Professional Documents
Culture Documents
Introduction
1172
array detector (HPLC-DAD).[14,15] Kirby also proposes that, in addition to the laccaic acids, the identification of erythrolaccin in a lac
paint may be considered as a marker for the possible presence of
shellac.[4] Over the years, the characterization of lake pigments
and dyes in works of art has greatly benefited from the advent of
new advanced techniques, such as microspectrofluorimetry and
surface-enhanced Raman spectroscopy (SERS),[1620] as well as from
the development of improved sample pretreatments, such as nonextractive gassolid hydrolysis procedures or mild extraction
methods.[2123] Recent articles have shown that SERS can be used
* Correspondence to: Maria Joo Melo, REQUIMTE-CQFB and Department of
Conservation and Restoration, Faculdade de Cincias e Tecnologia, Universidade
Nova de Lisboa, Campus da Caparica, 2829-516 Caparica, Portugal.
E-mail: mjm@fct.unl.pt
** Correspondence to: Marco Leona, Department of Scientific Research, The
Metropolitan Museum of Art, 1000 Fifth Avenue, New York, NY 10028, USA.
E-mail: Marco.Leona@metmuseum.org
This article is part of the special issue of the Journal of Raman Spectroscopy
entitled Raman in Art and Archaeology 2013 edited by Polonca Ropret and Juan
Manuel Madariaga.
Experimental
Historically accurate reconstructions
All reconstructions were carefully prepared with distilled water, reagent grade chemicals and sticklac from Kremer. Each recipe was
repeated several times. The lac paints were applied on parchment
wileyonlinelibrary.com/journal/jrs
1173
Figure 2. Examples of dark reds/carmine from Lorvo 5, f.6; Santa Cruz 20, f.86; and Alcobaa 419, f.98. Manuscripts at ANTT (Lisbon), BPMP (Porto) and BNP
(Lisbon), respectively.
R. Castro et al.
(acquired from the Muse du Parchemin, Rouillon), filter paper and/or
glass slides with parchment glue and/or egg white as binderswhich
were the most likely used in these collections.[25,35] Parchment glue
was prepared as described in LKFK, chapter 40;[30] egg white was
made based on the De Clarea treatise.[38]
Lac reference samples were prepared following documented
sources dating from the 11th to the 15th centuries,[2832] adapted
to laboratory conditions. Two types of recipes will be distinguished
for the purpose of this work, that is, free lac dyes (A) and alumlac
complexes (B):
(A) In the Ibn Badis recipe from chapter six,[28] 5 g of ground lac
was extracted in a 50 ml solution of 0.1 g of sodium carbonate and borax (pH 10), then heated, filtered and heated
again until the ink gained thick consistency; LKFK recipe,
from chapter 13,[30] uses 150 ml of urine filtered in a bed of
quicklime and ashes, in a proportion of 2:1 (pH 910);
10 g of ground lac is added into the solution during heating
and then filtered.
(B) In the Mappae Clavicula,[29] 10 g of sticklac is extracted in
100 ml of heated urine (pH 6); then, 0.6 g of alum (AlK
(SO4)212H2O) is added, and the solution filtered; in the
Bolognese 129 recipe,[31] 150 ml of one-week stale urine is
heated; then, ashes are added (pH 10). Five grams of
ground lac is mixed in the solution and filtered again before
0.6 g of alum is added; Bolognese 130[31] uses 60 ml of 20-day
stale urine (pH 67), 10 g of finely ground lac and 2 g of
alum in the same order but in different quantities (note: this
recipe also uses brazilwood, but we decided not to include it
in this study, as it would make spectral interpretation more
challenging at this stage); Strasbourg recipe named Bright
Paris red[32] starts by leaving 5 g of ground lac in a 50 ml
of lye of ashes (pH 11) overnight. The next day, the solution
is heated and 2 g of alum is added to the mix until the
lake precipitate is obtained. The solution is then filtered
and centrifuged.
In addition, we also used as reference materials for the analytical
approach a set of reconstructions previously made in the framework
of another study.[39] These were mainly prepared by extracting 3.6 g
or 7.2 g of ground sticklac with 30 or 60 ml of different acidic and
basic solutions, respectively. These reconstructions, not based on
medieval documentary sources, were applied on parchment with
parchment glue. For the purpose of this work, these will be referred
as lac 1, lac 2 and so on, as also part of type A free lacs, and lac dye
x0 precipitated, as type B complexed lacs.
For further information on the color of the reconstructions,
consult Table S1 (Supporting Information).
Other reference samples
SERS
1174
Laccaic acids A, B, C and E were collected after HPLC-DAD separation of a 10 3 mol/l H2O : MeOH 70:30 lac dye (Fluka) solution. The
instrumentation and solvent gradient were identical to those
described in the preceding texts, although perchloric acid was
replaced by a 10% aqueous solution of formic acid 99.9% (v/v). Each
laccaic acid was collected separately into a round-bottom flask
according to the retention time (several runs were performed to
collect a sufficient amount of each compound). The so obtained
solutions were rotavapped to remove the methanol and then
lyophilized. HPLC analysis of the samples was subsequently
wileyonlinelibrary.com/journal/jrs
Infrared spectra were acquired with a Nicolet Nexus spectrophotometer coupled to a Continum microscope (15 objective) with
a MCT-A detector cooled by liquid nitrogen. Spectra were obtained
in transmission mode, between 4000 and 650 cm 1, with a
resolution of 4 cm 1 and 128 or 256 scans. Samples were preliminarily compressed using a Thermo diamond anvil compression cell.
Spectra are shown here as acquired, without further manipulations,
except for the occasional removal of the CO2 absorption at ca.
23002400 cm 1.
Microspectrofluorimetry
The dark reds (carmine) and pink colors found in the manuscript
illuminations were historically applied as a single color or as a matiz;
the pink color was identified as a mixture of a dye with white lead or
wileyonlinelibrary.com/journal/jrs
1175
R. Castro et al.
white lead and vermilion (by Raman, infrared and microfluorescence
UV-Vis and supported by XRF[25,33,34]); some darker reds in Holy Cross
manuscripts were identified as the dye admixed with vermilion,
Fig. S4 (Supporting Information).[25,34] All of these paints were
applied with a proteinaceous binder, such as parchment glue or
egg white.[35] For these carmine paints, the FTIR spectra essentially
display the binders fingerprint as well as, in most cases, the signals
as a result of the fillers, Fig. S4 (Supporting Information). In some
samples, lac dye is indirectly detected upon identification of the
CH stretching absorption bands of the shellac resin, Fig. S2
(Supporting Information). Chalk or gypsum was most commonly
detected for Alcobaa samples but was also found in a few
specimens from Holy Cross. Because of the high fluorescence in
the Raman signal, these fillers were only detected by FTIR.
SERS results
Historically accurate reconstructions and standard laccaic acids
SERS spectra for the lacalum reconstructions could only be
obtained upon HF hydrolysis. Spectra of these reference samples
(Fig. 3) displayed the typical pattern of the main component of
lac dye, that is, laccaic acid A, with the most intense signals at circa
1578, 1464, 1368, 1326, 1287, 1227, 1188, 1098, 1052, 1010, 453 and
408 cm 1 (Fig. 4(d)). On the other hand, reference SERS spectra
from laccaic acids A, B, C and E as collected from HPLC are shown
in Fig. S5 (Supporting Information). The similarity between the
spectra of laccaic acid A and the commercial mixture (Figs S5a
and S5b) indicates that SERS is detecting primarily the former
compound, that is, the main dye chromophore. We experimentally
acquired the relative composition of lac dye chromophores (from
Fluka), by HPLC-DAD: laccaic acid A = 50%, laccaic acid B = 25%,
laccaic acid C = 20% and laccaic acid E = 5% (relative areas). Overall,
the best spectra were obtained for lakes that were in more acidic
conditions. This is in agreement with Caamares et al.[24] who found
that the SERS intensity for lac dye increases going from alkaline to
Figure 4. SERS spectra of the historical samples, at exc = 488 nm: (a)
Lorvo 5, f.6; (b) Santa Cruz 20, f.191; (c) Alcobaa 446, f.96v; (d) laccaic
acid A at pH = 2.
acidic pHs. This is likely because of the increase in the resonance Raman effect combined with a lower electrostatic repulsion between
laccaic acid and the negatively charged nanoparticle surface
(capped by the citrate ions) at acidic pH. Within this group of lac
lakes, minor shifts in wave number and slight changes in relative
intensities were observed, which may be attributed to pH effects
or to other components present in the matrix that may interfere
with the SERS signal.
Overall, a clear distinction was observed between the lac dyeAl3+
complexes analyzed after HF treatment and the free lac reconstructions examined without hydrolysis. The latter displayed intense
signals at circa 1622, 1568, 1534, 1460, 1346, 1192, 1100, 1058 and
1011 cm 1, as shown in Fig. 3. As expected from the literature,[24]
3+
1176
Figure 3. On the left, SERS spectra of Al lac complex reconstructions, at exc = 488 nm: (a) Mappae Clavicula; (b) Strasbourg; (c) Bolognese 130; (d) Bolognese
129. On the right, SERS spectra of non-complexed lac reconstructions, at exc = 488 nm: (e) Ibn Badis 2; (f) Libro de komo se fazen as kores; (g) lac dye 3
reconstruction (pH 3); (h) sticklac (raw material).
wileyonlinelibrary.com/journal/jrs
3+
Figure 5. Excitation and emission spectra from (a) the Al lac complex
.
5
() on filter paper; and its absorption spectra ( ) in solution (10 mol/l);
(b) Lorvo 5, f6 () with Mappae Clavicula recipe (---); (c) Alcobaa 412,
f.10v () with Ibn Badis recipe (---), respectively.
wileyonlinelibrary.com/journal/jrs
1177
R. Castro et al.
Table 1. Fluorescence emission and excitation maxima for the main paint reconstructions
Ibn Badis
Mappae Clavicula
LKFK
Bolognese
Strasbourg
589
472
590
520
No signal
No signal
589
526, 562
587
526, 562
em/nm
exc/nm
em/nm
exc/nm
Lv 5
ALC 238
ALC 249
ALC 412
ALC 419
ALC 421
ALC 446
SC 20
SC 21
58789
52326
563
522
589
554
587
474
587
556
593
470
593
476
58789
553
58991
520
excitation spectra, the bands are not resolved and the relative
intensities inverted, being the band around 520525 nm the most
intense. The shoulder at circa 470 nm may include the contribution
of erythrolaccin and can also be because of a lower ratio of alum
lac complex. For the Ibn Badis reproduction, the main band in the
excitation spectrum is shifted toward lower wavelengths and is
located at 472 nm. This may be ascribed to the presence of the
erythrolaccin as well as to the complete absence of alum. Given that
in the literature, the detection of erythrolaccin is usually associated
with the presence of shellac resin, we may conclude that
microspectrofluorimetry is indirectly sensing the presence of the
resin and indicating that the latter was incorporated in the final
pigment.[4] To confirm this hypothesis, HPLC-DAD was performed
on the lac reproductions. While the chromatogram of Bolognese
130 reconstruction showed only the main complexed laccaic acid
chromophores at 490 nm, the Ibn Badis was found to contain resin
as well, Figs S8 and S9 (Supporting Information), in accordance with
the microspectrofluorimetric data.
The emission and excitation maxima for the original medieval
samples are displayed in Table 2. The emission spectra show a band
around 589 nm, which is detected consistently for most of the lac
reproductions. As observed in the reconstructions, more pronounced shifts are observed for the excitation spectra. In addition,
some of the original samples display a band at 472 nm, particularly
prominent in the Alcobaa samples, which, as discussed in the
preceding texts, is likely because of the presence of higher amounts
of resin. In terms of signal intensity, more pronounced emissions
are observed in samples from Lorvo (circa 2), whereas spectra of
Alcobaa samples are characterized by lower intensities. When
comparing these with laccaic acids, lacAl3+, sticklac and shellac
spectra, it is possible to conclude that the original samples are
composed of laccaic acids and shellac, Figs 5 and S6 (Supporting
Information). Compared with the reconstructions, the most significant similarities were detected between the historical samples and
Mappae Claviculaalso in accordance with the SERS resultsand
Ibn Badis mock-ups (Fig. 5(b) and 5(c)). The color of these two reconstructions is also very similar to that observed in the manuscripts.
Conclusions
1178
In this work, for the first time, lac dye was unequivocally identified
in medieval illuminations. SERS proved to be an advantageous
technique for the identification of dyes in illuminated manuscripts,
because of the minute sample size required for analysis (1020 m).
By using the unambiguous molecular fingerprint of the dye
wileyonlinelibrary.com/journal/jrs
provided by SERS, we were able to assess the microspectrofluorimetry results and validate this technique as a robust analytical
tool for the detection of lac dye in situ. Work is in progress to further
explore the applications of microspectrofluorimetry.
The availability of historically accurate reconstructions was essential for this study, as it enabled to assemble a consistent spectral
database of mock-ups prepared according to historical recipes, to
be used for comparison in the analysis of samples from actual medieval illuminations. Additionally, comparison with paint reconstructions prepared in the lab shed new light on the fact that the
color shades observed in the medieval illuminations (pink, red
and violet, to brownish hues) may be the result of the processing
of the colorant as opposed to being as a result of degradation.
While SERS provides unequivocal identification of laccaic acid A
(the main component of lac dye), microspectrofluorimetry describes the chromophore in its environment. In this sense, these
two techniques, combined in this study for the first time, can be
seen as complementary. The microspectrofluorimetric data obtained so far suggest that it is also possible to detect the resin in
a lac sample indirectly by using erythrolaccin as a marker; this
was further confirmed by HPLC-DAD.
As for future research, we intend to explore the use of mixtures in
lac paints, such as fillers (gypsum, chalk), vermilion, lead white and
several dyes, in order to establish correlations with the historical samples. To elaborate the spectral information obtained by SERS and
microfluorimetry and ascertain whether it may carry details on the
process used to prepare the pigment or other hidden patterns and
features, a multivariate analysis approach will be followed.
In conclusion, our research sheds new light on the use of this
historical dye. Its systematic application in Portuguese medieval
manuscripts to obtain dark red colors shows the importance of
lac dye throughout the 12th and 13th centuries (much earlier than
the arrival to India of the Portuguese explorer Vasco da Gama, in
1498). There is no evidence, to date, that lac dye was being used
by medieval monasteries other than those cited in this work. Thus
far, it has only been identified by SERS in a French polychrome
wood sculpture[18] and in a Spanish crucifix,[22] both dating to
11501200. These three occurrences, in the Iberian Peninsula and
Provence, confirm the existence of an Arabic and Jewish trade
network in that region during the Romanesque period.
Acknowledgements
This work has been financially supported by Portuguese funds
through FCTFundao para a Cincia e a Tecnologia under the
projects PTDC/QUI-QUI/099388/2008 and PTDC/EAT-EAT/104930/
References
[1] M. J. Melo, in Handbook of Natural Colorants (Eds: T. Buchtold, R.
Mussak), John Wiley & Sons, Chichester, 2009, pp. 320.
[2] D. Cardon, Natural dyes. Sources, Tradition, Technology and Science,
Archetype Publications, London, 2007.
[3] P. K. Bose, Y. Sankaranaraynan, S. C. Sen Gupta, Chemistry of lac, Indian
Lac Research Institute, India, 1963.
[4] J. Kirby, in Dyes in History and Archaeology 21 (Ed. J. Kirby), Archetype
Publications, London, 2008, pp. 89108.
[5] M. P. Colombini, I. Bonaduce, G. Gautier, Chromatographia 2003, 58, 357.
[6] L. Wang, Y. Ishida, H. Ohtani, S. Tsuge, Anal. Chem. 1999, 71, 1316.
[7] M. Chairat, V. Rattanaphani, J. B. Bremner, S. Rattanaphani, D. F. Perkins,
Dyes Pigm. 2004, 63, 141.
[8] D. Cardon, Le Monde des Teintures Naturelles, Belin Sciences, Paris, 2003.
[9] R. A. Donkin, Anthropos 1977, 72, 864.
[10] M. Lombard, Les Textiles dans le monde musulman: du VIIe au XIIe sicle,
Mouton Editeur, Paris, 1978.
[11] (a)G. da Orta, Colquios dos Simples e Drogas da ndia, vol. 2, Imprensa
Nacional Casa da Moeda, Lisboa, 2011; (b)G. da Orta, Colloques des
Simples et des Drogues de LInde, Thesaurus, Actes Sud, 2004.
[12] J. H. van Linschoten, Itinerario: Voyage ofte schipvaert van Jan Huyghen
van Linschoten naer Oost ofte Portugaels Indien, 1579-1592, M. Nijhoff,
The Hague, 19101939.
[13] J. Wouters, A. Verhecken, Stud. Conserv. 1989, 34, 189.
[14] J. Kirby, R. White, National Gallery Tech. Bull. 1996, 17, 56.
[15] J. Kirby, M. Spring, C. Higgitt, National Gallery Tech. Bull. 2005, 26, 71.
[16] M. J. Melo, A. Claro, Acc. Chem. Res. 2010, 43, 857.
[17] A. Claro, M. J. Melo, S. Shefer, S. Seixas de Melo, F. Pina,
K. Jan van den Berg, A. Burnstock, Talanta 2008, 74, 922.
[18] M. Leona, Proc. Natl. Acad. Sci. 2009, 106, 14757.
[19] M. Leona, P. Decuzzi, T. A. Kubic, G. Gates, J. R. Lombardi, Anal. Chem.
2011, 83, 3990.
[20] F. Casadio, M. Leona, J. R. Lombardi, R. Van Duyne, Acc. Chem. Res. 2010,
43, 782.
[21] F. Pozzi, J. R. Lombardi, S. Bruni, M. Leona, Anal. Chem. 2012, 84, 3751.
[22] F. Pozzi, Development of Innovative Analytical Procedures for the
Identification of Organic Colorants of Interest in Art and Archaeology,
PhD thesis, Universit degli Studi di Milano, 2012.
[23] X. Zhang, R. A. Laursen, Anal. Chem. 2006, 77, 2022.
[24] M. V. Caamares, M. Leona, J. Raman Spectrosc. 2007, 38, 1259.
[25] M. J. Melo, A. Miranda, C. Miguel, R. Castro, A. Lemos, S. F. Muralha,
J. A. Lopes, A. P. Gonalves, Rev. Histria Arte 2011, 1(srie W), 152.
[26] C. Miguel, A. Miranda, J. A. Lopes, M. J. Melo, M. Clarke, in Preprints from
the 16th ICOM-CC Triennial Conference (Eds: J. Bridgland et al.),
International Council of Museums (ICOM), Lisbon, 2011, Paper 0107, 18.
[27] C. Miguel, A. Miranda, M. Oliveira, M. J. Melo, M. Clarke, in The artists
process. Technology and interpretation (Eds: S. Eyb-Green, J. H.
Townsend, M. Clarke, J. Nadolny, S. Kroustallis), Archetype
Publications, London, 2012, pp. 6065.
[28] M. Levey, Trans. Am. Philos. Soc. 1962, 52, 5.
[29] C. S. Smith, J. G. Hawthorne, Trans. Am. Philos. Soc. 1974, 64, 1.
[30] D. L. Strolovitch, Old Portuguese in Hebrew Script: convention, contact,
and convivncia, PhD Dissertation, Cornell University, 2005.
[31] M. P. Merrifield, Medieval and Renaissance Treatises on the Arts of
Painting, Dover Publications Inc., New York, 1967.
[32] V. Borradaile, R. Borradaile, The Strasbourg Manuscript: a Medieval
Painters Handbook, Transatlantic Arts, New York, 1966.
[33] MOLAB report from 2009. http://www.eu-artech.org/files/MEDMANUserReport.pdf. Retrieved October 2014.
[34] V. S. F. Muralha, C. Miguel, M. J. Melo, J. Raman Spectrosc. 2012, 43, 1737.
[35] C. Miguel, J. A. Lopes, M. Clarke, M. J. Melo, Chemom. Intell. Lab. Syst.
2012, 119, 32.
[36] A. Nevin, I. Osticiolli, D. Anglos, A. Burnstock, S. Cather, E. Castelluci,
J. Raman Spectrosc. 2008, 39, 993.
[37] C. Miliani, F. Rosi, I. Borgia, P. Benedetti, B. G. Brunetti, A. Sgamelotti,
Appl. Spectrosc. 2007, 61, 293.
[38] D. V. Thomson, Tech. Stud. 1932, 1, 919.
[39] V. Risdonne, Manoscritti Illuminati Portoghesi: Tecniche Pittoriche e
Materiali, Tesi di Laurea, Universit degli Studi di Perugia, Perugia, 2010.
[40] K. V. Casteele, H. Geiger, R. De Loose, C. F. Van Sumere, J. Chromatogr.
1983, 259, 291.
Supporting information
Additional supporting information may be found in the online
version of this article at the publishers web site.
1179
wileyonlinelibrary.com/journal/jrs