Organic/inorganic complex pigments: Ancient colors Maya Blue

Available online at www.sciencedirect.com JOURNAL OF

Inorganic Biochemistry Journal of Inorganic Biochemistry 101 (2007) 1958–1973 www.elsevier.com/locate/jinorgbio

Organic/inorganic complex pigments: Ancient colors Maya Blue Lori Ann Polette-Niewold 1, Felicia S. Manciu, Brenda Torres, Manuel Alvarado Jr., Russell R. Chianelli * Materials Research and Technology Institute, The University of Texas at El Paso, El Paso, TX 79912, United States Received 26 March 2007; received in revised form 29 June 2007; accepted 3 July 2007 Available online 18 July 2007 This report is dedicated to the memory of Edward I. Stiefel: mentor, friend, and scientist whose curiosity and love of science knew no bounds.

Abstract Maya Blue is an ancient blue pigment composed of palygorskite clay and indigo. It was used by the ancient Maya and provides a dramatic background for some of the most impressive murals throughout Mesoamerica. Despite exposure to acids, alkalis, and chemical solvents, the color of the Maya Blue pigment remains unaltered. The chemical interaction between palygorskite and indigo form an organic/inorganic complex with the carbonyl oxygen of the indigo bound to a surface Al3+ in the Si–O lattice. In addition indigo will undergo an oxidation to dehydroindigo during preparation. The dehydro-indigo molecule forms a similar but stronger complex with the Al3+. Thus, Maya Blue varies in color due to the mixed indigo/dehydroindigo complex. The above conclusions are the result of application of multiple techniques (X-ray diffraction, differential thermal analysis/thermal gravimetric analysis, high resolution transmission electron microscopy, scanning electron microscopy, infrared and Raman spectroscopy) to the characterization of the organic/inorganic complex. A picture of the bonding of the organic molecule to the palygorskite surface forming a surface complex is developed and supported by the results of density functional theory calculations. We also report that other organic molecules such as thioindigo form similar organic/inorganic complexes thus, opening an entirely new class of complex materials for future applications.  2007 Elsevier Inc. All rights reserved. Keywords: Maya; Pigments; Indigo; Palygorskite

1. Introduction The Maya, whose lands reached from the highlands of Guatemala and neighboring Mexico, El Salvador, and Honduras through Belize and the broad plains of Mexico’s Yucatan peninsula, boasted one of the greatest civilizations of antiquity between about 250 and 900 C.E. [1]. Mayan craftsmen created a striking, blue/turquoise pigment, resembling the Caribbean Sea known as Maya Blue. For more than 50 years, this particular blue hue has been the subject of much interest and debate among scientists [2,3]. In 1931, Merwin was the first person to publish an article describing the distinct blue paint found on the *

1

Corresponding author. Tel.: +1 915 747 7555; fax: +1 915 747 6007. E-mail address: [email protected] (R.R. Chianelli). Present address: Mayan Pigments Inc.

0162-0134/$ - see front matter  2007 Elsevier Inc. All rights reserved. doi:10.1016/j.jinorgbio.2007.07.009

remains of a Maya wall painting at Chitzen Itza, Yucatan [4]. Merwin described the paint as having fairly coherent films relative to the other colors found on the murals. In 1942, Gettens and Stout designated the term Maya Blue (today also known as Azul Maya) to describe the brilliant blue colorant because it was believed to exist exclusively in the Maya Yucatan region [5]. Since the first published description of Maya Blue found on the mural painting in Chitzen Itza, the color has also been identified in many other regions throughout Central America. A particularly fascinating aspect of Maya Blue, is that despite the environmentally harsh humidity and high temperatures, it doesn’t fade and has unprecedented stability when exposed to acids, alkalis, and chemical solvents, and is naturally resistant to biodegradation; properties that have contributed to its continued endurance throughout Meso-America. Fig. 1 shows remnants of a fresco at

L.A. Polette-Niewold et al. / Journal of Inorganic Biochemistry 101 (2007) 1958–1973

1959

Fig. 1. A photograph of a monument in Chitzen Itza, Yucatan (900 C.E.), Mexico painted with Maya blue. The photograph indicates the stability and long life of ancient Maya Blue. Photo by Lori Polette.

Chitzen Itza with Maya Blue still brilliant after 900 years although the other pigments have faded. Reconstruction of the method employed by the pre-Hispanic civilization to create exceptional paints has been a challenge. The pigment is not a copper mineral and has no relation to natural ultramarine or ground lapiz lazuli as originally thought. The composition consists of a mixture of natural white powdery clay, palygorskite, and an organic pigment, indigo, neither of which alone exhibit any of these characteristic properties [6]. While the exact technique that the Maya employed to synthesize such a sophisticated paint remains an unknown, the pigment can be reproduced in a laboratory today with all the properties of historical Maya Blue [7]. We now know it as an organic/inorganic complex and furthermore, the organic compound, indigo can be replaced with innumerable alternate organic dyes yielding a universe of beautiful and novel new materials [8]. The fact that Maya Blue also does not contain strategic or heavy metals makes it a promising candidate replacement pigment for the paint and pigment industry. These industries normally consume annually large amounts of strategically important and environmentally unfriendly metals. Replacement of these heavy metal containing pigments such as Co phthalocyanine with others based on the ancient Maya technology would greatly improve the current environmental situation and result in considerable savings in strategic metals. However, the feasibility of developing pigments such as Maya Blue and analogues of other hues necessitates first that the stability of the clay/ indigo complex be understood [9]. Samples of pigments used in Mayan painting and pottery have been studied for many years and by many authors [10–12]. In a previous study, Mayan samples were studied using X-ray absorption spectroscopy (XAS) and high-resolution transmission microscopy (HRTEM) [13].

In this report we review synthetic methods to produce Maya Blue and introduce novel organic/inorganic complexes. We also discuss current knowledge regarding the complex interaction of palygorskite and indigo from the macroscopic to the atomic scale. The physical and chemical properties are characterized using various experimental techniques (optical microscopy, scanning electron microscopy (SEM), (HRTEM), X-ray powder diffraction (XRD), infrared absorption and Raman scattering spectroscopy, thermogravimetric analysis (TGA) and surface analysis) and these results are complimented by density functional theory (DFT) calculations. Details of the analytical techniques not described in this report can be found in reference [14,15]. The solid-state chemical interaction phenomenon that gives rise to the pigments stability is shown to be a surface interaction between the organic and inorganic components. 2. Background In this section we describe the properties of the starting materials that are the components of Maya Blue and the organic/inorganic complex materials. 2.1. Palygorskite (Attapulgite) Palygorskite is a fibrous clay as seen in a SEM (scanning electron micrograph) (Fig. 2). The average fiber dimensions range between 0.1 and 2 lm, having a width between 100 ˚ , and thickness of 50 and 100 A ˚ . The crystal and 300 A structure of palygorskite based on Bradley’s original work is shown in Fig. 3 [16]. The clay has a anisotropic structure with porous channels along the C-axis with dimensions of ˚ (width) · 10.5 A ˚ (height) and 16.2 A ˚ (depth) as seen 4.1 A ˚, b= in Fig. 4. The cell parameters are a = 13.24 A

1960

L.A. Polette-Niewold et al. / Journal of Inorganic Biochemistry 101 (2007) 1958–1973

Fig. 2. SEM of palygorskite (Attagel 50 from Engelhard Corporation).

˚ , and c = 5.21 A ˚ and b = 105. Bradley was the first 17.89 A to deduce the structure and proposed it to be monoclinic with space group C2/m. Since Bradley presented his structural scheme, several other investigators have reportedly found alternative space groups with orthorhombic symmetry, although Bradley’s is still generally accepted [17]. Palygorskite is considered to be a hydrous hydrated magnesium phyllosilicate with varying amounts of Al3+, Fe3+ (R3+) and Mg2+, Ca2+, Fe2+ (R2+) cation substitutions [18]. Thus, the general structural formula is: ðR2þ ; R3þ Þ5 ðSi; R3þ Þ8 O20 ðOH2 Þ4 Mnþ ðH2 OÞ4 To maintain charge neutrality, if Al replaces Si in the tetrahedral position, one trivalent ion replaces one divalent ion in the octahedral position. Or for example, two trivalent ions may replace three divalent ions, in which case there is substitution without modification of the charge of the octahedral layer [19] Depending upon the extent of divalent

Fig. 4. 2 · 2 · 3 cell of palygorskite showing channel structure. Cerius2 model. The van der Waals surface is shown in light blue. The channel ˚ wide, 10.5 A ˚ high and surrounded by van der Waals surface is 4.1 A ˚ 16.2 A deep. Red = oxygen, orange = silicon, pink = magnesium.

or trivalent ion substitution, the clay is sometimes referred to as dioctahedral or trioctahedral. However, the octahedral sites are sometimes discontinuous and generally contain five cation sites that may or may not be occupied with exchangeable cations such as Ca2+ or Mg2+. Additional substitution can appear within the clay structure as follows: Al–Sitet ; Altet –Aloct ; Alþ -deficiency; Fe2þ –Mgoct ; Fe3þ –Al3þ

Fig. 3. Ideal accepted monoclinic structure of palygorskite (Attapulgite) projected along the fiber axis (13). Red = oxygen, grey = hydrogen, orange = silicon, pink = magnesium.

L.A. Polette-Niewold et al. / Journal of Inorganic Biochemistry 101 (2007) 1958–1973

The basic atomic framework consists of alternating ribbons of tetrahedral silicates (SiO4) with basal oxygen atoms that have apices alternately pointing up and down in adjacent sheets. The bonding within the structure occurs via covalent, ionic, metallic, and van der Waals bonds. However, it is the oxygen network that ultimately determines the crystal shape and symmetry [20]. A synchrotron X-ray diffraction pattern of the palygorskite used in these studies showed deviations from the ideal structure described previously. These discrepancies are likely due to the fact that palygorskite samples from different regions exhibit various structural differences. Some of the impurities inherent in the clay are Calcite (CaCO3), Dolomite (CaMg [CO3]2), and quartz. Calcite (CaCO3) and dolomite are the most common minerals found. An XRD pattern of a sample of palygorskite from the Yucatan, Mexico peninsula (Sacalum) mined near the Gulf of Mexico was obtained for comparative purposes and is shown in Fig. 5 along with a simulated diffraction pattern for the primary impurity, dolomite. The ubiquity of magnesium deposits near oceans increases the dolomite mineral content in palygorskite. Energy dispersive spectroscopy (EDS) was used to compare the ideal composition of palygorskite with the palygorskite used in the synthesis of Maya Blue (Attagel 50) and Sacalum. The ideal stoichiometry of Bradley’s model, palygorskite and Sacalum are compared in Table 1. There are three types of hydroxyl groups in palygorskite: crystalline OH, bound (also called coordinated) H2O, and zeolitic H2O. Eight water molecules are bound to octahedral cations, and other molecules that appear as zeolitic water can be bound to the exchange cations within the channels [21]. The free volume surface was calculated in

1961

Table 1 Ideal formula and calculated stoichiometry (from EDS) of palygorskite and Sacalum Sample

Formula

Ideal (from Bradley)

Mg5 Si8O20(OH)2(OH2)4 (Mg2,Al2)(Si7.8,Al0.2)O20 (OH)2(OH2)4

Palygorskite (purchased as Attagel 50)

Fe0.4Mg2Al2Si8O20(OH)2(OH2)4 Fe0.4 (Mg2,Al1.7)(Si7.7,Al0.3)O20(OH)2(OH2)4

Sacalum

Fe0.3Mg2Al2Si8O20(OH)2(OH2)4 Fe0.3(Mg2Al1.7)(Si7.7Al0.3)O20(OH)2(OH2)4

CERIUS2 as depicted in Fig. 4. The channel water was removed prior to calculation as these molecules are susceptible to dehydration under relatively low temperatures and can even be replaced by organics of similar dimension [22]. Loss of the water exposes the surface cations (Fe3+ or Al3+) and these cations play a role in the binding of indigo and related molecules. Because the water groups constitute an important role in surface adsorption a thermogravimetric analysis (TGA) and differential thermal analysis (DTA) study was performed. The results of the TGA and DTA are summarized in Table 2. From the TGA there is a 6% weight loss from 38 C to 111 C corresponding to the hygroscopic and zeolitic water. A 2% weight loss of coordinated water occurs between 111 C and 224 C. From 224 C to 500 C there is a 5% weight loss as some of the structural hydroxyls are lost. At temperatures higher than 500 C the silicate ribbons begin to fold, and an exothermic peak in the DTA at 712 C is due to transformation into collapsed, amorphous phases.

˚ ). Fig. 5. Synchrotron XRD pattern of Sacalum with simulated XRD of dolomite (wavelength k = 1.239 A

1962

L.A. Polette-Niewold et al. / Journal of Inorganic Biochemistry 101 (2007) 1958–1973

Table 2 Results of TGA and DTA of palygorskite Unit

Temperature (C)

TGA (wt% loss)

DTA

Hygroscopic, zeolitic water Coordinated water Hydroxyls Recrystallization

38–111 111–224 224–500 712

6% 2% 5.5% –

Endotherm Endotherm Endotherm Exotherm

2.2. Indigo/Thioindigo Adolf von Baeyer from Munich, Germany firmly established the structure of indigo in 1883 [23]. An indigo mol˚ ecule, which is depicted in has dimensions of 14.0 A ˚ ˚ (length), 5.3 A (width) and 3.4 A (thickness) as seen in Fig. 6. In the crystalline phase, indigo is monoclinic with space group symmetry P21. The lattice parameters are: ˚ , b = 5.89 A ˚ , and c = 12.28 A ˚ ; a = c = 90 and a = 10.84 A b = 130.02. The crystal structure obtained from CERIUS2 is shown in Fig. 6. Indigo is deeply colored blue in its natural form and is practically insoluble in any solvent. It is an important historical pigment and is used extensively in the textile dye industry. Indigo light absorption occurs at long wavelengths (approximately 600 nm, depending on the solute). This large bathochromic shift is likely due to hydrogen-bonded dimers or higher polymers present in the crystalline state of indigo [24]. The indigo monomer is known to possess a red color in the gas phase but is blue in the solid crystalline state [25].

The optical properties of indigo are due to the crossconjugated system of donor and acceptor groups linked by a C@C double bond [26]. The potentially acidic protons are held in close proximity to the potentially basic carbonyl oxygen atoms by means of hydrogen bonding in the ground state [27]. Hydrogen bonding in indigo causes indigo to sublime rather than melt at temperatures ranging from 360 to 390 C, depending on the purity and crystallinity. In addition, indigo can undergo the reactions as indicated in Fig. 7. In the first reaction indigo may be

dehydroindigo O2 + H2O Δ indigo nitrogen oxygen -

hydrogen leuco-indigo -

Fig. 7. Reactions of Indigo: bottom – reversible reduction to leuco-indigo, top – irreversible oxidation to dehydroindigo.

Fig. 6. Crystal structure of indigo from CERIUS2. Red = oxygen, grey = carbon, white = hydrogen, blue = nitrogen.

L.A. Polette-Niewold et al. / Journal of Inorganic Biochemistry 101 (2007) 1958–1973

1963

reduced with a reducing agent such as sodium dithionate (Na2S2O4) in NaOH to leuco-indigo which is water soluble and yellowish white. This reaction is easily reversed through exposure to air, which rapidly oxidizes the leucoindigo and produces the blue, water insoluble indigo. This reaction has been prevalent in the dye industry for thousands of years [28]. Fig. 7 also includes the molecule dehydroindigo. This molecule is produced irreversibly when indigo is heated in wet air at about 150 C [29]. It is now clear, as discussed below, that dehydroindigo is produced when Maya Blue is synthesized. This fact was first observed in historical Maya Blue by application of electrochemical techniques [30]. Fig. 8 shows the structure of thioindigo. The nitrogen of dehydroindigo is replaced by sulfur and thus its chemistry is analogous to dehydroindigo. The absence of the proton that in indigo is attached to the nitrogen makes the chemistry simpler for this study.

Maya blue, and authentic Maya blue were also obtained from Luis Rendon at Universidad Nacional Autonoma de Mexico [9]. The samples were prepared by grinding them in a small mortar, then they were suspended in an inert liquid, and put in an ultrasonic bath to disperse the particles. The samples were carbon coated and then dried. Using a small tubular capillary, a drop of the liquid suspension was placed in a 3 mm microscope slit holder. High-resolution images were obtained using a JEOL 4000-EX microscope with a point resolution of 0.17 nm, running at 400 keV. For the diffraction images, the camera constant was 2.83 nm/mm, after calibration. Thermal gravimetric analysis (TGA) and Differential thermal analysis (DTA): TGA and DTA were made simultaneously on a TA Instrument SDT 2960 DTA-TGA at CIMAV in Chihuahua, Mexico. Samples were run under dry nitrogen flow from 293 to 1073 K at a heating rate of 10 K/min. Approximately 20 mg of palygorskite, indigo, and a synthetic Maya blue sample were analyzed to monitor changes or examine any structural differences. Fourier-transform infrared spectrometer (FT-IR) and Raman spectroscopy: Experimental details of the FT-IR and Raman scattering spectroscopy have been previously reported [14,31]. X-ray powder diffraction (XRD): The X-ray diffraction data were collected at the dedicated powder diffractometer at Stanford Synchrotron Radiation Laboratory (Beam ˚ ) in symmetric Bragg–Bretano Line 2-1) at 10 keV (1.23 A geometry. A Si(1 1 1) crystal was used as an analyzer before the NaI detector to obtain a high (angular) resolution dataset. However, since most materials are poorly crystalline high angular resolution is unimportant. Surface area measurement: Specific surface area determination was done with a Quantachrome model AUTOSORB-1, by nitrogen adsorption at 77 K using the BET isotherm. Samples were degassed under flowing argon at 473 K for 2 h before nitrogen adsorption.

3. Experimental

4. Synthesis of the organic/inorganic complex materials

As previously mentioned, some of the experimental details not discussed below can be found in references [13–15]. Materials: Attagel 50 (palygorskite) used in synthetic samples was purchased from Engelhard Corporation and was used as received. Indigo and thioindigo were purchased from BASF Corporation and used as received. Scanning electron microscopy (SEM): SEM images were obtained from Luis Rendon at the Universidad Nacional Autonoma de Mexico [9]. Samples were gold coated prior to imaging. Scanning Electron Micrographs were also obtained at UTEP using a Philips Electroscan, ESEM model 2020 equipped with an EDS detector. In the case where EDS was employed, the samples were not gold coated and were imaged under 20 keV and approximately 2.4–3.0 Torr. High-resolution transmission electron microscopy (HRTEM): HRTEM images of palygorskite, synthetic

4.1. Synthesis of Maya Blue

Fig. 8. The structure of thioindigo. Red = oxygen, grey = carbon, white = hydrogen, yellow = sulfur. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

By examining the effects of indigo concentration and pH, a color remarkably similar and stable to that of ’’real’’ Maya Blue was developed. 4.2. Method 1 – reduction to Leuco-indigo A 0.05 g sample of indigo (BASF) was suspended in 50 ml of water. Sodium hydrosulfite, 0.03 g, and 15 drops of 1 M NaOH were added to the solution to reduce the indigo. The solution was heated to 90 C while stirring with a magnetic stirrer. The solution remained blue, at which point an additional 0.1 g of sodium hydrosulfite was added. The solution turned clear indicating that the indigo had been reduced and was now soluble leuco-indigo. The leuco-indigo was then poured over 5 g of palygorskite clay,

1964

L.A. Polette-Niewold et al. / Journal of Inorganic Biochemistry 101 (2007) 1958–1973

and was stirred. Immediately upon contact with the clay and being exposed to air, the solution turned dark blue. The solution was placed in the oven at 125 C for 4 days, at which point it resembled the color of Maya Blue. The hue was slightly lighter and less intense than that of original Maya Blue. 4.3. Method 2 – grinding in water This method determined that it was unnecessary to reduce the indigo prior to blending it with the clay. Thus, a series of Maya Blue samples were prepared without reducing the indigo first. Preparing a series of Maya Blue samples with varying concentrations was accomplished by grinding a 0.05 g, 0.1 g, 0.3 g, and 0.5 g of indigo with 5.00 g of palygorskite clay, respectively. Each concentration series was then placed in a blender with 100 ml of de-ionized water and blended for several minutes to ensure that a homogenous mixture was obtained. The corresponding solutions were placed in a 250 mL beaker in the oven at 125 C for 4 days. To determine whether or not the pH of the system was an important factor in obtaining the precise hue of Maya Blue, synthetic samples were prepared using either sodium hydroxide or hydrochloric acid to prepare the samples in either a basic or acidic solution. As will be discussed later, it was found that using P2% indigo was optimal for obtaining ancient Maya Blue. For the pH studies, 0.1 g of indigo was ground with 5.0 g of clay four times to obtain four samples. To each ground mixture was added 100 ml DI water, and drop-wise either 1 M NaOH, or 1 M HCl to obtain solutions with pH’s of 4, 7, 9, and 11. The pH of the system was monitored with a pH meter that had been calibrated with buffers of pH 4, 7, and 11. The various shades of Maya Blue produced as a function of altering the concentration of indigo and the pH of the solution are shown in Fig. 9. The colors arranged horizontally, prepared with increasing concentrations of indigo, visually range from a pale bluish-green to a darker

grayish blue-green. The variations of color seen in the vertical direction range from grayish blue-green under basic conditions to brighter, vibrant blues at neutral to acidic pH. Visually, the Maya Blue prepared with the highest weight percentage of indigo under neutral to acidic pH most resembles the ‘authentic’ Maya Blue. The sample prepared with nearly 1% indigo still lacked some brilliance in color suggesting that an even higher concentration of indigo is necessary. The surface of Palygorskite contains Al3+ sites that can act as Lewis acids, especially under neutral to acidic conditions. At high pH, the surface –Al–OH groups may only bond through hydrogen bonding. Under acidic conditions the following reaction is more likely: þ AlðOHÞ3ðsÞ þ 3Hþ ðaqÞ ! Al3ðaqÞ þ 3H2 OðlÞ

In this scenario, the aluminum can act as a Lewis acid and bind with the nitrogen and carbonyl on indigo. 4.4. Method 3 – dry grinding This method confirmed that Maya Blue could also be produced simply by grinding the indigo and clay without water, provided the two were subsequently heated. In similar manner a complex of thioindigo and palygorskite was prepared by mixing thioindigo and the palygorskite at room temperature and then heating for 24 h. Example: 0.3 g (6% weight) of indigo was blended with 4.7 g of palygorskite. One gram of this mixture was placed in a vial. The remaining sample mixture was heated for 24 h at 100 C at which point another gram was taken from the mixture and placed in a vial. Finally, the rest of the mixture was heated an additional 24 h at 140 C and then placed in a vial for later analyses. Mixing the beige clay (the color of palygorskite) with dark blue indigo at room temperature produces a blue tone. Upon heating the sample to 100 C the mixture turns a greenish blue and further heating to 140 C finally produces the brilliant Maya blue color. 4.5. Gettens test and DMSO extraction

Fig. 9. Horizontally shown are the color variations of synthetic Maya blue with increasing indigo. Vertically shown are the effects of pH.

The original method of testing the stability of Maya blue was developed by Gettens [5]. A modified version of the Gettens test was performed on the synthetic samples of Maya blue produced in the laboratory containing various weight percentages of indigo and prepared under neutral pH. Six 0.1 g samples of synthetic Maya blue were placed in six separate 70 mm long tubes having a diameter of 8 mm. Each sample was covered to a depth of 5 mm with HNO3, H2SO4, Aqua Regia, HCl, and a solution of Na2S2O4 + 5% NaOH, respectively. The color changes were monitored for a period of 24 h. For all of the test tube reactions, there was no immediate color change. Following a 24-h period, only the samples that had been exposed to HNO3 and Aqua regia had a slight color change in the solid and solution phase. The solution phase had an olive greenish tint, as did the solid sample. There are some

L.A. Polette-Niewold et al. / Journal of Inorganic Biochemistry 101 (2007) 1958–1973

1965

Fig. 10. Extraction of excess indigo and impurities from synthetic Maya Blue.

impurities in the indigo purchased from BASF that when extracted in any number of organic solvents have a green hue, and may be the cause of the colors in solution phase of the test tubes. It is accepted that Maya Blue is qualified as such if there are no color changes after it has been subjected to the Gettens test. As such, the preparation method of simply grinding and heating the indigo/palygorskite was proven successful. Additionally, samples of the synthetic pigment were extracted with DMSO. Samples were prepared in the traditional manner, including heat treatment. A control sample was also extracted that contained a mixture of indigo and clay that had not been heat-treated. Each sample was placed in a filter funnel and successively washed with DMSO until the filtrate became clear. There was very little change of color following DMSO extraction as shown in Fig. 10. The indigo from the control sample containing ground indigo and palygorskite, unheated, was easily extracted. After several washes with DMSO, most of the pigment was separated from the clay leaving the original Palygorskite in the filter with only the slightest hue of bluish gray. The results from the DMSO extraction experiment give convincing indication there is a strong chemical interaction only after heating the indigo and palygorskite. The color of the sample remained consistent even after attempted extraction with DMSO. The percent weight losses were calculated based on a standard calibration curve. 5. Organic/inorganic surface complexes 5.1. Color changes A key measure of the formation of the organic/inorganic complex is the color change seen in Fig. 11. On the left side of the Figure is the formation of the indigo/palygorskite, Maya Blue complex. The color changes occur as the temperature increases and the complex is formed. The deep blue of indigo changes to the classic Maya Blue color. Reproduction makes it difficult to see the actual colors, however the color changes seen (right side Fig. 11) upon

Fig. 11. Color changes as a function of temperature and heating time of organic/inorganic materials: left, indigo and right, thioindigo.

formation of the thioindigo complex are even more dramatic. The reddish thioindigo changes to dark blue upon forming the organic/inorganic complex. These color changes are a clear indication of chemical interaction between the clay and the organic molecules as confirmed by DFT calculations described later.

5.2. Synchrotron X-ray diffraction The XRD of the palygorskite clay (Attagel) and Maya Blue prepared at 25 C and 140 C are shown in Fig. 12. Also seen is the CERIUS2 simulated diffraction pattern of indigo. The diffraction pattern of the unheated mixture of clay and indigo (Maya blue 25 C) still has peaks due to the crystalline phase of indigo. Once the clay and indigo are heated together to produce the Maya blue 140 C sample, the indigo phase no longer gives diffraction peaks. Clearly, the crystal structure of indigo has been disrupted as a consequence of its binding to the clay surface. Fig. 13 shows the same effect in the formation of the thioindigo/palygorskite complex. The crystal structure of the thioindigo disappears. It should be further noted that no

1966

L.A. Polette-Niewold et al. / Journal of Inorganic Biochemistry 101 (2007) 1958–1973

˚ ). Fig. 12. XRD of the formation of Maya Blue as the temperature increases (wavelength k = 1.286 A

˚ ). Fig. 13. XRD of the formation of thioindigo/palygorskite complex as the temperature increases (wavelength k = 1.77 A

change in the lattice parameter or superlattice was observed in the synchrotron XRD of the palygorskite. If the organic molecules entered the channels completely, a change in the lattice parameters and/or the formation of a superlattice should be expected. The above results confirm the formation of an organic/ inorganic surface complex:  A strong color change upon heating.  Resistance to chemical extraction of the organic.

 Disappearance of the indigo crystal structure as determined by XRD.  No observable change in the lattice parameters of palygorskite.  Absence of any observable superlattice. Upon heating the organic molecule covers the surface of the palygorskite clay forming the stable organic/inorganic complex. Further evidence is provided in the following sections.

L.A. Polette-Niewold et al. / Journal of Inorganic Biochemistry 101 (2007) 1958–1973

1967

5.3. HRTEM of synthetic Maya Blue

16

Further evidence for the interaction between the organic molecule and the clay was obtained from HRTEM (highresolution electron microscopy) [13]. Palygorskite is unstable in the electron beam as originally reported by Vivaldi et al. [32]. Excessive beam exposure causes the ends of the particles to become rounded and details of the surface to be blurred and mottled. After prolonged exposure the clay becomes completely amorphous due to loss of water from the channels causing the channels to collapse. Even at low temperatures (90–110 C) the water molecules within the channels are subject to dehydration and more than half of the zeolitic water is lost. However, when indigo reacts with the clay through heating the situation changes dramatically. An HRTEM image of synthetically prepared Maya Blue shows that, when the organic/inorganic complex is formed through heating, the ensemble is now stable to the electron beam and lattice images can be seen in Fig. 14 [13]. Evidently the indigo molecules block the channels, preventing further water loss and channel collapse, imparting the clay/indigo stability. The indigo molecules are bound tightly in the channel ends and on the exposed channel surfaces.

14

intense broad peak (450–700 C) due to charring or degradation of the indigo is absent in Maya Blue, suggesting that a separate indigo phase is not present. Clearly, the composite structure is altered by the binding of indigo because with increasing temperature, there is no sharp peak due to the degradation of indigo.

5.4. Differential thermal analysis (DTA)

5.5. Surface area

Further data that support the previously described description of the organic/inorganic complex can be see in Fig. 15. The figure shows the DTA of Maya Blue and indigo. Indigo sublimes at 360 C and the sublimation point peaks of indigo are much less intense and broader in Maya Blue indicating that the bonding interaction between indigo and the clay surface is stronger than a hydrogen bond. The small peak at 390 C is still visible and is probably due dehydro-indigo sublimation. The

It was demonstrated by Shariatmadari et al. that the binding sites of palygorskite are located at the external surfaces of the fibrous clay particles so the external surface area is a major criterion controlling the contribution of these sites to the total sorption [33]. To further probe the possibility of pure surface effects, BET measurements of the pore size distribution for pure palygorskite clay and of synthetic Maya Blue prepared with 0.1 g, 0.3 g, and 0.7 g indigo per 5 g of Palygorskite (Attagel) clay were made. The results of the BET measurements and surface area changes between the pure palygorskite and Maya Blue are shown in Table 3. The most important region of pore ˚ because it represizes is in the range between 10 and 20 A sents the dimensions of the palygorskite channels; hence, if the indigo is locking into the channels, the surface area would drastically change as a function of increasing concentration of indigo. Initially it was believed that because ˚ · 4.5 A ˚ · the indigo molecule has dimensions of 3.35 A ˚ , it would fit into the 4.10 A ˚ · 10.5 A ˚ · 14.0 A ˚ chan14.0 A nels of the fibrous palygorskite filling each unit cell.

TGA

Weight Percent (mg)

12

DTA 10 8 6 4 2 0 0

100

200

300

400

500

600

700

800

900

Temperature (ºC)

Fig. 15. Combined DTA graph of indigo and synthetic Maya blue.

Table 3 BET surface area for palygorskite and synthetic Maya Blue

Fig. 14. HRTEM photo of palygorskite showing near atomic resolution [13].

Material

Surface area (m2/g)

%D(surface area)

Desorption peak ˚ maximum 10–20 A

Palygorskite 0.1 g Indigo/4.9 g clay 0.3 g Indigo/4.7 g clay 0.7 g Indigo/4.3 g clay

128.6 113.8 103.7 57.2

– 14.8 24.9 71.4

1.4 · 103 0.5 · 103 0.4 · 103 0.3 · 103

1968

L.A. Polette-Niewold et al. / Journal of Inorganic Biochemistry 101 (2007) 1958–1973

However, the palygorskite fiber can be thought of as a cylinder. The surface area of the end is considered negligible relative to the cylindrical surface, which is dominant. Furthermore, as discussed below upon bonding the molecule will no longer be flat and thus further penetration into the channels is limited. Thus, although the channel ends are probably filled with the indigo molecule, the penetration into the channels is minimal and the major interaction between the organic molecule and the clay occurs at the surface. The complex organic/inorganic material can be thought of as a surface compound. An average fibrous palygorskite particle has a length of 1000 nm and a radius of 6.5 nm. An average particle of this type with a density of 2.36 g/cm3 will have a surface area of 129 m2/g. The average particle size was confirmed by measuring the full-width at half maximum of several peaks in the X-ray patterns. This model means that approximately 1/3 of the atoms are on the surface and 2/ 3 of the atoms are in the interior and of the interior space ˚ pore region. The 1/3 is the channel volume in the 10–20 A surface area decreases only by 14.8% for 0.1 g of indigo and by 24.9% for 0.3 g of indigo. Based on the small change, the chemical interaction between the pigment and the clay occurs primarily at the clay surface. If any of the indigo molecules are intercalating into the channels, it is likely that they penetrate only within the first two or three unit ˚, cells. It is clear that in the pore size region of 10–20 A the surface area curve is decreasing with increasing concentration of indigo suggesting the pores or channels of the clay are blocked. The surface are of the palygorskite is 129 m2/g, which ˚ 2, and each indigo molecule occupies equals 1.29 · 1022 A 2 ˚ . For 1.29 · 1022 A ˚ 2/(55 A ˚ 2/indigo molecule) it takes 55 A 20 2.35 · 10 molecules indigo to cover the surface. EDS analysis of a 6% weight indigo sample resulted in the following calculated stoichiometry for a Maya Blue sample:

Fig. 16. FT-IR spectra of Maya Blue samples with 16% indigo heated for 3 h (MB6 –3 h), 6 h (MB6 –6 h), and 9 h (MB6 –9 h) at 170 C [31]. (a) including pure indigo and palygorskite, (b) expanded scale for the formed surface complexes.

ðIndigo0:2 ÞðFe0:4 ; Mg2:0 ; Al1:7 ÞðSi7:7 ; Al0:3 ÞO20 ðOHÞ2 ðOH2 Þ2 According to the experimental data, the mole occupancy of indigo is only 0.2 and cannot be layering the entire outer surface of the clay. Considering that 36% of the surface is composed of exposed channels occupied by indigo the mole ratio becomes 0.12, a value close to the experimental data. 5.6. FT-IR and FT-Raman FT-IR and FT-Raman spectra for indigo/palygorskite and thioindigo/palygorskite are reported in references [31,14] respectively. The FT-IR and Raman results for indigo loadings of 6% and 16% are summarized below. The former is representative of Maya Blue and the latter is representative of a new class of novel organic/inorganic complex materials unknown to the Maya [8]. The FT-IR and FT-Raman results are shown in Figs. 16 and 17, respectively. The partial elimination of the selection rules for the

Fig. 17. Raman spectra of Maya Blue samples with 16% indigo heated for 3 h (MB16 – 3 h), 6 h (MB16 – 6 h), and 9 h (MB16 – 9 h), at 170 C [31].

centro-symmetric indigo, indicating distortion of the molecule is observed. This distortion accounts for the observed color changes as the molecular orbital structure changes

L.A. Polette-Niewold et al. / Journal of Inorganic Biochemistry 101 (2007) 1958–1973

and the complex is stabilized. This data also shows the disappearance of the indigo N–H bonding, as the organic molecules incorporate into palygorskite material. A structural change of indigo to dehydroindigo during heating is indicated by this result. Evidence of bonding between cationic aluminum and dehydroindigo through oxygen and nitrogen at longer heating and at higher dye concentrations is also revealed by FT-Raman measurements. The infrared transmission spectra of indigo (gray line), palygorskite (dashed line), and of the Maya Blue samples with 6% wt dye concentration are presented in Fig. 16a. The most important change with increasing heating time is the disappearance of the indigo N–H stretching vibration at 3268 cm1. The intensity decrease of the broad bands around 3400 cm1 and 3560 cm1 indicates that water is removed from both the zeolitic and structural sites in the palygorskite [34]. Similar water removal results were previously reported by Giustetto et al. in their FT-IR experiments on outgassed palygorskite and freshly synthesized Maya Blue samples [35]. The spectra of Maya Blue samples with 16 wt% dye concentration (Fig. 16b) demonstrate a similar behavior in the 3000–4000 cm1 region, as previously observed for the analyzed sample with a lower dye concentration. However, there is a slight indication of less molecular water removal with the increase in heating time in these IR spectra. The obvious difference between the spectra of Maya Blue samples with 16 wt% dye concentration and the spectra of Maya Blue samples with 6% wt dye concentration is the presence of a weak feature corresponding to N–H vibration in the latter cases. Fig. 17 presents the FT-Raman spectra of Maya Blue samples with a 16% dye concentration, and heated for 3 h, 6 h and 9 h at 170 C, respectively. Intensity decrease, shifts toward higher wavenumber, and peak disappearance with increasing heating time are observed for the vibrational lines of indigo in these samples. The very weak presence of the cN–H vibration at 635 cm1 is still depicted in these spectra. The 1701 cm1 (mC@O) peak exhibits an evident decrease in intensity, and eventually a complete disappearance, in the spectra of MB16-6 h and MB16-9h samples. These vibrational lines are marked in the spectra by thin solid arrows. The spectra of MB16-3h, MB16-6h and MB16-9h show a definite presence of new features at 425 cm1 and 606 cm1. There are two possible assignments for the 606 cm1 peak: (i) a shift of the 598 cm1 feature (dC@O, dC–H, dC–NH–C), and (ii) a new vibrational line due to Al–O bonding [30]. The appearance of the 425 cm1 peak, which is absent in the indigo spectrum and is associated with Al–N bonding, removes previous doubts about the origin of the 606 cm1 vibration. Our explanation of potential interaction between the indigo and the clay is as follows. Palygorskite, the inorganic support for the synthesis of these pigments, contains tetrahedral ribbons consisting of silicon–oxygen–silicon and silicon–oxygen- (hydrogen, aluminum and magnesium). Thus, as the nitrogen loses a proton, it can bond to the exposed metal. Lewis acid sites

1969

(primarily aluminum) act as binding sites for the functional groups present in organic dyes such as indigo, causing a change in the electronic structure, color properties, and ultimately lead to excellent chemical and thermal stability of the hybrid pigment. In summary aluminum forms dative (Lewis acid/base) chemical bonds with oxygen and nitrogen. The appearance of a new peak at 1395 cm1, which is marked in the spectra by a dashed arrow and is corresponding to the Bu type IR active vibrational line at 1392 cm1 (dN–H, dC–H), is also observed in Fig. 17. The activation of the originally Raman-forbidden vibrational Bu mode may be due to perturbations in the planarity of the indigo molecules as previously attested by Witke et al. from micro-probe Raman measurements of a blue part of a Mayan clay ornament [36]. They stated that the observed shifts in the position of most bands in the Raman spectrum of Maya Blue, as compared with the spectrum of the synthetic indigo should be expected, since changes in the planarity of indigo molecules involves changes in bond distances and angles. The above observations such as disappearance of N–H and C@O vibrations together with the appearance of new features at 606 cm1 and 425 cm1, support the Al–O and Al–N bonding model as opposed to the hydrogen bonding model proposed by others. The progressive disappearance of the N–H vibration as a function of heating time indicates that indigo is converting to dehydroindigo. The conversion of indigo to dehydroindigo increases with increasing concentration of starting indigo and with increasing time of heating. There is also a visible change

Fig. 18. Insertion of indigo molecules into the channels of palygorskite [14].

1970

L.A. Polette-Niewold et al. / Journal of Inorganic Biochemistry 101 (2007) 1958–1973

in the color of the samples with heating time. The sample that were heated for 3 h have a deep blue color, as compared with the samples heated longer, which have a greenish-blue color. This color change is an indication of a strong interaction and structural change, possible due to an exchange of electron density at the surface. The presence of dehydroindigo can contribute to explain satisfactory the greenish-blue color of the samples heated for more than 3 h. It is known that indigo will transform to dehydroindigo upon heating in moist air [29]. These features will be included in the model described below. 6. Chemical model of synthetic Maya Blue There is still considerable controversy on the binding site of the indigo molecules and the binding site may be dye concentration dependent. For example, at low concentration, dye molecules may penetrate for a few layers into the channels. This effect is easily seen by subjecting the CERIUS2 structure seen in Fig. 4 to the Sorption Module in the CERIUS2 with indigo as the adsorbent [37]. The indigo molecules ‘‘tuck’’ into the channels as seen in

Fig. 18. However, it is well known in the field of insertion chemistry that diffusion in one dimensional channels is highly limited and at higher concentration surface coverage is preferred [37]. Additionally, the indigo molecule becomes not planar when binding further inhibiting diffusion. Some authors have considered that hydrogen bonding in the channels is the correct model [38]. However, our results show that higher concentrations surface binding is preferred and may even involve several layers. Further complicating the picture indigo undergoes oxidation at elevated temperatures in wet air producing dehydroindigo. It has been recently reported as described previously that mixtures of dehydroindigo and indigo are responsible for the various green/blue hues seen in ancient Mayan artifacts [29]. Thus, the presence of dehydroindigo is considered in describing the appropriate model. 6.1. DFT methods for calculation of optical spectra for Mayachrome pigments We modeled palygorskite-dye combinations using a plane-wave pseudopotential DFT (Density Functional

Fig. 19. Extended cell model for the organic/inorganic DFT calculation.

L.A. Polette-Niewold et al. / Journal of Inorganic Biochemistry 101 (2007) 1958–1973

Theory) code, CASTEP [39] followed by a calculation of optical and IR spectra using a molecular orbital method, VAMP [40]. The proposed structures were first built as molecules using 3 adjacent silica rings of the palygorskite structure (proposed surface structure) with a metallic impurity at a silicon site (aluminum, iron, and magnesium were used, with resulting optical spectra supporting the use of aluminum as the most promising candidate) chemically bonded to a dye molecule (indigo, thioindigo, etc.). The proposed surface complex was then used as the basis for a three dimensional periodic structure due to the inability of CASTEP to handle non-periodic structures. By allowing the unit cell to have a large lattice parameter along the axis perpendicular to the plane of the surface complex the problem is then reduced to that of modeling a series of noninteracting planes of surface complexes (this was confirmed by examination of the energies of the crystal structure as a function of spacing between planes of our model), as shown in Fig. 19. CASTEP was then used to provide an optimum geometry for the model. This is known as the empty cell method. CASTEP was then used to provide an optimum geometry for the model. Once optimized, the structure was reduced back to a non-periodic structure in order to calculate UV/Visible spectra. The calculations were performed on both the indigo and thioindigo/palygorskite surface complexes. However, the thioindigo due to the absence of the proton, is a simpler problem as described below. Many combinations of surface structures and bonding were calculated to determine the nature of the surface complex. The best fits to the optical spectra were gotten using the above empty cell method to optimize the structure followed by the application of VAMP to the resulting nonperiodic optimized structure. The following characteristics of the optimized surface structures gave the best results:  Octahedral bonding to aluminum surface defects.  Dimer structures involving two interacting organic molecules. This is consistent with the structure that the organic molecules take in their crystal structures.  The organic molecules distort form their planar structure. This fact is consistent with the observed color changes that require a change in the molecular orbital structure.

1971

Fig. 20. UV/Visible spectrum of the thioindigo/palygorskite surface compound.

Fig. 21. Relaxed structure of thioindigo/palygorskite complex.

6.2. Thioindigo DFT structure The optical results for the thioindigo/palygorskite complex are shown in Fig. 20. The excellent fit between the experimental and the calculated spectra is evident. The fit is particularly effective because both the UV and the visible spectra fit well. Although the UV portion is not visible in the pigment, the adsorption properties are crucial and related to the stability of the pigment. The optimized structure is shown in Fig. 21. The relaxation of the planar structure of the thioindigo is clearly seen. The molecular orbital

Fig. 22. UV–Vis spectra theoretical and experimental for indigo and palygorskite (Maya Blue).

1972

L.A. Polette-Niewold et al. / Journal of Inorganic Biochemistry 101 (2007) 1958–1973

Fig. 23. Cerius2 structure of the indigo/palygorskite calculated surface.

structure that causes the color change from the starting color (red) to the stabilized color (purple) will be calculated from these data. 6.3. Indigo/Dehydroindigo DFT structure As stated above, the indigo/palygorskite system is more complicated due to the reaction of indigo with air to produce dehydroindigo. Shown in Fig. 22 is the theoretical and experimental UV–Vis spectra for indigo/palygorskite. Indigo is significantly different from thioindigo because of the proton attached to the nitrogen. The previously discussed experimental results show that during the synthesis of Maya Blue the proton is lost and the indigo becomes dehydroindigo. Dehydroindigo is analogous to thioindigo with the sulfur replaced by nitrogen. The binding model to palygorskite is essentially the same as that of thioindigo though stronger. Confirming this is the fact that the best fit to the experimental visible spectra is this model as seen in Fig. 22 (blue line). We can also see that a model with indigo binding through the carbonyl oxygen (dotted red line) adds adsorption on the low energy side of the spectra changing slightly the overall color. The presence of adsorption in this region would make the color greener. Fig. 23 shows one of the calculated surface structures for the indigo/palygorskite complexes. The interaction of the indigo molecules with the surface and with other indigo molecules can be seen. These calculations now allow us to study and control the color differences in various preparations of indigo palygorskite and to ascertain when excess indigo is present. 7. Conclusions We have shown in this paper through a series of experimental and theoretical studies that a series of novel organic/inorganic surface compounds exist. These materials were inspired by the ancient pigment Maya Blue but go well beyond this one material into a new world of novel organic/inorganic complex materials with widely varying physical and chemical properties.

The organic/inorganic complexes are produced by heating a mixture of the organic and palygorskite clay at moderate temperatures above 100 C for varying times. During this period the clay loses water and the organic binds to Al3+or other metals substituting for Si in the crystal lattice thereby creating Lewis acid sites. At low levels of organic this binding may occur in the channels of palygorskite, but at higher concentrations surface binding is required. When the molecule binds to the surface the organic molecule distorts and thus the color changes as observed. In the case of thioindigo the molecules bind to the Al3+through the oxygen. Indigo, on the other hand, partially converts to dehydroindigo during the heating process and therefore the Maya Blue complex is usually a mixture of indigo and thioindigo depending on the time, temperature and atmosphere of preparation. We continue to study the conditions or preparation of the organic/inorganic complexes described above to better define synthesis conditions and the resulting physical properties. Additionally, the work is being extended to include a variety of different molecules not described in this report but with the ability to produce new organic/inorganic complex materials with useful and novel properties. Acknowledgements Robert A. Welch Foundation supported by DOE (BES) ‘‘Gateway Program’’, NSF-ADVANCE Grant # 0245071, and this work. We are also grateful to Mayan Pigments Inc., for additional support. References [1] G. Stuart, Nat. Geogr. 192 (1997) 68–93. [2] M.J. Yacama´n, M.C.S. Puche, Mater. Res. Soc. Symp. 352 (1995) 3– 11. [3] R. Kleber, L. Masschelein-Kleiner, J. Thissen, Stud. Conservat. 12 (1967) 41–56. [4] H.E. Merwin, in: Yucatan, H.E. Morris, J. Charlot, A.A. Morris (Eds.), Temple Warriors at Chitzen Itza, 406, Carnegie Institution of Washington, Washington, DC, 1931, p. 356.

L.A. Polette-Niewold et al. / Journal of Inorganic Biochemistry 101 (2007) 1958–1973 [5] R.J. Gettens, Am. Antiquity 27 (1962) 557–564. [6] R.J. Gettens, G.L. Stout (Eds.), Paint Materials: A short Encyclopedia, D. van Nostrand, New York, NY, 1946. [7] H.V. Olphen, Science 154 (1966) 645–646. [8] R.R. Chianelli, L.A. Polette, US Patent, 7,052,541, May 30, 2006. [9] L.A. Polette, N. Ugarte, M.J. Yacama´n, R.R. Chianelli, Discover. Archaeolog. (2000) 46–53. [10] D. Magaloni, R. Newman, L. Banos, V. Castano, An Analysis of Mayan Painting Techniques at Bonampak, Chiapas, Mexico, Materials Research Society Proceedings, 1995. [11] E.R. Littman, Am. Antiquity 47 (1982) 404–408. [12] A.O. Shepard, Am. Antiquity 27 (1962) 565–566. [13] L.A. Polette, G. Meitzner, M.J. Jose´ Yacama´n, R.R. Chianelli, Microchem. J. 71 (2–3) (2002) 167–174. [14] L.A. Polette, Maya Blue: An Ancient Organic/Inorganic Complex Pigment, Ph’D thesis University of Texas at El Paso, 2002. [15] R.R. Chianelli, M. Perez De la Rosa, G. Meitzner, M. Siadati, G. Berhault, A. Mehta, P. Pople, S. Fuentes, G. Alonso, L.A. Polette, J. Synchrotron Radiat. (2005) 129–134. [16] W.F. Bradley, Am. Mineral. 25 (1940) 405–410. [17] J.E. Chisholm, Can. Mineral. 28 (1990) 329–339. [18] G.W. Brindley, G. Brown, Crystal structures of clay minerals and their X-ray identification, Spottiswoode Ballantyne Ltd., Colchester, London, 1980. [19] S. Henin, S. Caillere, Soil Components, in: J.E. Gieseking (Ed.), Springer-Verlag, New York, 1975, p. 2. [20] B. Velde, Clay and Clay Minerals in Natural and Synthetic Systems, Plenum, New York, 1977. [21] M.S. Augsburger, E. Strasser, E. Perino, R.C. Mercader, J.C. Pedregosa, J. Phys. Chem. Solids 59 (1998) 175–180. [22] M.E. Fernandez, J.A. Ascencio, D. Mendoza-Anaya, V. Rodriguez Lugo, M.J. Jose-Yacama´n, J. Mater. Sci. 34 (1999) 5243–5255. [23] W. Beck, C. Schmidt, M.S. Weinold, B. Wagner, Angew. Chem. Int. Ed. Engl. 28 (1989) 1529–1531.

1973

[24] A.R. Monahan, J.E. Kuder, J. Org. Chem. 37 (1972) 4182–4184. [25] A.M. Bond, F. Marken, E. Hill, R.G. Compton, H. Hugel, J. Chem. Soc. Perkin Trans. 2 (1997) 1735–1742. [26] H. Bauer, K. Kowski, H. Kuhn, W. Luttke, P. Rademacher, J. Mol. Struct. 445 (1998) 277–286. [27] G.M. Wyman, Chem. Commun. (1971) 1332–1334. [28] J. Balfour-Paul, Indigo in the Arab World, Biddles Limited, Guildford and King’s Lynn, London, 1978. [29] L. Kalb, Berichte der Deutschen Chemischen Gesellschaft 42 (1910) 3642–3652. [30] A. Dome´nech, M.T. Dome´nech-Carbo´, P. Va´zquez de Agredos, J. Phys. Chem. B 110 (2006) 6027–6039. [31] F.S. Manciu, L. Reza, L.A. Polette, B. Torres, R.R. Chianelli, J. Raman, Spectrosc., doi:10.1002/jrs.1751. [32] R. Vivaldi, H.S. Robertson, in: J.A. Gard (Ed.), The Electron-Optical Investigation of Clays, vol. 3, Mineralogical Society, London, 1971, pp. 255–273. [33] H. Shariatmadari, A.R. Mermut, M.B. Benke, Clay Clay Miner. (1999) 47–48. [34] M. Klessinger, W. Luttke, Tetrahedron Lett. 19 (2) (1963) 315–335. [35] R. Giustetto, F.X. Llabres i Ximena, G. Ricchiardi, S. Bordita Al. Damin, R. Gobetto, M.R. Chierotti, J. Phys. Chem. B 109 (2005) 19360–19368. [36] K. Witke, K.W. Brzinka, I. Lamprecht, J. Mol. Struct. 661 (2003) 235–238. [37] R.R. Chianelli, M. Perez De la Rosa, G. Meitzner, M. Siadati, G. Berhault, A. Mehta, J. Pople, S. Fuentes, G. Alonso, L.A. Polette, J. Synchrot. Radiat. (2005) 129–134. [38] W. Jost, Diffusion in Solids, Liquids, Gases, Academic Press, New York, 1960. [39] M.J. Gillan, Future Generation Comput. Syst. 10 (1994) 213. [40] K.B. Stark, J.M. Gallas, G.W. Zajac, M. Eisner, T.J. Golab, J. Phys. Chem. B 107 (13) (2003) 11558–11562.