Elsevier

Journal of Cultural Heritage

Volume 10, Issue 2, April–June 2009, Pages 198-205
Journal of Cultural Heritage

Original article
Discrimination of painting binders subjected to photo-ageing by using microspectrofluorometry coupled with deconvolution analysis

https://doi.org/10.1016/j.culher.2008.04.011Get rights and content

Abstract

Organic binding media found in paintings exhibit characteristic fluorescence properties that strictly correlate with their chemical composition and may vary as a function of the ageing time. The aim of this work was to investigate the capability of microspectrofluorometry to distinguish between different binders. Linseed oil and protein-based media, deposited as thin films on microscope glass slides both in the presence and in the absence of inorganic pigments, were examined before and after artificial photo-ageing. Cross-sections of some paint layers were also examined. The article points out that microspectrofluorometry coupled with deconvolution analysis can be a useful tool for distinguishing between oil- and protein-based media. The curve-fitting analysis furnished a fine characterization of each binder/pigment combination, and highlighted the small spectral differences between their fluorescence signals.

Introduction

Paint layers typically consist of inorganic and organic pigments dispersed in an oil- or protein-based binding medium and applied over a preparation layer. Other organic substances are usually present in contiguous layers, and comprise animal glues in the preparation layers and natural resins in the varnish layers [1]. The difficulty in the mechanical separation of the various contiguous layers, as well as the reduced amount of the sample to be collected for the analysis, make identification of the binding media one of the most challenging steps in the diagnostic investigation of paintings.

Oil is a natural product that can be pressed out of a variety of plant seeds. Some types of oils, known as drying oils (e.g. linseed oil), form a solid film when exposed to air due to oxidation and cross-linking reactions, thus they assume good working properties as binding media. The protein-based materials mainly employed include: egg, used whole or after separation of the yolk from the glair; animal glue, extracted from skin or bones; milk, which is usually treated in order to isolate the casein fraction.

Binding media are subjected to ageing processes that are typically represented by physical, chemical and biological phenomena [2]. Photo-oxidation is one of the degradation mechanisms most involved in the ageing of painting media [3]. Some proteinic binders having an heterogeneous composition, such as egg, can also experience a variety of reactions among the various components of the medium [4]. Lastly, humidity, pH gradients and chemical interactions with the pigments can play important roles during ageing processes [2]. All these factors can lead to dramatic changes in the chemical composition of binding media within the paint layer, thus making their identification quite difficult. Hence, a further analytical problem adds to the previously mentioned ones for studying the original techniques of old paintings, as well as for planning their correct restoration.

Fluorescence spectroscopy is a well-developed and powerful technique that was proposed in the past as a practical tool for investigating the chemical composition of painting materials [5], [6]. Indeed, when exposed to ultraviolet excitation, many compounds found in oil- and protein-based paints can exhibit different colours and intensities of fluorescence, which are closely correlated to their chemical nature.

Ultraviolet sources have been applied to the examination of old pictures ever since they became commercially available around the 1920s. Nevertheless, no studies of interest on the fluorescence of paintings have been published since the pioneer works of Eibner [7] and Wolf and Toeldte [8] appeared around 1930. The most significant progress was made in the 1980s by de la Rie [9], [10], [11]. He studied the fluorescence properties of linseed oil and natural resins submitted to different periods of artificial ageing using a fluorescence spectrometer equipped with a mercury lamp. In further studies, several laser sources were employed to excite the endogenous fluorophores of the paints [12], [13], [14]. An excimer and a Nd:YAG laser coupled with a spectrometer to collect emission spectra was recently proposed as a non-destructive and potentially portable system [15], [16].

A different approach is represented by microspectrofluorometric analysis. This technique has been applied to investigations of organic binders as a promising tool for coupling the sensitivity of fluorescence measurements with the spatial resolution of a microscope [17], [18]. By using this technique, it is possible to carry out stratigraphic analyses of paint samples in order to obtain important information on the nature of the components, in addition to their distribution within the layers.

However, it has been observed that the fluorescence properties of organic binders vary depending on their physical state and ageing time. Protein-based binding media in wet form are poorly fluorescent when excited in the UV A (320–400 nm range) (P. Matteini, personal communication). The same behaviour has been observed for linseed oil, the most popular drying oil [10]. After drying, several alterations in the components of the medium take place, leading to an increase in the fluorescence signal that can be easily acquired by a microspectrofluorometric apparatus [17].

Fluorescent products found in aged proteins may derive from the oxidation of amino acids [19]. Nevertheless, in the presence of oxidative lipids (e.g. in egg yolk), other reactions associated with fluorescence products, can take place [20].

Under UV excitation, dried linseed oil shows a bright signal, which grows and shifts to longer wavelengths during ageing [11]. This behaviour has been attributed to the formation of soaps between free fatty acids and the metal ions of the pigments [7]. Furthermore, lipid oxidation can also lead to the formation of several conjugated Shiff bases [10], some of which have been suggested to have fluorescence properties [21].

The aim of this work was to assess the capability of microspectrofluorometry to identify and distinguish between different binding media used in works of art, by investigating their endogenous fluorescence properties in the presence or in the absence of mineral pigments. Particular attention has been devoted to the effects produced on fluorescence by accelerated photo-ageing process. This was done in order to evaluate the response of this technique on old paintings and “fill in” the lack of knowledge in this field. To achieve this objective, the emission responses of various organic binders, both on thin film cast onto microscope glass slides and on cross-section of paint layers, were analysed. A standard epifluorescence microscope equipped with a spectrum analyser was used to detect the emission spectra. Moreover, a curve-fitting deconvolution of the fluorescence spectra was performed in order to better discriminate the responses obtained from the different binders investigated.

Section snippets

Samples preparation and photo ageing procedure

Binding media were prepared following artist's accounts and historical recipes [1] as much as possible. Animal glue was swollen in water overnight and then warmed at 40 °C to give a 20% solution (w/w). Whole egg and commercial partially defatted milk were diluted in water (1:1). Linseed oil (boiled with catalytic quantities of cobalt salts) was purchased from Zecchi (Florence, Italy) and used without dilution.

Pure pigments (azurite, malachite, red ochre and white lead) were purchased from Zecchi

Assessment of fluorescence and re-absorption effect of the pigments

Preliminary tests were performed to verify the fluorescence properties of the inorganic pigments considered in the study, namely azurite, malachite, red ochre and white lead. All the pigments lacked fluorescence or showed a very weak signal when excited at 365 nm wavelength, even at the longest acquisition time used for the measurements (5 seconds). Hence, their contribution to the fluorescence of paint samples was considered to be negligible. These observations are in agreement with an earlier

Conclusions

This work points out that microspectrofluorometry coupled with deconvolution analysis can be a useful tool for distinguishing between oil- and protein-based media found in art works. In addition, accurate curve fitting can provide suggestions with regard to the different proteinaceous materials by highlighting the small differences in their fluorescence properties upon ageing. Due to its easy use, short times required for the analysis, and low cost, the proposed technique could be employed as a

Acknowledgements

We wish to thank Cinzia Mancuso and Maria Antonietta Gallone for helpful discussions.

References (30)

  • J.A. Radley et al.

    Fluorescence analysis in ultraviolet light

    (1959)
  • A. Eibner

    Zum gegenwartigen Stand der naturwissenschaftlichen Bilduntersuchung

    Angewandte Chemie

    (1932)
  • H. Wolff et al.

    Leinol im Ultraviolettlicht der Analysenquarzlampe

    Farbe und Lack

    (1926)
  • E.R. de la Rie

    Fluorescence of paint and varnish layers (Part I)

    Stud. Conserv.

    (1982)
  • E.R. de la Rie

    Fluorescence of paint and varnish layers (Part II)

    Stud. Conserv.

    (1982)
  • Cited by (21)

    • Unsupervised component analysis: PCA, POA and ICA data exploring - Connecting the dots

      2016, Spectrochimica Acta - Part A: Molecular and Biomolecular Spectroscopy
      Citation Excerpt :

      Thus, to evaluate the spectrum is necessary to make the spectral deconvolution of analytes, leading to a more reliable identification. Different authors have applied spectral deconvolution to better interpret the results [45–50]. These algorithms allow, with greater accuracy, to retrieve the individual contributions of each compound present in test sample.

    • Comparative analysis of eastern and western drying-oil binding media used in polychromic artworks by pyrolysis-gas chromatography/mass spectrometry under the influence of pigments

      2015, Microchemical Journal
      Citation Excerpt :

      But these techniques are commonly accompanied with very complicated preliminary scission of polymeric chain by hydrolysis in obtaining characteristic marked components [17]. Nevertheless, nuclear magnetic resonance (NMR) [3,18], infrared spectroscopy (IR) [19,20], Raman spectroscopy (RS) [4,21,22] and micro-fluorimetry [23] techniques are reported in the analysis of dry-oil, but they are actually inadequate in identifying the specific oil substances, since they are only effective in roughly distinguishing drying-oil from other binding media, such as polysaccharide [21], resin [21,22], wax [22] or protein-based materials [21,23]. Fortunately, pyrolysis coupled with gas chromatography–mass spectrometry (Py–GC/MS) technique, working by thermally cracking the macro-molecule materials into low-molecule compounds and followed by directly GC separating with MS detection, could overcome these shortages to provide a rapid and effective answer to the nature of organic materials [17,24–28].

    • Microanalytical method for studying paintings by use of fluorescence spectroscopy combined with principal component analysis

      2013, Microchemical Journal
      Citation Excerpt :

      Fluorescence spectroscopy is a well-developed and powerful technique for studying molecular properties and it has long been recognized that it may be possible to determine the composition of painting layers with this technique [31–34]. Surprisingly, detailed studies of the fluorescence properties of substances used on paintings have been limited [35,36] and, to the authors' knowledge, only few applications of the technique for the analysis of varnishes of musical instruments have been undertaken [37]. One significant disadvantage associated with fluorescence spectroscopy for the analysis of complex mixtures is that it is not generally possible to ascribe broad fluorescence emissions to specific molecules, because the observed spectrum is probably the sum of the spectra of many fluorophores present in the sample.

    • Protein identification and localization using mass spectrometry and staining tests in cross-sections of polychrome samples

      2013, Journal of Cultural Heritage
      Citation Excerpt :

      Furthermore, the capability of the MALDI-TOF MS technique to discriminate between different proteinaceous binders make it a powerful tool to be applied as a complement for identifying the already mapped organic constituents of a multilayered paint structure. Nowadays, the identification of organic (mainly proteinaceous) binders in works of art is still a challenging tasks in conservation science applications, although different mass spectrometry techniques (GC-MS, MALDI-TOF MS, DTMS, LC-MS/MS) [1,2,5–9], spectrofluorimetric [4,10] and spectroscopic [11] techniques, immunochemical methods (Elisa, IFM, nano-SERS antibodies, chemiluminescence etc.) [12–16] were reported as successful on model (mock) samples and real samples taken from cultural heritage artefacts (mural paintings, easel and panel paintings, polychrome sculptures). All the above-mentioned methods allow identifying and distinguishing at least between the egg, animal glue and milk proteins.

    • A simple protocol for Matrix Assisted Laser Desorption Ionization- time of flight-mass spectrometry (MALDI-TOF-MS) analysis of lipids and proteins in single microsamples of paintings

      2012, Analytica Chimica Acta
      Citation Excerpt :

      Moreover, binder identification is usually complicated by very low sample size, containing only about 10% of organic binder. In the last few years various analytical techniques capable of identifying organic materials in cross-sections have come into use, including immunofluorescence microscopy (IFM) [1], microspectrofluorimetry [2,3], chemiluminescence [4], mapping and imaging Fourier Transform Infrared (FTIR) spectroscopy [5–7], secondary-ion mass spectrometry (SIMS) imaging [8,9], and surface enhanced Raman scattering (SERS) [10]. With these techniques localization of different binders can be also accomplished.

    View all citing articles on Scopus
    View full text