摘要:

J . CHEM. SOC. DALTON TRANS. 1995 2577Svnthesis, Structural Characterisation and RamanSpectroscopy of the Inorganic Pigments Lead Tin YellowTvpes I and II and Lead Antimonate Yellow: TheirIdentification on Medieval Paintings and ManuscriptsRobin J. H. Clark,” Lucas Cridland, Benson M. Kariuki, Kenneth D. M. Harris andRobert Withnall tChristopher lngold Laboratories, University College London, 20 Gordon Street, London WC I H OAJ, UKLead tin yellow type I (Pb,SnO,) and type II (PbSn,-,Si,O,) have each been prepared in a high-temperature furnace, and the preparative conditions defined. The crystal structure of type II has beenrefined from powder X-ray diffraction data and indicates that the Sn and Si atoms are randomlydistributed over the same type of site with Sn/Si ratio z 3/1 (i.e., x z;).The Raman spectra of eachform and also that of lead antimonate yellow (Pb,Sb,O,) have been obtained. Raman microscopy isshown to be an effective technique whereby these inorganic yellow pigments as minute (zl pm)grains may be identified on medieval manuscripts and paintings.Raman microscopy is becoming an important technique forthe identification of pigments on medieval manuscripts owingto the facts that it may be used in situ, is non-destructive, hashigh spatial resolution ( < 1 pm), is highly specific, and is usuallyfree from interference from problems such as fluorescence of thebinder, etc.2 Thus the components of pigment mixtures canreadily be identified by focusing the microscope onto eachindividual grain in turn, and then comparing the resultingspectrum with that in an appropriate data base.There are several reasons for wishing to identify the pigmentsused on a manuscript or painting, the most important beingto do with art restoration.Restorers obviously should usethe original pigments or combination of pigments in theconservation of damaged or faded/degraded pictures in order toreproduce the intended effects. But the type of pigment used isalso important in art history, since the availability of particularpigments is linked to the existence or otherwise of particulartrading routes, to pigment costs, to contemporary tastes, andsometimes to the perceived importance of the figures illustrated.Moreover the identification of a pigment may aid in theauthentication of a painting since some pigments were preferredto others at particular periods in time, and synthetic ones wereobviously only available from the dates of their firstmanufacture.The usage of lead tin yellow as a pigment has a complicatedhistory, owing to the (not originally appreciated) existence oftwo forms, type I which is Pb,SnO, ’-* and the less studied typeI1 which is PbSn, -xSix03 9*10 (with a previously unspecifiedsilicon content), and also to the possibility of confusion of typeI1 with the structurally related compound lead antimonateyellow, which is Pb,Sb,O, (commonly known as Naplesyellow).I ‘ All three pigments are brilliant yellows (Munsellhue 5Y) with high refractive indices (all > 2) and thus have goodcovering power.They are alkali fast, insoluble in water and inorganic solvents, and (by contrast to PbO) light fast. Bothtypes of lead tin yellow are compatible with all other pigmentsand so is lead antimonate yellow, except when used in frescowith barium yellow (BaCrO,). Lead tin yellow of either type isessentially unaffected by inorganic acids and is thermally stable7 Presenf uddress; School of Biological and Chemical Sciences,University of Greenwich, Wellington Street, London SEI 8 6PF, UK.to ca. 900°C; lead antimonate yellow is soluble in inorganicacids and in 4 mol dm-3 NaOH. Lead tin yellow is blackened byhydrogen sulfide or soluble sulfides owing to the formation oflead sulfide. All three pigments are toxic to prolonged ingestionor inhalation or on contact with the skin.Following a brief survey of the history of the usage of thesepigments, this paper presents the Raman spectra of bothlaboratory-synthesised samples and samples on originalpaintings and manuscripts and so defines an effective methodbased upon Raman microscopy for the identification of thesepigments of similar hues (note, however, that lead tin yellowtype I is much paler than either lead tin yellow type I1 or leadantimonate yellow).In addition, the crystal structurerefinement of lead tin yellow type 11, to assess the extent ofincorporation of Si in this structure, is reported.Historical Background and NomenclatureThe long period of usage of lead-based yellow pigments datingback to 1600 B.C.has led to many different documents indifferent languages using particular words to mean the samepigment.6 European sources have regularly used (amongothers) a term ‘giallolino’ or ‘giallorino’, which is derived fromthe word ‘giallo’ meaning pale yellow; it is, however, notspecific, since it has referred to three different pigments atdifferent periods of time, viz. massicot [lead(II) oxide], lead tinyellow or lead antimonate yellow. lo*ll Kiihn 5 * 6 has concludedthat, on the evidence below, the terms ‘giallolino’ mentioned inItalian literary sources and ‘massicot’ (or ‘masticot’ of northernmanuscripts) referred to lead tin yellow. Thus:(1) Yellow lead oxide (PbO, i.e. genuine massicot) has onlyvery rarely been found in any painting made between the 13thand 201h centuries.(2) Lead tin yellow has been identified in Munich on over 150paintings of known date, provenance and authorship; similarresults have been found elsewhere.(3) In his investigations of Italian paintings, no other pigmentcould be found which could be convincingly identified with the‘giallolino’ frequently mentioned in the Italian manuscripts.(4) After analysing several old samples of lead oxide, the oneslabelled ‘massikot’ were found to be lead tin oxides mixed inwith lead oxide (PbO).Kiihn first demonstrated that there are in fact two types o2578 J .CHEM. SOC. DALTON TRANS. 1995lead tin yellow, types I and 11, and that they could be preparedby essentially 1 5Ih century Bolognese procedures; however, thelatter are not well defined, particularly with respect to optimumtemperatures and reaction times for their preparation.Lead tin yellow was the substance used to opacify ancient(Roman or immediate post-Roman) yellow glass and has beenused as an artists' pigment since around 1300 A.D.; it is not,however, the only lead-based yellow pigment to have been usedby artists.The main yellow pigments which have been used arelisted below, together with very approximate dates of theirusage.Pre 1300:1300-1 650:1450-1 750:1700-1850: Naples yellow (Pb,Sb,O,)1800-: lead chromate (PbCrO,)18%orpiment (As,S3), massicot (PbO)lead tin yellow type 11 (PbSn -xSi,03)lead tin yellow type I (Pb,SnO,)cadmium sulfide (CdS), cobalt yellow(K3[Co(NO2)dThere are, however, exceptions to this somewhat oversimpli-fied chronology of the pigments' usage.Moreover, since there isconsiderable overlap between the starting and terminal dates ofusage of each pigment, it is difficult to make an identification onthis basis. Type IT has been identified in a late-Roman shardof glass (4th--51h century A.D.) by X-ray fluorescence anddiffraction. l 1The rise and fall in the usage of lead tin yellow type 11 hasbeen linked to the economic success of particular geographicregions. This pigment is thought to have been first used as aglaze (opacifier) in the glass industry before becoming a genuineartists' pigment and it is significant that, at the time of the twomain periods of its usage ( 14Ih century in Florence and 1 61hcentury in Venice and Bohemia), there were successful localglass industries in existence.Although lead tin yellow had been used in European art sincearound 1300 none of the studies on its occurrence has revealedany example which dated after 1750.There is no reference toit in any of the art literature sources on colours from the 191hand 20Ih centuries up to 1941, when it was rediscovered byJacobi."-14 de Wild identified yellow lead oxide in 39 paintingsdating from 1400 to 1700 and found that the term 'massicot' inthe corresponding literature had probably related to lead tinyellow instead. I Naples yellow has been identified in somepaintings from the Renaissance but only tested for the presenceof lead and not for antimony; these too, were probably lead tinyellow.Since their rediscovery in 1941, lead tin yellow types I andI1 have been identified at the Doerner Institute (Munich) in overI54 paintings and four polychrome statues using emissionspectrometric studies and by Martin and Duval' in 114paintings using energy dispersive X-ray analysis of samples in ascanning electron microscope. Both the studies concluded thatlead tin yellow type I was the more commonly used, despite thefact that it first came into usage later than type 11; this isdoubtless a consequence of the higher price for type 11, owing toits more involved synthesis.Both types of lead tin yellow were replaced by leadantimonate yellow around 1750. This pigment, commonlyknown as Naples yellow, had been manufactured as early as1600-1400 B.C.at which time it was the only yellow colourantand opacifier available to Egyptian glass makers. Thus it hadbeen found in glass fragments from a site in Thebes (1450--1422B.C.). It then fell into disuse until it came back into favour in ca.1750. However, it too was replaced towards the end of the 1 91hcentury by other synthetic pigments, notably lead chromate andcadmium sulfide.ExperimentalPrepuration of Compounds.-Reactions were carried out in aplatinum crucible using a Eurotherm furnace programmable atdifferent ramp-up and ramp-down rates. Lead tin yellow type Iwas prepared by heating a finely ground intimate mixture ofPb304 and SnO, in air at 900 "C for 3 h, [equation (l)].2Pb30, + 3Sn02 - 3Pb2Sn0, + 0, (1)Heating above 900 "C leads to decomposition of the product(and loss, by volatilization, of PbO), but heating below 800 "Cleads to incomplete reaction, contrary to the recommendationsof Jacobi (650-800 "C for 5 h);14 medieval recipes do not refereither to the temperature or to the length of the heating period.A stoichiometric excess of Pb,O, leads to the presence of PbO(massicot, a thermal decomposition product of Pb304) in thesample.Lead tin yellow type I1 was a by-product of the glass industry,but enjoyed less usage than its precursor and has been identifiedin significantly fewer paintings.It cannot, apparently, beprepared directly from its binary oxide components, but wasprepared optimally by heating a 1 : 1 mixture of type I and silicaat 900°C for 5 h [equation (2)].Continued heating abovePb,SnO, + SiO, - PbSn, -xSi,03950°C causes the product to decompose to its originalconstituents, tin(1v) oxide, lead(rr) oxide and silica, which forma glassy mass. The exact extent to which silicon replaces tin isinvestigated in this paper. The structure is not known toundergo a phase transition on change of temperature.Lead antimonate yellow was supplied by Kremer Pigments; itcan be prepared by heating Pb304 with Sb,03 at 900-1000 "Cfor 5 h. A product deficient in antimony can also be prepared;this may be Pbii,Sbv2-x07, where 0 < x < 1 and 2 <.y < 3.1SInstrumentution. -Raman spectra were recorded on aDILOR XY spectrometer or on a Ramalog 14018 (R6) Spexspectrometer, in conjunction with Coherent argon-ion orkrypton-ion lasers operating at powers at the sample of 2-5mW.Spectra were recorded both at room temperature using theRaman microscope and at liquid-nitrogen temperature usingconventional macro sampling modes; the latter equipment gavethe better spectrum in each case.The Fourier-transform (FT) Raman spectra were taken onsamples at room temperature with a Nicolet 910 FT Ramanspectrometer, in conjunction with a Nd:YAG laser operating at1064 nm.The UVjVIS spectra were recorded with an Oriel InstaspecI11 0.25 m polychromator equipped with a photodiode arraydetector and 200 groove mm-l grating. The electronic spectra ofthe pigments are similar (Fig. l), type 11 absorbing morestrongly at longer wavelengths than type I.and 11, lead antimonatey e l l ~ w , ' ~ Pb304,8 Pb0,I6 Si0217 and SnO,'* are highlycharacteristic and were all recorded for analytical purposes andin order to determine the optimum preparative conditions forthe pigments.Spectra of lead tin yellow types I.-.'.*I Type I -700 6 0 0/ -' q,I I5 0 0 400Wavelength 1 nmFig. 1 Electronic spectia of lead tin yellow types I and IJ . CHEM. SOC. DALTON TRANS. 1995 2579Powder X-ray diffraction was used for phase identification ofthe materials prepared in this work, as well as for an assessmentof the Sn/Si ratio in the lead tin yellow type 11. For theseexperiments, powder X-ray diffraction data were recorded atroom temperature in transmission mode on a Siemens D5000diffractometer using Ge-monochromatised Cu-Kml radiation.The diffraction data were collected in the 26 range 13-66' insteps of A(28) = 0.02', with a total data measurement time ofcu.6 h. In each experiment, the polycrystalline sample wasground, and the powder X-ray diffractogram recorded for afine layer of the ground material mounted between two layers oftransparent tape.Results and DiscussionStructural Characterization.--Leud tin vellow. type I:Pb,SnO,. The crystal structure (Fig. 2) of lead tin yellow typeI is reported to be orthorhombic, with space group Pbam. Thisstructure is very closely related to that of Pb304, which istetragonal (space group P4,lmbc) at ambient temperature (butundergoes a tetragonal - orthorhombic transition at 170 Kon being cooled).The Pb304 structure consists of chains ofPbiV06 octahedra joined by pyramidally co-ordinated Pb"atoms; when treated with SnO,, the Sn" substitutes for the Pb"in the PbiVO, octahedral chains. Both Pb304 and Pb2Sn0, maybe regarded as mixed-valence materials ,'q2 of general formulaM,l'MIVO,. Powder X-ray diffraction confirmed that thestructural identity of the lead tin yellow type 1 samplesynthesised in this work was the same as the reported crystalstructure of Pb,Sn04.Lead tin yellow type II: PbSn, -xSixO,. The crystal struc-ture,, of PbSnO, (space group F d h , a = 10.719 A) can bedescribed as a defect pyrochlore (A,B&X') structure in whichA = Pb, B = Sn, X = 0 and X' = vacant 0 site.It has beenreported that the colour becomes deeper yellow onincorporation of Si into this structure.For lead tin yellow type 11, PbSn, -xSi,O,, we have carriedout a crystal structure refinement (in space group F d h ) fromits powder X-ray diffractogram, using the Rietveld refinementprocedure embodied within the GSAS program package.,, Forthese Rietveld refinement calculations, the reported crystalstructure of PbSnO, was used as the initial structural model.After convergence of the refinement assuming the initialstoichiometry PbSnO,, a variable amount of Si was allowed tooccupy the Sn sites in this structure, such that the sum of the SnFig. 2 Crystal structure of lead tin yellow type I (Pb,SnO,) viewedapproximately along [OOI]; structural parameters are taken from ref.19[Sn, filled circles; Pb. large open circles; 0, small open circles]and Si occupancies on these sites was unity (with the isotropicthermal parameters constrained to be equal for these Sn and Siatoms). This corresponds to refinement of the parameter x inthe formula PbSn ~ xSixO,. Isotropic thermal parameterswere assumed for the Pb and 0 atoms. In the fully refinedcrystal structure (corresponding to R,, = 0.0673 and R , =0.0532, where R,, and R , are the weighted-profile and profileR-factors, respectively), the stoichiometry is PbSn,,,,Si,,,,O,(it.., x = 0.24); other structural parameters relating to the fullyrefined structure are given in Table 1. [Note that a significantimprovement in fit was obtained on adding Si to the Sn sites;structure refinement using the same diffraction data butassuming the stoichiometry PbSnO, (ie., x = 0) correspondedto R,, = 0.0794 and R , = 0.0574.1 The crystal structure isshown in Fig.3, and the powder X-ray diffractogram calculatedfor this structure is compared with the experimental powderX-ray diffractogram in Fig. 4. The comparatively lowestimated standard deviation [o(x) 2 0.0121 in the value of.Y is noteworthy, and gives considerable confidence to our assign-ment of the stoichiometry of this material, and clarifiesthe uncertainty concerning whether Si is indeed incorporatedinto the structure of lead tin yellow type 11.It is relevant to note that the powder X-ray diffractogram ofthe lead tin yellow type I1 sample used for this Rietveldrefinement calculation also contained a subset of peaks due tocontamination by a minor phase, which was identified asSnO,(one of the original reagents).The presence of this secondphase was included in the Rietveld refinement calculation.Lead anfinzonate yel/o\li: Pb,Sb,O,. Lead antimonate yellowalso has the pyrochlore structure 24*25 (space group FdSrn, a =10.47 A) but, unlike lead tin yellow type 11, it is not defective.In this case A = Pb, B = Sb and X = X' = 0. I t is theisostructural anhydrous analogue of the mineral Bindheimite.Table Istructure of PbSn, *Si,O, [space group Fd%n, u = 10.715 (4) A]Atom X'(I Y ' h Z/c Occupancy U,50/A2Pb 0 0 0 1 0.039 (4)Sn 0.5 0.5 0.5 0.758 ( 12) 0.038Si 0.5 0.5 0.5 0.242 (12) 0.0380 0.4240 ( 2 2 ) 0.125 0.125 1 0.082 (9)Structural parameters relating to the fully refined crystalFig.3 Crystal structure of lead tin yellow type I1 (PbSn, -xSi,O,)viewed approximately along [ 1011; structural parameters are specifiedin Table 1 [Sn/Si, filled circles: Pb, large open circles; 0, small opencircles2580 J . CHEM. SOC. DALTON TRANS. 19956 .Ocnc30-< 4.0 .- 0= 2.07cn ca3 *0v0 .oI I I I I12.0 3.0 4.0 5.0 6 .Olo-' 2810Fig. 4 Experimental (+ ), calculated (solid line) and difference (bottom trace) powder X-ray diffraction profiles for the Rietveld refinement of leadtin yellow type I1 (PbSn, -xSix03) containing an impurity amount of SnO,. Marks represent positions of reflections (upper marks, SnO,; lowermarks, PbSn, - xSixO,).The calculated powder X-ray diffraction profile is for the final refined crystal structure; details of the refined structure ofPbSn, -xSixO, are specified in Table 1. Agreement factors for this structure refinement are R,, = 0.0673 and R, = 0.0532Table 2metal oxidesCharacteristic Raman bands of the pigments and relatedCompoundPbzSn04Pb304PbSn , - xSixO,PbzSb,O,PbO massicot(orthorhom bic.yellow)PbO litharge(tetragonal,reddish)I,/ n m Waven umber/cm-514.5 35wm, 58w, 80m, 129vs, 196m, 274w,291wm, 379w, 454wm, 524w, 613w647. I 121vs, 152m, 223w, 232w, 313w,39 1 w, 477w, 549s5 14.5 40m, 66m, 85 (sh), 138vs, 324wm (br),444w (br)514.5 76s, 147vs, 343s, 464m, 5 13wm647. I 87s, 144vvs, 171 (sh), 217vw, 289vs,385m.424w647.1 81s, 147vvs, 322vw, 338sPowder X-ray diffraction confirmed that the crystal structure ofthe sample of Pb2Sb207 used in this work was the same as thatreported p r e v i o ~ s l y . ~ ~ . ~ ~ More than one phase can beproduced, depending on the ratio of reactants, reaction timeand temperatures used.A second modification l 2 of Pb,Sb,O, (not known to havebeen used as a pigment) has a slightly distorted non-centrosymmetric structure related to that of Weberite (spacegroup 12cm, L( = 7.484, b = 7.857, c = 10.426 A). The co-ordination numbers of Pb(1) (distorted prism +2) andPb(2) (distorted hexagonal bipyramid) are both eight, with thesepolyhedra linked by Sb06 octahedra. It is not known whetherthis modification can be induced to undergo a structural phasetransition to the pyrochlore modification.Ruman Spectroscopy.-The principal bands in the Ramanspectra of lead tin yellow types I and I1 (Figs.5 and 6), Pb,O,,Ph2Sb207 and PbO are listed in Table 2, together with theintensity designators relevant to 514.5 or 647.1 nm excitation.Comparison of Figs. 5 and 6 indicates that there is no evidencefor resonance effects on changing from 1064 to 514.5 nmexcitation for lead tin yellow types I and 11, nor is there forthe other compounds mentioned. However, Pb,O, must behandled with low power excitation only, otherwise it decom-poses into PbO. The FT Raman spectra of these pigmentsare of high quality, with very good signal-to-noise ratios.The Raman spectrum of Pb,SnO, is similar to that recordedby Vigouroux et aL8 except that the band they reported at 540cm-' is observed in this work at 524 cm-' both by conventional(514.5 nm excitation) as well as by FT (1064 nm excitation)Raman spectroscopy.The similarity of the Raman spectra oflead tin yellow type I1 and lead antimonate yellow reflects boththe close similarity of the structures of these two pigments andthe similarity of the atomic masses of tin and antimony.The very strong bands in the Raman spectra of Pb,O,, Pb,Sn-0,, PbSn, - xSixO, and Pb2Sb207 at 12 1,129,138 and 147 cm-',respectively, are assigned to a lattice Pb-0 stretching mode.Applications.-Raman spectra taken of a painting in situ or ofa pigment section removed therefrom for analysis are unlikelyto relate to the pigment alone, but may also include bands dueto the binding medium (traditionally egg tempera or an oil suchas poppy oil) as well as bands due to any varnish which mayhave been applied to the painting. For this reason, samples oflead tin yellow types I and I1 were prepared in a poppy oilmedium and painted onto pieces of card and canvas; the canvashad already been covered in a brilliant white paint which wasidentified as containing rutile (TiO,) by the intense Ramanbands at 448 and 612 cm (Fig.7).26,27 The yellow pigmentscould easily be identified in this test, whether they were paintedout on canvas or card.The yellow pigments were then identified on the followingitems:(1) Painting by Titian, 'Death of Actaeon' (National Gallery:accession no.6420): sample taken from the yellow-leaved bushin the foreground [see Plate l(a)] was shown to be lead tinyellow type I (the principal identifying features are thestrong/very strong bands at 80, 129 and 196 cm-'). Theidentification of this pigment on the painting is consistent withthe date of its execution, which is known to have been towardsthe end of Titian's life.(2) Painting by Paulo Veronese, 'Allegory of Love,' IV(National Gallery: accession no. 1326): sample taken frombright yellow of man's cloak [see Plate l(b)] was shown to belead tin yellow type I1 (the principal identifying features arethose at 138 and 324 cm '). This painting originates from thelast quarter of the 16'h centuryJ CHEM soc DALTON TRANS 1995(0)Plate 1 ( r r ) ldcntitication of lead tin yellow type I on 'The Death of Actaeon'.by Titian at the National Gallery. Highlight on foreground foliage.Rcproduccd by Courtesy of the Trustees. the National Gallery, London. (h) Identification of lead tin yellow type I1 on 'Allegory of Love', IV, by PauloVcroncsc at the National Gallery. Bright yellow ol' man's cloak. Reproduced by Courtesy of the Trustees. the National Gallery. LondoJ . CHEM. SOC. DALTON TRANS. 1995 258 1I I I I 15 0 0 3 0 0 100Wavenumber / cm-'Fig. 5 Raman spectra of lead tin yellow types I and I1 with 514.5 iimexcitation4 Ix +Jv) CC.-+YType I1I 1 I I J6 0 0 4 0 0 200Wavenumber / cm-IFig. 6nm excitationFT Raman spectra of lead tin yellow types I and I1 with 1064(3) Unattributed oil painting on panel, 'The Lamentationwith Saints Peter, John, Mary Magdalene and Jerome'(Sotheby's Old Master Paintings auction, 21"' April 1993):sample taken from bright yellow of man's cloak in theforeground on the left was shown to be lead tin yellow type I.The identification of this pigment is consistent with the date ofc'u.1500 which is given in Sotheby's catalogue.28(4) Portrait of Lady Spenser, by John Betts ( 1 590).* A sample* This sample was supplied by courtesy of Lane Fine Art. The paintingis now in private ownership.Wavenumber 1 cm"Fig. 7 The Raman spectra of the rutile (a) and anatase (b) forms oftitanium(1v) oxidewas taken from the fringe border at the top of the painting; theRaman spectrum of this sample is thought to provide the firstidentification of lead tin yellow on a piece of English art; bothtypes I and I1 were identified as being present on this painting.(5) Historiated initial 'R' from a 16Ih century German choirbook, illustrated in ref.3. The top of the grey column in thisillustration consists of a mixture of seven different pigments,one of which is lead tin yellow type I.ConclusionLead tin yellow type I1 was successfully synthesised by fusion ofan intimate mixture of lead tin yellow type I and silica at a 1 : Imole ratio. A higher value for this ratio gave a product withRaman bands attributable to both types I and 11, whereas alower value gave a product with Raman bands attributable totype I1 and unreacted silica.The extent of incorporation of Si inthe structure of lead tin yellow type I1 has been determined bycrystal structure refinement from powder X-ray diffractiondata.Raman microscopy is at the forefront of methods availablefor the in situ analysis of pigments. The technique is easy toapply, and readily permits a distinction to be drawn betweenlead tin yellow types I and 11; the former has a very strong bandat 129 cm-' and a medium strength band at 196 cm-' , whereasthe latter has a very strong band at 138 cm and a weaker,broad band at 324 cm-'. Both may readily be distinguishedfrom lead antimonate yellow since this has a Raman spectrumdominated by a very strong band at 147 cm-' and a strong bandat 343 cm-'.AcknowledgementsThe authors thank the SERC, the Leverhulme Trust, theULIRS and the Royal Commission for the Exhibition of 1851for financial support, and Dr.A. Roy, Sam Fogg, Sotheby'sand Lane Fine Art for samples of the pigments and access toworks of art.References1 G. J. Rosasco, in Advances in Infrared and Raman Spectroscopy,eds. R. J. H. Clark and R. E. Hester, Heyden, London, 1980, vol. 7,pp. 223-282.2 J. Corset, D. Dhamelincourt and J. Barbillat, C'hem. Br., 1989, 25,612.3 S. P. Best, R. J. H. Clark and R. Withnall, Endeavour, 1992, 16,6625824 S. P. Best, R. J. H. Clark, M. A. M. Daniels and R. Withnall, Chem.5 H. Kuhn, Stud. Conserv., 1968,13, 7, and unpublished work.6 H. Kuhn, in Artists’ Pigments, ed. A.Roy, Oxford University Press,7 E. Martin and A. R. Duval, Stud. Conserv., 1990,35, 11 7.8 J. P. Vigouroux, E. Husson, G. Calvarin and N. Q. Dao,9 W. E. S. Turner and H. P. Rooksby, Glastech. Ber., 1959, 32, 17;Br., 1993, 118.1994, vol. 2.Spectrochim. Acta, Part A, 1982,38, 393.H. P. Rooksby, Phys. Chem. Glasses, 1964,5,20.10 N. J. Eastaugh, PhD Thesis, University of London, 1988.11 I. N. M. Wainwright, J. M. Taylor and R. D. Harley, in Artists’Pigments, ed. R. L. Feller, Cambridge University Press, 1986, vol. 1,pp. 219-254.12 S. A. Ivanov and V. E. Zavodnik, Sou. Phys. Crystallogr., 1990,35,494.13 B. Guineau, in lCOM Committee for Conservation, 71h TriennialMeeting, Copenhagen, 1994, p. 1429.14 R. Jacobi, Angew. Chem., 1941,54,27.15 G. Friedrich and R. Marx, Erzmetall, 1962, 15, 72.16 D. M. Adams, J. Chem. Sac., Dalton Trans., 1977, 1096.17 R. J. Hemly, H. K. Mao, P. M. Bell and B. 0. Mysen, Phys. Rev.Lett., 1986, 57, 747.J. CHEM. SOC. DALTON TRANS. 199518 J. F. Scott, J. Chem. Phys., 1976,33, 77.19 J. R. Gavarri, J. P. Vigouroux, G. Calvarin and A. W. Hewat,20 R. J. H. Clark, Chem. SOC. Rev., 1984, 13,219.21 R. J. H. Clark, J. Mol. Struct., 1984, 113, 117.22 I. Morgenstern Badarauand M. A. Michel, Ann. Chim. (Paris), 1971,23 A. C. Larson and R. B. Von Dreele, Los Alamos Laboratory Report24 G. Natta and M. Baccaredda, 2. Kristallogr., Kristallgeom.,25 M. Gasperin, C. R. Hebd. Seances Acad. Sci., 1955,240,2340.26 W. P. Griffith, in Advances in Spectroscopy, eds. R. J. H. Clark and27 S. P. S. Porto, P. A. Fleury and T. C. Damen, Phys. Rev., 1967, 154,28 Sotheby’s catalogue, Sale of Old Master Paintings, April 1993,J. Solid State Chem., I98 1,36, 8 1.109.NO. LA-UR-86-748, 1987.Kristallphysik, Kristallchem., 1933,05, 27 1.R. E. Hester, Wiley, Chichester, 1987, vol. 14, p.135.522.p. 129.Received 30th March 1995; Paper 5/02002

ISSN:1477-9226    

DOI:10.1039/DT9950002577

出版商:RSC    

年代:1995

数据来源: RSC