Anal. Chem. 2007, 79, 2812-2821
Chemometric Study of Maya Blue from the Voltammetry of Microparticles Approach Antonio Dome´nech,*,† Marı´a Teresa Dome´nech-Carbo´,‡ and Marı´a Luisa Va´zquez de Agredos Pascual‡,§
Departament de Quı´mica Analı´tica, Universitat de Vale` ncia, Dr. Moliner, 50, 46100 Burjassot, Vale` ncia, Spain, Institut de Restauracio´ del Patrimoni/Departament de Conservacio´ i Restauracio´ de Bens Culturals, Universitat Polite` cnica de Vale` ncia, Camı´ de Vera 14, 46022, Vale` ncia, Spain, and Departament de Histo` ria de l’Art, Universitat de Vale` ncia, Passeig al Mar, Vale` ncia, Spain
The use of the voltammetry of microparticles at paraffinimpregnated graphite electrodes allows for the characterization of different types of Maya Blue (MB) used in wall paintings from different archaeological sites of Campeche and Yucata´ n (Mexico). Using voltammetric signals for electron-transfer processes involving palygorskite-associated indigo and quinone functionalities generated by scratching the graphite surface, voltammograms provide information on the composition and texture of MB samples. Application of hierarchical cluster analysis and other chemometric methods allows us to characterize samples from different archaeological sites and to distinguish between samples proceeding from different chronological periods. Comparison between microscopic, spectroscopic, and electrochemical examination of genuine MB samples and synthetic specimens indicated that the preparation procedure of the pigment evolved in time via successive steps anticipating modern synthetic procedures, namely, hybrid organic-inorganic synthesis, temperature control of chemical reactivity, and template-like synthesis. Maya Blue (MB), a pigment produced by the ancient Mayas, has claimed considerable attention because of its enormous stability and peculiar hue, ranging from a bright turquoise to a dark greenish blue. The use of MB in Central America, mostly in Mexico, is well documented in archaeological sites between the 8th and the 16th centuries,1 but its use is prolonged even to recent times.2 MB can be considered as a hybrid organic-inorganic material3 resulting from the attachment of indigo, a blue dye extracted from Indigofera suffruticosa and other species, and a local clay, palygorskite.4,5 A Yucata´n site, Sak lu’um, served as the main and almost unique source of palygorskite in Mesoamerica. Natural indigo is * Corresponding author. E-mail:
[email protected]. † Departament de Quı´mica Analı´tica, Universitat de Vale`ncia. ‡ Universitat Polite`cnica de Vale`ncia. § Departament de Histo`ria de l’Art, Universitat de Vale`ncia. (1) Magaloni Kerpel, D. Materiales y Te´cnicas de la Pintura Maya. Facultad de Filosofı´a y Letras, Universidad Nacional Auto´noma de Me´xico, 1996. (2) Tagle, A.; Paschinger, H.; Richard, H.; Infante, G. Stud. Conserv. 1990, 35, 156-159. (3) Romero, P.; Sa´nchez, C. New J. Chem. 2005, 29, 57-58. (4) Torres, L. M. Mater. Res. Soc. Symp. Proc. 1988, 123.
2812 Analytical Chemistry, Vol. 79, No. 7, April 1, 2007
formed by indigotin (3H-indol-3-one, 2-(1,3-dihydro-3-oxo-2H-indol2-ylidene)-1,2-dihydro), a quasi-planar molecule of approximate dimensions 4.8 × 12 Å.6 The inorganic component of MB, palygorskite, is a fibrous phyllosilicate of ideal composition (Mg,Al)4Si8(O,OH,H2O)24‚nH2O. Palygorskite appears usually as a mixture of two different polytypes, one monoclinic and one orthorhombic, whose structure can be described in terms of continuous, two-dimensional tetrahedral and octahedral sheets.7,8 Both polytytpes show discontinuous 2:1 tetrahedral/octahedral layers with one 2:1 unit joined to the next by inversion of the SiO4 tetrahedra along Si-O-Si bonds. The tetrahedral and octahedral mesh gives rise to a series of rectangular tunnels of dimensions 6.4 × 3.7 Å. Such clays are, therefore, crossed by zeolite-like channels and permeated by weakly bound, nonstructural (zeolitic) water. Magnesium and aluminum cations complete their coordination with tightly bound water molecules (structural water). Although the procedure of extraction of indigo from plants has remained in use until current times, the procedure for preparing MB is unknown.1,4,5 It is believed that the Mayas prepared the pigment by crushing indigo and palygorskite with a moderate thermal treatment, but the details of the process and the way in which the ancient Mayas modulated the hue of the pigment have not been elucidated. The nature of the indigo-palygorskite association has claimed considerable attention in the last years.9-22 Thus, Jose´-Yacama´n et al.9,10 discovered that most palygorskite particles exhibit a superstructure along the a-axis about 14 Å, which roughly (5) Reyes-Valerio, C. De Bonampak al Templo Mayor: el azul Maya en Mesoame´rica; Siglo XXI, Madrid, 1993. (6) Gordon, P. F.; Gregory, P. Indigoid Dyes in Organic Chemistry in Colour; Springer-Verlag: Berlin 1983; pp. 208-211. (7) Chisholm, J. E. Can. Miner. 1990, 28, 329-339. (8) Chisholm, J. E. Can. Miner. 1992, 30, 61-73. (9) Jose´-Yacama´n, M.; Rendo´n, L.; Arenas, J.; Serra Puche, M. C. Science 1996, 273, 223-225. (10) Polette, L. A.; Meitzner, G.; Jose´-Yacama´n, M.; Chianelli, R. R. Microchem. J. 2002, 71, 167-174. (11) Hubbard, B.; Kuang, W.; Moser, A.; Facey, G. A.; Detellier, C. Clays Miner. 2003, 51, 318-326. (12) Chiari, G.; Giustetto, R.; Riccihiardi, G. Eur. J. Mineral. 2003, 15, 21-33. (13) Fois, E.; Gamba, A.; Tilocca, A. Microporous Mesoporous Mater. 2003, 57, 263-272. (14) Witke, K.; Brzezinka, K.-W.; Lamprecht, I. J. Mol. Struct. 2003, 661-662, 235-238. (15) Reinen, D.; Ko ¨hl, P.; Mu ¨ ller, C. Z. Anorg. Allg. Chem. 2004, 630, 97-103. 10.1021/ac0623686 CCC: $37.00
© 2007 American Chemical Society Published on Web 03/08/2007
corresponds to three times the original lattice constant and might well be explained by the presence of indigo in the channels. Such authors reported the presence of iron nanoparticles outside the lattice of the crystallites of palygorskite as well as inside the channels accompanying iron oxide and an amorphous phase of FeO(OH) in MB samples10 and suggested that light dispersion on metal nanoparticles contributes significantly to the optical properties of the pigment. Sa´nchez del Rı´o et al.,14,16,19 however, did not find neither iron in metallic form nor goethite in archaeological MB while Chiari et al.,12 Reinen et al.,15 and Giustetto et al.,17,18 suggested that the MB hue is mainly attributable to bathochromic shift of the indigo spectrum originated when single indigo molecules become attached to the palygorskite matrix. Molecular modeling and spectral data indicate that hydrogen bonding between indigo and the hydroxy groups in the palygorskite channels12,13,22 or edge silanol groups11 occurs. In previous reports, we have studied the solid-state electrochemistry of genuine MB samples in contact with aqueous23 and nonaqueous24 electrolytes using the voltammetry of microparticles approach. This methodology, developed by Scholz et al.,25,26 provides information on the composition of solid micro- and nanosamples of solid materials. Application of solid-state electrochemistry combined with electron microscopy, atomic force microscopy, FT-IR, and visible spectroscopies and nuclear magnetic resonance data allowed us to conclude that not only indigo but also its oxidized form, dehydroindigo, is attached to palygorskite in MB, contributing significantly to modulate the hue of the pigment.23 The voltammetry of microparticles allowed detection of the presence of MB in wall paintings of the substructure IIC in the archaeological site of Calakmul, dated in the Early Classical period (440-450 B.C.), thus providing evidence on the use of this pigment 700-750 years prior to the date currently accepted.23 In order to explore the capabilities of the voltammetry of microparticles for acquiring analytical information, a chemometric study was performed on a series of 32 MB samples from 12 archaeological sites in Yucata´n (Chacmultu´n, D’zula, Ek Balam, Acanceh, Kuluba´, and Mulchic, all of the Late Classical period, Chiche´n Itza´, corresponding to the Terminal Classical period, and Mayapa´n, dated at the Postclassical period) and 3 sites in Campeche (Dzibilnocac and El Tabasquen˜o, Late Classical period, (16) Sa´nchez del Rı´o, M. S.; Martinetto, P.; Somogyi, A.; Reyes-Valerio, C.; Dooryhe´e, E.; Peltier, N.; Alianelli, L.; Moignard, B.; Pichon, L.; Calligaro, T.; Dran, J.-C. Spectrochim. Acta, B 2004, 59, 1619-1625. (17) Vandenabeele, P.; Bode, S.; Alonso, A.; Moens, L. Spectrochim. Acta, A 2005, 61, 2349-2356. (18) Chianelli, R. R.; Perez de la Rosa, M.; Meitzner, G.; Siadati, M.; Berhault, G.; Mehta, A.; Pople, J.; Fuentes, S.; Alonzo-Nun ˜ez, G.; Polette, L. A. J. Synchrotron Radiat. 2005, 12, 129-134. (19) Sa´nchez del Rio, M. S.; Sodo, A.; Eeckhout, S. G.; Neisius, T.; Martinetto, P.; Dooryhee, E.; Reyes-Valerio, C. Nucl. Instrum., Methods Phys. Res., Sect. B 2005, 238, 50-54. (20) Giustetto, R.; Llabres i Xamena, F. X.; Ricchiardi, G.; Bordiga, S.; Damin, A.; Gobetto, R.; Chierotti, M. R. J. Phys. Chem. B 2005, 109, 19360-19368. (21) Sa´nchez del Rı´o, M.; Martinetto, P.; Reyes-Valerio, C.; Dooryhee, E.; Sua´rez, M. Archaeometry 2006, 48, 115-130. (22) Giustetto, R.; Levy, D.; Chiari, G. Eur. J. Mineral. 2006, 18, 629-640. (23) Dome´nech, A.; Dome´nech, M. T.; Va´zquez, M. L. J. Phys. Chem. B 2006, 110, 6027-6039. (24) Dome´nech, A.; Dome´nech, M. T.; Va´zquez, M. L. J. Phys. Chem. C, submitted. (25) Scholz, F.; Meyer, B. In Electroanalytical Chemistry, A Series of Advances; Bard, A. J., Rubinstein, I., Eds.; Marcel Dekker: New York, 1998; Vol. 20, pp 1-87. (26) Grygar, T.; Marken, F.; Schro ¨der, U.; Scholz, F. Collect. Czech. Chem. Commun. 2002, 67, 163-208.
and the substructure A-6 of the site of Calakmul, corresponding to the Early Classical period, and the substructure II-C of the same archaeological site, dated in the Late Preclassical period). The response in contact with aqueous acetate buffer of genuine MB samples was compared with that of synthetic specimens prepared by crushing palygorskite and indigo under different thermal treatments. Multivariate chemometric methods have been applied to voltammetric data for MB samples and synthetic specimens in order to identify possible regularities in the electrochemical response of such material and obtain analytical information concerning two unsolved questions: (i) the method of preparation of the pigment by the ancient Mayas and (ii) the existence of a possible geographical or chronological variation in the Maya Blue technology. Multivariate statistical techniques, including pattern recognition and hierarchical cluster analysis have claimed considerable attention for analytical purposes. Such methods have been recently applied to different electroanalytical problems.27-29 EXPERIMENTAL SECTION Microsamples (∼1 µg) and eventually nanosamples (∼20-50 ng) of MB were taken from blue or greenish-blue regions of wall paintings in the listed archaeological sites, with the help of a microscalpel, during their routine examination and restoration. Figure 1 shows a typical Maya wall painting. A light microscope Leica DMR (×25-×400) was used for selecting the samples to be analyzed and for morphological examination of them. Palygorskite was collected from the Sak lu’um classical deposits in Yucatan. Synthetic indigo (Fluka) was used as a reference material. To mimic possible preparation procedures of MB employed by the ancient Mayas, a series of synthetic specimens (S-1 to S-6) were prepared by crushing palygorskite with synthetic indigo (3% w/w) plus calcite (10% w/w) and heating at temperatures of 25, 70, 160, 260, 320. and 400 °C for periods of 12 h. With except sample S-6, which acquired a brown hue, all Maya Blue synthetic specimens became greenish-blue, the greenish hue increasing on increasing temperature along the S-1 to S-5 series. After thermal treatment, a portion of the specimens was modified by adding ochre from Sak lu’um (10% w/w), (series SC-1 to SC-6). A third series of synthetic specimens (ST-1 to ST-6) were heated at 200 °C during 1, 2, 4, 6, 12, and 24 h. A fourth series (SO-1 to SO-6) was prepared by adding ochre and then heating at the above listed temperatures. Pictorial samples were examined with a Jeol JSM 6300 scanning electron microscope operating with a Link-Oxford-Isis X-ray microanalysis system. The analytical conditions were as follows: accelerating voltage 20 kV, beam current 2 × 10-9 A, and working distance 15 mm. Samples were carbon coated to eliminate charging effects. Quantitative microanalysis was carried out using the ZAF method for correcting interelemental effects. The counting time was 100 s for major and minor elements. Voltammetry of microparticles experiments were performed using paraffin-impregnated graphite electrodes (PIGEs) immersed into acetic acid/sodium acetate buffer (total acetate concentration (27) Scampicchio, M.; Mannino, S.; Zima, J.; Wang, J. Electroanalysis 2005, 17, 1215-1221. (28) Gute´s, A.; Iba´n ˜ez, A. B.; Ce´spedes, F.; Alegret, S.; del Valle, M. Anal. Bioanal. Chem. 2005, 382, 471-476. (29) Richards, E.; Bessant, C.; Saini, S. Electroanalysis 2002, 14, 1533-1542.
Analytical Chemistry, Vol. 79, No. 7, April 1, 2007
2813
Figure 1. Image of Maya wall paintings in the substructure I of Calakmul (Campeche, Me´ xico), Early Classical Maya period.
0.50 M, pH 4.85). PIGEs consisted of cylindric rods of 5-mm diameter prepared as described by Scholz et al.25,26 Measurements were performed in a thermostated three-electrode cell under argon atmosphere using a AgCl (3 M NaCl)/Ag reference electrode and a platinum wire auxiliary electrode. Square wave voltammograms (SQWVs) were obtained with CH I420 equipment. Statistical analyses were performed with the Minitab 14 software package. For modified electrode preparation, the samples were accurately powdered in an agate mortar and pestle and extended, forming a spot of finely distributed material. The lower end of the graphite electrode was pressed over that spot of sample to obtain a sample-modified surface. Transmission electron microscopy (TEM) was carried out with a Philips CM10 with Keen view camera. Soft imaging system was used operating at 100 kV. MB samples were prepared by grinding a few micrograms of the samples in an agate mortar and then dispersing them by the help of an ultrasonic bath in CH2Cl2. A drop of the dispersion was poured on TEM grids pretreated with a polymer film layer with holes in order to improve the quality of the images. FT-IR-ATR spectra of samples and reference materials were obtained with a Vertex 70 Fourier transform infrared spectrometer with a fast recovery deuterated triglycine sulfate temperaturestabilized coated detector. Number of coadded scans, 32; resolution, 4 cm-1. Visible spectra were obtained with a Minolta CM-503i spectrophotometer using a Xe arc lamp and a Si photodiode detector. 2814
Analytical Chemistry, Vol. 79, No. 7, April 1, 2007
The instrument was calibrated with a standard white (chrom coordinates Y ) 95.8, x ) 0.3167, y ) 0.3344). RESULTS AND DISCUSSION Voltammetry of Maya Blue Samples. In Figure 2 is compared the SQWV response of PIGEs modified with (a, d) indigo microparticles and (b, c) Maya Blue sample from the substructure II-C of Calakmul. This last corresponds to the oldest Maya Blue sample, dated in the Late Preclassical period. In both cases, two indigo-localized electrochemical processes, characterized by sharp peaks at +475 (I) and -285 mV (II), occur. These processes correspond to the almost-reversible two-electron, two-proton oxidation of indigo to dehydroindigo and the two-electron, twoproton reduction of indigo to leucoindigo,30-32 depicted in Scheme 1. These processes can be represented as
{H2IN} f {IN} + 2H+ (aq) + 2e-
(1)
{H2IN} + 2H+ (aq) + 2e- f {H4In}
(2)
where { } represents palygorskite-anchored species. Notice that charge conservation is ensured in both cases by the expulsion/ insertion of two protons and two electrons from/to the solid. (30) Bond, A. M.; Marken, F.; Hill, E.; Compton, R. G.; Hu ¨ gel, H. J. Chem. Soc., Perkin Trans. 2 1997, 1735-1742. (31) Grygar, T.; Kuckova, S.; Hradil, D.; Hradilova, D. J. Solid State Electrochem. 2003, 7, 706-713.
Scheme 1. Electrochemistry of Indigo (H2IN) Oxidation to Dehydroindigo (IN) and Reduction to Leucoindigo (H4IH)
Figure 2. SQWVs of (a, c) indigo microparticles and (b, d) Maya Blue sample from Calakmul, substructure II-C, immersed into 0.50 M acetate buffer, pH 4.85. (a, b) Potential scan initiated at +0.85 V in the negative direction; (c, d) potential scan initiated at -750 mV in the positive direction. Potential step increment 4 mV; square wave amplitude 25 mV; frequency 5 Hz.
Current data confirmed previous results,23 showing that the relative intensity of peaks I and II differs substantially from that determined for indigo microparticles, a feature attributable to the presence of a significant amount of dehydroindigo (5-15%) accompanying indigo in MB. Coupled electrochemical/AFM experiments indicate that extensive electrooxidation of the playgorskite-indigo complex produces a significant alteration in the clay structure.24 (32) Dome´nech, A.; Dome´nech, M. T. J. Solid State Electrochem. 2006, 10, 459468.
Apart from indigo-centered peaks I and II, voltammograms of MB exhibit two peaks at +200 (III) and -15 mV (IV). These last two peaks, although with different peak profile, were also recorded in blank experiments performed with PIGEs modified with pristine palygorskite. As can be seen in Figure 3, the profile of voltammetric peaks III and IV varied significantly from one sample to another. Examination of SQWVs for the studied samples allowed us to distinguish four main morphological types, labeled a-d. Chemometric analysis (vide infra) allowed us to distinguish five subtypes (A1-A5) within the type A. Representative voltamograms are presented in Figure 2 for MB samples from (a) Chacmultu´n (type A3), (b) El Tabasquen ˜o (type B), (c) D’zula (type C), and (d) Mayapa´n (type D). Type A4, corresponding to the oldest MB sample, is shown in Figure 1, while Figure 4 shows SQWVs of samples from (a) Acanceh (type A1), (b) El Tabasquen˜o (type A2), and (c) Dzibilnocac (type A5). These voltammograms were taken as “holotypes” of the groups of MB samples. Diagnostic criteria for grouping voltammetric profiles are provided as Supporting Information. In agreement with literature data,33-37 peaks III and IV can be attributed to quinone functionalities produced in the graphite surface. Such processes are promoted by scratching of the graphite surface, thus yielding relatively intense voltammetric peaks. The observed differences in the shape of peaks III and IV is attributable to the differences in the mineralogy and textural properties of the sample, so that the variability in their profile should be representative of differences, not only in mineralogy, but also in the mechanical properties of the sample. Since these differences may result from differences in the preparation procedure, the analysis of voltammetric profiles can provide information on the procedure of preparation of MB. Chemometrics. Chemometric analysis was performed by defining eight quantitative parameters representative of the shape of the voltamograms. On first examination, peak potentials, peak currents, and peak areas were used. The variables concerning (33) Evans, J. F.; Kuwana, T. Anal. Chem. 1977, 49, 1632-1635. (34) Gunasingham, H.; Fleet, B. Analyst 1982, 107, 896-902. (35) Engstrom, R. C.; Strasser, V. A. Anal. Chem. 1984, 56, 136-141. (36) Barisci, J. N.; Wallace, G. G.; Baughman, R. H. Electrochim. Acta 2000, 46, 509-517. (37) Barisci, J. N.; Wallace, G. G.; Baughman, R. H. J. Electroanal. Chem. 2000, 488, 92-98.
Analytical Chemistry, Vol. 79, No. 7, April 1, 2007
2815
Figure 4. SQWVs of PIGEs modified with Maya Blue samples from (a) Acanceh (type A1), (b) El Tabasquen˜o (type A2), and (c) Dzibilnocac (type A5), immersed into 0.50 M acetate buffer, pH 4.85. Potential scan initiated at -850 mV in the positive direction. Potential step increment 4 mV; square wave amplitude 25 mV; frequency 5 Hz.
Figure 3. SQWVs of PIGEs modified with Maya Blue samples from (a) Chacmultu´ n (type A3), (b) El Tabasquen˜o (type B), (c) D’zula (type C), and (d) Mayapa´ n (type D), immersed into 0.50 M acetate buffer, pH 4.85. Potential scan initiated at -750 mV in the positive direction. Potential step increment 4 mV; square wave amplitude 25 mV; frequency 5 Hz.
peak current and peak area were found to be strongly coupled, a common situation in chromatography, voltammetry, and electrophoresis,27 so that in subsequent data handling only peak currents were used. Since absolute peak currents varied slightly from one sample to another, peak current ratios were used. Pertinent data for the holotypes are summarized in Table 1. Different two-variable diagrams were constructed for several pairs of electrochemical parameters. This can be seen in Figure 5, in which the peak III/peak current IV ratio, I(III)/I(IV), is plotted as a function of the separation between the peak potential of peaks III and IV, Ep(III)-Ep(IV). As can be seen in this figure, 2816 Analytical Chemistry, Vol. 79, No. 7, April 1, 2007
data points for the different main morphological types (A-D) cover different regions of the diagram but subtypes A1-A5 become overlapped. The principal component method provided similar results, thus reinforcing the idea that the different electrochemical types correspond to differently prepared MB samples. Distribution of samples into the different electrochemical types suggests that there is a chronological variation in the properties of MB, as can be seen in Table 2 (see also Supporting Information). In a second step, hierarchical cluster analysis was carried out to reveal hidden relationships between voltammograms and the 32 MB samples. The used parameters were as follows: (i-iv) peak potentials of peaks I-IV, (v) peak III/peak current IV ratio (I(III)/I(IV), (vi) the quotient between the current at the shoulder between peaks III and IV (located ∼ +100 mV) and the peak current of peak IV (I(100)/I(IV)), (vii) the peak current I/peak current II ratio (I(I)/I(II)), and (viii) the quotient between the sum of peak currents for peaks I and II and the sum of the peak
Table 1. Voltametric Data for MB Holotypes Defined in This Studya sample
Ep(I)b
Ep(III)b
Ep(IV)b
Ep(II)b
I(III)/I(IV)
I(I+II)/I(III+IV)
I(I)/I(II)
I(100)/I(IV)
∆Ep2(IV)b
type
Acanceh-2 El Tabasquen ˜o-2A Kuluba´-9 Calkmul IIC Dzibilnocac-9A Mayapa´n 2A D’zula-2 Dzibilnocac-9
+460 +455 +485 +515 +475 +470 +450 +450
+210 +190 +195 +180 +175 +180 +185 +190
0 0 -30 -45 0 +40/0 +10 +40
-285 -265 -265 -250 -265 -285 -285 -275
0.938 0.320 0.536 0.320 0.132 0.172 1.400 1.539
3.250 0.485 0.556 0.242 0.225 0.034 0.354 0.188
1.524 1.909 0.326 0.389 3.667 1.143 2.200 1.308
0.438 0.330 0.446 0.329 0.150 0.146 0.900 1.129
130 130 163 167 113 135 216 288
A1 A2 A3 A4 A5 B C D
a From SQWVs recorded at sample-modified PIGEs immersed into 0.50 M acetate buffer, pH 4.85. Potential step increment 4 mV; square wave amplitude 25 mV; frequency 5 Hz. bPotentials in mV.
Table 2. Chronological Distribution of Samples into the Different Archaeological Sitesa site postclass. Mayapa´n terminal class. Chiche´n Itza´ D’zula late Ek Balam Dzibilnocac classical Kuluba´ Chacmultu´n Mulchic El Tabasquen ˜o Acanceh early class. Calakmul A6 late preclass. Calakmul IIC a
Figure 5. Two-dimensional I(III)/I(IV) vs Ep(III)-Ep(IV) (in mV) diagram for MB samples grouped into electrochemical types A (solid rhombs), B (rhombs), C (squares), and D (triangles). Froms SQWVs of sample-modified electrodes; conditions such as described in precedent figures.
currents for III and IV (I(I + II)/I(III + IV)). This last parameter can be taken as a measure of the “indigo/clay” ratio, i.e., of the concentration of indigo in the palygorskite. Confirming prior results,23 the values of the I(I)/I(II) ratio in MB samples, representative of the existing relation dehydroindigo/indigo, were clearly larger than that determined for indigo microparticles under the same conditions. A matrix was constructed with rows representing 32 MB samples and columns corresponding to the above parameters. Pertinent data are included as Supporting Information. Dendrogram building using this analysis is shown in Figure 6a. First, two branches, corresponding to D and A + B + C types are separated. In this second branch, C first separates from the A + B group, further divided into A and B branches. The former subsequently displays a separation in A1-A5 types. This separation scheme is reproduced by building the dendrogram from the eight electrochemical holotypes alone, as shown in Figure 6b. In Table 2, the sample distribution along chronologically ordered archaeological sites is depicted. Here, electrochemical types are distributed following the similarity sequence derived from cluster analysis in Figure 6. MB sample grouping in Table 2 suggests that there is a progressive replacement of some electrochemical types by other ones along time following a ramified scheme.
A2
A1
A4
A3
A5
B
C
•
•• •• •
•• ••
D •
• •
•
••• •
•
• • • ••
• •• •
••
• •
Each point corresponds to one sample.
Electrochemical Types and MB Preparation. The existence of net linkages between the different “electrochemical” types acts in support of the idea that voltammetric profiles are representative of mineralogical and textural differences and, ultimately, of the existence of different preparation procedures of MB samples. In order to verify this idea, electrochemical data were crossed by microscopic and spectroscopic data on genuine MB samples. First, MB contains calcite grains, as confirmed by ATR/FT-IR data.23,24 Calcite appears as a result of the preparation of indigo from macerated leaves of I, suffruticosa in quick lime. As shown in Figure 7a, spectra of MB samples show a prominent band a 1408 cm-1, attributable to the stretching vibration of the carbonate group. This peak is entirely absent in the spectra of pristine palygorskite and MB specimens prepared by crushing the clay with synthetic indigo. MB overlapped bands in the 1600-1750cm-1 region correspond to the carbonyl frequency of indigo (theoretical 1629 cm-1),38 dehydroindigo (theoretical 1736 cm-1),38 and zeolitic water.20,22 Such bands can be clearly seen of removing calcite by treating the sample with diluted HCl, as can be seen in Figure 7b. Second, MB often appears in paint layers accompanied by ochre grains, as indicated by SEM/EDX data and microscopical examination of cross-section stratigraphies (see Supporting Information). This ochre can, in principle, be attributed to a deliberated addition, “in situ” of ochre for modulating the hue of the pictorial layer. The presence of ochre in these samples is clearly reflected in their visible spectrum, as can be seen in Figure (38) Klessinger, M.; Lu ¨ ttke, W. Tetrahedron Lett. 1963, 19 (Suppl. 2), 315335.
Analytical Chemistry, Vol. 79, No. 7, April 1, 2007
2817
Figure 7. ATR/FT-IR spectra of MB sample from Mulchic (a) before and (b) after treatment with diluted HCl for removing calcium carbonate.
Figure 6. Hierarchical cluster analysis with euclidean distance of the autoscaled variables applied to voltammetric parameters recorded for MB samples. Data matrix using the electrochemical parameters described in the text for (a) 32 MB samples and (b) the holotypes listed in Table 1.
8. Diffuse reflectance spectra of many MB samples (curve A in Figure 8) exhibit typically two bands at ∼420 and ∼540 nm, close to those described for indigo and dehydroindigo in solution phase.38 In contrast, samples of paint layers containing additions of ochre (curve B in Figure 8) exhibit a diffuse reflectance spectrum close to that of goethite,39 shown in curve C of Figure 8. It should be noted that palygorskite contains iron ions. Although such ions do not influence the hue of the pristine clay (white in all samples from Sak lu’um), their participation in chemical processes (acting, for instance, as catalytic redox sites or modulating water/indigo coordination) involved in the preparation of MB cannot be entirely discarded. Remarkably, TEM examination of MB samples revealed significant differences between pristine palygorskite crystals from the classical site of Sak lu’um and palygorskite in MB samples, as well as between the different types of MB. As can be seen in Figure 9A, pristine palygorskite consists of aggregates of elongated crystals 0.5-1-µm sized having fine fiber structures with thicknesses from 300 to 600 Å, as shown in Figure 9A. In contrast, palygorskite from MB samples showed narrow fibbers with a corrugated structure. Remarkably, a clear correlation between “electrochemical” types and TEM images was obtained. Thus, A-type MB samples presented irregular-shaped palygorskite crystals, whereas B-type samples exhibited numerous 10-20-nmsized pores irregularly distributed on the surface of the crystals (39) Elias, M.; Chartier, C.; Prevot, G.; Garay, H.; Vignaud, C. Mater. Sci. Eng. B 2006, 127, 70-80.
2818 Analytical Chemistry, Vol. 79, No. 7, April 1, 2007
Figure 8. Diffuse reflectance spectra of (A) MB sample from El Tabasquen˜o, (B) MB sample from Kuluba´ , and (C) goethite (from ref 40).
(see Figure 9B). In contrast, MB samples of the C and D types showed a dense array of fine (2-5-nm-sized) pores covering the entire surface of palygorskite crystals (see Figure 9C). See details in Supporting Information. Such textural differences can be attributed to differences in the preparation of MB via crushing indigo and palygorskite with application of a given thermal treatment and, eventually, with addition of other components. Similarly to other minerals,40 pores observed in palygorskite crystals can be attributed to segregation and evacuation of loosely bound (physisorbed) water at and,
Figure 10. Variation of I(100)/I(IV) ()W) with I(III)/I(IV) for MB samples (solid figures, type D, triangles; type C, squares; types A3, A4, and B, rhombs; A1 and A2, circles) and synthetic specimens prepared from synthetic indigo and palygorskite submitted at different thermal treatments (squares, temperatures in the figure) during 24 h.
Figure 9. TEM images of (A) pristine palygorskite crystals from Sak lu’um, (B) MB sample from Chiche´ n Itza´ (electrochemical type B), and (C) MB sample from Mayapa´ n (electrochemical type D).
mainly, zeolitic or structural water. Thermal data11,17,18 indicate that such processes occur at 110-120 and 220-230 °C, respectively, temperatures able to be handled by the ancient Mayas. The upper limit of usable temperatures must be situated at 360 °C, where thermal decomposition of indigo occurs.11 Release of the gross structural water at 460-480 °C and dehydroxylation and phase transformation to clinoestatite at 700 °C complete the (40) Ruan, H. D.; Frost, R. L.; Kloprogge, J. T. Spectrochim. Acta, A 2001, 57, 2575-2586.
thermal evolution of palygorskite.17,18 Our data on MB samples and synthetic specimens (vide infra) suggest that some types of MB were prepared with no or smooth thermal treatment, in agreement with Hubbard et al.11 This implies that indigo adsorption in the palygorskite system does not necessarily follow the release of zeolitic water. In order to study the influence of the thermal treatment on the preparation of MB, different synthetic specimens were prepared from synthetic indigo combining different thermal treatments with the addition of calcite and, in some cases, ochre. Specimens were prepared by crushing palygorskite from the classical site of Sak lu’um, synthetic indigo (3% w/w) plus calcite (10% w/w), and, eventually, ochre (10% w/w) and heating at temperatures between room temperature (25 °C) and 400 °C during times between 6 and 48 h. Ochre was selected as a possible additive in view of its use as a pigmenting agent by the ancient Mayas and its availability as a common mineral in the region. In a second series of samples, ochre was added after the thermal treatment acting merely as a pigmenting agent. Experimental data indicated that the response of the specimens varied mainly with the duration of the crushing process and the temperature of the thermal treatment, while the duration of the thermal treatment produced minor variations in the voltammetric response. Increasing the duration of the crushing process produced a decrease in the indigo-localized peaks I and II with respect to peaks III and IV. This suggests that palygorskite channels are partially empty in MB specimens prepared with no or minimal Analytical Chemistry, Vol. 79, No. 7, April 1, 2007
2819
Scheme 2. Possible Chronological Scheme for the Evolution of MB in the Late Preclassical-Postclassical Maya Periodsa
a Relationships between electrochemical types are crossed with the accepted chronology of archaeological sites. Three possible technological revolutions are localized.
thermal treatment and are progressively filled on increasing temperature, so that indigo molecules become increasingly attached to the inorganic host. Additionally, the I(I)/I(II) ratio increases on increasing temperature, denoting the concomitant increase in the relative amount of dehydroindigo, a feature consistent with prior thermochemical data.23,24 Using the above preparation procedures, voltammetric profiles of types A3 and A4 were reproduced satisfactorily by specimens S1-S3, prepared by crushing indigo and palygorskite with no or smooth thermal treatment. A5-type voltammograms were obtained for heated synthetic specimens S4 and S5. Types A1 and A2, however, were approached by nonheated or moderately heated synthetic specimens after adding ochre (SO1-SO3). The electrochemical type C was approached by specimens SC3-SC5, which were heated after adding ochre. However, the electrochemical types B and D were not completely attained. Comparing data for selected pairs of electrochemical parameters for genuine MB samples and synthetic specimens indicated that the temperature of the thermal treatment probably increased along the sequence A-B-C-D. This can be seen in Figure 10, where the variation of I(100)/I(IV) with I(III)/I(IV) is given for genuine MB samples (solid figures) and a series of synthetic specimens prepared without addition of ochre at different temperatures. In this figure, the different electrochemical types are sequentially grouped along an almost linear representation. Data points for nonheated synthetic specimens fall within the region of A3-A4 types and, remarkably, on increasing the temperature of the thermal treatment, synthetic specimens approach, successively, the regions of A5, B and C-D genuine MB samples. Interestingly, synthetic specimens prepared with ochre without thermal treatment approached the region of A1-A2 types. Implications. Crossing microscopical, spectral, and voltammetric data for MB samples and synthetic specimens suggested the following: (i) electrochemical types A3 and A4 correspond 2820
Analytical Chemistry, Vol. 79, No. 7, April 1, 2007
to MB prepared by crushing indigo and palygorskite with no (or smooth) thermal treatment and no significant addition of ochre; (ii) types A1 and A2 were prepared similarly but incorporating a smooth thermal treatment and, frequently, a postheating addition of ochre-type minerals; (iii) preparation of type A5 MB required the use of a thermal treatment between 160 and 200 °C; (iv) preparation of types B-D required the use of more strong thermal treatments, probably with temperatures above 200 °C and additives, ochre being one of the possibly used materials. Crossing the linkages derived from the chemometric analysis of voltammetric data with chronological data for the studied archaeological sites allows presentation of a scheme in which the preparation of MB evolved in time. Accordingly, a seminal preparation procedure, dated, at least, from the Late Preclassical period, consisted of crushing palygorskite with indigo (and, eventually, adding ochre in situ for modulating the hue of the paint layer). From the actual perspective, this step can be viewed as a first “technological revolution”, where the modern synthesis of hybrid organic-inorganic materials was precluded. This preparation procedure, corresponding to A3 and A4 electrochemical types, was maintained until the Postclassical Maya period with local variations (A1, A2 types) attributable to the use of additives such as ochre. During the Late Classical period, a second technological revolution consisting of the performance of vigorous thermal treatments was introduced, yielding MB samples of types A5, B, C, and D. Differences between such types are attributable to the use of different temperatures and the introduction of ochre or other unknown additives prior to the thermal treatment for types B-D. Iron ions and other species existing in the parent palygorskite can eventually influence MB preparation. This scheme should explain the conflict between the observations concerning the absence9,11,14 and the presence4,5 of iron oxide materials in some MB samples. The use of additives during the crushing/
thermal treatment of indigo and palygorskite should indicate that the ancient Mayas anticipated to any extent the philosophy of contemporary template synthesis, thus yielding a third technological revolution also during the Late Classical period. This is reflected in Scheme 2, where a tentative fine chronology of the Late Classsical period is offered. FINAL CONSIDERATIONS In contact with aqueous electrolytes, genuine Maya Blue microsamples attached to graphite electrodes provide well-defined electrochemical responses, which can be described in terms of the oxidation and reduction of palygorskite-associated indigo to palygorskite-associated dehydroindigo and palygorskite-associated leucoindigo, respectively. Such indigo-centered processes are accompanied by electron-transfer steps involving quinone functionalities generated in the graphite surface. The corresponding voltammetric signals provided information on the composition and texture of the palygorskite/indigo complex. Application of chemometric methods allows for characterizing different electrochemical types of MB samples that can be correlated with different preparation procedures. Comparison between microscopic, spectroscopic, and electrochemical examination of genuine MB samples and synthetic specimens indicated that the preparation procedure of MB evolved in time via successive steps anticipating modern synthetic procedures, namely, hybrid organic-inorganic synthesis, temperature control of chemical reactivity, and template-like synthesis. These results provide a new picture of Maya culture, where the “static” view of a tradition-dominated ancient civilization needs
to be replaced by a more “dynamic” one. In this view, ancient Mayas were concerned with innovation and technological development, local workshops probably competing to obtain more stable MB pigments. These results illustrate the inherent capabilities of intersecting voltammetry of microparticles with chemometric methods and the possibility of acquiring analytical information from solid-state electrochemical techniques. ACKNOWLEDGMENT Financial support is gratefully acknowledged from the Generalitat Valenciana GVAE06/131 Project and the MEC Projects CTQ2005-09339-C03-01 and 02, which are also supported with FEDER funds. The authors thank Prof. Ramo´n Carrasco Vargas, Director of the Calakmul Archaeological Project, Dr. Jose´ Luis Moya Lo´pez, and Mr. Manuel Planes Insausti (Microscopy Service of the Polytechnical University of Valencia) for their technical support. SUPPORTING INFORMATION AVAILABLE Additional information as noted in text. This material is available free of charge via the Internet at http://pubs.acs.org.
Received for review December 14, 2006. Accepted February 5, 2007. AC0623686
Analytical Chemistry, Vol. 79, No. 7, April 1, 2007
2821
