Spectroscopic characterization of a sepiolite-based …tetide.geo.uniroma1.it/riviste/permin/testi/V79.SpecialIssue/2010... · sites, such as Chichén itzà, bonampak, Cacaxtla, El

abstRaCt - Maya blue, an ancient pigment used byMayas in Mesoamerica, can be considered an ancestorof modern nano-composites as it forms by grindingand heating the guest indigo molecule with a hostingclay framework (palygorskite or sepiolite). specificbonds form between the guest dye and the hostingmatrix, conferring the pigment an exceptional stabilityto both acid and alkali attacks. While most works havebeen focused on the study of the palygorskite/indigocomplex, this project is aimed to unveil the structuralfeatures of Maya blue when sepiolite is the hostingframework. a freshly-synthesized sepiolite-basedMaya blue was prepared accordingly to ancientMayas recipes and investigated with variousspectroscopic techniques (uv-visible, FtiR, Raman).all evidences show that by strongly crushing andheating (190°C in air or 150°C in vacuum) thesepiolite + indigo (2 wt%) composite, the dyeaggregates dissociate to monomers favouringdiffusion inside the clay tunnels emptied from looselybound zeolitic H2o. once encapsulated indigo partlydecomposes to dehydroindigo, its oxidized form,whose flexible molecule can easily diffuse inside thetunnels slightly changing the pigment colour (fromdeep blue to blue/green). applied heating, however,does not cause loss of Mg-coordinated oH2 whichforms H-bonds with the indigo reactive groups (C=oand n–H) after encapsulation, thus stabilizing the

composite. spectroscopic evidences supportingpresence of H-bonds are less evident in sepiolite-basedrather than in palygorskite-based composites withindigo. in fact while indigo is perfectly juxtaposed inthe narrower tunnels of palygorskite (6.4 x 3.7 Å),receiving H-bonds on both sides of the molecule, inthe wider tunnels of sepiolite (10.6 x 3.7 Å) itsmolecule has to get near to one border in order to forma single bond. in a sepiolite-based Maya blue pigmentthe global number of possible host/guest interactionsis therefore dramatically halved with respect to apalygorskite-based one, inevitably reducing resistanceto chemical attacks. Consequently, the extent of Mayablue renowned stability is directly related to thecomposition of its main clay constituents.

Riassunto - il blu Maya, un antico pigmento uti-lizzato dai Maya nell’america centrale, può essereconsiderato un antesignano dei moderni compostid’inclusione, prodotto della macinazione e riscalda-mento dell’indaco con un minerale argilloso (palygor-skite o sepiolite). specifici legami si formano tra lamolecola ospite e la matrice ospitante, conferendo alpigmento un’eccezionale stabilità chimica nei con-fronti di attacchi acidi e basici. dal momento che lamaggior parte dei lavori sin qui pubblicati è stata in-centrata sullo studio del complesso palygorskite/in-daco, il presente progetto è stato finalizzato alla

PerIodIco di MIneralogIa

established in 1930

Period. Mineral. (2010), Special Issue, 21-37 doi: 10.2451/2010PM0019http://go.to/permin

An International Journal of

MINERALOGY, CRYSTALLOGRAPHY, GEOCHEMISTRY,ORE DEPOSITS, PETROLOGY, VOLCANOLOGYand applied topics on Environment, Archeometry and Cultural Heritage

Spectroscopic characterization of a sepiolite-based Maya Blue pigment

RobERto GiustEtto*1, Kalaivani sEEnivasan2 and silvia boRdiGa2

1department of Mineralogical and Petrologic sciences and nis Centre of Excellence,università degli studi di torino via valperga Caluso 35, 10125 torino (italy),

2department of inorganic, Physical and Materials Chemistry, instM Centro di Riferimento and nis Centre of Excellence,università di torino, via Quarello 11, 10135 torino, italy

Submitted, September 2010 - Accepted, November 2010

*Corresponding author, E-mail: [email protected]

giustetto:periodico 30/11/2010 15:23 Pagina 21

22 R. GiustEtto, K. sEEnivasan and s. boRdiGa

descrizione delle caratteristiche strutturali del bluMaya quando il reticolo ospitante è costituito dalla se-piolite. un blu Maya di sola sepiolite è stato sintetiz-zato basandosi sulle antiche ricette Maya ecaratterizzato mediante svariate tecniche spettrosco-piche (uv-visibile, FtiR, Raman). le risultanzehanno dimostrato che macinando e scaldando (190°Cin aria o 150°C in vuoto) una miscela di sepiolite + in-daco (al 2% in peso), gli aggregati grossolani di colo-rante si dissociano in monomeri favorendone ladiffusione nei canali dell’argilla, svuotati dell’H2ozeolitica debolmente legata. l’indaco, una volta inca-psulato, in parte si trasforma in deidroindaco, la suaforma ossidata, la cui molecola più flessibile può fa-cilmente diffondere nei canali provocando un lieve vi-raggio di colore del pigmento (da blu scuro ablu/verdastro). il riscaldamento applicato, comunque,non provoca la perdita dell’oH2 coordinata al Mg, chea seguito dell’incapsulamento forma legami-H con igruppi funzionali dell’indaco (C=o e n–H), stabiliz-zando il composto. le evidenze spettroscopiche a ca-rico dell’esistenza di questi legami-H sono menopronunciate in un blu Maya di sola sepiolite rispettoad un analogo composto di palygorskite. Mentre l’in-daco, infatti, si incastra perfettamente nei canali piùstretti della palygorskite (6.4 x 3.7 Å), ricevendo le-gami-H su entrambi i lati della molecola, nei più ampicanali della sepiolite (10.6 x 3.7 Å) la sua molecoladeve avvicinarsi ad un bordo per formare un singololegame. in un blu Maya di sepiolite il numero globaledi possibili interazioni tra la molecola ospite e la ma-trice ospitante è perciò drasticamente dimezzato ri-spetto ad un analogo pigmento di palygorskite,riducendosi di conseguenza la resistenza agli aggres-sivi chimici. Pertanto l’effettiva entità della rinomatastabilità del blu Maya risulta essere direttamente cor-relabile alla composizione dei suoi principali costi-tuenti argillosi.

KEy WoRds: Maya Blue; sepiolite; indigo;spectroscopy; host/guest interaction.

intRoduCtion

Maya blue is a famous blue pigment used byancient Mayas in Pre-Columbian america,mainly in the yucatan peninsula (Mexico), fromvi to Xvi century a.d. Wonderful artefacts

decorated with Maya blue such as pottery, statuesand mural paintings (Fig. 1) can still be admirednowadays in several Mexican archaeologicalsites, such as Chichén itzà, bonampak, Cacaxtla,El tajìn and others.

Maya blue is the result of a combinationbetween a microporous clay mineral (palygorskiteand/or sepiolite, which ancient Mayas caved fromlocal outcrops) and the indigo dye (extracted fromthe leaves of the Indigofera Suffruticosa plant;Reyes-valerio, 1993). the pigment forms bystrongly crushing the clay with the dye andmoderately heating (120-190°C) the resultingblue mixture. Maya blue immediately attractedthe interest of the scientific Community due to itsastounding stability, as the pigment is virtuallyunaffected by both acids or alkali attacks andimmune to light exposition. in order to acquirestability, however, the clay/dye mixture must beadequately heated: an unheated mixture, thoughsimilar in aspect to Maya blue, lacks resistanceto chemical agents.

sepiolite and palygorskite are fibrous,colourless, phyllosilicate clay minerals. in theirstructure octahedral (o) and tetrahedral (t)sheets alternate along the x crystallographic axis:o sheets, however, are broken in ribbonselongated along z whereas t sheets, in order tomaintain their continuity, show a waving coursedue to a periodic inversion in the orientation ofapical oxygens, which point alternatively up anddown bonding to the o strips. Projection of thestructure on (001) can be described aschessboard-like disposition of tot units(Ferraris et al., 2004) (Fig. 2). this arrangementcauses micro-tunnels, filled by weakly-boundzeolitic H2o and exchangeable cations, to crossthe framework running along [001]. tightly-bound structural oH2, on the other hand,completes the coordination of cations at theedges of o ribbons, protruding in the tunnels. acomplicated web of H-bonds links zeolitic H2omolecules among them and to the Mg-coordinated oH2. both zeolitic H2o and


23Spectroscopic characterization of a sepiolite-based Maya Blue pigment

Fig. 1 - Mural painting from the archaeological site of bonampak (Mexico); background is painted with Maya blue pigment(taken from http://www.regiosfera.com/resuelto-el-misterio-del-azul-maya/).

Fig. 2 - the crystal structure of sepiolite projected on (001) (fractional coordinates by Post et al., 2007). unit cell is outlinedwith continuous lines.


structural oH2 can leave the structure byprogressively heating the clay, but while theformer is lost at moderate temperatures (120-200°C), higher ones are requested for the latter(> 300°C).

Main difference between sepiolite andpalygorskite stands in the dimensions of themicro-tunnels, which are larger in the former(10.6 x 3.7 Å) and narrower in the latter (6.4 x3.7 Å). as a consequence, each tot ribboncounts eight o sites in sepiolite whereas only five(central one usually empty) appear inpalygorskite. While this last can be considered anintermediate di/tri-octahedral phyllosilicate, dueto contextual presence of Mg and al, sepiolite isexclusively tri-octahedral [ideal crystal-chemical

formula: Mg8si12o30(oH)4(oH2)4·8H2o]. due totheir peculiar structural features, both sepioliteand palygorskite were investigated with regard totheir ability to exchange cations and absorb smallorganic molecules in the clay micro-tunnels or onthe fibres surface (akyuz and akyuz, 2004;Polette-niewold et al., 2007; Giustetto et al.,2006) and as support for catalysts (occelli et al.,1992; suarez et al., 2008; Zheng et al., 2010).

indigo is a well-known blue dye, both naturaland synthetic, whose molecule (Fig. 3) containsa cross-conjugated chromophore composed of acentral C=C bond carrying two donor (amineunits: n-H) and two acceptor groups (carbonylunits: C=o). distribution of electric charge in thebasic chromophore is strongly influenced by


Fig. 3 - the molecule of indigo, with indication of the donor and acceptor groups (a); importance of inter-molecular forces(dashed lines) in dimer formation (b).


inter-molecular forces between reactive groupsof separate molecules, responsible for aggregatesformation (dimers or crystalline indigo). in spiteof the brilliant colour of Maya blue, the amountof indigo expected to react with the activatedclay during the pigment synthesis is rather low,not exceeding 2 wt% (Kleber et al., 1967; Chiariet al., 2003; Giustetto et al., 2005; Giustetto etal., 2006, Chiari et al., 2008).

according to the most quoted hypothesis, inthe synthesis of Maya blue indigo is adsorbedwithin the micro-tunnels as a result of theheating, which causes loosely bound zeoliticH2o to leave the clay thus favoring dyediffusion. in addition, specific host/guestinteractions are likely to form between the clayand indigo, thus stabilizing the dye inside theclay structure and conferring the pigment itsrenowned stability. Maya blue has therefore tobe considered a forerunner of modernnanocomposite materials as encapsulation withinthe pores guarantees shielding from externalenvironment, while formation of host/guestinteractions assures permanence in situ of thecromophore. the real nature of such bonds,however, is still disputed, whether H-bondsbetween structural oH2 and indigo reactivegroups (Chiari et al., 2003; Fois et al., 2003;Reinen et al., 2004; Giustetto et al., 2005;Giustetto et al., 2006) or direct metal-oxygenbonds between octahedral cations at the edges ofo strips and the dye carbonyl group (Polette-niewold et al., 2007; Manciu et al., 2008;tilocca and Fois, 2009).

several analytical methods have been used sofar to investigate the structural features of Mayablue and the sorption of indigo in the hostingclay framework. Most studies, however, focusedon the interaction between indigo andpalygorskite. basing on such premises andfollowing-up the in deep structural studiesperformed on the palygorskite-based pigment,this work is aimed to investigate the dyeenvironment and the nature of the chemical

bonds when sepiolite is the hosting framework.this in order to check whether and how presenceof wider tunnels may influence both the natureand/or the strength of the host/guest interactions,affecting the composite stability. as inpalygorskite, the guest indigo molecule issupposed to enter the wider sepiolite nano-tunnels due to consistent H2o loss during heating(Hubbard et al., 2003; ovarlez et al., 2006;domenech et al., 2009; Giulieri et al., 2009).Presence of larger tunnels (10.6 x 3.7 Å) shouldfacilitate diffusion.

Experimental conditions adopted in the currentstudy were modeled, as far as possible, on theancient Mayas procedures, in order to re-create acomposite analogue to original Maya bluestarting from pure precursors. it is generallyreputed that moderate temperatures were reachedduring the pigment preparation, as heatingbetween 100 and 200°C allow formation of astable compound (sanchez del Rio et al., 2009).it cannot be excluded, however, that more severethermal treatments (> 300°C) may affect thenature and strength of the host/guest interactions,implying formation of similar although alternativesepiolite + indigo adducts, characterized bypeculiar structural features (Giulieri et al., 2009).

MatERials and EXPERiMEntal

the perspective of studying archaeologicalMaya blue specimens, although appealing, wasimmediately discarded as such samples alwaysinclude traces of the below substratum (i.e. plasteror mortar), lacking the requested quality standards.besides, authentic Maya blue usually contains amixture of palygorskite or sepiolite, although otherclays (i.e. illite and montmorillonite) may bepresent too. For all these reasons, the current studywas performed on freshly-synthesized sepiolite +indigo (2 wt%) composites, a procedure alreadyadopted in past works (Chiari et al., 2003;Hubbard et al., 2003; Giustetto et al., 2006).



natural sepiolite and synthetic indigo wereprovided by sigma-aldrich (cod. nos. 70253 and56980 respectively).

in order to prepare freshly-synthesizedsepiolite-based Maya blue specimens, the well-established procedure codified in literature forthe laboratory preparation of Maya blue wasadopted (van olphen, 1966; sanchez del Rìo etal., 2006; Polette-niewold et al., 2007). suchprocedure mainly consists of three phases: i)mixing and dry grinding sepiolite with syntheticindigo (2 wt.%); ii) heating up to 190°C for 22hours; iii) purification of the powders throughsoxhlet extraction in CHCl3, in order to removethe exceeding dye unfixed to the hosting matrix.a different procedure was adopted for samplesanalyzed by means of iR techniques, for whichsepiolite was previously activated (heating invacuum at 120°C, removing most zeolitic H2o),ground under glove-box in ar atmosphere with2 wt% indigo and further heated in vacuum at150°C (dehydration effects analogous to heatingin air at 190°C).

uv-visible spectra in transmittance mode werecollected using a varian Cary 300 biospectrophotometer in the 800-200 nm range onan indigo CHCl3 solution. uv-visible-niRspectra in diffuse reflectance mode were collectedusing a varian Cary 5000 spectrophotometer inthe 2500-200 nm range, on samples previouslydiluted with baso4 (1:3/1:20).

infrared (iR) absorption spectra werecollected, under controlled atmosphere, on aFtiR bruker iFs 66 spectrometer, with aresolution of 2 cm-1 and collecting 64 scans foreach spectrum. all samples, in the form of thinfilms supported on si plates, were outgassedunder vacuum at 150° C for 1.5 hours in order toeliminate zeolitic H2o.

Raman spectra were collected on a Renishawmicro-Raman system 1000 on powder samples,excited with a He-Cd green laser beam operatingat 532 nm in the range 1100-1800 cm-1,addressed with an optical microscope. the

scattered signal was collected on a CCd detectorlocated at 180° with respect to the incident beam.

REsults and disCussion

UV-Vis spectroscopy

uv-visible spectra were collected both on pureindigo and on the sepiolite + indigo (2 wt%)adduct in all different stages of the synthesisprocedure (Fig. 4).

the absorption spectrum of indigo in a CHCl3

solution (curve a) shows a well-definedmaximum at 603 nm, related to the p⇒p*transition. the deep minimum between 500 and370 nm is responsible for the deep blue-colourof the solution and consistent with the presenceof a double cross-conjugated system formed byboth donor and acceptor groups (Klessinger andlüttke, 1963).

the isolated indigo molecule (gaseous) isknown to have its lower energy band at 540 nm.variations in the location of absorption maximaare related to the particular chemicalenvironment of the dye molecule, either insolution and in the solid-state (Reichardt, 2003).Perturbations due to interactions cause sensibleshifts towards lower energies, whose magnitudeis related to both their strength and number. assuch, position of the absorption band in theindigo CHCl3 solution (603 nm) is related to thefairly weak connections - if any - between thedye molecules and a non-polar solvent.

the diffuse reflectance spectrum of solid-stateindigo dispersed in an inert matrix (baso4: curve1) shows a significant shift of the absorptionmaximum towards higher wavelengths, reaching660 nm (weak shoulder at 554 nm). such shiftimplies dye molecules in solid-state are stronglylinked one another by means of inter-molecularhydrogen bonds, responsible for aggregateformation (dimers and/or crystalline indigo) andabsorption of lower energy light. analogouseffect exerted by polar solvents (such as ethanol



and acetone) appears to be weaker (Reinen et al.,2003).

a similar outcome occurs when indigo issolvated in the host matrices of microporous clayminerals. a mixed and ground sepiolite + indigo(2 wt%) composite (curve 2) shows anabsorption maximum at 648 nm, related to thesame p⇒p* transitions, slightly shifted (-12 nm)with respect to pristine indigo. Presence of suchshift implies the dye molecules do not formaggregates when solvated on sepiolite, but ratherspecific interactions with the hosting matrix,whose nature and strength are different from

inter-molecular bonds occurring among indigomolecules. it is remarkable how suchinteractions tend to form already during grinding(curve 2) and progressively stabilize afterheating and soxhlet extraction (curves 3 and 4),as confirmed by the further slight shift of theabsorption maximum.

Careful examination of the mixed and groundcomposite spectrum (curve 2) suggests presenceof an additional broad shoulder at 603 nm, aposition typical of the quasi-isolated indigomolecule, which is consistent with resultsobtained on analogous palygorskite-based


Fig. 4 - uv-visible absorption spectrum of indigo in CHCl3 solution (a) and diffuse reflectance spectra of solid-state indigo(1) and sepiolite + indigo (2 wt%) composite after crushing (2), heating (3) and soxhlet extraction in CHCl3 (4). spectra wereshifted on y axis for sake of clarity.


composites by Reinen et al. (2003). Presence ofsuch shoulder might be explained as biggerindigo particles, when strongly crushed to thehosting matrix, tend to dissociate in smaller unitscausing progressive disappearance of inter-molecular bonds and formation of indigomonomers which, as such, might easily diffuseinside the sepiolite tunnels. subsequent heating(curve 3) and soxhlet extraction in CHCl3 (curve4) cause progressive disappearance of suchshoulder, as indigo monomers link to the claymatrix due to formation and stabilization ofhost/guest interactions.

the progressive decrease of absorption in the550-600 nm region in subsequent stages of thesynthesis procedure causes the colour of theresulting composite to vary from deep togreenish-blue.

FTIR spectroscopy

Ft-iR spectra were collected both on isolateprecursors (sepiolite; indigo) and on the sepiolite+ indigo (2 wt%) adduct while heating thepowders in vacuum at 150°C, in order to observevariations of the iR-active vibrational modes inthe exact moment of the pigment synthesis.

in the hydroxyl stretching region (2800-3800cm-1) of the iR spectrum of sepiolite evacuatedat 150°C (Fig. 5, curve 1) two separate doubletscan be observed at high wavenumbers. the firstone (3738 and 3726 cm-1) is related to superficialsilanols whereas the second (3691 and 3680 cm-

1) to hydroxyls in the octahedral sheet (Mg-oH).these last appear to be moderately perturbed,presumably as the result of a limited loss ofcoordinated oH2, causing partial transformationto sepiolite di-hydrated (Hayashi et al., 1969;Giulieri et al., 2009). Most relevant features,however, are two broad maxima located at 3600and 3531 cm-1. such bands can be observed onlyafter heating, as at room temperature (not shownhere) they are overlapped by the broad signalsrelated to physisorbed and zeolitic H2o, and can

be related to the anti-symmetric and symmetricstretching frequencies of the tightly-boundcoordinated oH2 (Cannings, 1968; Frost et al.,2001). Most coordinated oH2 is therefore stilltightly bound to the sepiolite framework at theadopted heating experimental condition (150°Cin vacuum).

addition of indigo to sepiolite (Fig. 5, curve2) modifies selected vibrational modes related tothe hosting framework. in addition, indigo bandsare easily identifiable in the composite spectra inspite of its low amount (2 wt%). doublet relatedto superficial silanols (3738 and 3726 cm-1) is notaltered by indigo addition, implying interactionwith the dye does not occur on the fibres surface.severe intensity decrease of the 3691 cm-1

maximum, on the other hand, suggests fixationof indigo increases the stability of the hostingstructure, preventing folding due to oH2 loss(Giulieri et al., 2009). as already seen for puresepiolite, also in the pigment the applied heatingcauses structural oH2 n(oH) maxima to appearin the very same positions (3600 and 3531 cm-1),implying presence of indigo does not favour butprevent loss of oH2. additionally, a new broadfeature appears at 3408 cm-1, which was notvisible in the spectrum of sepiolite. Most feasibleexplanation is that such feature accounts forperturbation of a certain amount of structuraloH2, possibly involved in H-bond formation. aszeolitic H2o is known to have abandoned theclay structure due to applied heating, it is likelythat such bonds may be formed with the guestindigo molecule, aptly diffused inside thetunnels. the scarce dye quantity (2 wt%) impliesnot all structural oH2 can form H-bonds withindigo, thus explaining persistence of the 3600and 3531 cm-1 maxima related to unperturbedhydroxyls.

the only band directly related to presence ofindigo in this region is a broad maximumappearing at 3212 cm-1, related to the dye n(n-H). such band is significantly red-shifted withrespect to the one appearing in pristine indigo



(3270 cm-1), suggesting the amine group issomehow involved in formation of host/guestinteractions with the inorganic matrix.

in the hydroxyl bending region (1150-1800cm-1; Fig. 6) a major broad feature appears at1615 cm-1 in the spectrum of pristine sepioliteevacuated at 150°C (curve 1). Emergence ofsuch band is due to loss at previous temperaturessteps of loosely bound H2o and has to be relatedto structural oH2 d(oH) (Hayashi et al., 1969;Frost et al., 2001). once again, addition of indigoto sepiolite and heating at the same level causesan analogous band to appear (curve 2): carefulexamination, however, reveals non-symmetry of

such feature if compared to the one typical ofpure sepiolite, as at least two weak but definiteshoulders appear at higher wavenumbers (1623and 1640 cm-1 respectively). Presence of suchshoulders is therefore an indirect evidence of thedye addition, being related with all dueprobability to perturbations affecting a fractionof oH2 molecules. again, magnitude of observedshifts suggests possible formation of H-bondsbetween oH2 and the guest indigo molecule.

the main iR-active vibrational modes of pureindigo (Fig. 6, curve a) appear in the 1000-1650cm-1 interval, in a region relatively free from clayrelated maxima. as stated above, in solid-state


Fig. 5 - iR spectra in the hydroxyl stretching region (2800-3800 cm-1) of pristine sepiolite (1) and sepiolite + indigo (2 wt%)adduct (2) outgassed in vacuum at 150°C. spectra were shifted on y axis for sake of clarity.


R. GiustEtto, K. sEEnivasan and s. boRdiGa30

indigo molecules in trans-form are bound oneanother through intra- and inter-molecularbonds, formed by reactive groups (n-H/C=o)within a single molecule or between differentmolecules respectively (Fig. 3). the anti-symmetric stretch of the C=o group forms adoublet at 1626 and 1614 cm-1, whereas twoseparate features (shoulder at 1412 and band at1393 cm-1) are related to in-plane d(n-H), thefirst related to intra-molecular and the second tointer-molecular bonds.

addition of indigo to the clay causes selectedshifts and intensity variations to occur in thevibrational modes attributed to the dye, probable

consequences of host/guest bond formation andstrengthening. unfortunately no considerationcan be made about the behavior of the C=ogroup, as the related features are hidden by thestronger maximum related to oH2 bending.luckily no overlapping affects n–H groupfeatures, whose band related to inter-molecularbonding (1393 cm-1) disappears in the heatedsepiolite + indigo composite (Fig. 6, curve 2).such evidence is consistent with uv-visiblespectroscopy results and confirms that duringpigment synthesis dissociation of indigoaggregates occurs, forming isolated monomerswhich, as such, can easily diffuse inside the host

Fig. 6 - iR spectra in the hydroxyl bending region (1150-1800 cm-1) of pure indigo (a), pristine sepiolite (1) and sepiolite +indigo (2 wt%) adduct (2) outgassed in vacuum at 150°C. spectra were shifted on y axis for sake of clarity.


Spectroscopic characterization of a sepiolite-based Maya Blue pigment 31

micro-tunnels. breaking of inter-molecularbonds, besides, allows indigo reactive groups (n-H/C=o) to interact with the clay framework andestablish new bonds. Feature related to intra-molecular bonds (1412 cm-1), alternatively,shows moderate shift towards higherwavenumbers (reaching 1420 cm-1) and intensityincrease. Previous experiences (ovarlez et al.,2006) proved that such interactions are due tosplit only at temperatures sensibly higher thanthose applied (> 300°C in air).

in the pigment spectrum (curve 2), however,several vibrational modes related to the moleculebenzene rings ν(C-C) significantly vary, as perposition and/or intensity, with respect to pureindigo (curve a). such variations can be resumedas follows: i) disappearance of the 1586 cm-1

feature; ii) growth of the 1569 and 1552 cm-1

bands; iii) splitting of the 1483 cm-1 maximum,with appearance of a new band at 1500 cm-1; iv)shift towards higher wavenumbers of the 1317cm-1 band, which reaches 1332 cm-1. all thesevariations can tentatively be related to interveneddeformations of the molecule due toencapsulation in the hosting matrix. loss of bothplanarity and centre-symmetric geometry may beexplained by partial transformation of indigo toits oxidized form, dehydroindigo. Presence ofdehydroindigo in Maya blue was reported byseveral authors (sanchez del Rio et al., 2009;domenech et al., 2009) and is reputed to favorthe guest diffusion within the tunnels, as thecentral double bond (C=C) in the chromophoreis replaced by a single one, increasing flexibilityof the molecule. in addition, presence ofdehydroindigo is consistent with a slight colorchange of the pigment from deep blue to blue-green, as already detected by uv-visspectroscopy. Weak new maxima at 1700 and1374 cm-1 (curve 2) have also to be related topossible thermal decomposition of indigo andpartial transformation to dehydroindigo (Giulieriet al., 2009).

Raman spectroscopy

Raman spectra were collected on pure indigo,on a sepiolite + indigo (2 wt%) ground mixtureand on a sepiolite-based Maya blue pigmentafter heating and soxhlet extraction in CHCl3

(Fig. 7).no spectrum could be collected on pristine

sepiolite due to strong fluorescence effects, asalready reported in previous works (McKeownand Post, 2002). straight comparison betweencollected and literature data (Witke et al., 2003;leona et al., 2004; Giustetto et al., 2005,vandenabeele et al., 2005) shows severalanalogies, though interpretation is complicatedby resonance effects. intensity of Raman signalis very sensitive to the excitation wavelength andvariations sometimes depend from particularexperimental conditions rather than differentchemical surroundings of the dye molecule.

assignment of Raman-active vibrationalmodes of indigo (curve a in Fig. 7) was based onthe results of tatsch and schrader (1995). allfeatures appearing in the spectra of the sepiolite-based composites with indigo (curves b and c)are related only to the dye molecule, being theclay contribution irrelevant.

Most significant considerations emerge bycomparing the pristine indigo (curve a) andsepiolite-based Maya blue (curve c) spectra. theground but unheated sepiolite + indigo (2 wt%)mixture (curve b) represents an intermediate butfundamental step, as it is commonly acceptedthat host/guest interactions start to form evenduring crushing (Hubbard et al., 2003; sanchezdel Rìo et al., 2006).

Most relevant considerations can besummarized as follows: i) feature related to thesymmetric ν(C=o) gradually red-shifts passingfrom 1701 (pristine indigo; curve a) to 1695(ground sepiolite + indigo mixture; curve b) andthen to 1690 cm-1 (Maya blue; curve c). theobserved red-shift (-11 cm-1) is consistent,though slightly less marked, with the theoretical



value computed by means of ab-initio methodsfor a C=o perturbed by a H-bond (-15 cm-1;Giustetto et al., 2005). an analogous shift, albeitmore evident, was observed for similarpalygorskite/indigo adducts (Witke et al., 2003;leona et al., 2004; Giustetto et al., 2005;vandenabeele et al., 2005). Raman spectroscopyexhibits therefore direct evidence aboutinvolvement of indigo carbonyl group information of host/guest interactions, being suchgroup the effective acceptor of H-bonds donatedby the nearest clay structural oH2, as shown byFtiR data. ii) maximum at 1365 cm-1, related toδ(n-H) coupled with δ(C-H) modes, showssensible intensity decrease and partial shift

towards higher wavenumbers passing fromindigo (curve a) to the ground mixture (curve b)and finally to Maya blue (curve c), where a newcomponent appears at 1378 cm-1. such evolutionimplies the indigo amine unit, as well as the C=ogroup, is perturbed and involved in formation ofmutual clay/dye interactions. an analogousvariation was described by leona et al. (2004),though here it is exalted by resonance effects(Giustetto et al., 2005). iii) more or lesssignificant shifts and intensity variations affectselected maxima passing from pure indigo to thesepiolite-based pigment [significant increase andshift towards higher wavenumbers of the 1631cm-1 band of indigo (curve a), related to

Fig. 7 - Raman spectra of pure indigo (a) and sepiolite + indigo (2 wt%) adduct after crushing (b) and heating/soxhlet extractionin CHCl3 (c). spectra were shifted on y axis for sake of clarity.



collective n(C-C) modes, which reaches 1640cm-1 in Maya blue (curve c); intensity decreaseand shift towards higher wavenumbers ofmaxima at 1575 and 1365 cm-1 (curve a), relatedto collective n(C-C) modes and in-plane d(n-H)coupled with C-H respectively, which graduallymove to 1590 and 1378 cm-1 in the pigment afterheating (curve c)]. such variations are consistentwith previous literature data (Witke et al., 2003)and can be explained by intervened activation ofRaman-inactive vibration modes of the bu type,due to loss of planarity of the indigo moleculeafter adsorption in the hosting sepiolite matrix.once again, such behavior might be explainedby partial transformation of indigo todehydroindigo, a phenomenon reported byseveral authors (Giustetto et al., 2006;domenech et al., 2006; 2007; 2009). iv) at lowerfrequencies (not shown in Fig. 7), the maximumrelated to vibrations in the five-member ringshifts towards higher wavenumbers, passingfrom 549 (pure indigo) to 553 cm-1 (sepiolite-based Maya blue). such variation, observed alsoby vandenabeele et al. (2005), is related tocombination of charge and symmetry effects (ascommented by Coupry et al., 1997, studyingdyes spectra fixed on cotton fibers) and again canbe justified by activation of Raman-forbiddenmodes caused by modifications in the planarityof the indigo molecule, due to partialtransformation to dehydroindigo.

ConClusions

a freshly-synthesized sepiolite-based Mayablue pigment was characterized by means ofvarious spectroscopic techniques (uv-visible,FtiR and Raman), in order to study theadsorption process of indigo inside the claytunnels and the nature of the host/guestinteractions responsible for the pigment stability.

Collected evidences proved that at the adoptedexperimental conditions (heating at 190°C in airor 150°C in vacuum), analogous to those adopted

by the ancient Mayas while preparing theoriginal pigment, the tightly-bound Mg-coordinated oH2 does not leave the sepiolitestructure, being the related vibrational modeswell visible in the iR spectra. addition of indigoto the clay, besides, perturbs such featuressuggesting direct involvement of oH2 in theformation of mutual bonds between the clay andthe dye.

Fixation on the hosting sepiolite matrix causesindigo aggregates (dimers and crystalline indigo)to dissociate, due to progressive disappearance ofinter-molecular bonds favoured by strongcrushing and moderate heating of the clay/dyemixture. dissociation of aggregates impliesdiffusion inside the clay tunnels of indigomonomers, whose functional groups (C=o; n–H)become available to react with the clay matrixgiving way to formation of host/guest interactions.

diffusion within sepiolite tunnels is furtherencouraged by partial thermal decomposition ofindigo to its oxidized form - dehydroindigo,evidenced by intervened deformation and loss ofplanarity of the guest molecule after heating. themore flexible dehydroindigo molecule favoursdeeper encapsulation inside the sepiolite poresand explains minor colour changes (from deepblue to blue/greenish) observed during thepigment preparation.

all spectroscopic evidences show that theencapsulated indigo molecules are stabilizedwithin the sepiolite tunnels by specific clay/dyeinteractions, presumed responsible for thecomposite stability. such interactions, consideringthe reactive groups involved and the magnitudeof the related vibrational shifts, have beenidentified as H-bonds formed between thesepiolite structural oH2 and the indigo acceptorand donor groups (C=o; n–H). this kind ofinteraction, albeit usually considered weak, whenformed within narrow micropores can be toughto reach and therefore difficult to break. Whilethe C=o groups receive H-bonds from thenearest oH2 molecules protruding in the tunnels,



it is still doubtful whether the n-H groups maydonate analogous bonds to the same oH2.

no evidence was found about existence ofdirect carbonyl/cation (C=o···Mg) interactions.it cannot be excluded, however, that such bondsmay form if higher temperatures are reachedduring heating (> 300°C), favoured by loss ofcoordinated oH2.

Presence of H-bonds in a sepiolite-based Mayablue pigment is consistent with previous studiescentered on palygorskite-based analogues (Foiset al., 2003; Reinen et al., 2004; Giustetto et al.,2005; Giustetto et al., 2006; Chiari et al., 2008).

in spite of this, intensities of the relatedvibrational modes are severely weakened whensepiolite is the hosting framework, as if the globalnumber and/or strength of these interactions couldbe diminished if compared to a palygorskite-based analogue. in order to justify such evidence,details of indigo encapsulation in the sepioliteframework need to be further discussed.

While in the narrower palygorskite pores (6.4x 3.7 Å) indigo can be exactly juxtaposed in themiddle of each tunnel, receiving H-bonds fromcoordinated oH2 on both sides of the molecule,in the larger tunnels of sepiolite (10.6 x 3.7 Å)

Fig. 8 - Encapsulation of indigo molecule inside a sepiolite tunnel; dashed lines indicate H-bonding between the indigo C=ogroup and the clay oH2. Zeolitic H2o not shown for sake of clarity.



the dye molecule must get close to one tunnelborder in order to adequately interact with theclay framework (Fig. 8). only one side of theindigo molecule, namely the one facing theborder of the tunnel, can therefore anchor itselfto the clay framework, whereas the opposite sidefacing the tunnel lumen is bond-free. suchsituation cannot be modified by adding anothermolecule next to the first, as the dye stericimpediment is such that two molecules cannotbe paired side by side in the same tunnel portion.

the described accommodation for indigo insepiolite causes the number of host/guestinteractions possibly existing in a sepiolite-basedMaya blue to be drastically halved with respectto a palygorskite-based one. this severereduction in the global number of H-bonds,considered on equivalent indigo amounts (2wt%), justifies the observed intensity decrease ofthe spectroscopic features, as a minor fraction ofboth coordinated oH2 and dye reactive groupsare involved in formation of interactions.

a direct and macroscopic consequence of suchsituation, coupled with the more enhancedfragility of sepiolite structure compared topalygorskite, is that a palygorskite-based Mayablue pigment is sensibly more resistant tochemical agents than its sepiolite-basedanalogue, as performed stability tests seem tocorroborate (sanchez del Rio et al., 2006). therenowned stability of Maya blue pigment cantherefore be directly related to the compositionof its clay constituent: samples mainly formedby palygorskite are more resistant than analogueswhere sepiolite is the predominant clay.

aCKnoWlEdGEMEnts

the authors wish to thank Gabriele Ricchiardi,simona Ferrando and serena Corallini for theirprecious help in collecting spectroscopic data.

We would also like to thank the referees of thispaper for their succinct but precious advices.

special thanks go to Giacomo Chiari, who firstintroduced us to the marvelous world of Maya blue,

and Giovanni Ferraris, who helped supporting thisstudy.

REFEREnCEs

aKyuZ a. and aKyuZ t. (2004) - Study on theinteraction of nicotinamide with sepiolite, loughliniteand palygorskite by IR spectroscopy. J. Molec.struct., 744, 1-3, 47-52.

CanninGs F.R. (1968). An infrared study of hydroxylgroups on sepiolite. J. Phys. Chem., 72, 1072-1074.

CHiaRi G, GiustEtto R and RiCCHiaRdi G. (2003) -Crystal structure refinements of palygorskite andMaya Blue from molecular modelling and powdersynchrotron diffraction. Eur. J. Mineral., 15, 21-33.

CHiaRi G., GiustEtto R., dRuZiK J., doHEnE E. andRiCCHiaRdi G. (2008) - Pre-columbian nano -technology: reconciling the mysteries of the MayaBlue pigment. appl. Phys. a, 90, 3-7.

CouPRy C., saGon G. and GoRGuEt-ballEstERos P.(1997) - Raman spectrocopic investigation of bluecontemporary textiles. J. Raman spectr., 28, 85-89.

doMénECH a., doMénECH-CaRbò M.t. and vàsQuEZ

dE aGREdos PasCual M.l. (2006) -Dehydroindigo: a new piece into the Maya Bluepuzzle from the voltammetry of microparticlesapproach. J. Phys. Chem. b, 110, 6027-6039.

doMénECH a., doMénECH-CaRbò M.t. and vàsQuEZ

dE aGREdos PasCual M.l. (2007) - Indigo/dehydroindigo/ palygorskite complex in Maya Blue:an electrochemical approach. J. Phys. Chem. C,111, 4585-4595.

doMénECH a., doMénECH-CaRbò M.t., sanCHEZ dEl

Rio M., vaZQuEZ dE aGREdos PasCual M.l. andliMa E. (2009) - Maya Blue as a nanostructuredpolyfunctional hybrid organic-inorganic material:the need to change paradigms. new J. Chem., 33,2371-2379.

FERRaRis G., MaKoviCKy E. and MERlino s. (2004) -Crystallography of Modular Materials, iuCr/oxford university Press, oxford.

Fois E., GaMba a. and tiloCCa a. (2003) - On theunusual stability of Maya Blue paint: moleculardynamics simulations. Microp. Mesopor. Mater.,57, 263-272.

FRost R.l., loCos o.b., Ruan H. and KloPRoGGE

J.t. Near-infrared and mid-infrared spectroscopicstudy of sepiolites and palygorskites. vibrationalspectroscopy 2001, 27, 1-13.



GiuliERi F., ovaRlEZ s., CHaZE a.M., dElaMaRE F.,Raya J. and HiRsCHinGER J. (2009) - Theincorporation of indigo molecules in sepiolitechannels. Chem. Eur. Jour., 15, 11326-11332.

GiustEtto R., llabREs i XaMEna F., RiCCHiaRdi G.,boRdiGa s., daMin a., CHiERotti M. and GobEtto

R. (2005) - Maya Blue: a computational andspectroscopic study. J. Phys. Chem. B, 109 (41),19360-19368.

GiustEtto R., lEvy d. and CHiaRi G. (2006) - Crystalstructure refinement of Maya Blue pigmentprepared with deuterated indigo, using neutronpowder diffraction. Eur. J. Mineral., 18, 629-640.

HayasHi H., otsuKa R. and iMai n. (1969) - Infraredstudy of sepiolite and palygorskite on heating. am.Mineral., 53, 1613-1624.

HubbaRd b., WEnXinG K., MosER a., FaCEy G.a. anddEtElliER C. (2003) - Structural study of MayaBlue: textural, thermal and solid-state multinuclearmagnetic resonance characterization of thepalygorskite-indigo and sepiolite-indigo adducts.Clays and Clay Minerals, 51, 3, 318-326.

KlEbER R., MassCHElEin-KlEinER l. and tHissEn J.(1967) - Étude et identification du Bleu Maya.studies in Conservation, 12, 41-55.

KlEssinGER M. and lüttKE W. (1966) - Theoretischeund spektroskopische Untersuchungen anIndigofarbstoffen. Part 3, Der Einflußzwischenmolekularer Waserstoffbrücken auf dieSpektren von Indigo im festen Zustand. Chemischeberichte, 99, 2136-45.

lEona M., Casadio F., baCCi M. and PiCollo M.(2004) - Identification of the Pre-Columbianpigment Maya Blue on works of art by non invasiveUV-vis and Raman spectroscopic techniques. J.amer. insti. Conserv., 43, 1, 39-54.

ManCiu F.s., RaMiREZ a., duRRER W., Govani J. andCHianElli R.R. (2008) - Spectroscopic analysis ofa dye-mineral composite - a Raman and FT-IRstudy. J. Raman spectr., 39, 1257-1261.

MCKEoWn d.a. and Post J.E. (2002) - Vibrationalanalysis of palygorskite and sepiolite. Clays andClay Minerals, 5, 667-680.

oCCElli M.l., MaXWEll R.s. and ECKERt H. (1992)- The interaction of vanadia with sepiolite.Structural studies by 51V Solid-State NMR andRaman spectroscopy. J. Catal., 137, 36-50.

ovaRlEZ s., CHaZE a.M., GiuliERi F. and dElaMaRE

F. (2006) - Indigo chemisorption in sepiolite.

Application to Maya Blue formation. ComptesRendu Chimie, 9, 1243-1248.

PolEttE-niEWold l.a., ManCiu F.s., toRREs b.,alvaRado JR M. and CHianElli R.R. (2007) -Organic/inorganic complex pigments: Ancientcolors Maya Blue. J. inorg. biochem., 101, 1958-1973.

Post J.E., bisH d.l. and HEanEy P.J. (2007) -Synchrotron powder X-ray diffraction study of thestructure and dehydration behavior of sepiolite.am. Mineral., 92, 91-97.

REiCHaRdt C. (2003) - solvents and solvent effects inorganic chemistry. 3rd Ed., Wyley-vCH, Weinheim,chapter 6.

REinEn d., KöHl P. and MullER C. (2004) - Colourcentres in ‘Maya Blue’ - the incorporation oforganic pigment molecules into the palygorskitelattice. Zeitsch. anorg. allgem. Chemie, 630, 97-103.

REyEs-valERio C. (1993) - De Bonampak al TemploMayor: el Azul Maya en Mesoamerica, siglo XXiEditores.

sanCHEZ dEl Rìo M., MaRtinEtto P., REyEs-valERio

C., dooRyHéE E. and suáREZ M. (2006) - Synthesisand acid resistance of Maya Blue pigment.archaeometry, 48, 1, 115-130.

sanCHEZ dEl Rio M., boCCalERi E.G.d., Gionis v.,KaCandEs G.H., suaREZ M. and GaRCia-RoMERo

E. (2009) - A combined synchrotron powderdiffraction and vibrational study of the thermaltreatment of palygorskite-indigo to produce MayaBlue. J. Mater. sci., 44, 5524-5536.

suaREZ s., CoRonado J.M., PoRtEla R., CaRlos

MaRtin J., yatEs M., avila P. and sanCHEZ b.(2008) - On the preparation of TiO2 - Sepiolitehyhrid materials for the photocatalytic degradationof TCE: Influence of TiO2 distribution in themineralization. Env. sci. techn., 42, 16, 5892-5896.

tatsCH E. and sCHRadER b. (1995) - Near-InfraredFourier transform Raman spectroscopy ofindigoids. J. Raman spectr., 26, 467-473.

tiloCCa a. and Fois E. (2009) - The color andstability of Maya Blue: TDDFT calculations. J.Phys. Chem. C, 113, 8683-8687.

van olPHEn H. (1966) - Maya Blue: a clay-organicpigment? science, 154, 645-646.

vandEnabEElE P., bodé s., alonso a. and MoEns l.(2005) - Raman spectroscopic analysis of the Mayawall paintings in Ek’Balam, Mexico. spectro -



chimica acta part a, 61, 2349-2356.WitKE K., bRZEZinKa K.W. and laMPRECHt i. (2003)

- Is the indigo molecule perturbed in planarity bymaterices? J. Molec. struct., 661-662, 235-238.

ZHEnG s., Han y., HuanG X., dai y., Qian d., ZHanG

J. and REn s. (2010) - Acid and aluminiummodification of sepiolite and its application in FCCcatalysis. Clay Minerals, 45, 1, 15-22.



Documents

Spectroscopic characterization of a sepiolite-based …tetide.geo.uniroma1.it/riviste/permin/testi/V79.SpecialIssue/2010... · sites, such as Chichén itzà, bonampak, Cacaxtla, El