Azurite in medieval illuminated manuscripts: a reflection-FTIR study concerning the characterization of binding media
In illuminated manuscripts, a reliable identification of oxyanion pigments such as azurite by rFTIR is simple, as several combination and overtone bands are strongly enhanced compared to transmission mode. However, the characterization of the used binding media is rather difficult, as the analysis of four medieval manuscripts from the late thirteenth to the fifteenth century (e.g. Cod. slav. 8 in the collection of the Austrian National Library), as well as the earliest known map of Vienna (Albertinischer Plan from 1421, Wien Museum) showed. According to the literature, mainly glair (egg white) and plant gums were applied as binding media for azurite. Moreover, both were used in many cases also as “varnishes” in order to improve optical and mechanical properties of the paint layer. In order to assess the possibilities and to distinguish between proteinaceous and carbohydrate binders, mock-ups with azurite were prepared on parchment support with various quantities of binders. Additionally, some of the specimen were varnished using the binders mentioned above. Furthermore, mock-ups on aluminium foil were prepared to evaluate the influence of the support on the reflection spectra. The results showed that the binding medium content in the mock-ups usually was too low for a reliable determination by rFTIR (except the ones with the highest contents), whereas it was possible to characterize the varnish materials. Only an insignificant influence of the support on the spectra from the mock-ups was observed. However, the spectra obtained from three manuscripts suggested a certain influence of the parchment support, which indicates thinner paint layers.
KeywordsIlluminated manuscripts Non-invasive Reflection-FTIR Azurite Binding medium Glair Egg white Arabic gum
Centre of Image and Material Analysis in Cultural Heritage
National Research & Development Institute for Textiles and Leather, Bucharest
Infrared and Raman Users Group
Fourier Transform Infrared spectrometry in the reflection mode
Azurite (Cu3(CO3)2(OH)2), is composed of basic copper carbonate. The mineral azurite is found in many parts of the world in the upper oxidized portions of copper ore deposits. Azurite mineral is usually found in association with malachite (Cu2CO3(OH)2), the green basic carbonate of copper, and often with other copper-rich minerals (e.g. cuprite, tenorite, and chrysocolla) . Depending of the region and the time period, the pigment was known under different names such as azzurro della magna, azzurro citramarina (Italy) [2, 3], Bergblau (Germany), bleu de montagne or bleu d’Allemagne (France) , blew bice (England) , lapis armenius (Pliny) or Berglasur (Agricola) , as well as several other names referring to the origin (German azure, Spanish blue, Hungarian blue, Ragusa blue, Lombard blue) . The natural pigment, which was widely used in manuscripts, as well as panel and wall paintings, was made by grinding it to a powder and washing with plain water or solutions containing soap, gum and lye . Furthermore, solutions containing honey, fish glue or gum were used . The optical quality of the resulting pigment depends on the content of impurities and even more on the particle size. Deep blue qualities, which were favored in medieval times, are relatively coarse and the pigment turns pale if it is ground too fine. A larger particle size leads to longer absorption paths of the incident light before reflection and thus a darker appearance, but on the other hand the applicability is becoming worse and it might be too gritty to be used as a pigment . The most important binding media for manuscript illuminations were clarified egg white or glair (clare, albumen, glarea, albuginea ovi) and gums, e.g. Arabic gum (gumma). Furthermore, fish glue (ichtyocollon), parchment size (cola pergamena), and casein glue (glutine casei) were used as well. In most cases only a single binder was applied, but depending on the technique or pigment, mixtures were also prepared in different proportions [6, 7] and it is reported that glair and occasionally Arabic gum  were used for varnishing in order to improve the optical and mechanical properties of the paint layers.
The availability of mobile analytical instruments, which are applicable in a non-invasive way, as well as an enhanced collaboration between natural sciences, humanities, and computer sciences has led to an increased analytical interest in illuminated manuscripts in the last years [8, 9, 10, 11, 12, 13, 14]. These investigations aim to characterize the materials in the manuscripts and to elucidate the techniques which were applied for their production. Among the analytical techniques used for these purposes, Fourier transform infrared spectrometry in the reflection mode (rFTIR) has proved to be suitable for the characterization of several inorganic and organic materials encountered in manuscripts, such as the parchment support, pigments, contaminants, or degradation products [9, 10, 11, 12, 13, 14, 15, 16], although the evaluation of reflection data is often complicated by the contribution of both, specular and diffuse reflection . Nevertheless, the literature shows that azurite can easily be identified by several combination bands in the rFTIR spectrum, which are strongly enhanced compared to spectra obtained in transmission mode, and that the binding medium usually cannot be characterized, except in one case .
We were also able to identify azurite in several medieval manuscripts on parchment during investigations carried out within the framework of CIMA (Centre of Image and Material Analysis in Cultural Heritage)  without receiving significant information concerning the binding medium. Although some of the spectra showed spectral features of proteins, it was not clear whether they derive from a proteinaceous binding medium or the parchment support. For this reason, we examined mockups with paint layers of azurite with either egg white or Arabic gum on parchment in order to assess the possibilities for a reliable characterization of proteinaceous and polysaccharide binding media using rFTIR. In addition, the resulting paint layers were partially varnished by egg white or Arabic gum to test the influence of this treatment and furthermore, mockups on aluminum foil were prepared in order to evaluate the influence of the support on the resulting rFTIR spectra. The present study should contribute to the knowledge about historical painting techniques and a better understanding of FTIR spectra obtained in reflection mode.
Azurite with a particle size of approximately 14 µm (Order number: 1673112) was purchased from Kremer Pigmente (Aichstetten, Germany); Arabic gum solution (Sennelier) with 20% dry mass content (Order number 8-33577) was obtained from Gerstaecker (Germany). The Arabic gum solution contains less than one per cent of a mixture of the preserving agents 5-chloro-2-methyl-1,2-thiazol-3(2H)-one and 2-methyl-1,2-thiazol-3(2H)-one, which did not lead to visible spectral features in transmission FTIR spectra of dried Arabic gum. Glair was produced from the egg white of fresh chicken egg, which was whipped to a stiff froth by use of a handheld electric mixer with stainless steel beaters. The froth was allowed to stand for 5 h and the watery liquid on the bottom of the vessel was collected for the experiments, which had a dry mass content of 12%. Calf parchment was provided by the National Research & Development Institute for Textiles and Leather, Bucharest (INCDTP). The paint layers were applied to the flesh side of the parchment, where only negligible contents of calcium carbonate were detected by rFTIR (in contrast to the hair side).
Although the mentioned ratios between the pigment and the binders provided paint layers with very good mechanical and optical properties, the influence of various binding medium contents was tested. For this purpose, the Arabic gum solution was allowed to dry and the residue was dissolved in water to produce solutions with 50% and 35%. 300 µl of each solution were mixed with 0.5 g azurite to prepare mockups in the same way as described above. Glair was concentrated to 20% dry mass content by evaporation of a corresponding quantity of the water, and 400 µl of the resulting solution were mixed with 0.5 g azurite for the mockups.
For comparison, powder samples without binder were analyzed on glass microscope slides. The weight of a cover slide (0.17 mm thickness) was used to flatten the surface of the powder without additional pressure. Furthermore, glair and Arabic gum without pigment were applied on the parchment in order to assess their spectral features.
rFTIR measurements were performed using a spectrometer Alpha operated with software Opus 7.5 (Bruker Optics, Ettlingen, Germany), equipped with the module for external reflection. The instrument has a beam diameter of approximately 5 mm (roughly circular) and the background was acquired using a gold mirror. Reflection spectra were collected in the range between 4000 and 375 cm−1 (manuscripts), as well as 5000 and 375 cm−1 (mockups) with a resolution of 4 cm−1 over 64–128 scans. The reflection spectra were either directly evaluated, or occasionally after Kramers–Kronig algorithm, followed by baseline correction.
Results and discussion
Characterization of the azurite powder
In addition to the bands of azurite, also the O–H stretching bands of kaolinite could be detected at 3696, 3655 and 3620 cm−1, whereas the Si–O stretching bands in the range of 1000–1100 cm−1, which potentially could interfere with the C–O stretching vibrations of the Arabic gum binder, were almost missing in the spectra obtained [20, 21]. However, a slight influence on the results from the mockups cannot be completely excluded. As described in the literature , natural azurite frequently contains impurities, particularly malachite and silicates, although kaolinite is not mentioned. The microscopic evaluation of the purchased azurite showed contents of white, greenish and red particles, which were not further analyzed.
Compared to the rFTIR spectrum of azurite powder, the addition of both, glair and Arabic gum resulted in additional spectral features in the range of 3700–3100 cm−1 (N–H and O–H stretching), as well as 1750–1600 cm−1 (Fig. 4). In case of Arabic gum as binder, no additional bands can be determined in the ν C–O range of 1160–900 cm−1.
The differences between the rFTIR spectra with either glair or Arabic gum are only marginal, and do not allow a reliable differentiation between both types of binders. Small differences are observed between 4800–4590 cm−1, 3400–3000 cm−1, and 1700–1600 cm−1. Carlesi and coworkers  used an instrument similar to ours (Alpha) to analyze model films of pure binding media on glass slides in the NIR range (7500–3900 cm−1) and were able to differentiate egg white and Arabic gum by application of multivariate analysis. This could be an interesting approach for further studies, although the situation is more complex when binders are mixed with pigments in mockups, or if various other components contribute to the spectra obtained from illuminations in historic manuscripts. Moreover, ageing processes of the materials might also complicate chemometric analyses. It has been further reported that a chemometric approach was successfully applied to discriminate transmission FTIR spectra of egg white and parchment glue that were used in reconstructions of medieval paints, exploiting the C–H stretching absorption regions . However, the rFTIR spectrum of glair on parchment only shows one uncharacteristic, relatively broad C–H feature and that is also the case for Arabic gum. Furthermore, we frequently detected calcium soaps on the surface of paint layers in manuscripts, which also exhibit strong C–H bands. An example is presented in the following section, where the results obtained for Cod. slav. 8 are discussed. For these reasons, the application of a similar approach does not seem to be very promising in terms of the differentiation of glair and Arabic gum binders in illuminated manuscripts.
The application of the paints on aluminum foil indicates that Arabic gum paint layers show a greater tendency towards shrinking than glair samples (Fig. 2). This might have implications on the mechanical properties of the illuminations in manuscripts, although Arabic gum was considered as stronger binding medium, and glair as rather brittle . However, the stability of paint layers in illuminations is not only determined by the binding medium, but also by the support.
In the case of glair with 20% dry mass content, slightly enhanced features deriving from the binder were observed around 4600 cm−1, corresponding to overtone ν C=O amide I and amide III deformation (ν C–N/δ N–H in plane) , and the amide I band around 1690 cm−1  (Fig. 7b). However, according to the literature and our preliminary application tests we suppose that illuminators rather used glair undiluted or mixed with water. A comparable content of binder is mainly not expected in historic manuscripts.
The analytical results from the mockups demonstrate the limited possibilities for a characterization of binding media applied for azurite in illuminated manuscripts. In the case of glair the protein contents of pigment/binder mixtures are usually too low for a reliable detection and also a clear distinction from Arabic gum is not possible. Furthermore, it has to be considered that in historic manuscripts thinner paint layers may have been applied and an interference of the protein bands with those of parchment cannot be excluded, as described in the next chapter. On the other hand, Arabic gum could only be detected if applied in very high concentrations or if used as varnish. However, a differentiation of these possible applications by use of rFTIR is hardly possible and would require analyses of cross-sections.
Azurite was detected in four outstanding illuminated manuscripts on parchment and a colored map on paper, which were investigated non-invasively by rFTIR within the framework of CIMA. As it was impossible to determine whether a proteinaceous or a polysaccharide binder was used, we examined mockups with azurite in combination with either glair or Arabic gum aiming to assess the possibilities for a reliable characterization. Both materials were additionally used to varnish the mockup samples, according to historical sources. The results obtained from the mockups allowed us to state that the used method usually does not enable a reliable characterization of the used binding media, but is capable of determining the varnish materials in certain cases. The investigations further revealed that paint layers in manuscripts are usually more complex, as they may contain additives or reaction products, such as calcium soaps, which may interfere in the spectral ranges, which are important for the characterization of the binding media.
WV conceived the study, performed the analyses of the manuscripts, carried out interpretation of the results, and drafted the manuscript. IL developed the methods for the preparation of the mockups and performed the analyses of the mockups. MS participated in the coordination of the study, and helped to draft the manuscript. All authors have read and approved the final manuscript.
This project was realized in the framework of CIMA, an interuniversity research institution with an interdisciplinary approach to the investigation of cultural heritage. The authors are grateful to their partners in the project, especially to Prof. Dr. Heinz Miklas. The authors further wish to thank Lucretia Miu (INCDTP) for providing the parchment used for this study.
The authors declare that they have no competing interests.
Availability of data and materials
The Austrian Science Fund (FWF) and the Austrian Ministry for Science, Research and Economy (BMWFW) are gratefully acknowledged for the financial supports: P23133 and P29892 as well as Hochschulraum-Strukturmittelfonds (HRSM)-project “Analysis and Conservation of Cultural Heritage-Modern Imaging and Material Analysis Methods for the Visualization, Documentation and Classification of Historical Written Material (Manuscripts)”.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
- 1.Eastaugh N, Walsh V, Chaplin T, Siddall R. Pigment compendium, a dictionary and optical microscopy of historical pigments. Oxford: Butterworth-Heinemann; 2008. p. 39–40.Google Scholar
- 2.Delamare F. Blue pigments: 5000 years of art and industry. London: Archetype Publications Ltd; 2013. p. 119.Google Scholar
- 3.Cennini C. Il libro dell’arte. A cura di F. Frezzato, Neri Pozza, Collana I colibrì; 2009.Google Scholar
- 4.Gettens RJ, West Fitzhugh E. Azurite and Blue Verditer. In: Roy A, ed. Artists´ pigments, a handbook of their history and characteristics. Vol 2. London: London Archetype Publications; 1993. p. 23.Google Scholar
- 5.Thompson DV. The materials and techniques of medieval paintings. New York: Dover Publications; 1956. p. 130–5.Google Scholar
- 6.Thompson DV. The materials and techniques of medieval paintings. New York: Dover Publications; 1956. p. 50–61.Google Scholar
- 7.Kroustallis S. Binding media in medieval manuscript illumination: a source research. In: Medieval Colours: between beauty and meaning, Revista de història de arte, no. 1, Serie W. 2011. p. 112-125.Google Scholar
- 17.CIMA is an interuniversity research institution with an interdisciplinary approach aiming at the investigation of cultural heritage and was founded in 2014. http://hrsm.caa.tuwien.ac.at/. Accessed 27 Nov 2018.
- 19.Price BA, Pretzel B, Lomax SQ, eds. Infrared and Raman Users Group Spectral Database. 2007 ed. Vol. 1 & 2.Google Scholar
- 20.Vaculikova L, Plevova E, Vallova S, Koutnik I. Characterization and differentiation of kaolinites from selected Czech deposits using infrared spectroscopy and differential thermal analysis. Acta Geodyn Geomater. 2011;8(1):59–67.Google Scholar
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.