Confocal Raman spectroscopic study of painted medieval manuscripts

Confocal Raman spectroscopic study of painted medieval manuscripts

Journal of Cultural Heritage 2 (2001) 191−198 © 2001 Éditions scientifiques et médicales Elsevier SAS. All rights reserved S1296207401011219/FLA Conf...

436KB Sizes 23 Downloads 63 Views

Journal of Cultural Heritage 2 (2001) 191−198 © 2001 Éditions scientifiques et médicales Elsevier SAS. All rights reserved S1296207401011219/FLA

Confocal Raman spectroscopic study of painted medieval manuscripts Francesca Magistroa, Domenico Majolinoa*, Placido Migliardoa, Rosina Ponterioa, MariaTeresa Rodriquezb a

Dipartimento di Fisica and Gruppo Operativo di Fisica Applicata - Sezione Beni Culturali, Università di Messina, P.O. Box 55, Papardo 98166 S. Agata-Messina, Italy b Biblioteca Regionale di Messina, Sezione Fondi Antichi, Via dei Verdi 98122 Messina, Italy

Abstract – In the present work we show, together with a historiographic research, the results obtained by the application of Raman confocal microspectroscopy on the miniatures contained in three membranous medieval codices (Fondo Vecchio 18 IX–X century, Messanensis S.Salv. Graec. 51 XIII century, Messanensis S.Salv.Graec. 83 XII century). The analysis, which is non-destructive and performed ‘in situ’ on micrometre-sized spots, allowed us to characterize the pigments used, in particular to detect the presence of inorganic substances. © 2001 Éditions scientifiques et médicales Elsevier SAS Raman / microspectroscopy / pigments

1. Research aim In the past, the knowledge of a work of art used to be gaised essentially through historical-artistic approaches, which alone contributed to collocate temporally and geographically the work itself and to characterize the various schools of production. Nonetheless, a work of art is made of matter, therefore scientific studies greatly contribute towards the historical and technical knowledge of the system under investigation (analysis of constituting materials and execution techniques, dating, authentication and so on). Moreover, investigations of this kind provide a remarkable contribution to the preservation operations: in fact it is possible to execute specifically directed restorations, using original materials. The use of coloured materials has represented, since the appearance of the first forms of figurative expression, the deep need for a better defined representation of reality and for an enrichment of images. The search for substances or means suited to this aim is typical of every civilization, from the most primitive ones which had to draw upon natural sources, mainly mineral and vegetable, to the most evolved ones which have *Correspondence and reprints. E-mail address: [email protected] (D. Majolino).

produced substances and realized means fit for chromatic expression through suitable synthesis proceedings. In modern technology [1], there are few experimental methods available that allow the non-destructive analysis and the identification in situ of the pigments used by the old illuminators in painting the manuscripts. Modern optical micro-spectroscopies, such as confocal Raman and IR microscopies, are undoubtedly the most refined physical methods in order to lay open any secret de arte illuminandi. Clark R.J.H. and co-workers [2, 3] critically examined the strengths and weaknesses of the main techniques used for pigment analysis and recently [4] described the Raman microscopy potentiality, performing a careful analysis of the inorganic pigments present in five medieval books. Our research group [5] performed, by FT-IR microscopy, a study ‘in situ’ of pigments contained in miniatures of medieval texts: the Gospel Messanensis Gr. S. Salvatore 88 (XIII century) and Neapolitan S. Giovanni’s (XV century). Very different results are obtained in the two cases: as far as the first manuscript is concerned, the mainly organic nature, together with a palaeographic analysis, allow us to recognize a typical style of a Byzantine school. The exclusively inorganic nature of the pigments analysed has been revealed in the second case.


F. Magistro et al. / J. Cult. Heritage 2 (2001) 191–198

2. Experimental section In this section we show the results obtained by using a new Raman spectrometer DILOR LABRAM, operating in a microconfocal configuration, that allowed us to identify, in a totally non-destructive way, the pigments used in making the illuminations of three precious medieval manuscripts presently kept in the Regional Library of Messina: the Fondo Vecchio 18, a tetragospel of the first Macedonian period, coming from the Library of the Jesuiti College of Messina; the Messanensis Graecus 83 and the Messanensis Graecus 51, coming from the monastery of S. Salvatore, founded by Roger II in 1130 in Messina [6]. 2.1. Materials The three parchment manuscripts come from geographical areas distant from each other: Istanbul, Calabria, Palestine and Cyprus. Yet, from a cultural point of view all of these are culturally Greek areas, even though they can be placed chronologically in different periods. The choice of the manuscripts to be examined was conditioned by the possibility of advancing hypotheses of a homogeneous use of books of receipts and techniques in the art of illumination in different provinces of the Byzantine world. Brief historical and palaeographic hints about the manuscripts are given here, and the illuminations which underwent microspectroscopic investigation are also described. The Fondo Vecchio 18, which today contains only the Matteo and Marco Gospels, was originally a tetragospel, preceded by a letter from Eusebius to Carpiano and by the tables of canons; it was also accompanied by the illuminations of the four evangelists with epigrams dedicated to them [7]. Three hands collaborate in the laying out of the code, one of which uses a handwriting that makes it datable between the two last decades of the IX century and the first 30 years of the X century [8] and originating from the area of Istanbul. The decorative apparatus shows the miniature of evangelist Matteo, stylistically classic [9], together with the tables of canons of oriental style [10]. The original nucleus of the manuscript begins with sheet 1, on whose verso there is a full page illumination, representing a niche supported by two columns surmounted by small domes; the internal space, veiled at the sides by dark green curtains, is occupied by a cross of gold and gems. With a rectangular shape, Matteo’s pinax is enclosed by a frame with a double golden listel decorated by a stylised classical motif with flower. The

right half of the composition is occupied by the figure of the sitting evangelist. Covered with gold are the sumptuous furnishings of the small study we see on the left: a large writing-desk on which writing tools lie. The background of the scene is occupied by a peripteral temple of grey-azure marble, provided with a triangular pediment. The Messanansis S. Salv. Graecus 83 is part of a group of manuscripts studied by Leroy J., [11], datable within the first 30 years of the XII century and coming from the same writing centre located by Leroy in the monastery of S. Mary Odigitria in Rossano [12]. This manuscript, that ends with a subscription dating it to 1104–1105, is one of the best-known of the Fondo [13], and is decorated with frames and polychrome initials (blue, yellow, red lead, ochre, violet). There are numerous letter ‘A’s of the word ‘adelphoi’ (brothers) with which the exhortations of S. Theodore Studita start. For example the initial letter A of sheet 234 v (figure 1). Illuminations are

Figure 1. The initial A (f.234v) from the XII century Messanensis S. Salvatore Gr. 83.

F. Magistro et al. / J. Cult. Heritage 2 (2001) 191–198


in all condensed phases of matter. The relevant part in Raman scattering is represented by the cross-section r共 Q,x 兲 that is connected to inelastic energy 共 hx 兲 and moment 共 hQ 兲 transfer process. The scattered intensity is proportional to the double differential 2 cross-section d r共 Q,x 兲/dx dX :: I = d r/dx dX ≈ I0 rk NXTk sk 2

Figure 2. Illuminated miniature (f.52v) from the XIII century Messanensis S. Salvatore Gr. 51.

where I0 is the laser radiance on the sample, rk represents the differential cross section of the Raman line analysed, N the number of the molecules, Ο the solid angle of collection of Raman light, Tk is the instrument brightness and sk is the detector sensitivity at k wavelength. More specifically the double differential crosssection defines the dynamic structure factor S共 Q,x 兲 that can be written: S共 Q,x 兲 =

noteworthy both for their design and for the special shade of colour, which is remarkable both for the quality of the pigments and for the way it is drawn. The Octoecos Messanensis S. Salvatore Graecus 51 [14, 15] is a liturgical-musical text which is part of a group of manuscripts datable around the second half of the XIII century, which the palaeographic studies carried out by P. Canart [16] and the researches of the illuminations historian Weyl Carr A. [17] have allotted to a Palestinian/Cyprian zone. At the beginning of each of the eight ‘echoi’ of the manuscript there are pictured illuminations placed inside panels, in which the disposition and the choice of the colours sharply oppose the human to the divine; at the top there are golden and azure areas detached by means of bands from the dark of the earthly word, or of Hell, represented in the inferior part. In the illumination laying on the verso of page 18, at the top, inside a semicircle, Christ in glory appears; below there is a mountain representing Hell. The illumination on the verso of page 52 (figure 2) is divided into two scenes. In the upper scene Christ stands out inside half an almond between two groups of angels in bluish and pink-lilac clothes. In the lower scene, on the right Hell (partly erased) can be seen, while from the left the Damasceno comes forth, indicating the apparition to the monks following him. 2.2. Confocal Raman microspectroscopy Raman spectroscopy is based on the spectral distribution of inelastically scattered light and is a very useful technique for the study of the molecular species


兺 共q 兲

m 2


Qm共 x + Xm 兲^F2m共 x + Xm 兲 (2)

This equation states that the Raman spectrum is constituted by vibrational peaks Gm共 x + Xm 兲 convoluted with the rotational second order Legendre polynomial F2m共 x + Xm 兲. In our case, measurements performed by Raman confocal microspectroscopy has solved absolutely the characterization of inorganic pigments. We have used a new advanced experimental apparatus: the Raman confocal microspectrometer DILOR LABRAM [18]. The laser is focused on the sample by means of the notch filter and the microscope lens. Raman light collected by the lens is reflected by the notch filter to the confocal diaphragm and input iris of a monochromator and Raman spectra are collected on a CCD detector (figure 3). When, as in the case of interest, the analysis is performed on a small spot, the number N of molecules contained in the scattering volume will be reduced. In this case the relevant parameters to compensate the decrease of N are I0 and Ο. The use of microscope lenses in order to illuminate the sample and collect the diffuse radiance is the best experimental technique for the scattering surface reduction and for the Ο optimisation. The use of a microscope, in particular, allows the focusing of a visible laser on spots of 0.5–1 µm of diameter. The diffuse radiance is collected in back scattering through the microscopy lens with a great efficiency. Spatial resolution of sampling can be improved by use of confocal geometries: a small mobile diaphragm (confocal) is placed on the lens microscope imaging plane and selects only the diffuse radiance


F. Magistro et al. / J. Cult. Heritage 2 (2001) 191–198

2.3. Results and discussion

Figure 3. Scheme of confocal Raman apparatus.

coming from the portion of sample illuminated by the laser spot. In the case of homogenous samples the conical diaphragm diameter can be increased with the aim of collecting the larger quantity of Raman signal. The Raman signal selected by the confocal diaphragm is then focused through the monochromator input aperture. Since the Rayleigh diffusion intensity covers the Raman signal, we have to operate a filtering system before the signal enters the detector. In order to do this we used a holographic notch filter placed in the beam-splitter site. The laser is completely transmitted towards the microscope, and only Raman radiation is reflected onto the monochromator. The notch filter allows us to use a single monochromator with a great percentage of gain; the low frequency Raman, instead, is not reflected by notch, and therefore is not observable. Also in the case of interest, a LABRAM system has been used, equipped with a HeNe laser, operating with a mean power of 10 mW, a spectrograph with two interchangeable gratings and a rugged, compact, air-cooled CCD.

The different Raman spectra, obtained by the various illuminated decorations, allow for the identification of colours shown in the table I. As can be seen, they unambiguously evidenced the presence of inorganic substances. A restriction of Raman microscopy is that some pigments (mainly organic ones) may still only give rise to a weak signal and others may fluoresce, probably because of their extremely small grain sizes; it’s possible to resolve this problem by changing the excitation wavelength in order to identify all the substances, especially concerning mixtures. In the Middle Age, moreover, illuminators commonly used pure or mixed organic compounds; in this way, hence, it may be acceptable to hypothesize the simultaneous presence of organic substances (binders, additives and pigments) and the revealed ones by performed Raman measurements. In fact our research group discovered, by FTIR microspectroscopy, that coloured pigments used by illuminators consist also of organic chromophores, usually alkaloids [19]. These substances, as reported in the literature [20], are contained in animal fluids or vegetable provenance and have in the past been commonly used as correctives to delay the drying up of the mixture during the miniature painting. We report in some detail an analysis of the various pigments unambiguously identified as follows. 2.3.1. Red pigment Two red tonalities were analysed: dark and pale red. In the Fondo Vecchio 18 the dark red pigment is located in the cupola on the left of the sheet 1v, as can be seen in figure 4; the pale red pigment is located in the pediment of the temple of the sheet 11r, as can be seen in figure 5, in the Messanensis S. Salvatore Graecus 83 the (pale) red pigment is located on sheet 234 v (figure 1). The Raman analysis reveals that the main pigment responsible for the dark red colour is the lead oxide or minium (Raman bands centred at 119, 151 and 547 cm–1) known and used since antiquity (figure 6).

Table I. The different Raman spectra, obtained by the various illuminated decorations.

Dark red Pale red Blue Black Brown Green 1 Green 2 Violet

F.V. 18

S. Salv. 51

S. Salv. 83

minium cinnabar – lampblack (amorphous carbon) iron oxide lead tin yellow + lazurite copper compound red ochre

minium cinnabar lapis lazuli – – – – red ochre

– cinnabar – – – – – –

F. Magistro et al. / J. Cult. Heritage 2 (2001) 191–198


Instead the pigment mainly used for the pale red colour is vermilion (Raman bands centred at 250 and 343 cm–1) as shown in figure 7. The Raman spectra of the pigments of Messanensis S. Salvatore Graecus 51, both located in the miniature in the sheet 52v (figure 2), show the presence of the same substances used in the painting of the miniatures of Messanensis Fondo Vecchio 18. 2.3.2. Blue pigment The blue coloration analysed is referred to the parallel waves that represent the firmament in the upper scene of the illumination on sheet 52 v of the Messanensis Graecus 51. The main pigment used to obtain the blue colour is an inorganic pigment unambiguously identified as lapis lazuli (3Na·3Al2O3· 6SiO2·Na2S) by Raman microspectroscopy. This precious mineral has characteristic Raman bands centred at 552 and 1 095 cm–1 (figure 8) that are all present in our spectra. There is extensive literature concerning the methods of manufacture of this material. It is very

Figure 5. Illuminated miniature (f.11r) of Matthew’s pinax from the IX–X century Tetragospel Fondo Vecchio 18.

well known that this sodium and aluminium silicate includes other minerals as sulphides, carbonates and pyrite in a small quantity. This latter brings about a change of the colour towards a grey-blue tonality. For this reason the illuminator probably mixed acids with the lapis lazuli in order to eliminate the pyrite and the sulphides and to dissolve the carbonates, so that the pigment possesses a very bright colour [21].

Figure 4. Illuminated miniature (f.1v) from the IX–X century Tetragospel Fondo Vecchio 18.

2.3.3. Black pigment The analysis performed on the black pigment, located in the Saint’s sandal on the sheet 11r of the Fondo Vecchio 18 manuscript, showed the presence of lampblack, a pigment essentially constituted by elementary carbon (88.3–99.5%) and several impurities such as sulphur, oxygen etc. Raman experimental spectrum contains two characteristic broad bands centred at ca. 1 345 and 1 660 cm–1 (figure 9). The use of this pigment was very common in the Middle Ages and in general in all the ancient world, because of its high covering power, its photostability and its reparability. In fact, this deep black colour was real


F. Magistro et al. / J. Cult. Heritage 2 (2001) 191–198

ized by collecting the soot produced by an oil lamps flame on a cold surface.

Figure 6. Raman spectrum of dark red pigment in Tetragospel Fondo Vecchio 18.

Figure 7. Raman spectrum of pale red pigment in Mess. S. Salv. Gr. 83 manuscript.

Figure 8. Raman spectrum of blue pigment in Mess. S. Salv. Gr. 51 manuscript.

2.3.4. Brown pigment The brown coloration analysed is referred to the writing-desk of the sheet 11r of the Fondo Vecchio 18 manuscript (figure 5). Experimental spectrum shows a number of sharp bands centred at about 262, 130, 82 and 60 cm–1(e.g., < 400 cm–1) as shown in figure 10, probably attributable to metallic oxides. In particular it is plausible that the pigment responsible for this deep brown coloration is Brown Sienna (burnt). Brown Sienna, that in nature shows an orange coloration, is a mixture of iron hydroxide, manganese oxide and clay (Al2O3·2SiO2·2H2O). The pigment after a calcinations procedure and hence a dehydration, put on a deep brown tonality. 2.3.5. Green pigment The pigment commonly used in the Middle Age to realize the green coloration was malachite; nonetheless in all the analysed manuscripts this substance is absent. It is relevant to evidence in particular the situation of the Fondo Vecchio 18 manuscript. In this case, in fact, we analysed two tonalities of green: one located in the curtain of the sheet 1 and the other in sheet 10r. In the first case, by an inspection of the spectrum it is evident that a mixture of inorganic substances rather a single inorganic one was used [22]. The coloration was obtained by mixing a yellow and a blue compound; in particular the Raman bands centred at 130, 200 and 347 cm–1 are characteristic of lead tin yellow (Pb2SnO4), and the bands centred at ca. 550 and 1 100 cm–1 are justified by the use of Lazurite (figure 11). The spectrum shows a weak

Figure 9. Raman spectrum of black pigment in Tetragospel Fondo Vecchio 18.

F. Magistro et al. / J. Cult. Heritage 2 (2001) 191–198


Figure 10. Raman spectrum of brown pigment in Tetragospel Fondo Vecchio 18.

Figure 12. Raman spectrum of pale green pigment in Tetragospel Fondo Vecchio 18.

background of fluorescence, probably connected with the residue of some organic substance used during the manufacture of the pigment and during the fixing operation of the pigment itself to the matrix. The analysis performed on the second green tonality shows a spectrum immune to fluorescence effects with a series of sharp bands centred at about 340, 700 and 1 180 cm–1 (figure 12). In this case it wasn’t possible to characterize unequivocally the pigment responsible for the coloration because of the lack of a similar reference standard spectrum. In any case, by an inspection of bibliographic fonts, all the evidenced peaks can be connected with the presence of a some copper compound, probably chrysocolla: a copper

silicate (CuSiO3·nH2O), greatly used in the Middle Ages, that allows one to obtain a very brilliant green [23], as in our case, permanent and light-fast. 2.3.6. Violet pigment This pigment is located in the sheet 18v of the Messanensis S. Salvatore Graecus 51. An inspection of experimental data shows the presence of three bands centred at about 220, 290 and 400 cm–1, characteristic of iron oxide (Fe2O3) commonly named red ochre (figure 13). The presence of a fluorescence background may be due to organic substances that were probably mixed (as binders or correctives) to the pigment responsible of the coloration.

Figure 11. Raman spectrum of dark green pigment in Tetragospel Fondo Vecchio 18.

Figure 13. Raman spectrum of violet pigment in Mess. S. Salv. Gr. 51 manuscript.


F. Magistro et al. / J. Cult. Heritage 2 (2001) 191–198

3. Conclusions In this paper we have applied confocal Raman microspectroscopy to a non-destructive identification of ancient pigments. The illuminated miniatures of the Gospels Fondo Vecchio 18, Messanensis Graecus S. Salvatore 51 and Messanensis Graecus S. Salvatore 83 (XII century) have been subjected to careful study. The Raman spectral results reveal that the nature of the pigments employed was not very different in all three manuscripts, although they came from distant geographical areas and belong to different historical periods (IX–X, XIII and XII centuries, respectively). Such a circumstance can be connected with the fact that all the manuscripts are set into the same Greek culture. Our data confirm the attribution to historical periods and production schools based on stylistic and palaeographic considerations. This experimental technique used for the pigment analysis, in our opinion, allows us an almost precise and final assignment in situ of the ancient pigments in a microscopic scale and furnishes a useful and versatile tool for the conservation of our very precious cultural heritage.

Acknowledgements We are indebted to Dr Sandra Conti, regent of the Regional Library of Messina for the friendship, constant helpfulness and enthusiasm for this project. The authors thank the Assessorato Regionale per i Beni Culturali, Ambientali e per la Pubblica Istruzione for allowing us rapid access to the precious Fondo Vecchio 18, Messanensis Gr. S. Salvatore 51 and Messanensis Gr. S. Salvatore 83. The authors wish to thank the Instruments S.A. Italia staff (Dr Oddo, Dr Valisa, Dr Marchione) for the measurements with the DILOR LABRAM apparatus, for the optimization of spectrometer and for the useful discussion. Furthermore the authors wish to thank Prof. Irene Davì for her critical reading of the English language of this paper.

References [1] Raggi A., Bertocchi G., Tecniche Analitiche ChimicoFisiche e Radiochimiche applicate all’arte e all’archeologia, Cooperativa Libraria Universitaria Editrice, Bologna, 1975. [2] Best S.P., Clark R.J.H., Withnall R., Non-destructive pigment analysis of artefacts by Raman microscopy, Endeavour 16 (1992) 66–73. [3] Best S.P., Clark R.J.H., Daniels M., Withnall R., A Bible laid open, Chemistry in Britain (1993) 118–122.

[4] Clark R.J.H., Raman microscopy: application to the identification of pigments on medieval manuscripts, Chem. Soc. Rev. (1995) 188–195. [5] Majolino D., Migliardo P., Ponterio R., Rodriquez M.T., Ars illuminandi and FTIR microspectroscopy: an advanced way to disclose ancient secrets, Sci. Technol. Cult. Heritage 5 (1996) 57–74. [6] Foti M.B., Il monastero del Salvatore in lingua Phari, Proposte scrittorie e coscienza culturale, Messina, 1989. [7] Perria L., Iacobini A., Il Vangelo di Dionisio. Il codice F.V. 18 di Messina, l’Athous Stavronikita 43 e la produzione libraria costantinopolitana del primo periodo macedone, Rivista di studi bizantini e neoellenici 31 (1994) 82–163. [8] Perria L., Arethaea. Il codice Vallicelliano di Areta e la Ciropedia dell’Escorial, Rivista di studi bizantini e neoellenici 25 (1988) 41–56. [9] Friend A.M., The portraits of evangelists in the Greek and Latin manuscripts, Art Studies 5 (1927) 115–146. [10] Nordenfalk K., Die spatantiken Kanontafeln Kunstgeschitliche Studien uber die eusebianische EvangelienKonkordanz in den vier ersten Jahrhunderten threr Geschichte, Goteborg, and, successively, Wessel K., kanontafeln, in Reallexicon zur byzantinische Kunst, Stuttgart,1978 and P. Réfice, Canoni, tavole, die, Enciclopedia dell’Arte Medievale, IV, Rome, 1993. [11] Leroy J., Bisanzio e l’Italia. Raccolta di studi in memoria di Agostino Pertugi, Milan, 1982. [12] Lucà S., Rossano, il Patir e lo stile rossanese. Note per uno studio codicologico-paleografico e storico-culturale, Rivista di studi bizantini e neoellenici (1985–1986) 22–23. [13] Leroy J., Les patitos catéchèses de S. Teodoro Studita, Le Mugeon 71 (1958) 319–358. [14] Daneu Lattanzi, I Manoscritti e incunaboli miniati della Sicilia, Accademia di Scienze, lettere ed arti, Palermo, 1984. [15] Tarlo L., I manoscritti greci di musica bizantina nella Biblioteca universitaria di Messina, Archivio storico per la Calabria e la Lucania 23 (1954) 188–190. [16] Canart P., Les écritures livresque chypriotes du milieu du IX siècle au milieu du XIII et le style palestino- cypriote “epsilon”, Scrittura e civiltà 5 (1981) 17–76. [17] Weyl Carr A., Illuminated musical manuscripts in Byzantium: A note on the late twelfth century, Gesta 28 (1989) 41–52. [18] Valisa P., I nuovi sviluppi della spettrometria Raman. Instruments S.A. Italia, Laboratorio, 1997. [19] Gibbs P.J., Seddan K.R., The in situ identifications of pigments on ancient papers by mass spectrometry, Spectroscopy Europe 7 (1995) 10–18. [20] Matteini M., Moles A., La chimica nel restauro, Nardini Editore-Centro Internazionale del Libro, Florence, 1989. [21] Linzi C., Tecnica della pittura e dei colori, Milan, 1966. [22] Burgio L., Ciortman D.A., Clark R.J.H., J. Raman Spectrosc. 28 (1997) 79–83. [23] Brunello F., De Arte Illuminandi e altri Trattati sulla Tecnica della Miniatura Medievale, Neri Pozzi, Vicenza, 1975.