approximately 0.5 mm thick [91, 92]. The alloy layer in
a Zn5% Al0.1% Mg coating was reported to be 50 nm and was composed of
AlFeZn [93]. The breakdown of the initial inhibiting layer by the formation of
outbursts was rst detected at 4508C for dipping times of 8 and 4 s in Zn3 wt%
A.R. Marder / Progress in Materials Science 45 (2000) 191271 227
Al and Zn6 wt% Al baths, respectively [91]. Ichiyama et al. [94] suggested that
the intermetallic layer was Fe
2
Al
5
or a composite layer of Fe
2
Al
5
/FeZn
7
/FeZn
13
and Makimattila [95] suggested that two separate FeAlZn phases exist. The
initial formation of intermetallic phases after inhibition layer breakdown at 4508C
Fig. 28. A schematic representation of FeZn phase layer formation in a 0.20 wt% AlZn galvanizing
bath. t
0
corresponds to zero time, and development occurs according to time such that t
1
<t
2
<t
3
<t
4
.
From Ref. [90].
A.R. Marder / Progress in Materials Science 45 (2000) 191271 228
is a characteristic outburst at the substrate/melt interface consisting mainly of
Fe
2
Al
5
Zn
x
phase, with the outer part next to the eutectic alloy coating being
FeAl
3
Zn
x
[96]. The Fe
2
Al
5
Zn
x
has a preferred orientation such that the (001)
plane is parallel to the substrate/coating interface and is similar to the growth of
Fe
2
Al
5
during aluminizing where the preferred orientation is also well
developed [96]. The eect of mischmetal (0.050.2 wt% rare earth) was to
postpone the formation of a detectable interfacial inhibition layer at the interface
from 4 s to approximately 64 s [97]. It was reported that the rare earth elements
segregated at the coating/substrate steel interface [97], suggesting a major role of
the rare earth elements in stabilizing the interfacial layer. As the dipping
temperature is increased in the range of 4505758C the rates of intermetallic
formation and growth were found to increase markedly [98] and growth was
attributed to the formation of an Fe
3
Al porous layer [99].
4.3.2. 55 wt% Al (galvalume)
Galvalume sheet coating consists of nominally 55 wt% Al, 43.4 wt% Zn and
1.6 wt% Si. Si is added to the bath to prevent a very strong exothermic reaction
between the AlZn bath and the steel sheet. Without Si, rapid attack of the iron
panel by the high Al containing Zn baths was controlled by the rate of aluminum
diusion in the liquid bath in contact with the iron panel [100]. The interfacial
layers exhibited two morphologies. In the rst morphology, the interfacial layers,
Fe
2
Al
5
and FeAl
3
, ake o from the iron panel and allow linear attack by the
molten bath being in direct contact with the iron. The reaction proceeds very
rapidly, limited only by the diusion of Al in the liquid AlZn bath to the
substrate iron surface. In the second morphology the interfacial layers are
Table 8
Individual FeZn phase layer growth-rate time constant, n, values
for the steels hot-dip galvanized in a 0.20 wt% AlZn bath [90]
Sample/layer Growth-rate time constant, n
Gamma-phase layer
ULC 0.05020.049
ULCP 0.03220.003
Ti IF 0.1620.01
TiP IF 0.528*
TiNb IF 0.1920.05
TiNbP IF 0.03620.158
Delta-phase layer
ULC 0.3820.02
ULCP 0.3820.05
Ti IF 0.5320.07
TiP IF 0.5020.07
TiNb IF 0.5720.07
TiNbP IF 0.6220.09
* No error determined due to t over just two data points
A.R. Marder / Progress in Materials Science 45 (2000) 191271 229
adherent to the iron substrate. However, aluminum again diuses to the iron
substrate through liquid channels in the interfacial layers and the reaction
proceeds rapidly. In both cases the exothermic reaction produced by forming
Fe
2
Al
5
and FeAl
3
consumes the panel in less than 2 min. Ghuman and
Goldstein [49] found a similar eect in baths above 6008C containing 110 wt%
Al, which was related to the formation of an Fe
2
Al
5
Zn
x
(Z) phase, hypothesized
to be an BrewerEngle compound, which forms by a large heat of evolution.
Thus, in this case an inhibition layer never has an opportunity to form.
Silicon suppresses the exothermic reaction by forming a solid interfacial layer
that acts as a diusion barrier or inhibition layer for the reactive species, which is
probably Al [101]. Therefore, in order for a reaction between the bath and the
iron substrate to occur, the Al has to diuse through the solid phase(s) in the
interfacial layer, which is considerably slower than liquid state diusion. The
composition of the interfacial layer of commercial galvalume was reported to be
in one study a single-phase layer containing 55.8 wt% Al, 33.7 wt% Fe, 6.7 wt%
Si and 3.8 wt% Zn, and in another study a two-phase layer containing 53.1 wt%
Al, 25 wt% Fe, 8.9 wt% Si and 13 wt% Zn, and 56.6 wt% Al, 36.7 wt% Fe, 3
wt% Si and 3.7 wt% Zn [102]. Table 9 lists the types of phases that constitute the
interfacial layer [101] as a function of dipping time at 6108C. In the 1.7 wt% Si
bath, T
5C
is the rst phase formed at the interface at short times (<4 s).The T
5C
phase has a composition of 57 wt% Al, 30 wt% Fe, 6 wt% Si and 7 wt% Zn.
Fig. 29. The eect of Ti** on the growth rate time constant, n. From Ref. [80].
A.R. Marder / Progress in Materials Science 45 (2000) 191271 230
T
a
b
l
e
9
P
h
a
s
e
s
p
r
e
s
e
n
t
i
n
t
h
e
r
e
a
c
t
i
o
n
z
o
n
e
o
f
g
a
l
v
a
l
u
m
e
5
5
w
t
%
A
l
c
o
a
t
i
n
g
s
h
o
t
-
d
i
p
p
e
d
a
t
6
1
8
C
a
s
a
f
u
n
c
t
i
o
n
o
f
d
i
p
p
i
n
g
t
i
m
e
[
1
0
1
]
S
i
l
i
c
o
n
c
o
n
t
e
n
t
o
f
t
h
e
b
a
t
h
(
w
t
%
)
D
i
p
t
i
m
e
(
s
)
0
.
7
1
.
7
3
.
0
5
.
0
6
.
8
8
4
?
T
5
C
?
?
T
4
2
5
F
e
2
A
l
5
+
?
?
?
?
F
e
2
A
l
5
+
?
1
0
0
F
e
2
A
l
5
+
?
F
e
2
A
l
5
+
?
F
e
2
A
l
5
+
?
F
e
2
A
l
5
+
?
F
e
2
A
l
5
+
?
2
5
6
F
e
2
A
l
5
+
?
F
e
2
A
l
5
+
T
1
+
?
F
e
2
A
l
5
+
T
1
+
T
5
H
F
e
2
A
l
5
+
T
1
+
T
5
H
F
e
2
A
l
5
+
?
4
0
0
F
e
2
A
l
5
+
?
F
e
2
A
l
5
+
T
1
+
?
F
e
2
A
l
5
+
?
F
e
2
A
l
5
+
?
F
e
2
A
l
5
+
?
9
0
0
F
e
2
A
l
5
+
F
e
A
l
3
+
T
1
F
e
2
A
l
5
+
T
1
F
e
2
A
l
5
+
?
F
e
2
A
l
5
+
T
1
+
?
F
e
2
A
l
5
+
T
1
F
e
A
l
3
+
T
5
C
F
e
A
l
3
+
T
2
+
T
4
1
8
0
0

F
e
2
A
l
5
+
T
+
1
F
e
2
A
l
5
+
T
1

F
e
2
A
l
5
+
T
1
+
F
e
A
l
3
+
T
2
+
T
3
+
T
4
F
e
A
l
3
+
T
5
C
+
T
5
H
F
e
A
l
3
+
T
2
+
T
5
H
2
7
0
0

F
e
2
A
l
5
+
T
1

F
e
2
A
l
5
+
T
1
+
F
e
A
l
3
+
T
2
+
T
3
+
T
4
F
e
A
l
3
+
T
5
C
+
T
5
H
3
6
0
0

F
e
2
A
l
5
+
T
1

F
e
2
A
l
5
+
T
1
+
F
e
A
l
3
+
T
2
+
T
3
+
T
4
F
e
A
l
3
+
T
5
C
+
T
5
H
A.R. Marder / Progress in Materials Science 45 (2000) 191271 231
As Al diusion continues with longer dipping times, Fe
2
Al
5
and FeAl
3
phases
containing 23 wt% Si form and continue to slowly consume the entire coating.
Apparently, the T
5C
phase acts as the initial inhibition layer with the Fe
2
Al
5
phase
containing Si forming next to the substrate interface. Others have reported that
the intial inhibiting phase was a cubic structure based on the hexagonal phase
Fe
5
Si
2
Al
20
with small amounts of a transition metal [103] and that these layers
form in as little as 0.3 s at 6008C [104]. It was found that alloy growth increased
from an immersion temperature of 590 to 6058C. However, no further increase
alloy growth was observed when the temperature was increased to 6208C. This
eect was related to an accompanied change in the structure and composition of
the interfacial alloy from Fe
2
Al
5
/cubic phase to Fe
2
Al
5
/FeAl
3
. It was also reported
that signicant alloy growth occurs during cooling from 6008C at rates as high as
308C/s [104]. Air cooled specimens were shown to produce faceted crystals on the
alloy layer that were prevented by the addition of 0.026 wt% V to the bath.
Localized outbursts are observed only at immersion times greater than 1800 s [101].
This reaction produces Fe
2
Al
5
and FeAl
3
phases, both of which ake o the
substrate and oat into the bath. Si was not found in Fe
2
Al
5
and FeAl
3
phases
formed in the burst region, whereas the Fe
2
Al
5
and FeAl
3
layers that adhered to
the panel contained Si. This may indicate a buildup of stresses at the interface in
the absence of Si in the Fe
2
Al
5
and FeAl
3
phases [100].
5. Overlay coating formation
In order to understand overlay coating microstructure formation, it is rst
necessary to determine the thermal conditions encountered during solidication of
the coating so that the nucleation and growth of the overlay may be ascertained.
Recently, a solidication model for hot-dip galvanized zinc coatings was
developed [105]. The amount of undercooling and the temperature gradient in the
liquid zinc lm prior to solidication was evaluated in a hot dip simulator using
superimposed cooling rates from natural convective cooling in air (28 K/s) to jet
cooling in a gas ow of 10 m
3
/h5% H
2
/N
2
atmosphere (158 K/s). The measured
undercooling and solidication period is given in Table 10. Previously reported
research showed that the undercooling in the Zn coating ranged from 1 to
108C [106108]. More recent research showed the undercooling to be less than
18C [109]. Since heat can only be removed by conduction through the liquid zinc
lm and considering a 25 mm zinc coating, the approximate dierence in
temperature between the steel/zinc interface and the zinc surface was estimated as:
DT
min
= 2:2 10
3
K at 2K=s (6)
DT
max
= 1:6 10
2
K at 15K=s (7)
As a results of these small temperature gradients and the slight undercooling, the
authors [109] proposed that heterogeneous nucleation occurs at the rough steel/
A.R. Marder / Progress in Materials Science 45 (2000) 191271 232
zinc interface which oers preferred nucleation sites for zinc dendrites. It may be
assumed that similar Zn alloy coatings will behave in the same manner in that
nucleation of the overlay coating will occur at the zinc/substrate interface.
The importance of aluminum additions to the bath to control the type and
amount of inhibiting layer(s) has been reviewed above. Actually, the eect of
aluminum additions to interfacial layer control is a result of the addition of
aluminum to produce a specic as-cast microstructure in the overlay portion of
the coating in order to induce specic properties in the coating such as corrosion
resistance. As a result of the aluminum additions, interface reactions require a
knowledge of ternary and quaternary equilibrium and metastability for the
elements involved. On the other hand, binary and ternary equilibrium interactions
are sucient to explain the as-cast overlay structure. This section will rst review
the important ZnAl equilibrium phase diagram that will be followed by a
discussion of the three commercially important coatings: galvanized (<1 wt% Al),
galfan (5 wt% AlZn) and galvalume (55 wt% AlZn).
5.1. ZnAl equilibrium
Aluminum is added to the zinc bath to improve corrosion resistance by either
allowing for the formation of a pure zinc overlay for galvanic protection by
inhibiting the formation of FeZn phases or by introducing multiphase
microstructures in the overlay coating. In each case, the resulting microstructures
produce coatings that oer corrosion resistance by a combination of galvanic and
barrier protection with corrosion products that in some cases are passive. An
examination of the ZnAl equilibrium phase diagram is helpful to understand the
overlay cast structure, Fig. 30 [110]. The solid phases present in the ZnAl system
include b
/
, b and Z. b
/
and b represent the Al- and Zn-rich portions, respectively,
of the aluminum fcc solid solution. The solubility of Zn in the Al solid solution
increases with temperature to 32.4 wt% Zn (16.5 at% Zn) at the eutectoid
temperature (2778C) and 83.1 wt% Zn (67 at% Zn) at the eutectic temperature
(3818C). The Zn terminal solid solution, Z, is hexagonal close packed with a
maximum Al solubility of 1.2 wt% (2.8 at%) at the eutectic temperature. This
solubility decreases to 0.7 wt% Al (1.6 at% Al) at the eutectoid temperature and
0.03 wt% Al (0.07 at% Al) at room temperature. The transformations involving
Table 10
Correlation between cooling rate, solidication period and undercooling [105]
Cooling rate (K/s) Atmosphere Solidication period
a
(s) Undercooling
a
(K)
2 Air 10 <1
4 5% H
2
/N
2
5 <1
15 5% H
2
/N
2
2 <2
a
Results obtained from two dierent bath compositions (Zn+0.2 wt% Al +0.0025 wt% Pb and
Zn+0.2 wt% Al +0.05 wt% Pb) did not vary signicantly; thus, an average value is given
A.R. Marder / Progress in Materials Science 45 (2000) 191271 233
F
i
g
.
3
0
.
T
h
e
A
l

Z
n
e
q
u
i
l
i
b
r
i
u
m
d
i
a
g
r
a
m
.
F
r
o
m
R
e
f
.
[
1
1
0
]
.
A.R. Marder / Progress in Materials Science 45 (2000) 191271 234
these phases, Table 11, are a eutectic at 95.0 wt% Zn (88.7 at% Zn), a eutectoid
at 77.7 wt% Zn (59.0 at% Zn) and the existence of a miscibility gap with a
critical temperature of 351.58C at 61.3 wt% Zn (39.5 at% Zn).
5.2. Zinc coating types
5.2.1. Galvanized (<1 wt% Al)
Aluminum is probably the most important alloying element added to the hot-
dip galvanizing bath, with dierent levels required in order to produce dierent
properties in the bath [111]. Al levels of 0.0050.02 wt% are added to brighten the
initial coating surface. The eect is related to the formation of a continuous Al
2
O
3
layer on the coating surface that inhibits further oxidation by acting as a
protective barrier layer. This eect is also responsible for the reduced atmospheric
oxidation of the zinc bath. In addition, aluminum in the range of 0.10.3 wt% is
added to the zinc bath to suppress the growth of brittle FeZn intermetallic
phases at the steel coating interface by forming the Fe
2
Al
5
(Zn) inhibition layer as
discussed above. The end of this incubation period is marked by the disruption of
the initial layer, followed by rapid attack of the substrate steel. Thus, during
commercial production, the immersion time is kept below the incubation period in
order to obtain a highly ductile product.
Galvanized hot-dip coatings often have a structure consisting of very large
grains called ``spangles''. In cross-section, an Fe
2
Al
5
(Zn) inhibition layer develops
rst, preventing any FeZn intermetallic phase formation. The overlay layer is
made up of dendrites of pure zinc eta (Z) phase and appears as a polycrystalline
structure. The surface of as-galvanized coatings are commonly characterized by
spangles, shown in Fig. 31. Spangle size is inuenced by the cooling conditions
during solidication. The three surface nishes commonly produced are:
. Regular spangle, where the coating solidies from the dipping temperature by
air cooling, producing the well-known spangle nish.
. Minimum spangle, where the coating is quenched using water, steam, chemical
solutions, or by zinc powder spraying.
. Extra-smooth temper roll nish carried out as an additional operation with
regular and minimum spangle material.
Table 11
Phase transformations in the AlZn system
Phase transformation Composition wt% (at % Zn) Temperature (8C) Transformation type
L= b+ Z 95.0 (88.7) 381 Eutectic
b= b
/
+ Z 77.7 (59.0) 277 Eutectoid (monotectoid)
(Al) = b
/
+ b 61.3 (39.5) 351.5 Critical
L=(Al) 0 (0) 660.45 Congruent
L+ Z 100 (100) 419.58 Congruent
A.R. Marder / Progress in Materials Science 45 (2000) 191271 235
Hot-dip galvanized coatings often show a strong (0001) basal plane texturing of
the spangles (zinc crystals) upon solidication. The proportion of basal plane
orientation increases with decreasing coating grain size [112] (i.e. minimized
spangle and temper rolled coatings) and correlate well with improved paint
adherence [113]. With preferred nucleation sites at the liquid Zn/substrate
interface, growth of the Zn spangles occurs in the following manner [105]:
. The initial growth stage (0.1 s) involves sideways growth of Zn basal planes
(0001) parallel to the interface in the [1100] growth direction. The entire steel
surface is covered by a solid layer of zinc dendrites, whereas the surface remains
liquid.
. The second stage involves slow thickening of the solid zinc grain which is
dependent on grain orientation.
. The third stage, or continuous thickening of the zinc grain, causes solute
enrichment of the remaining liquid and nally Pb precipitation between Zn
dendrite arms.
Large spangles are generally associated with solute additions to the galvanizing
bath and Pb, at concentrations greater than 0.04 wt%, is the solute generally
added to molten Zn to produce spangles, Fig. 32 [114]. Al, Mg, Sn and Cd
additions do not inuence spangle growth [109, 115], but Sb and Bi additions
Fig. 31. Typical spangle structure of a galvanized coating.
A.R. Marder / Progress in Materials Science 45 (2000) 191271 236
produce spangles [116]. Fasoyino and Weinberg [109] found that alloy additions
such as Pb and Sb have lower surface tension and promote spangle formation.
Strutzenberger and Faderl [105] proposed that these same elements segregate
ahead of the growing dendrite, lowering surface tension and allowing the growth
velocity of the zinc dendrite to increase, resulting in larger grains or spangles.
Fig. 32. Eect of lead content on spangle size. From Ref. [114].
A.R. Marder / Progress in Materials Science 45 (2000) 191271 237
5.2.2. Galfan (5 wt% AlZn)
Galfan is a Zn5% Al alloy coating, which is near the eutectic point in the Al
Zn equilibrium phase diagram, Fig. 30. Two compositions have been reported
based on additions to the eutectic composition: small mischmetal additions
containing lanthanum and cerium contents up to about 0.5% [117] and additions
of 0.5% magnesium [118]. These additions are made to improve the wettability
and uidity of the molten bath without aecting the corrosion resistance of the
coating [119]. They also suppress bare spot formation [120] as well as producing a
typical ``minimized spangle'' structure [93]. The microstructure of galfan is
characterized by a two-phase structure, a zinc-rich eta (Z) proeutectoid phase
surrounded by a eutectic type phase consisting of beta (b) aluminum and eta (Z)
zinc lamellae, Fig. 33. However, the microstructure can be varied depending upon
cooling rate [120].
In the early stages of solidication, proeutectic zinc particles nucleate rst and
the solidication of the eutectic matrix proceeds from these particles [121]. The
amount of the proeutectic was calculated to be 18% [122]. These results indicate
nonreciprocal nucleating characteristics, i.e. one primary phase will act as an
eective heterogeneous nucleation site for the other phase, but not vice versa, and
the nucleation characteristics and microstructure in o-eutectic AlZn alloys have
been detailed [122]. The galfan overlay surface structure is shown in Fig. 34 and a
schematic of the solidication sequence is found in Fig. 35. Although the nodules
are on the order of 4 mm in diameter, each grain is divided into many smaller
subgrains or colonies, approximately 22.5 mm in size. Each colony contains
lamellae of alternate layers of beta (b) aluminum and eta (Z) zinc, Fig. 33. In the
range of normal bath temperatures, 4204408C, it is reported that there is no
visible intermetallic layer or at least an extremely thin layer (<0.5 mm) at the
interface between the steel substrate and the overlay coating [121]. This would
Fig. 33. Planar view of the lamellar microstructure of a galfan coating. From Ref. [111].
A.R. Marder / Progress in Materials Science 45 (2000) 191271 238
indicate the presence of an inhibition eect in preventing FeZn alloy
formation [92, 93]. The alloy layer in a Zn5 wt% Al0.1 wt% Mg coating was
reported to be 50 nm and is composed of AlFeZn [95]. Thus, galfan coatings
have excellent formability and cut edge corrosion protection.
The surface of galfan coatings was shown to have solidication defects in
association with nodule boundaries [123]. Surface depressions, or dents, were
found to occur at eutectic nodule boundaries and triple points, and were typically
1015 mm deep. Substrate interactions were not expected to be the cause of these
dents or shrinkage cavities, since the these surface defects were reproduced in the
solidication of Zn5% Almischmetal alloy samples on an inert substrate. It was
also found that impurities, particularly Pb, strongly segregate to eutectic nodules
and triple points. Solidication cracking was found to be associated with both the
shrinkage cavities (dents) and segregated impurities [123, 124]. A mechanism for
galfan solidication and denting/cracking was proposed, as schematically
illustrated in Fig. 36. During the solidication process, the liquid metal is quickly
consumed due to the relatively large volume changes associated with the
solidication of Zn and Al. As a result, there is a shortage of liquid between the
two or three adjacent growing eutectic nodules and, upon impingement of the
nodules, the interface at the surface will be curved, creating a surface depression.
Cracking can occur within weakened boundaries due to superimposed
solidication stresses and the precipitation of Pb particles. Galfan coating surface
Fig. 34. Planar view showing (a) nodules and (b) colonies of a galfan coating. From Ref. [111].
A.R. Marder / Progress in Materials Science 45 (2000) 191271 239
appearance has been enhanced by decreasing the Pb content in the bath alloy and
by an increase in skin-pass reduction [124].
5.2.3. Galvalume (55 wt% AlZn)
Galvalume, is a 55%Al alloy coating containing about 1.5% Si added for the
purpose of preventing an exothermic reaction at the coating overlay/substrate steel
interface [101]. During the coating process an interfacial FeAlZn intermetallic
alloy layer forms at the interface between the steel substrate and the overlay
coating [125]. The surface of the Galvalume coating contains characteristic
spangles, Fig. 37a, that consist of aluminum dendrites with a clearly measurable
dendrite arm spacing (DAS), Fig. 37b. In cross-section, three features of the
overlay (or solidied bath) are dened, Fig. 37c: the coating contains beta (b)
aluminum dendrites, Zn-rich interdendritic regions and a ne dispersion of Si
particles. The Al dendrites were reported to contain approximately 18 wt% Zn
and up to 1.8 wt% Si which is in good agreement with the AlZn phase
diagram [102].
Clearly [126] has shown that increased cooling rates after dipping can
signicantly rene the DAS of the coating, Fig. 38, and Selverian et al. [102] have
reported that spangle size, dendrite grain size and DAS decrease as coating
thickness decreases. The number of silicon particles in the overlay increased with
cooling rate, Fig. 39 [126], and are primarily responsible for rening the
microstructure by either constraining the growth of the Al dendrites or by
providing numerous heterogeneous nucleation sites for dendrite formation. Si
particles appear to nucleate at the alloy layer/bath interface and grow into the
bath as shown schematically in Fig. 40 of the growth sequence of Galvalume
solidication. The zinc-rich interdendritic region is heavily populated by
precipitates, Fig. 41. The large precipitates could be the metastable R phase [111],
whereas the small precipitates are possibly coarsened coarsened G.P. zones or
spinodal phase precipitates [102].
Fig. 35. Schematic of the solidication of a galfan coating. From Ref. [111].
A.R. Marder / Progress in Materials Science 45 (2000) 191271 240
F
i
g
.
3
6
.
M
e
c
h
a
n
i
s
m
f
o
r
g
a
l
f
a
n
f
o
r
m
a
t
i
o
n
a
n
d
d
e
n
t
i
n
g
/
c
r
a
c
k
i
n
g
.
F
r
o
m
R
e
f
.
[
1
2
4
]
.
A.R. Marder / Progress in Materials Science 45 (2000) 191271 241
Fig. 37. Microstructure of galvalume coating: (a) spangle nish; (b) dendrite arm spacing; (c) cross-
sectional view. From Ref. [111].
A.R. Marder / Progress in Materials Science 45 (2000) 191271 242
Fig. 38. The eect of cooling rate on DAS in galvulume [126].
Fig. 39. The eect of cooling rate on Si particles in the galvalume coating. From Ref. [126].
A.R. Marder / Progress in Materials Science 45 (2000) 191271 243
6. Heat treatmentgalvannealing
Galvanneal coatings are essentially diusion coatings that expose the zinc
galvanized steel to an annealing temperature around 5008C to produce a fully
alloyed coating containing FeZn intermetallic phases. This is accomplished by
inserting heating and cooling capacity above the liquid zinc pot in order for the
galvannealing process to be continuous, Fig. 1. The variables involved in
producing the desired galvanneal microstructure and properties are complex.
Good process control requires that the eects of heating rate, hold temperature
and time, and cooling rate on the FeZn reaction kinetics be well understood so
that the optimum coating for the desired properties can be obtained. Besides
processing variables, variations in bath chemistry and substrate composition all
Fig. 40. Schematic of the solidication of galvalume, t
1
<t
2
<t
3
<t
4
. From Ref. [111].
A.R. Marder / Progress in Materials Science 45 (2000) 191271 244
contribute to the nal microstructure. The galvanneal structure will rst be
reviewed and then the eects of processing variables and alloy additions to the
bath and substrate steel will be discussed.
6.1. Galvanneal microstructure
The cross-section microstructure of galvanneal coatings, Fig. 42, has been
classied as follows [127]:
. Type 0underalloyed coating containing predominantly zeta (z) phase.
. Type 1optimum alloyed coating with less than a 1 mm interfacial gamma (G)
layer and an overlay containing delta (d) phase interspersed with a small
amount of zeta (z) phase.
. Type 2overalloyed coating with a gamma layer >1 mm and an overlay
containing a delta (d) phase with basal plane cracks perpendicular to the
coating/substrate interface and an occasional top layer of zeta (z) phase.
The topography of the galvanneal coatings is also important. Van der Heiden et
al. [128] studied the surface of typical production galvanneal coatings by SEM
stereophotography and found the existence of ``craters''. Crater formation was
observed in-situ in a hot-stage environmental scanning electron microscope [129].
Craters, Fig. 43, were found to be inherent to the galvanneal coating layer,
causing major uctuations in coating thickness [128] and appear to be associated
with outburst formation [129, 130]. A mechanism was proposed that can explain
the relationship between outburst formation and the appearance of craters [128].
Fig. 41. Precipitates in the interdendritic boundaries of galvalume. From Ref. [111].
A.R. Marder / Progress in Materials Science 45 (2000) 191271 245
Fig. 42. Morphology of galvanneal coatings: (a) type 0, (b) type 1, and (c) type 2. From Ref. [127].
A.R. Marder / Progress in Materials Science 45 (2000) 191271 246
After an outburst of FeZn intermetallic compounds occurs at the substrate grain
boundary, rapid growth of these solid phases occurs into the liquid. The liquid Zn
will remain present between the crystals, which enables very fast transport of iron
from the substrate [74]. Capillary eects of the pores and channels between the
crystals drain the nearby regions where no alloy reaction takes place, whereby
``dry spots'' or craters are formed. A similar phenomenon has been used to
describe denting in galfan coatings [124].
These results have led to a phenomenological model that describes the sequence
of events that occur during the microstructure development of a fully alloyed
galvanneal coating, Fig. 44 [129]. The kinetics of formation will greatly depend on
Fig. 43. Crater formation in galvanneal coatings: (a) cross-section, (b) planar surface. From Ref. [129].
A.R. Marder / Progress in Materials Science 45 (2000) 191271 247
F
i
g
.
4
4
.
S
c
h
e
m
a
t
i
c
o
f
t
h
e
p
h
e
n
o
m
e
n
o
l
o
g
i
c
a
l
m
o
d
e
l
o
f
g
a
l
v
a
n
n
e
a
l
m
o
r
p
h
o
l
o
g
y
d
e
v
e
l
o
p
m
e
n
t
.
t
0
c
o
r
r
e
s
p
o
n
d
s
t
o
z
e
r
o
t
i
m
e
o
f
t
h
e
a
s
-
g
a
l
v
a
n
i
z
e
d
s
t
r
u
c
t
u
r
e
,
a
n
d
d
e
v
e
l
o
p
m
e
n
t
o
c
c
u
r
s
a
c
c
o
r
d
i
n
g
t
o
t
i
m
e
s
u
c
h
t
h
a
t
t
1
<
t
2
<
t
3
<
t
4
<
t
5
.
F
r
o
m
R
e
f
.
[
1
2
9
]
.
A.R. Marder / Progress in Materials Science 45 (2000) 191271 248
the processing parameters and composition of the bath and substrate steel,
however it was proposed that the time sequence was as follows:
. t
0
the steel develops an FeAl interfacial layer during hot-dip galvanizing
which, depending on Al content in the Zn bath, inhibits the formation of Fe
Zn phases [49, 131]. The eectiveness of the inhibition will depend on the Al
content in the bath [59], as well as time of immersion and bath temperature [63].
. t
1
FeAl inhibition layer breaks down during annealing causing nucleation
and growth of delta (d) phase at the coating/steel interface [66]. Outburst
formation can also occur with accelerated growth of FeZn phases [74].
Simulated annealing of a galvanized coating in an environmental SEM shows
the formation of outbursts at the interface, Fig. 45 [132]. The rate of Zn attack
will depend on the substrate alloy addition as well as the temperature prole in
the annealing process. This microstructure is referred to as Type 0
galvanneal [127].
. t
2
As annealing continues, diusional growth of delta (d) phase in a columnar
growth morphology occurs. Previously nucleated zeta (z) phase transforms to
delta (d) phase. Additional zeta (z) phase may nucleate due to oversaturation of
Fig. 45. Sequence of outburst formation during simulated in-situ galvannealing. From Ref. [132].
A.R. Marder / Progress in Materials Science 45 (2000) 191271 249
iron in the liquid eta (Z) phase or upon cooling. An interfacial gamma (G) phase
forms at the coating/steel interface [127].
. t
3
Zn depletion occurs with longer temperature exposure resulting in the
complete consumption of eta (Z) phase at the coating surface. Fe concentration
increases in the coating as the delta (d) phase continues to form, pushing the
zeta (z) phase to the surface and maintaining a constant 1 mm thickness gamma
(G) phase [133]. This has led many to rely on an Fe concentration measurement
(approximately 10 wt%) as a requirement for the optimum galvanneal coating,
see e.g. [134]. This microstructure is referred to as Type 1 galvanneal [127].
. t
4
With longer times at temperature, delta (d) phase diusional growth
continues towards the surface of the coating consuming the zeta (z) phase, while
maintaining a constant 1 mm thickness gamma (G) phase [133].
. t
5
Once the delta (d) phase reaches the surface it serves as the Zn rich side of
the delta (d) phase FeZn/steel diusion couple, allowing for the continued
growth of gamma (G) phase at the expense of the delta (d) phase. Cracking
occurs along delta (d) phase basal planes parallel to the coating/steel
substrate [135]. This microstructure is referred to as Type 2 galvanneal [127].
6.2. Process variables
In continuous galvannealing, heating, holding and cooling sections are added to
the up-leg of the hot-dip coating line, Fig. 1, above the jet nishing nozzles that
are used for coating thickness control. A gas-red furnace or an induction coil are
used to heat the strip and in some designs a holding zone is utilized prior to
cooling. Besides the variables associated with hot-dip galvanizing, e.g. bath
temperature, substrate alloy content and bath composition, a major variable of
the galvannealing process is the line speed, which will control the dependent
galvannealing variables such as (1) heating rate, (2) peak or holding temperature
and time and (3) cooling rate. Initial studies were only concerned about the
holding temperature and time, and the eect on galvaneal mechanical properties
such as bend tests or powdering [35].
The concept of processing windows for optimum microstructure was rst
introduced to determine the kinetics of coating microstructure development during
galvannealing [136]. These maps, Fig. 46, were able to illustrate the isothermal
temperature and times necessary to produce Type 0, Type 1 and Type 2
morphologies, as well as iron content and powdering properties. In these studies,
a Gleeble thermomechanical test unit was used to simulate the galvannealing
temperature and time, while the heating rate was set at 5008C/s and the cooling
was xed at 258C/s [127]. Similar test conditions were used to determine the eect
of the galvannealing isothermal temperature and time on coating microstructure,
galvanneal TTT diagrams [41] and Fe content [134]. Still others produced
processing windows for powdering [128, 137].
A.R. Marder / Progress in Materials Science 45 (2000) 191271 250
Fig. 46. Morphology development for the TiIF steel (0.10 wt% Al bath). From Ref. [136].
Fig. 47. Typical temperature prole of the cooling tower. From Ref. [138].
A.R. Marder / Progress in Materials Science 45 (2000) 191271 251
Characterizing the temperature prole of the entire cooling tower is extremely
important. A typical calculated temperature prole of the strip moving through
the cooling tower portion of a modern hot-dip galvannealing line is seen in
Fig. 47 [138]. It is apparent from this prole that the temperature exposure of the
strip is never isothermal. The galvannealing furnace takes the strip from the air
knives at 4508C to a peak temperature of about 5058C at a heating rate of
approximately 158C/s. The strip is held in the soaking furnace at the peak
temperature of around 5008C for 89 s, where some slight cooling occurs (28C/s),
and in the air coolers the strip cools down to below 3508C at a rate of over 178C/s.
Kanamura and Nakayama [63] shows that the heating rate during galvannealing
should be as high as possible to limit the growth of both zeta (z) and gamma (G)
phase. Hayes [139] modeled the galvanneal temperature prole and studied the
eect of heating rate, peak temperature and cooling rate on morphology
development in a series of alloy substrates and bath Al compositions. He found
that during galvannealing it is entirely possible for alloying to occur on heating
and cooling, as well as during isothermal holding. For example, heating rates
necessary to prevent alloying, ranged between 25 and 508C depending on the
substrate composition and peak temperature. These results were based on the
transition from Type 0 to Type 1 coating morphology, Fig. 48, and Fe pick-up in
the coating. Likewise, cooling rates less than 508C/s resulted in coatings that were
more alloyed than the peak temperature microstructure. The isothermal and non-
isothermal kinetics of iron pick-up in galvanneal coatings has also been
quantitatively modeled [140].
Fig. 48. The eect of heating rate on the type 0 to type 1 microstructure transition. From Ref. [139].
A.R. Marder / Progress in Materials Science 45 (2000) 191271 252
6.3. Alloy additions
The control of Al bath content in galvanneal production is critical [59]. The Al
content should be high enough so that the intermetallic particles formed in the
bath oat to its surface and line stoppages for bottom dross removal are avoided.
Ideally, there should be no partial or continuous inhibition layer in the as-
galvanized coating, so that during annealing FeZn alloy formation is not
impeded and a uniform high quality coating is obtained. Tang [59] suggested that
the optimum Al composition for galvannealing is the transition from delta (d)
phase (or gamma (G) phase) to the Fe
2
Al
5
Zn
x
eta (Z) phase point at 0.15 wt%. At
this point, the Fe
2
Al
5
Zn
x
eta (Z) becomes an equilibrium compound in the bath.
Al contents marginally higher than this level could prevent bottom dross
formation, whereas the formation of the Fe
2
Al
5
phase at the interface is
practically impossible because of severe competition from the zeta (z) and delta (d)
phases.
In practice, the stability of the Fe
2
Al
5
interfacial layer is enhanced by high Al
levels (>0.15 wt% Al) inhibiting FeZn phase formation in the short term,
whereas FeZn reactions are enhanced by low Al bath levels (<0.15 wt% Al) [59].
In some cases, TEM results reported the formation of FeZn compounds at the
FeAl inhibition layer/liquid Zn interface [141] and the Al-depletion model may
be an additional mechanism that is expected to control FeAl inhibition layer
break down [142]. However, there is considerable evidence in the literature to
support the outburst formation mechanism of inhibition layer break
down [43, 57, 78, 128]. More craters were seen on the surface of galvanneal
coatings for low Al baths which were related to outburst formation [129]. In low
Al baths, the formation of Type 1 morphology in a DQSK steel occurs at much
shorter times for an isothermal holding temperature range of 4505508C,
Fig. 49 [127] and with heating rate [139]. Similar Al eects have been found for
TiIF steel [83, 127, 129, 139, 143], PIF [132, 139, 143] and Ti/NbIF [139]. Al
added to the bath is eective in inhibiting the formation of zeta (z) phase in the
bath and during the heating period and also in suppressing the growth of gamma
(G) phase, Fig. 50 [64]. Urai et al. [144] have shown that a higher Al bath content
of 0.16 wt% produces a coating with a higher zeta (z)/delta (d) phase ratio than
that in a bath with 0.12 wt% Al.
Galvannealing using various substrate steels, especially IF steels [38], has
demanded a lot of attention in the literature. Hisamatsu [74] reported that Ti
added sheet steel will alloy at 50708C lower than a low carbon steel, both being
galvanized in a 0.13 wt% Al bath at 4708C. Mercer [83] found that Ti-stabilized
steels were more reactive than Al killed steels for two dierent galvannealing
temperatures and Maki [145] using a 0.11 wt% Al bath, showed that the gamma
(G) phase thickness was greater than Al killed steel for every annealing
temperature studied. Similar studies on a DQSK and a TiIF steel showed that
the Ti steel inhibited the formation of a Type 1 microstructure using a 0.10 wt%
Al galvanizing bath, but accelerated the formation of Type 1 with a 0.15 wt% Al
A.R. Marder / Progress in Materials Science 45 (2000) 191271 253
bath, especially at higher galvannealing temperatures [127]. Similarly, Ti +Nb IF
steel, galvanized in a 0.14 wt% Al bath, accelerated alloy formation [38]. As
discussed previously, Ti added steels generally produce a clean substrate grain
boundary for faster reactions. Phosphorous, on the other hand, segregates to the
grain boundary and slows the formation of FeZn phases. Mercer [83] found that
a rephosphorized steel was less reactive than an Al killed steel at two dierent
galvannealing temperatures and P was found to delay Type 1 alloy formation over
a range of annealing temperatures [146]. P was reported to restrain the formation
of gamma (G) phase [147] and it was shown by in-situ annealing that P retards the
Fig. 49. Summary plot of temperature/time morphology transition for DQSK steel with (a) 0.10 wt%
AlZn, and (b) 0.15Zn wt% Al. From Ref. [127].
A.R. Marder / Progress in Materials Science 45 (2000) 191271 254
onset of outbursts [132]. Con and Thompson [41] were able to show that 0.07
wt% P can shift the phase eld boundary stability to longer times.
7. Coating properties
In general, the coating microstructure consists of the substrate, the interfacial
alloy layer and the overlay cast structure. Depending upon the type of coating, the
microstructure and composition of these constituents will control the desired
properties. The substrate must meet the design requirements of the component
and therefore is usually selected based on mechanical properties such as strength,
ductility, formability, etc. However, the substrate composition can play a major
role in the type of coating obtained and can aect the growth kinetics of the
phases formed in the coating. Alloy additions to the steel to improve sheet
formability, e.g. interstitial free (IF) steels with Ti, Ti/Nb and P inuence the
microstructure of the FeZn phases in galvanized and galvanneal steels. Substrate
grain size was shown to greatly aect the nucleation of the FeZn phases [79].
Considerable work has been reported about Zn coating properties and has
recently been reviewed [148151]. The important properties that concern the use of
zinc coatings are primarily corrosion and formability and other properties involve
weldability and paintability. These properties will be reviewed with respect to the
microstructural development of each coating. However, it should be noted that
the corrosion resistance of any zinc coating can drastically change depending upon
the specic corrosion environment and whether the coating is welded, contains a
paint system or is deformed, and the extent to which galvanic protection is
required for nearby uncoated areas.
Fig. 50. Eect of bath aluminum content on the growth of G
1
phase as a function of Fe content in the
coating. From Ref. [64].
A.R. Marder / Progress in Materials Science 45 (2000) 191271 255
7.1. Corrosion resistance
Zinc coatings add corrosion resistance to steel in several ways. As a barrier
layer, a continuous zinc coating separates the steel from the corrosive
environment. By galvanic protection, zinc acts as a sacricial anode to protect
the underlying steel at voids, scratches and cut edges of the coating. The
sacricial properties of zinc can be seen in a galvanic series where the potential
of zinc is less noble than steel in most environments at ambient temperatures.
A porous zinc oxide supercial layer forms on the surface by a mechanism of
dissolution/reprecipitation, leading to preferential corrosion pathways across the
high porosity areas, which accounts for the linear corrosion rate [152]. In
addition, after dissolution of the zinc metal, zinc hydroxide can precipitate at
the cathodic areas of the exposed steel, forming a secondary barrier layer.
Thus, the zinc coating will corrode at a slower rate than the steel substrate,
although the corrosion rate of zinc will vary depending upon the exposure,
Fig. 51 [148].
As in all coatings, the corrosion resistance the galfan coating is dependent on
coating thickness. In non-marine (industrial) environments, the galfan corrosion
mechanism occurs by a two-step process: in the rst step, a temporary protective
aluminum oxide passive layer forms on the surface, followed by a zinc galvanic
step in which a zinc sulfate layer forms on top of the oxide layer [152]. The last
step is much slower than the rst step because of the necessity for diusion
through the oxide layer and explains why galfan coatings corrode slower than
pure zinc galvanized coatings. It has also been shown that the zinc-rich Z phase in
both the proeutectic and eutectic lamella microstructure corrodes
preferentially [153]. Additions of Mg (up to 0.8 wt%) in galfan increase the
amount of proeutectic Al phase, which improves the corrosion resistance [154]. It
was found that corrosion in galfan proceeded parabolically, whereas galvanized
coatings showed linear behavior [155]. After 10-year outdoor exposure tests, Zn5
mass% Al alloy coated sheet steel had at least twice the corrosion resistance of
galvanized steel sheet, Fig. 52.
Initially the atmospheric corrosion of the galvalume coating takes place in
the zinc rich interdendritic regions, enabling the coating to exhibit galvanic
protection [156]. As the coating continues to corrode, the zinc corrosion
products become trapped in the interdendritic regions and act as a further
barrier to corrosion. Palma et. al. [157] have shown that some aggressive
corrosion species can move through the interdendritic regions and attack the
substrate steel without the Al dendrites in the coating being totally consumed.
Eventually, the aluminum dendrites, which also act as a barrier layer, add to
the corrosion protection, as does the FeAlZn intermetallic alloy layer. This
results in a parabolic type of corrosion as shown in Fig. 53 [148]. Although
the galvanic protection is greater in galvanized coatings, after 30 year
exposures in marine, rural and industrial environments, 55 wt% AlZn alloys
were found to be more than twice as durable as an equal thickness zinc
A.R. Marder / Progress in Materials Science 45 (2000) 191271 256
coating [159]. Galvalume is generally adequate to protect against rust staining
at scratches and cut edges of the steel sheet. The corrosion resistance of
galvanneal coatings can be slightly reduced because of the increased Fe in the
coating from the FeZn phases and the galvanic potential is not as great as
pure zinc.
Fig. 51. Service life (time to 5% rusting of steel surface) versus thickness of the coating for various
atmospheres. From Ref. [148].
A.R. Marder / Progress in Materials Science 45 (2000) 191271 257
7.2. Formability
The deformation and fracture behavior of Zn-based coatings on sheet steel can
alter the performance of steels in stamping operations [160]. During deformation,
increased frictional conditions at the sheet/tool steel interface and the substrate
mechanical properties can change the formability response of the material.
Fig. 52. Corrosion loss of coating layer after atmospheric exposure test. From Uchima Y, Hasaka M,
Koga H. Eect of structure and mischmetal addition on the corrosion behavior of Zn-5 mass%Al
alloy. GALVATECH '89. Tokyo: The Iron and Steel Institute of Japan. 1989. p. 545. # 1989. The
Iron and Steel Institute of Japan. [155].
A.R. Marder / Progress in Materials Science 45 (2000) 191271 258
Coating ductility depends on factors such as grain size, crystallographic
orientation, temperature, coating thickness and phase composition of the
intermetallic layer. Coating particulate buildup on die surfaces can lead to changes
in frictional behavior as well as poor appearance on surfaces of formed parts.
Proper lubrication is essential in the design of any forming process, especially
when forming zinc coated parts. Zinc coatings fail as a result of particle removal
during forming. Coating failures have been classied as follows [161]:
. Powderingparticle formation by intracoating failure to produce particles with
dimensions less than the coating thickness.
. Flakingformation of at particles by decohesion of the coating substrate
interface to produce particles with sizes similar to the coating thickness.
. Gallingdamage resulting from particles that bond to the tool surface. The
adhered particles plough through the coating or bond to the coating resulting in
additional coating damage.
Fig. 53. Corrosion losses of hot-dip coatings in the industrial environment of Bethlehem, PA. From
Ref. [148].
A.R. Marder / Progress in Materials Science 45 (2000) 191271 259
. Crackingthrough-thickness fracture of the coating without particle decohesion
from the substrate.
A schematic of powdering and aking is shown in Fig. 54 [162]. Deformation and
fracture mechanisms which lead to particle formation include basal plane cleavage
of pure zinc coatings, brittle fracture of intermetallic phases, intergranular
separation, ductile rupture and twinning [161], and a specic failure mode can be
produced by more than one mechanism.
There are distinct microstructural dierences in zinc coatings and their response
to deformation will be dierent. In hot-dip galvanized coatings, the deformation
of a regular spangle coating at low strain levels (<5%) occurs by twinning and at
higher strains intergranular cracking associated with cleavage on the Zn hcp basal
plane (0001) predominates [163]. Cracking is spangle specic [164], due to the
crystallographic nature of the cracks and the various spangle orientations. In
galfan coatings, virtually no cracking was observed at low levels of strain and
deformation proceeded by slip and twinning [163]. Intergranular cracking at
higher levels of strain were associated with the cleavage of basal planes. The
deformation of galfan coatings depended on the microstructure, texture and the
extent of dent formation at the grain boundaries. The eutectic morphology of the
coating and the almost undetectable intermetallic layer present at the coating/
substrate interface provide for the superior deformation response of the galfan
coating. Cracking in the galvalume coating was found at zinc interdendritic
regions [163, 164], spherical voids [163] and Si particles [163165]. The distinct
1 mm intermetallic alloy layer is extensively cracked, even in a 6 T bend diameter;
however there are only occasional indications that these cracks act as sites to
initiate cracking in the coating [164].
Fig. 54. Schematic model for the coating exfoliation process. From Ref. [162].
A.R. Marder / Progress in Materials Science 45 (2000) 191271 260
Formability is an important property in galvanneal coatings since FeZn
intermetallic phases are considered brittle. As a result, powdering and aking of
the coating can occur during the forming operation, resulting in reduced corrosion
resistance and paintability. Formability has been evaluated through many tests
including: Double-Olsen, 1808 U-bend, Swift-cup, Draw bead and 608 V-
bend [166]. Each test measures a dierent type of performance since the various
tests evaluate dierent modes of deformation. Generally, it has been found that
the degree of powdering increases with both coating iron
content [53, 134, 138, 167, 168] and coating weight [167, 168] and failure has been
observed at the gamma (G) phase layer/substrate interface [158, 169, 170] and the
delta (d) phase layer/gamma (G) phase layer interface [169, 171, 172]. It has also
been reported that the degree of cracking depends on the presence of a gamma (G)
phase layer [133, 173], the zeta (z)/delta (d) phase ratio [35, 167, 174], or the iron
gradient within the layers [168]. Delta (d) and gamma (g) are the brittle phases in
the FeZn coatings, while zeta (z) phase is the most ductile. The zeta (z) phase
may relieve some of the strains in the coating and yield better resistance to
powdering as a result of its plasticity [172, 173].
The Type 1 coating, Fig. 42, was found to have the optimum formability
properties [136, 171], but as in most forming operations lubrication to improve
metal ow is essential [175]. Galvanneal processing variables control the kinetics
and growth of the FeZn intermetallic phases and therefore dene the
morphology. However, it is the morphology of the coating, i.e. interfacial layer
and the size, shape and distribution of the various FeZn phases, that dictates the
powdering characteristics [136]. Furthemore, it is the morphology that will dictate
the secondary eects that are related to powdering such as gamma (G) phase layer
thickness [133], Fe content [127], the zeta (z)/delta (d) phase ratio [35] and the zeta
(z) phase [127, 134, 174].
7.3. Weldability
Weldability of zinc coatings is an important property of the coating, since most
galvanized product is joined in this manner. Arc welding of galvanized steel sheet
produces defects such as gas cavities (blowholes) and spatters. Recent research has
shown that the formation mechanism of gas cavities was due to vaporization of
the zinc, and methods for the reduction of blowholes were developed to improve
welding quality [176]. In spot weldability zinc coatings reduce the life of welding
electrodes due to alloying of the copper electrode with zinc. In the case of
galvanized steel, the electrode life may be as little as 15002000 welds, compared
to a tip life for bare steel of 10,000 welds [177]. This eect leads to higher
resistance, localized heating and increased pitting and erosion of the electrode tip.
As a result, manufacturing costs increase because lower tip-life reduces
productivity due to frequent down time in the welding operation to redress tips.
The mechanism of spot weld degradation is due to the fact that the temperature
at the interface between the electrode and the coating (8008C) is higher than the
A.R. Marder / Progress in Materials Science 45 (2000) 191271 261
zinc melting point (4198C) [178]. Zn alloys with the Cu of the Cu/Cr/Zr electrode
to produce brass. The mechanical properties of brass are lower than the electrode
alloy, leading to mushrooming of the active face of the electrode, Fig. 55. As the
active face increases, the current density decreases, lowering the nugget size.
Spot weldability of galvanneal coatings is improved over galvanized coating
since it is more dicult for these FeZn phases to alloy with the copper electrode,
thus improving electrode life. White et al. [179] studied the weldability of
galvanneal coatings and found that electrode life was primarily controlled by
welding current. The improvement in electrode life is due to the special ability of
the coating to form a convex shape without cavities on the electrode
tips [180, 181].This phenomenon allows a high current density to be maintained,
enhancing electrode life. Electrode life is also a function of steel substrate
composition. Boron additions to the base steel have been found to improve the
life of galvanneal coated IF steels [182, 183]. Electrode life can vary by a factor of
23 between ULC and HSLA steel, conrming the fact that the electrode life is
governed by the ability of the coating to form intermetallic phases on the
electrode tip. These phases do not stick or diuse into the weld surface and
remain on the electrode tip [184].
7.4. Paintability
Although zinc coatings are often used in the as-coated state, some applications
call for a painted surface and therefore paintability is an important design
property of the coating. It has been shown that large spangle material is dicult
to paint [113], therefore most painted products are either minimum spangle or
Fig. 55. Increase of active face diameter with the number of welds. From Ref. [178].
A.R. Marder / Progress in Materials Science 45 (2000) 191271 262
temper rolled. In most painted products, a complex coating composite is used for
corrosion protection. In addition to the Zn coated sheet steel, the composite
includes a zinc phosphate pretreatment or complex oxide thin coating, the primer
and various top coats [185]. Much work has been done on chemical
pretreatments [186] and a detailed discussion of painting technology of zinc
coatings is beyond the scope of this review, see e.g. [187]. In the automotive
industry, following the pretreatment, most automobile bodies are primed with an
electrophoretic paint (e-coat) and as a result, resistance to e-coat cratering is an
important property. At high e-coat voltages, sparking as a result of exceeding the
dielectric properties of the deposited paint lm, causes localized heat generation,
lm disruption and premature curing of the paint [188]. After paint curing these
sparked areas form pin-point craters that result in a paint surface with a
detrimental appearance [113]. Therefore, resistance to e-coat cratering, expressed
in cratering threshold voltage, is an essential paintability property, Table 12 [148].
These results show that zinciron galvanneal have a greater tendency to produce
cratering defects than galvanized steels, although galvalume coatings are superior
to both. However, galvanneal coating paintability is better than galvanized
coatings because of the microscopically rough surface formed as a result of the
FeZn alloy phases throughout the coating.
Acknowledgements
The author would like to acknowledge the numerous discussions and invaluable
help given by his colleagues A.O. Benscoter, M.R. Notis, J.N. DuPont, M.
Dubois, J. Wegria, J. Faderl and F. van Loo. The review of the manuscript and
the enlightening conversations of F.E. Goodwin is greatly appreciated. The
specic nancial support through the years by the International Lead Zinc
Research Organization, British Steel, Cockerill Sambre, Union-Miniere, Voest
Alpine, National Steel, LTV Steel, Dofasco, Rouge Steel and Noranda is greatly
acknowledged. Thanks to the Technion, Israel Institute of Technology, for the
ability to initiate this eort on my sabbatical visit. Most of all, gratitude is due to
all my former students, S. Bluni, C. Drewien, S. Hayes, R. Fraley, K. Goggins,
Table 12
Eect of hot dip coatings on threshold voltages for cratering
of cathodic electrophoretic primer [148]
Type of surface Cratering threshold, V
Uncoated bare steel >400
Zinc 275
Zinciron 225
Zn55 Al 375
Aluminium >400
A.R. Marder / Progress in Materials Science 45 (2000) 191271 263
M. Gu, C. Jordan, W. Maschek and J. Selverian, who have contributed their
research eorts to the vast body of knowledge on zinc coatings.
References
[1] Horstmann D. Reaction between iron and zinc. London: Zinc Development Association, 1978.
[2] Mackowiak J, Short NR. Metallurgy of galvanized coatings. Int Met Reviews 1979;1.
[3] GALVATECH '89. Tokyo: The Iron and Steel Institute of Japan, 1989.
[4] GALVATECH '92. Amsterdam: Stahl and Eisen, 1992.
[5] GALVATECH '95. Chicago, IL: Iron and Steel Society, 1995.
[6] GALVATECH '98. Tokyo: The Iron and Steel Institute of Japan, 1998.
[7] Krauss G, Matlock DK, editors. Zinc-based steel coating systems: metallurgy and performance.
Warrendale, PA: TMS, 1990.
[8] Marder AR, editor. The physical metallurgy of zinc coaated steel. Warrendale, PA: TMS, 1994.
[9] Goodwin FE, editor. Zinc-based steel coating systems: production and performance. Warrendale,
PA: TMS, 1998.
[10] General Galvanizing Practice. London: Hot Dip Galvanizers Association, 1965.
[11] Sendzimer T. 1938. U.S. Patent No. 2,110.893.
[12] Olefjord I, Leijon W, Jelvestam U. Selective surface oxidation during annealing of steel sheets in
H
2
/N
2
. Appl Surf Sci 1980;6:241.
[13] Krauss G, Wilshynsky DO, Matlock DK. Processing and properties of interstitial free steels. In:
Proceedings of Conference on Interstitial Free Steel Sheet: Processing, Fabrication and
Properties, vol. 1. Ottawa: Metallurgical Society of CIM, 1991.
[14] Kato C, Koumura H, Mochizuki K, Morito N. Dross formation and ow phenomena in molten
zinc bath. In: GALVATECH '95. Chicago, IL: Iron and Steel Society, 1995. p. 801.
[15] Yamaguchi H, Hisamatsu Y. Tetsu-To-Hagane 1974;60:96.
[16] Tang NY, Adams GR, Kolinyk PS. On determining eective aluminum in continuous galvaniz-
ing baths. In: GALVATECH '95. Chicago, IL: Iron and Steel Society, 1995. p. 177.
[17] Yamaguchi S. Thermochemical stability and precipitation behavior of dross phases in CGL bath.
In: GALVATECH '98. Tokyo: The Iron and Steel Institute of Japan, 1998. p. 84.
[18] Schramm J. Z Metallkde 1938;30:122.
[19] Hansen M, Anderko K. Constitution of Binary Alloys. 2nd ed., New York: McGraw Hill, 1958.
p. 738.
[20] Ghoniem MA, Lohberg K. Metallurgy 1972;10:1026.
[21] Bastin GF, van Loo FJJ, Rieck GD. A New Compound in the Iron Zinc System. Z Metallkde
1977;68:359.
[22] Gellings PJ. Z Metallkde 1980;71:70.
[23] Kubaschewski O. Massalski T, editors. Binary alloy phase diagrams. Metals Park, OH: ASM,
1986. p. 1128.
[24] Massalski TB. Phase Diagrams. ASM Metals Handbook 1992;3:206.
[25] Brown PJ. The structure of transition metalzinc alloy systems. Acta Crystall 1962;15:608.
[26] Gellings PJ, de Bree EW, Gierman G. Synthesis and charactertization of homogeneous interme-
tallic FeZn compounds. Z Metallkde 1979;70:312.
[27] Jordan CE, Marder AR. FeZn phase formation in interstitial-free steels hot-dip galvanized at
4508C, Part I 0.00 wt% AlZn baths. J Mater Sci 1997;32:5593.
[28] Gellings PJ, Gierman G, Koster D, Kuit J. Synthesis and charactertization of homogeneous
intermetallic FeZn compounds. Z Metallkde 1980;71:70.
[29] Horstmann D. Formation and growth of ironzinc alloy layers. In: Proceedings of 14th
International Hot Dip Galvanization Conference. London: Zinc Development Association, 1986.
p. 6/1.
A.R. Marder / Progress in Materials Science 45 (2000) 191271 264
[30] Rangarajan V, Giallourakis NM, Matlock DK, Krauss G. The eect of texture and microstruc-
ture on deformation of zinc coatings. J Mater Shap Technol 1989;6:217.
[31] Allen C, Mackowiak J. J Inst Met 19623;91:369.
[32] Onishi M, Wakamatsu Y, Miura H. Formation and growth kinetics of intermediate phases in
FeZn diusion couples. Trans JIM 1974;15:331.
[33] Allen C. Ph.D. Thesis. University of London, 1963.
[34] Horstmann D, Peters FK. The reaction between iron and zinc. Proceedings of 9th International
Hot Dip Galvanization Conference. London: Zinc Development Association, 1971, p. 75.
[35] Hisamatsu Y. Science and technology of zinc and zinc alloy coated steel sheet. In:
GALVATECH '89, Tokyo: ISIJ. 1989. p. 3.
[36] Sebisty JJ, Ruddle GE. Hydrogen-atmosphere galvanizing of iron-base alloys, Research Report
No. R 255. Ottawa: DOE, Mines Branch, 1972.
[37] Elias JA, Hook RE. Interstitial free sheet steel applications and performance, SAE Paper No.
720018, 1972.
[38] Marder AR. Zinc coating of interstitial-free steel sheet. In: Proceedings of Conference on
Interstial-Free Steel Sheet: Processing, Fabrication and Properties. Ottawa: Metallurgical Society
of CIM, 1991. 157.
[39] Toki T, Oshima K, Nakamori T, Saito Y, Tsuda T, Hobo Y. 1994. Eect of P content in ultra-
low carbon Ti stabilized steel on the rate of FeZn alloy formation through ferritic grain bound-
ary diusion during hot-dip galvanizing. In: Marder AR, editor. The Physical Metallurgy of Zinc
Coated Steel. Warrendale, PA: TMS. p. 169.
[40] Jordan CE, Marder AR. A model for galvanneal morphology development. In: Marder AR, edi-
tor. The Physical Metallurgy of Zinc Coated Steel. Warrendale, PA: TMS, 1994. p. 197.
[41] Con C, Thompson SW. The eect of 0.07 weight percent phosphorous on the galvannealing
behavior of a niobium/titanium-stabilized interstitial-free sheet steel. In: Hartmann JE.
Warrendale, editors. GALVATECH '95. PA: ISS. p. 121.
[42] Allegra L, Hart RG, Townsend HE. Intergranular zinc embrittlement and its inhibition by phos-
phorous in 55 pct AlZn coated sheet. Met Trans 1983;14A:401.
[43] Guttmann M, Lepretre Y, Aubry A, Roche M-J, Moreau T, Drillet P, Mataigne JM, Baudin H.
Mechanism of the galvanizing reaction. Inuence of Ti and P contents in steel and of its surface
microstructure after annealing. In: GALVATECH '95. Chicago, IL: Iron and Steel Society, 1995.
p. 295.
[44] Ko ster W, Go decke T. Das Dreistosystem Eisen-Aluminum-Zink. Z Metallkde 1970;61:642.
[45] Urednicek M, Kirkaldy JS. Mechanism of iron attack inhibition arising from additions of alumi-
num to liquid Zn(Fe) during galvanizing. Z Metallkde 1987;64:649.
[46] Osinski K. The inuence of aluminum and silicon on the reaction between iron and zinc.
Doctoral Thesis. Technical University, Eindhoven, 1983.
[47] Perrot P, Tissier J-C, Dauphin J-Y. Stable and metastable equilibria in the FeZnAl system at
4508C. Z Metallkde 1992;83:11.
[48] Chen ZW, Sharp RM, Gregory JT. FeAlZn ternary phase diagram at 4508C. Mater Sci
Technol 1990;6:1173.
[49] Ghuman ARP, Goldstein JI. Reaction mechanisms for the coatings formed during hot dipping of
iron in 0 to 10 Pct AlZn baths at 4507008C. Met Trans 1971;2A:2903.
[50] Belisle S, Lezon V, Gagne M. The solubility of iron in continous hot-dip galvanizing baths. J
Phase Equilibria 1991;12:259.
[51] Dauzat M, Stouvenot F, Moreau T. Zinc rich corner of the FeZnAl revised phase diagram,
Amsterdam: Stahl and Eisen, 1992. p. 449.
[52] Tang N-Y, Adams GR. Studies on the inhibition of alloy formation in hot-dip galvanized coat-
ings. In: Marder AR, editor. The Physical Metallurgy of Zinc Coated Steel. Warrendale, PA:
TMS. 1994. p. 41.
[53] Tang N-Y. Comments on FeAlZn diagram. J Phase Equilibria 1994;15:237.
[54] Tang N-Y, Adams GR, Kolisnyk PS. On determining eective aluminum in continous galvaniz-
ing baths. GALVATECH '95. Chicago, IL: Iron and Steel Society, 1995. p. 777.
[55] Tang N-Y. Modeling of enrichment in galvanized coatings. Met Mater Trans 1995;26A:1669.
A.R. Marder / Progress in Materials Science 45 (2000) 191271 265
[56] Raynor GV. Institute of metals annotated equilibrium diagram series. vol. 8. London: The
Institute of Metals, 1951.
[57] Guttmann M. Diusive phase transformations in hot dip galvanizing. Mater Sci Forum
1994;155(156):527.
[58] Lepre tre Y, Mataigne JM, Guttmann M, Philibert J. Reactive interdiusion in the FeAlZn sys-
tem: reaction mechanisms during hot-dip galvanizing. In: Goodwin FE, editor. Zinc-based steel
coating systems: production and performance. Warrendale, PA: TMS, 1998. p. 95.
[59] Tang N-Y. Thermodynamics and kinetics of alloy formation in galvanized coatings. In: Goodwin
FE, editor. Zinc-based steel coating systems: production and performance. Warrendale, PA:
TMS, 1998. p. 3.
[60] Yamaguchi S, Fukatsu N, Kimura H, Ueda J, Iguchi Y, Ohashi T. Development of an Al sensor
for molten zinc bath using composite salt electrolyte. CAMP-ISIJ 1991;4:669.
[61] Isobe M. Initial alloying behavior in galvannealing process. CAMP-ISIJ 1992;5:1629.
[62] Faderl J, Pimminger M, Scho nberger L. Inuence of steel grade and surface topography
on the galvannealing reaction. GALVATECH '92. Amsterdam: Stahl and Eisen, 1992.
p. 194.
[63] Kanamaru T, Nakayama M. Alloying reaction control in production of galvannealed steel.
Mater Sci Res Int 1995;1:150.
[64] Nakayama M, Kanamaru T, Numakura Y. Eect of alloying conditions and phase structure on
alloying behavior and formability of galvannealed steel sheet. CAMP-ISIJ 1992;5:1665.
[65] Borzillo AR, Hahn WC. Growth of the inhibiting aluminum-rich alloy layer on mild steel during
galvanizing in zinc that contains aluminum. Trans ASM 1969;62:729.
[66] Yamaguchi H, Hisamatsu Y. Reaction mechanism of the sheet galvanizing. Trans ISIJ
1979;19:649.
[67] Jordan CE, Goggins KM, Benscoter AO, Marder AR. Metallographic preparation technique for
hot-dip galvanized and galvanneal coatings on steel. Mater Characteriz 1993;31:107.
[68] Ohtsubo H. Crystallography of intermetallic interface layers in hot-dip galvanizing steel sheets.
ISIJ Int 1996;36:317.
[69] Inagaki J, Sakurai M, Watanabe T. Alloying reactions in hot-dip galvanizing and galvannealing
processes. ISIJ Int 1995;35:1388.
[70] Inagaki J, Sakurai M, Sagiyama M. FeZn alloying reactions in CGL process. GALVATECH
'92. Amsterdam: Stahl and Eisen, 1992. p. 220.
[71] Kiusalaas R, Engberg G, Klang H, Schedin E, Scho n L. Control of texture and formation of
intermetallic phases in continuously hot-dip galvanized coatings. In: GALVATECH '89. Tokyo:
The Iron and Steel Institute of Japan, 1989. p. 485.
[72] Kato T, Nunome K, Morimoto Y, Nishimura K, Kato N, Saka H. In situ TEM observations of
reaction between Fe and molten Zn(Al). GALVATECH '98. Tokyo: The Iron and Steel Institute
of Japan, 1999. p. 803.
[73] Jordan CE, Marder AR. Eect of phosphorous surface segregation on ironzinc reaction kine-
ticsduring hot-dip galvanizing. Met Mater Trans 1997;28A:2695.
[74] Hisamatsu Y. Science and technology of zinc and zinc alloy coated sheet steel. GALVATECH
'89. Tokyo: The Iron and Steel Institute of Japan. 1989. p. 3.
[75] Saito M, Uchida Y, Kittaka T, Hirose Y, Hisamatsu Y. Tetsu-to-Hagane 1991;77:947.
[76] Nishimoto A, Inagaki J, Nakaoka K. Eects of surface microstructure and chemical composition
of steels on formation of FeZn compounds during continuous galvanizing. Trans ISIJ
1986;26:807.
[77] Abe M, Kanbara S. Tetsu-to-Hagane 1983;69:S1061.
[78] Jordan CE, Marder AR. The eect of iron oxide as an inhibition layer on ironzinc reactions
during hot-dip galvanizing. Met Mater Trans 1998;29B:479.
[79] Jordan CE, Marder AR. The eect of subsrate grain size on ironzinc reactions during hot-dip
galvanizing. Met Mater Trans 1997;28A:2683.
[80] Jordan CE, Marder AR. Inhibition layer break down and outburst FeZn alloy formation during
galvanizing. In: Goodwin FE, editor. Zinc-based steel coating systems: production and perform-
ance. Warrendale, PA: TMS, 1998. p. 115.
A.R. Marder / Progress in Materials Science 45 (2000) 191271 266
[81] Abe M, Kanbara S, Okuyama T. Iron zinc alloy behavior in galvannealing P and PSi contain-
ing steel. Trans ISIJ 1985;25:B15.
[82] Allegra L, Hart RG, Townsend HE. Intergranular zinc embrittlement and its inhibition by phos-
phorous in 55Pct-AlZn coated sheet steel. Met Trans 1983;14A:401.
[83] Mercer PD. Factors which aect the alloy growth rate in galvannealed coatings. GALVATECH
'92. Amsterdam: Stahl and Eisen, 1992. p. 204.
[84] Lin CS, Chiou WA, Meshi M. Eect of phosphorous content in base-steel on the formation of
alloy layer of hot-dip coated steel sheets. In: Marder AR, editor. The physical metallurgy of zinc
coated steel. Warrendale, PA: TMS, 1994. p. 31.
[85] Lin CS, Meshi M. The eect of steel chemistry on the formation of FeZn intermetallic com-
pounds of galvannealed-coated steel sheets. Met Mater Trans 1994;25A:721.
[86] Jordan CE, Zuhr R, Marder AR. Eect of phosphorous surface segregation on ironzinc reac-
tion kinetics during hot-dip galvanizing. Met Mater Trans 1997;28A:2695.
[87] Lepre tre Y, Mataigne JM. Reaction mechanisms during hot dip galvanizing: eect of phosphor-
ous on coating development. In: GALVATECH '98. Tokyo: The Iron and Steel Institute of
Japan, 1998. p. 133.
[88] Kodras MS, Niessen P. Silicon induced destabilization of galvanized coatings in the sandelin
peak region. Metallography 1989;22:253.
[89] Lepre tre Y, Mataigne JM. A new interpretation of the sandelin eect. In: Goodwin FE, editor.
Zinc-based steel coating systems: production and performance. Warrendale, PA: TMS, 1998.
p. 303.
[90] Jordan CE, Marder AR. FeZn phase formation in interstitial-free steels hot-dip galvanized at
4508C, Part II 0.20 wt% AlZn baths. J Mater Sci 1997;32:5603.
[91] Caceres PG, Hotham CA, Spittle JA, Jones RD. Mechanisms of formation and growth of inter-
metallic layers during hot dipping of iron in Zn3Al and Zn6Al baths. Mater Sci Technol
1986;2:871.
[92] Pelerin J, Servais JP, Coutsouradis D, Herrschaft DC, Radke SF. Proceedings of 13th
International Hot Dip Galvanization Conference. London: Zinc Development Association, 1982.
p. 49/I.
[93] Takahashi A, Kato T, Funaki S. CAMP-ISIJ 1989;2:1662.
[94] Ichiyama K, Kobayashi J, Sugimoto S. Proceedings of 14th International Hot Dip Galvanization
Conference, Munich: Zinc Development Association, 1985. p. 9/1.
[95] Makimattila SJ. On the production possibilities of a deep drawing quality Zn5% Al coated steel
sheet. Scan J Metall 1986;15:224.
[96] Chen ZW, Gregory JT, Sharp RM. Intermetallic phases formed during hot dipping of low car-
bon steel in a Zn5%Al melt at 4508C. Met Trans 1992;23A:2393.
[97] Yang Y, Yu Z. Eect of mischmetal on reaction between solid iron and liquid Zn+5 wt% Al
alloy. Acta Met Sin 1993;29:A540.
[98] Sharp RM, Gregory JT, Chen ZW. Intermetallic phases formed during hot dipping of
low-carbon steel in a Zn5%Al melt at temperatures up to 5758C. Mater Forum 1992;
16:205.
[99] Lin KL, Ho JK, Jong CS, Lee JT. The formation of intermetallics and its eect on the micro-
structure of the hot dip 5%AlZn coatings on steel. In: Marder AR, editor. The physical metal-
lurgy of zinc coated steel. Warrendale, PA: TMS, 1994. p. 89.
[100] Selverian JH, Marder AR, Notis MR. The reaction between solid iron and liquid AlZn baths.
Met Trans 1988;19A:1193.
[101] Selverian JH, Marder AR, Notis MR. The eect of silicon on the reaction between solid iron
and liquid 55 wt pct AlZn baths. Met Trans 1989;20A:543.
[102] Selverian JH, Marder AR, Notis MR. The microstructure of 55 w/o AlZnSi (galvalume) hot
dip coatings. J Mater Engng 1987;9:133.
[103] Mercer PD. The inuence of temperature on the growth rate and composition of the interfacial
alloy in 55% Al, 43.5% Zn, 1.5% Si hot-dipped coatings. In: Marder AR, editor. The physical
metallurgy of zinc coated steel. Warrendale, PA: TMS, 1994. p. 129.
A.R. Marder / Progress in Materials Science 45 (2000) 191271 267
[104] Zhou ZF, Tilden GM, Liu QY, Mercer PD. The formation and growth of the interfacial layer in
Zn55% Al1.5% Si hot-dipped coatings. GALVATECH '95. Chicago, IL: Iron and Steel
Society, 1995. p. 289.
[105] Strutzenberger J, Faderl J. Solidication and spangle formation of hot-dip-galvanized zinc coat-
ings. Met Mater Trans 1998;29A:631.
[106] Cameron DI, Harvey GJ. Proceedings of 8th International Hot Dip Galvanization Conference,
London: Zinc Development Association, 1967. p. 86.
[107] Kim Y-W, Patil RS. Solidication characteristics of hot dip galvanized steel. In: Proceedings of
1st International Conference on Zinc Coated Steel. Munich: Zinc Development Association,
1985. p. D/1.
[108] White DWG. J Inst Met 1971;99:287.
[109] Fasoyino FA, Weinberg F. Spangle formation in galvanized sheet steel coatings. Met Trans
1990;21B:549.
[110] Murray JL. In: Massalski TB, editor. Binary alloy phase diagram. Materials Park, OH: ASM,
1986. p. 185.
[111] Marder AR. Microstructural characterization of zinc coatings. In: Krauss G, Matlock DK, edi-
tors. Zinc-based steel coating systems: metallurgy and performance. Warrendale, PA: TMS, 1990.
p. 55.
[112] Wall NJ, Spittle JA, Jones RD. Proceedings of 1st International Conference on Zinc Coated
Steel. Munich: Zinc Development Association, 1985. p. C/1.
[113] Leidheiser H, Kim DK. J Met 1976;28:19.
[114] Fraley RE. The eect of lead on the solidication and preferred orientation of the zinc coating
on continously hot dipped galvanized sheet steel. M.S. Thesis. Lehigh University, 1994.
[115] Sebisty JJ, Palmer RH. Proceedings of 6th International Hot Dip Galvanization Conference,
Interlaken. London: Zinc Development Association, 1962. p. 215.
[116] Radeker W, Friehe W. Proceedings of 7th International Hot Dip Galvanization Conference,
Paris. Oxford: Zinc Development Association, 1964. p. 167.
[117] Herrscaft DC, Radke SF, Coutsouradis D, Pelerin J. SAE Paper No. 830517, Warrendale, PA:
SAE, 1983.
[118] Harvey GJ, Richards RN. Zinc-based coatings for corrosion protection of steel sheet and strip.
Met Forum 1984;6:234.
[119] Radtke SF, Herrschaft DC. Role of mischmetal in galvanizing with a Zn5%Al alloy. J Less
Common Met 1983;93:253.
[120] Skenazi AF, Davin D, Coutsouradis D, Goodwin FG. Proceedings of 1st International
Conference on Zinc Coated Steel, Munich, London: Zinc Development Association, 1985.
[121] Makimattila SM, Ristolainen E. Proceedings of 3rd International Hot Dip Galvanization
Conference, London: Zinc Development Association, 1982.
[122] Bluni ST, Notis MR, Marder AR. Nucleation characteristics and microstructure in o-eutectic
AlZn alloys. Acta Metall Mater 1995;43:1775.
[123] Bluni ST, Marder AR, Goldstein JI. Surface characterization of hot-dip galfan coatings. Mater
Characteriz 1994;33:93.
[124] Bluni ST, Marder AR. Improvement of galfan surface appearance for coil coating uses.
4th International Zinc Coated Sheet Conference. Paris: Zinc Development Association, 1994.
p. SC6/1.
[125] Selverian JH, Marder AR, Notis MR. J Electron Micro Technol 1987;5:223.
[126] Cleary HJ. Microstr Sci 1985;12:103.
[127] Jordan CE, Marder AR. Morphology development in hot-dip galvanneal coatings. Met Mater
Trans 1994;25a:937.
[128] van der Heiden A, Burghardt AJC, van Koesveld W, van Perlstein EB, Spanjers MGJ.
Galvanneal microstructure and anti-powdering process windows. In: Marder AR, editor. The
physical metallurgy of zinc coated steel. Warrendale, PA: TMS, 1994. p. 251.
[129] Jordan CE, Marder AR. A model for galvanneal morphology development. In: Marder AR,
editor. The physical metallurgy of zinc coated steel. Warrendale, PA: TMS, 1994.
p. 197.
A.R. Marder / Progress in Materials Science 45 (2000) 191271 268
[130] Con C, Thompson SW. 1994. Galvannealing of interstitial-free sheet steels strengthened by
manganese, silicon, or phophorous: an initial study. In: Marder AR., editor. The physical metal-
lurgy of zinc coated steel. Warrendale, PA: TMS, 1994. p. 181.
[131] Kikaldy JS, Urednicek M. Mechanism of iron attack inhibition arising from additions of alumi-
num to liquid zinc(Fe) during galvanizing at 4508C. Z Metallkde 1973;64:2903.
[132] Maschek W, Hayes SP, Marder AR. Cross sectional studies of zinc iron phase growth in an en-
vironmental scanning electron microscope. In: GALVATECH '95. Chicago, IL: Iron and Steel
Society, 1995. p. 309.
[133] Jordan CE, Goggins KM, Marder AR. Interfacial layer development in hot-dip galvanneal coat-
ings on interstitial free (IF) steel. Met Mater Trans 1994;25a:2101.
[134] Claus G, Dilewijns J, De Cooman BC, Meers U. Determination of the process window for opti-
mal galvannealing of TiIF steel. GALVATECH '95. Chicago, IL: Iron and Steel Society, 1995.
p. 107.
[135] Goggins KM, Marder AR. Crack initiation and propagation in hot-dip galvanneal steel sheet
during bending. 3rd International Conference on Zinc Coated Sheet, Barcelona 1991;S4I:111.
[136] Jordan CE, Goggins KM, Marder AR. Kinetics and formability of hot-dip galvanneal coatings.
GALVATECH '92. Amsterdam: Stahl and Eisen, 1992. p. 137.
[137] van Koesveld W, Lamberigts M, van der Heiden A, Bordignon L. Coating microstructure assess-
ment and control for advanced product properties of galvannealed IF steels. GALVATECH '95.
Chicago, IL: Iron and Steel Society, 1995. p. 343.
[138] Stadlbauer A, Rubenzucker F, Zeman K, Fuhrmann E, Faderl J, Scho nberger L. A new galvan-
nealling process controller. GALVATECH '95. Chicago, IL: Iron and Steel Society, 1995. p. 81.
[139] Hayes S. Morphology development of galvannealed coatings based on steel substrate chemistry
and processing parameters. M.S. Thesis. Lehigh University, 1994.
[140] Rios PR, Seixas UR. Modeling of isothermal and nonisothermal kinetics of iron enrichment in
hot-dip galvanneal coatings on IF steel sheets. GALVATECH '98. Tokyo: The Iron and Steel
Institute of Japan, 1998. p. 215.
[141] McDevitt ET, Morimoto Y, Meshii M. Microstructural evolution during galvannealing. Part 1:
formation and break down of the FeAl inhibition layer. GALVATECH '98. Tokyo: The Iron
and Steel Institute of Japan, 1998. p. 153.
[142] Kato T, Nunome K, Saka H. Formation of z phase at an interface between Fe and Zn(Al)
GALVATECH '98. Tokyo: The Iron and Steel Institute of Japan, 1998. p. 797.
[143] Arai M. Eects of steel type and Al content in zinc bath on alloying behavior of galvannealing
steel sheet. CAMP-ISIJ 1992;5:1649.
[144] Urai M, Terada M, Yamaguchi M, Nomura S. CAMP-ISIJ 1988;1:651.
[145] Maki J. Eect of steel type alloying behavior of galvannealed steel sheets. CAMP-ISIJ
1992;5:1641.
[146] Jordan CE. Structure/properties of galvannealed coated steel. M.S. Thesis. Lehigh University,
1991.
[147] Lin CS, Meshii M. Eect of steel chemistry on the microstructure and mechanical properties of
the commercial galvanneal coatings. GALVATECH '95. Chicago, IL: Iron and Steel Society,
1995. p. 477.
[148] Marder AR. Eects of surface treatments on materials performance. Materials selection and de-
sign. ASM Handbook, vol. 20. 1997. p. 470.
[149] Nevison DCH. Corrosion of zinc. Corrosion. ASM Handbook, vol. 3. 1987. p. 755.
[150] Townsend HE. Continuous hot dip coatings. Surface engineering. ASM Handbook, vol. 5. 1994.
p. 339.
[151] Zhang XG. Corrosion and electrochemistry of zinc, New York: Plenum Press, 1996.
[152] Goodwin FE. Mechanisms of corrosion of zinc and zinc5% aluminum steel sheet coatings. In:
Krauss G, Matlock DK, editors. Zinc-based steel coating systems: metallurgy and performance.
Warrendale, PA: TMS, 1990. p. 183.
[153] Uchima Y, Hasaka M, Koga H. Eect of structure and mischmetal addition on the corrosion
behavior of Zn5 mass% Al alloy. GALVATECH '89. Tokyo: The Iron and Steel Institute of
Japan, 1989. p. 545.
A.R. Marder / Progress in Materials Science 45 (2000) 191271 269
[154] Tajiri Y, Shimada S, Yamaji T, Adaniya T. Eect of coating structure on corrosion resistance of
hot dipped Zn5%AlMg alloy coated steel sheet. GALVATECH '89. Tokyo: The Iron and
Steel Institute of Japan, 1989. p. 553.
[155] Aoki T, Miyoshi Y, Kittaka T. Results of 10-year atmospheric corrosion testing of hot dip Zn5
mass% Al alloy coated sheet steel. GALVATECH '95. Chicago, IL: Iron and Steel Society, 1995.
p. 463.
[156] Zoccola JC, Townsend HE, Borzillo AR, Horton JB. In: Coburn SK, editor. Atmospheric factors
aeacting the corrosion of engineering metals. ASTM STP 646. Philadelphia, PA: ASTM, 1978.
p. 165.
[157] Palma E, Puente JM, Morcillo M. The atmospheric corrosion mechanism of 55% AlZn coating
on steel. Corrosion Sci 1998;40:61.
[158] Townsend HE, Zoccola JC. Mater Perform 1979;10:13.
[159] Townsend HE, Borzillo AR. Thirty-year atmospheric corrosion performance of 55 wt% alumi-
numzinc alloy coated sheet steel. GALVATECH '95. Chicago, IL: Iron and Steel Society, 1995.
p. 171.
[160] Opbroek JB, Granzow WG. A deep drawing, hot-dipped galvanized steel for dierent forming
applications, SAE Paper No. 850275. Warrendale, PA: SAE, 1985.
[161] Deits SH, Matlock DK. Formability of coated sheet steels: an analysis of surface damage mech-
anisms. In: Krauss G, Matlock DK, editors. Zinc-based steel coating systems: metallurgy and
performance. Warrendale, PA: TMS, 1990. p. 297.
[162] Arimura M, Urai M, Iwaya J, Iwai M. Eects of press forming factors and ash plating on coat-
ing exfoliation of galvannealed steel sheets. GALVATECH '95. Chicago, IL: Iron and Steel
Society, 1995. p. 733.
[163] Shah SRH, Dilewijns JA, Jones RD. The structure and deformation behavior of zinc-rich coat-
ings on steel sheet. GALVATECH '92. Amsterdam: Stahl and Eisen, 1992. p. 105.
[164] Willis DJ. Cracking characteristics of zinc and zincaluminum alloy coatings. GALVATECH
'89. Tokyo: The Iron and Steel Institute of Japan, 1989. p. 351.
[165] Cape TW, Gomersall DW, Denner SG. Tension bend staining of prepainted galvalume. In:
Goodwin FE, editor. Zinc-based steel coating systems: production and performance. Warrendale,
PA: TMS, 1998. p. 271.
[166] Gallo E, Matlock DK. The importance of microstructure on the formability of galvannealed I.F.
sheet steel. GALVATECH '95. Chicago, IL: Iron and Steel Society, 1995. p. 739.
[167] Bae DC, Choi YM, Chang S, Shin JC. Improved powdering resistance of hot-dip galvannealed
steel sheets. GALVATECH '92. Amsterdam: Stahl and Eisen, 1992. p. 132.
[168] Jagannathan V. The inuence of coating characteristics on the powdering and corrosion resist-
ance of galvanneal coated steel sheet. GALVATECH '92. Amsterdam: Stahl and Eisen, 1992.
p. 127.
[169] Martin P, Handford MA, Packwood R, Dignard L, Moore V. Mechanical and structural study
of ZnFe coatings. GALVATECH '92. Amsterdam: Stahl and Eisen, 1992. p. 112.
[170] Saka H, Kato T, Hong MH, Sasaki K, Kuroda K, Kamino T. Cross-sectional TEM observation
of interfaces in a galvannealed steel. GALVATECH '95. Chicago, IL: Iron and Steel Society,
1995. p. 809.
[171] Lucas PL, Quantin D, Brun CG. Eects of alloying parameters on formability of galvan-
nealed sheets. GALVATECH '89. Tokyo: The Iron and Steel Institute of Japan, 1989.
p. 138.
[172] Shi MF, Smith GM, Moore M, Meuleman DJ. Galvannealed coating optimized for coating ad-
hesion through dies. In: Krauss G, Matlock DK, editors. Zinc-based steel coating systems: metal-
lurgy and performance. Warrendale, PA: TMS, 1990. p. 387.
[173] Cheng C, Rangarajan V, Franks L, L'Ecuyer J. The eect of steel substrate on the microstructure
and powdering of galvanneal coatings. GALVATECH '92. Amsterdam: Stahl and Eisen, 1992. p.
122.
[174] Kato C, Koumura H, Uesugi Y, Mochizuki K. Inuence of phase composition on formability of
galvannealed steel sheet. In: Marder AR, editor. The physical metallurgy of zinc coated steel.
Warrendale, PA: TMS, 1994. p. 241.
A.R. Marder / Progress in Materials Science 45 (2000) 191271 270
[175] Goggins KM, Marder AR. Coating formability evaluation method for hot-dip galvannealed
sheet. SAE Paper No.920175. Warrendale, PA: SAE, 1992.
[176] Matsui H, Oikawa H. Arc welding technologies of galvanized steel sheet for automotive under-
body. GALVATECH '98. Tokyo: The Iron and Steel Institute of Japan, 1998. p. 778.
[177] Howe P, Kelley SC. A comparison of the resistance spot weldability of bare, hot-dipped, galvan-
nealed and electrogalvanized DQSK sheet steels. SAE Technical Paper No. 880280. Warrendale,
PA: SAE, 1988.
[178] Gobez P. Improvement of electrode life in resistance spot welding. GALVATECH '92.
Amsterdam: Stahl and Eisen, 1992. p. 96.
[179] White CL, Lu F, Kimchi M, Dong P. Resistance welding electrode wear on galvannealed steels.
In: Zinc-based steel coating systems: production and performance. ed. Goodwin FE. Warrendale,
PA: TMS, 1998. p. 219.
[180] Sakiyama T. Electrode tip life of hot-dip galvannealed steel. CAMP-ISIJ 1990;3:1508.
[181] Ikeda R. Resistance spot weldabing of galvannealed steel sheets. CAMP-ISIJ 1990;3:1506.
[182] Tobiyama Y, Kato C, Yasuda A. Improvement in formability and spot weldability of galvan-
nealed steel sheet. SAE Technical Paper No. 940593. Warrendale, PA: SAE, 1994.
[183] Kato C. Eect of B addition in ultra-low carbon steel substrate on spot weldability of galvan-
nealed steel sheet. CAMP-ISIJ 1992;5:1633.
[184] Guth J, Mataigne JM. Comparing spot-weldability of galvanneal coated steel sheets: mechanism
and base metal inuence. GALVATECH '95. Chicago, IL: Iron and Steel Society, 1995. p. 709.
[185] Scho CK. Interactions between automotive electrodeposition primers and pretreatments/zinc
coated steels. GALVATECH '89. Tokyo: The Iron and Steel Institute of Japan, 1989. p. 230.
[186] Rausch W. Chemical surface treatment of steel coated with zinc and zinc alloys. GALVATECH
'89. Tokyo: The Iron and Steel Institute of Japan, 1989. p. 197.
[187] Surface Engineering. ASM Handbook, vol. 5, ASM International, 1994.
[188] Katayama T, Kume M. Relations between coating failures and compositions of cathodic electro-
coat primer. GALVATECH '92. Amsterdam: Stahl and Eisen, 1992. p. 245.
A.R. Marder / Progress in Materials Science 45 (2000) 191271 271

You might also like