Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 195 (2018) 103–112
Contents lists available at ScienceDirect
Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy journal homepage: www.elsevier.com/locate/saa
Capabilities and limitations of handheld Diffuse Reflectance Infrared Fourier Transform Spectroscopy (DRIFTS) for the analysis of colourants and binders in 20th-century reverse paintings on glass Simon Steger a,⁎, Heike Stege b, Simone Bretz c, Oliver Hahn a,d a
BAM Federal Institute for Materials Research and Testing, Division 4.5, Unter den Eichen 44-46, Berlin, Germany Doerner Institut, Bayerische Staatsgemäldesammlungen, Barer Str. 29, Munich, Germany c Conservator for Reverse Paintings on Glass, Garmisch-Partenkirchen, Germany d Centre for the Study of Manuscript Cultures, University of Hamburg, Hamburg, Germany b
a r t i c l e
i n f o
Article history: Received 18 October 2017 Received in revised form 16 January 2018 Accepted 21 January 2018 Available online xxxx Keywords: DRIFT spectroscopy Reverse painting on glass Non-invasive analysis Synthetic organic pigments FTIR
a b s t r a c t A non-invasive method has been carried out to show the capabilities and limitations of Diffuse Reflectance Infrared Fourier Transform Spectroscopy (DRIFTS) for identifying of colourants and binders in modern reverse glass paintings. For this purpose, the reverse glass paintings “Zwei Frauen am Tisch” (1920–22), “Bäume” (1946) (both by Heinrich Campendonk), “Lofoten” (1933) (Edith Campendonk-van Leckwyck) and “Ohne Titel” (1954) (Marianne Uhlenhuth), were measured. In contrast to other techniques (e.g. panel and mural painting), the paint layers are applied in reverse succession. In multi-layered paint systems, the front paint layer may no longer be accessible. The work points out the different spectral appearance of a given substance (gypsum, basic lead white) in reverse glass paintings. However, inverted bands, band overlapping and derivative-shaped spectral features can be interpreted by comparing the spectra from the paintings with spectra from pure powders and pigment/linseed oil mock-ups. Moreover, the work focuses on this method's capabilities in identifying synthetic organic pigments (SOP). Reference spectra of three common SOP (PG7, PY1, PR83) were obtained from powders and historical colour charts. We identified PR83 and PY1 in two reverse glass paintings, using the measured reference spectra. The recorded DRIFTS spectra of pure linseed oil, gum Arabic, mastic, polyvinyl acetate resin and bees wax can be used to classify the binding media of the measured paintings. © 2018 Elsevier B.V. All rights reserved.
1. Introduction The technique of painting on the reverse side of glass was rediscovered by artists in the early 20th century and gained great popularity, especially in Germany. The artist group “Der Blaue Reiter” (the Blue Rider) around Wassily Kandinsky and Franz Marc got in touch with this technique in the summers of 1908 and 1909. Heinrich Campendonk (1889–1957) joined the group in 1911 and conducted his first reverse paintings on glass [1]. He created N70 paintings with this technique during his career. Campendonk met the Belgian artist Edith van Leckwyck (1899–1987) in 1929, and married her in 1935. Painting on the reverse side of glass became popular among artists in Germany during the first half of the 20th century, and many known and unknown artists tried this technique. The Bavarian artist Marianne Uhlenhuth created a single reverse glass painting in 1954. However, no further information is available about her, using this technique.
⁎ Corresponding author. E-mail address:
[email protected] (S. Steger).
https://doi.org/10.1016/j.saa.2018.01.057 1386-1425/© 2018 Elsevier B.V. All rights reserved.
In contrast to other paint techniques (e.g. panel and mural painting), the paint layers are applied in reverse succession starting with the foremost paint layer and ending with the primer (backmost layer). The paintings are viewed in reflected light, thus revealing an impressive gloss, luminosity and depth of colour. The glass plate behaves like a varnish for the front layers. Therefore, compared with other techniques, the viewing side (front) is better protected against climate-induced damages. Due to the lack of early references for recommendations of certain materials, there are no strict rules about which materials should preferably be used for reverse glass paintings. Until the 19th century, paints were produced by the artists themselves by mixing pigment powders with binders, additives and siccatives. In most cases, they used drying oil (linseed or nut spike oil), proteins (e.g. egg, casein), resins (e.g. dammar) and gums (e.g. gum Arabic). However, mixtures of several binding media (e.g. protein & oil) were also common. In the 20th century, paint tubes were highly recommended for reverse glass paintings, because of their good quality and low price [2,3]. In his publication, Stahl mentions about 40 common colourants for oil colours that can be used in reverse glass paintings [4]. However, there are no pigment recommendations specifically for reverse glass paintings [3].
104
S. Steger et al. / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 195 (2018) 103–112
Another characteristic technique for reverse glass paintings is etching a dried paint layer, a metal foil or a gold leaf and then overpainting the etched structure with another colour [3,5,6]. Metal foils (e.g. tin, aluminium), coloured paper or cartons are common protective layers. Reverse glass paintings comprise a non-porous glass substrate and multi-layered paint system; hence, delamination and flaking of the paint layer are the most common conservation problems [7]. However, fractures of the glass panel can also be observed. Scientific investigation of the material provides important information for appropriate conservation concepts. Moreover, thorough research on reverse glass paintings will lead to a better understanding of the artistic technique. A comparison of the materials used in reverse glass paintings and in easel paintings in classic modern art reveal certain differences. Recently, several studies of medieval reverse glass paintings using minimal-invasive methods were published [7–10]. After several extraction steps, samples were analysed using gas chromatography (GC), gas chromatography–mass spectrometry (GC–MS) and amino acid analysis (AAA), allowing the identification of binding media used in the artworks [10]. Moreover, Hahn et al. reported on non-destructive analyses of colourants of reverse glass paintings from the 14th to the 16th century using in-situ techniques like digital microscopy, visible reflectance spectroscopy (VIS), and X-ray fluorescence analysis (XRF) [11]. Using a mobile reflectance IR device, Miliani et al. found bassanite (CaSO4·0,5H2O) in an Italian reverse glass painting from the 17th century [12]. Only the study by Baumer et al. deals with a reverse glass painting from the classical modern period (1905–1955) [10]. The study presents an overview of the binding media analysed using minimal-invasive analytical techniques. In particular, the results obtained for Franz Marc's – “Landschaft mit Tieren und Regenbogen” (1911) reveal the complexity of this kind of painting. The painting is a collage of reverse painting on glass, paper, and metal foil and was executed with gouache distemper (animal glue, egg, polysaccharide gum) as binding media [10]. The literature on the chemical analysis of modern reverse glass paintings using non-destructive techniques is limited. To the best of our knowledge, this is the first work that presents the study of modern reverse glass paintings using non-destructive techniques. For several years, portable reflectance Fourier Transform Infrared Spectroscopy (FTIR) has been a convenient tool for the non-destructive analysis of cultural heritage objects. Powerful and robust spectrometers became available at lower cost. Thorough in-situ reflection, IR analyses of various cultural heritage objects, like easel paintings [13], canvas painting [14], panel paintings [15,16], mural paintings [17], wall paintings [18,19], illuminated manuscripts [20,21], painted metal sculptures [22], stringed musical instruments [23,24] and built heritage materials [25–27] have been conducted, showing the high potential of this technique. Generally, when light irradiates a sample material, it can be directly reflected from the surface, or it penetrates the material and can then be absorbed, refracted, reflected or scattered before reaching the surface again. In the case of DRIFT, the total amount of reflected light that passes the detector consists not only of diffuse reflected light (volume reflection), but also of surface-reflected light (specular reflection), as they cannot be optically separated. Specular reflection causes two main distortions in diffuse reflection spectra: derivative-like features [12,28] and inverted bands (reststrahlen bands) [12,29,30]. The specular reflection contribution can be corrected with the Kramers-Kronig (KK) algorithm [13,31–33]. It can correct the relation between the refractive index and the absorption coefficient when the surface reflection is dominant, but it gives inconclusive results when both specular and diffuse-reflected components are present [25,34]. Diffuse reflection originates from an absorption process, as IR light refracts through each particle and is scattered by the combined process of reflection, refraction and diffraction [28]. Compared with transmission spectra, undistorted diffuse-reflection spectra yield no significant
band shifting (as occurs in the ATR technique); however, some differences in relative band intensities can be observed [35,36,51]. Moreover, diffuse reflection leads to an enhancement of weak absorption bands because the light can travel a longer distance in a material with small absorption coefficient by repeated refractions [12]. Diffuse reflection lacks an exact theoretical explanation, but the Kubelka-Munk theory tends to be the most practical [37]. For reflection FTIR measurements of artworks, both diffuse reflection and specular reflection are present, and their proportion cannot be predicted. The amount of diffuse-reflected light strongly depends also on the sample surface, i.e. optically flat surfaces (with particles larger than the IR wavelength) will increase the amount of specular reflection, whereas rougher surfaces (with particle dimensions similar to the IR wavelength) will generate mostly diffuse-reflected light [12]. In the literature, three modifications of portable reflection FTIR spectrometers are generally reported: reflectance FTIR spectroscopy without fibre optics [16,21,22,24,33,42,43] (also called “external” reflectance FTIR); Infrared Fibre Optic Reflectance spectroscopy (IR-FORS) [12,38– 42,47,48,52–54] and handheld Diffuse Reflectance Fourier Transform Infrared Spectroscopy (DRIFTS) [14,18,19,25,44,45]. The external reflectance FTIR is able to measure in a range of 7500– 375 cm−1 and preferably collects specular reflected light [15]. For this reason, most external-reflection FTIR studies have used the KK transform to get absorption-like spectra. Zaffino et al. claim that the quality of the specular reflection IR spectrum depends on the sample surface. A highly reflective surface will yield a better spectrum, which can be easily corrected by the KK transform [21]. However, when diffuse reflection is more active (i.e. the surface is rougher), the KK correction cannot be applied with this setup [22,43]. The optical layout for reflection measurements is 22°/22° with a working distance of 15 mm [22,24]. The width of the measured area is ~6 mm [21]. The optical path is vertical. Thus, paintings can be measured only when they are hanging on the wall. The device (~7 kg) can be placed on a tripod for this purpose [16]. The IR-FORS setup works in 0°/0° geometry (angle of incidence and angle of collection are perpendicular to the sample surface) [12]. It has been successfully applied for both the mid-infrared (MIR) range (7000–900 and 4000–900 cm−1) and the near-IR range (12,500–4000 cm−1) [40,52]. The spectra show strong contributions of both diffuseand specular-reflected IR light, depending on the surface roughness of the sample [12]. Thus, KK corrections are often not feasible [12]. However, Monico et al. showed that applying the KK correction only to certain areas of the spectrum may give positive results [33]. The chalcogenide glass fibres are orientated perpendicularly to the sample surface by a mechanical arm. The device (~35 kg) shows a working distance of ~5 mm and a sampling spot of ~5 mm width [12]. Due to the Se\\H stretching absorption of the fibres, the 2200–2050 cm−1 region is not accessible with this set up [38]. The handheld-DRIFTS device (3.2 kg) works in the MIR range (4000–650 cm−1). It emits IR radiation at 0° (i.e. perpendicular to the sample) and collects the reflected light inside an imaginary cone of 45° around the emission beam. This configuration supports the collection of diffuse-reflected light [25]. However, it needs to be emphasized that both specular- and diffuse-reflected IR light are collected and their ratio depends on the surface roughness of the sample. The device measures contact-free close to the sample surface (working distance ≤1 mm) with a sampling spot of ~10 mm width. It can be used handheld or on a tripod. We want to study the advantages, capabilities and limitations of the handheld DRIFTS device for the analysis of modern reverse glass paintings. The over-arching objectives of this study are (1) to examine the different spectral appearances of a given substance; (2) to review critically the interpretation of distorted bands and the applicability of correction algorithms; and (3) to study the limitations of the device for identifying pigments and classifying binders in modern reverse glass paintings. To attain these objectives, a detailed study of reference materials, mock-ups properly prepared and careful examination of each
S. Steger et al. / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 195 (2018) 103–112
spectrum in relation to the spot analysed is necessary to avoid overinterpretations and wrong band assignment. Measurements of pigment powders, pure binders and pigment-binder mock-ups are compared with in-situ studies of paintings by Heinrich Campendonk, Edith Campendonk-van Leckwyck and Marianne Uhlenhuth. 2. Experimental Procedure 2.1. Reference Materials 2.1.1. Raw Materials The powdered pigments of interest here include: basic lead carbonate (from Aldrich), phthalocyanine green (Pigment green 7, PG7, from Rasquin), gypsum, alizarin crimson dark (Pigment red 83, PR83) (all from Kremer) and from Siegle's colour chart: Siegleechtgrün G and Universalechtgrün 763 (PG7), Sieglegelb G & Universalgelb 7930 F (Pigment yellow 1, PY1) and Rotlack (PR83). Linseed oil, gum Arabic, mastic and polyvinyl acetate resin (polyvinyl acetate 50) were purchased from Kremer. The bees wax was bought from a local supplier. 2.1.2. Colour Charts DRIFTS measurements on the colour chart “organic pigments” from G. Siegle & Co. GmbH Stuttgart-Feuerbach (Table 1), which is included in the book H. Kittel “Pigmente” (1960; in German) were conducted and supplemented by Raman spectroscopy and, where needed, by Xray fluorescence (the results are not presented here). The organic paints were mixed with an unknown binder and applied on paper strips, which were glued into the book. The paint layer is unvarnished. 2.1.3. Measurements We analysed commercially available pigments as pure powders compressed on a glass slide. Pure liquid binders (linseed oil, egg), mastic, bees wax and the pigment-binder mock-ups were dried on glass slides. Solid polyvinyl acetate was dissolved in toluene, then spread on a glass slide. Gum Arabic was dissolved in distilled water and, after filtration, dried on a glass slide. Getting the right sample thickness (to avoid any spectral contribution of the glass) was the crucial point for all mock-ups.
105
particles in the red-grey part. The single, white, semi-glossy paint layer (lead white) reveals distinct brush strokes (Fig. 1c). The bluishgreen head was overpainted with a matte light blue layer. The red head was covered with tin foil that was overpainted with red iron oxides (ochre or synthetic hematite) (Fig. 1d). However, the overpainting of an opaque metal foil does not yield any aesthetic advantages. It may act as a protective layer against oxidation of the metal foil [1,5]. The reverse glass painting ‘Bäume’ (Fig. 2) by Campendonk's painting ‘Bäume’ comprises 2–3 paint layers, which were partly etched and overpainted. The paint was applied partly wet-on-wet. Campendonk used whitish and yellow bronze particles in a few white areas (e.g. Fig. 2b). The paint layers are backed by an olive-green glazed paper. The painting ‘Lofoten’ (Fig. 3a, b) reveals 1–2 paint layers that are partly etched and overpainted. Thick brush strokes are visible in the whitish and blue areas, whereas the matte red of the sailing boat (Fig. 3b) reveals very weak brush strokes. Marianne Uhlenhuth's painting ‘Ohne Titel’ (Fig. 3c) consists of a single paint layer. The matte, thin paint layer is partly delaminated (e.g. in the violet area) and was partly fixed during a previous restoration treatment. Brush strokes are hardly visible in the well-defined areas. 2.3. Instrumentation 2.3.1. DRIFT Spectroscopy In-situ, diffuse-reflectance spectra were recorded, using a 4100 Exoscan FTIR spectrometer (Agilent Technologies, Santa Clara CA, USA) fixed on a tripod in perpendicular geometry to the sample. In this configuration, the reflected signal is collected inside an imaginary cone of 45° around the emission beam, which increases the proportion of diffuse-reflected light and reduces specular reflection. The sample-tospectrometer distance was ≤1 mm without touching the surface. The instrument is equipped with a ZnSe beam splitter, a Michelson interferometer and a thermoelectrically cooled DTGS detector. The spectral range is 4000–650 cm−1 with an energy resolution of 4 cm−1. A gold reference cap is used for background calibration. The spot size is about 10 mm. For every spectrum, 264 or 500 scans were recorded. The Thermo Scientific™ OMNIC™ Specta software (Version 9.7, Madison, WI, USA) was used for spectral treatment and for comparison with internal databases. The spectrum intensity was defined as the pseudo-absorbance A′ = log (1/R).
2.2. Paintings Four original paintings were chosen for this study. The paintings ‘Zwei Frauen am Tisch’ (1920–22) and ‘Bäume’ (1946) are by the German painter Heinrich Campendonk. The other paintings, ‘Lofoten’ (1933) and ‘Ohne Titel’ (1954), were created by Edith Campendonkvan Leckwyck and Marianne Uhlenhuth, respectively. The painting “Zwei Frauen am Tisch” reveals a multi layered paint system (2–3 paint layers). Campendonk used fine-grained bronze Table 1 Pigments and binders used as reference materials. Sample
Description
Reference
Basic lead white Gypsum Phthalocyaningrün rein RG (PG7) Siegleechtgrün G (PG7) Universalechtgrün 763 (PG7) Sieglegelb G (PY1) Universalgelb 7930 F (PY1) Alizarine crimson dark (PR83) Rotlack (PR83) Linseed oil Gum Arabic Mastic Polyvinyl acetate
PbCO3·Pb(OH)2 CaSO4·2H2O C32Cl16CuN8, phthalocyanine green Phthalocyanine green Phthalocyanine green C17H16N4O4, monoazo yellow Monoazo yellow C14H8O4, anthraquinone Anthraquinone Linseed oil, cold-pressed
Aldrich 24854621 Kremer 58300 Rasquin
Polyvinyl acetate 50
Colour charts Colour charts Colour charts Colour charts Kremer 23610 Colour charts Kremer 73054 Kremer 63320 Kremer 60050 Kremer 67040
2.3.2. Infrared Spectroscopy in Transmission Mode The transmission spectra were recorded, using a Nicolet iS5 FTIR spectrometer. The spectral range was 4000–400 cm−1 with an energy resolution of 2 cm−1. For every spectrum, 512 scans were recorded. 2.3.3. Raman Spectroscopy Raman measurements were performed on a portable i-Raman®Plus spectrometer (B&W Tek Inc. Newark, DE, USA), equipped with a 785 nm diode laser, a handheld fibre optic probe and a CCD detector. The recorded spectra range from 100 to 3300 cm−1 at a maximum spectral resolution of 4 cm−1. 3. Results and Discussion We first want to discuss the IR spectral features of lead white and gypsum as powders and compare them with the spectra of the powder/binder mock-ups. We want to point out how the binder affects the spectral appearance of the pigment in terms of potential binder/pigment band overlapping and the appearance of inverted bands. With this knowledge, we will discuss the spectral appearance of these pigments in modern reverse glass paintings. Recently, Manfredi et al. reported a comprehensive database of DRIFT spectra from mainly inorganic substances [45]. Hence, our study will focus on gypsum and lead white, which were not discussed in detail in this previous study. Moreover, it is important for our purpose that the selected pigments yield an intense
106
S. Steger et al. / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 195 (2018) 103–112
Fig. 1. Heinrich Campendonk, ‘Zwei Frauen am Tisch’, 1920–22; front side (a), reverse side (c). The detail in figure b (front) is covered by an overpainted tin foil (d). The × marks the measured spots. (For interpretation of the references to colour in this figure, the reader is referred to the web version of this article.) All images courtesy of Simone Bretz. ©VG Bildkunst Bonn, 2017.
IR spectrum in the measured range. Furthermore, we discuss the spectra of three important synthetic organic pigments (SOP): phthalocyanine green (PG7), pigment yellow 1 (PY1) and alizarin crimson (PR83). We want to learn whether this technique can identify them in reverse glass paintings and how binders and fillers (e.g. chalk, barium sulphate, talc) may hamper their proper identification. Finally, we discuss the IR spectral features of selected binders and use the obtained spectra to classify binding media in the paintings. 3.1. Pigments 3.1.1. White Pigments The DRIFT spectrum (Fig. 4 I) of lead white powder reveals bands of the fundamental vibrations of the carbonate group at 686 (ν4; in-plane bending), 835 (ν2 out-of-plane bending motion; weak shoulder) and 1044 cm−1 (ν1; symmetric stretching) [12,55]. We interpret the intense signal at ~1370–1500 as a reststrahlen band with a minimum at ~1398 cm−1 (ν3; antisymmetric stretching). The inverted band from the powder originates from particles that are coarser than the IR wavelength (λ = 2,5–15 μm). The intense band at 3536 cm−1 is the fundamental OH stretching mode [55]. Enhancement of combination modes and overtones leads to significant bands at 1735 (ν1 + ν4 combination) and
~2427 cm−1 (2ν2 + ν4 and/or ν1 + ν3 combination) [12]. The bands are in good agreement with the data in the published literature (Table 2). The band shapes changed massively for the lead white/linseed oil mock-up (Fig. 4 II). The spectrum here is dominated by a reststrahlen band with a minimum at 1392 cm−1. The fundamental modes ν2 and ν4 also yield inverted bands with minima at 835 and 679 cm−1, respectively. Additionally, the ν1 is weakly present with a derivative shape (1047 cm−1). The typical OH band at 3540 cm−1 and the combination band at ~2415 cm−1 are slightly shifted but still visible. The ν1 + ν4 combination band overlaps with the C_O stretching mode of the oil. The resulting band at 1748 cm−1 represents a combination of both. The pure powder is an optically rough surface. The rough surface, together with the 45° collection geometry of the instrument, promotes diffuse reflection and limits specular reflection. Hence, the powder spectrum is free of reststrahlen bands and the peaks are sharp and less derivative. However, when the grain size is coarser than the MIR light, inverted bands can still appear. Binding the powder in oil optically flattens the surface. Therefore, the specular component dominates, and even with the instrument's 45° collection geometry, reststrahlen bands and derivative-shaped bands occur. This outlines the role of the surface in the spectral appearance of the material. In the case of paintings, the latter scenario can be assumed, because pigments are mixed
S. Steger et al. / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 195 (2018) 103–112
107
Fig. 4. DRIFT spectra of basic lead white: (I) powder, (II) lead white/linseed oil mock-up, (III) lead white spectrum from Heinrich Campendonk, ‘Bäume’, (IV) lead white spectrum from Heinrich Campendonk, ‘Zwei Frauen am Tisch’. RSB = reststrahlen band, ■ = binder (oil).
Fig. 2. Heinrich Campendonk, ‘Bäume’, 1946: reverse side (a); detail of the reverse side revealing the measured spot (marked with ×) (b). All images courtesy of Simone Bretz. ©VG Bildkunst Bonn, 2017.
with binders. However, we found both cases in reverse glass paintings. The lead white spectrum of Heinrich Campendonk's painting ‘Bäume’ (Fig. 4 III) reveals bands at 681 (ν4), 1045 (ν1), 1432 (ν3) and 3538 cm−1 (OH stretching), while the ν2 and the 2ν2 + ν4 (and/or ν1 + ν3) combination mode at 835 and ~2410 cm−1, respectively, are missing. The ν1 + ν4 combination band at 1730 cm−1 overlaps with the C_O stretching mode of the oil binder. In the measured area, Campendonk spread fine bronze particles on the paint layer [1]. Therefore, the surface appears rougher, the specular component is minimized and no reststrahlen bands can be observed anymore. We also found lead white in different areas in Campendonk's painting ‘Zwei Frauen am Tisch’ (Fig. 1). The white zone yields a distorted spectrum, which is dominated by reststrahlen bands with minima at 680 (ν4) and 1399
Fig. 3. Edith Campendonk-van Leckwyck, ‘Lofoten’, 1933: reverse side (a); reverse side, detailed photograph of the red sailing boat (b); Marianne Uhlenhuth, ‘Ohne Titel’, 1954: reverse side (c). Documentation of measurement spots (×). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.) © Simone Bretz.
108
S. Steger et al. / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 195 (2018) 103–112
Table 2 Selection of reflectance FTIR data of lead white (PbCO3·Pb(OH)2) with different set-ups and samples in the 4000–650 cm−1 range. Reference
Own data
Own data
Arrizabalaga et al. [25]
Miliani et al. [12]
Zaffino et al. [21]
Zaffino et al. [21]
Technique Sample O\ \H stretching ν1 + ν3 combination ν1 + ν4 combination Antisymmetric stretching ν3 Symmetric stretching ν1 Out-of-plane bending ν2 In-plane bending ν4
DRIFTS, raw data Powder 3536 ~2427 1735 RSB 1398 1044 835 686
DRIFTS, raw data In linseed oil 3540 2411 1748 RSB 1392 1047 RSB 835 RSB 679
DRIFTS, KM In-situ, building material 3542
FORS, raw data Powder 3535 ~2428 1740 RSB 1390–1500 RSB ~1000
Ext. refl. KK In gum Arabic 3537 2384 1725 1447 1047
Ext. refl. KK In egg white 3522 2380 1725 1441 1042
684
682
1737 RSB 1400–1500
682
Band assignment: cf. [12] and references therein. Technique: KM = Kubelka Munk transform; RSB = reststrahlen band; Ext. refl. = external reflection FTIR; KK = Kramers Kronig transform.
cm−1 (ν3) (Fig. 4 IV). The prominent OH stretching band at 3543 cm−1 and the sharp ν1 fundamental mode at 1047 cm−1 are observable, whereas the ν2 mode at 835 cm−1 is missing again. The 2ν2 + ν4 (and/or ν1 + ν3) combination band appears at 2430 cm−1. The ν1 + ν4 combination band at 1747 cm−1 overlaps with the C_O stretching mode of the binder. Comparing the lead white results with various studies (Table 2) we conclude that: (1) different set-ups yield similar results for a given material, (2) Kramers-Kronig corrections were successfully only when applied to external-reflection spectra, (3) different samples (i.e. pure powder, powder/binding media mixture, in-situ detection in artwork…) can lead to different results for the same material, (4) KKcorrected spectra cannot be directly correlated with raw data. Powdered gypsum yields inverted bands (reststrahlen effect) at 677 and ~1165 cm−1 (Fig. 5 I), which can be ascribed to the bending mode ν4 and the asymmetric stretching mode ν3, respectively. We conclude that the inverted bands from the powder originate from particles that are coarser than the IR wavelength. The symmetric stretching mode ν1 is present as a weak feature at 1004 cm−1 and is partly hidden by a sharp band at 1014 cm−1. This band corresponds to the IR spectrum of anhydrite (CaSO4) [56], which is common in naturally mined gypsum. Enhancing combination modes and overtones leads to significant bands at 2131 (ν1 + ν3 SO4) and 2235 cm−1 (bending and libration modes of H2O; ν2 + νL) and a weak spectral feature at 2009 cm−1 (2ν1 SO4) [12,46]. Additionally, the powder spectrum shows vibrational modes of the water molecules in the 3200–3600 cm−1 range (O\\H stretching), at 1621 and at 1684 cm−1 (O\\H bending) [56]. Please note that the bands at 876, 1793 and 2519 cm−1 originate from calcite (CaCO3) impurities [12]. The gypsum/binder mock-up yields different
3.1.2. Green Pigment The powder spectrum of phthalocyanine green (PG7) (Fig. 6 I) reveals various intense and sharp bands in the fingerprint region (1700– 650 cm−1), originating from fundamental vibrations. Detailed band assignment is reported elsewhere [57,58]. There are no signals that can be attributed to fillers (e.g. barium sulphate, chalk etc.). Comparing the
Fig. 5. DRIFT spectra of: powdered gypsum (I); gypsum/linseed oil mock-up (II); gypsum from Heinrich Campendonk's painting ‘Zwei Frauen am Tisch’ (III). RSB = reststrahlen band, ■ = binder (oil), ● = anhydrite, ★ = chalk.
Fig. 6. DRIFT spectra of phthalocyanine green (PG7) as powder (I) and from the colour charts: “Siegleechtgrün G” (II) and “Universalechtgrün 763” (III). ■ = binder, ● = kaolinite, ★ = chalk, + = barium sulphate.
spectral features (Fig. 5 II). Intense reststrahlen bands with minima at 690 (ν4), 1100 and 1205 (ν3) dominate the spectrum. The overtones and combination bands are completely lacking. The inverted band with a minimum at 3420 cm−1 can be ascribed to O\\H vibrational modes. However, the binder dominates the spectrum, yielding prominent inverted bands with minima at 1720, 2845 and 2914 cm−1 and a sharp band at 1467 cm−1. We found gypsum in a reddish area of Heinrich Campendonk's painting ‘Zwei Frauen am Tisch’ (Fig. 1d). Surprisingly, the spectrum is free of distortions such as derivative-shaped features or inverted bands (Fig. 5 III). The following bands originate from gypsum: 669 (ν4), 1005 (shoulder; ν1) 1099, 1180 (ν3), 1620, 1685 (O\\H bending), 2131 (ν1 + ν3 SO4), 2235 (ν2 + νL) and 3200– 3600 cm−1 (O\\H stretching). The binder can be classified as oil, yielding bands at 1460, 1710, 2851 and 2919 cm−1. XRF analysis (not presented here) showed that the reddish paint covers a tin foil and was executed with red iron oxides (ochre or synthetic hematite) mixed with gypsum. After refracting through the paint layer, the IR light is reflected from the foil surface and passes through the paint layer a second time. This transflection effect yields transmission-like spectra with intense fundamental modes and weaker combination and overtone bands. Rosi et al. reported the same effect from a metal colour that was covered by a yellow paint layer [52].
S. Steger et al. / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 195 (2018) 103–112
109
powder spectrum with the various PG7 spectra of the colour chart, we conclude that (1) spectrum 6 II shows several bands that are in good agreement with the powder spectrum, but the binder yields an intense band at 1749 cm−1, (2) the spectral quality decreases (in terms of PG7 bands) from spectrum I to spectrum III using the same measuring conditions, owing to sample parameters and (3) chalk (874, 1796 cm−1), barium sulphate (984 cm−1), kaolinite (913 cm−1) [12] and the binder (1740 cm−1) hamper spectrum III, so that hardly any signal can be ascribed to PG7. We conclude that PG7 yields a distinct spectrum as powder, but its identification with DRIFTS becomes difficult when it is mixed with fillers and extenders (e.g. barium sulphate, gypsum, chalk etc.). We discovered PG7 mixed with PR81 in Marianne Uhlenhuth's painting ‘Ohne Titel’, using Raman spectroscopy [59]. However, the resulting DRIFT spectrum reveals no signal that can be assigned to any organic pigment. The spectrum (Fig. 11) is reported in the Binding Media section, showing gum Arabic. 3.1.3. Yellow Pigment Pigment Yellow 1 (PY1) (introduced in 1910) was one of the first synthetic organic pigments widely used by artists. However, this monoazo yellow is no longer commercially available, having been replaced by more stable pigments. We measured two PY1 colour charts (Fig. 7), because we had no access to PY1 powder. The sample “Sieglegelb G” (spectrum I) yields the most intense bands and shows the fewest distortions. Several sharp bands appear at b1700 cm−1 (e.g. 776, 809, 954, 1143, 1182, 1275, 1300, 1347, 1498, 1519, 1567, 1606 and 1672 cm−1). The spectrum is distorted by the presence of the binder (1750 cm−1). In spectrum II (Universalgelb 7930), additional peaks arise from the presence of chalk (e.g. 715, 874, 1795 cm−1), barium sulphate (984 cm−1) and the binder (1741 cm−1). However, there are enough PY1 bands left at b1700 cm−1 to identify PY1 properly. We discovered PY1 in Marianne Uhlenhuth's painting ‘Ohne Titel’ (spectrum III). The bands are in good agreement with the spectra from the colour chart: 779, 813, 954, 1142, 1182, 1275, 1299, 1498, 1528, 1570, 1606, 1617 and 1672 cm−1 for PY1 and 877, 1793 cm−1 for chalk. However, it is difficult to distinguish the pigments from the monoazo yellow group (e.g. PY1, PY3, PY65) only by their IR spectra. The identification of PY1 becomes unambiguous using a combination of Raman and DRIFT analysis (Fig. 7). The Raman spectrum of the yellow area of Uhlenhuth's painting ‘Ohne Titel’ shows plenty bands which are in good agreement with the PY1 reference spectrum [67]. 3.1.4. Red Pigment Alizarin crimson (PR83) was introduced in about 1870 and started to replace natural madder lake. The powder (Fig. 8 I) yields bands at 840, 1110, 1156, 1191, 1277, 1299, 1350, 1364, 1470, 1530, 1590, and 1640 cm−1¸ which could originate from the colourant. The assignment of the bands at ~930 and 1021 cm−1 is not unambiguous. There are no signals in the area at N1700 cm−1, that can be ascribed to the pigment. Siegle's “Rotlack” (spectrum II) from the colour chart yields similar results. The bands at 839, 1156, 1187, 1270, 1288, 1347, 1361, 1465, 1525, 1588 and 1636 cm−1 are slightly shifted, but they are still in good agreement with the powder. The sharp signal at 984 cm−1 originates from barium sulphate. Moreover, we found PR83 in a red area of Edith Campendonk-van Leckwyck's painting ‘Lofoten’. The bands at 843, 1190, 1275, 1292, 1350, 1364, 1466 and ~1638 cm−1 correspond very well with the previous results from the powder and the colour chart. Additional signals from barium sulphate and the binder are observable at 984 and 1746 cm−1, respectively. We want to point out that we couldn't confirm the presence of PR83 in the painting by Raman spectroscopy. This is mainly owed to the low Raman sensitivity of PR83 and the technical limitations of our mobile spectrometer compared with lab instruments. However, the identification of PR83 with DRIFTS outlines the advantage of the complementary use of both techniques.
Fig. 7. a): DRIFT spectra of the PY1 colour charts ‘Sieglegelb G’ (I) and “Universalgelb 7930” (II) compared with the PY1 spectrum of Uhlenhuth's painting ‘Ohne Titel’ (III). ■ = binder, ★ = chalk, + = barium sulphate. b) Raman spectrum of PY1 from Uhlenhuth's painting ‘Ohne Titel’ along a PY1 reference spectrum (grey) [67].
3.2. Binding Media Various methods can be used for the non-invasive identification of pigments and dyes in artworks. However, there are few options for the non-invasive characterisation of binding media. Portable reflectance FTIR spectroscopy provides useful information for classifying the binder and was successfully used in all three configurations for this purpose e.g. [14,42,49]. The main disadvantages and challenges are (1) the spectral contribution of pigments and fillers can hide the bands of the binder, (2) the binder's spectrum can be affected by spectral distortions like reststrahlen bands or derivative band shapes, depending on the optical properties of the sample and the surface morphology, (3) paint systems with mixed binding media often yield poorly resolved bands due to band overlapping and (4) ageing and pigment-binder reactions may also affect the proper identification of binding media. However, to get a comprehensive characterisation of the binding media, micro-sampling is needed. GC-MS analysis coupled with other methods (FTIR in transmission, AAA etc.) e.g. [10,61,60], immunological methods such as
110
S. Steger et al. / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 195 (2018) 103–112
Fig. 8. DRIFT spectra of PR83: powder (I), colour chart ‘Rotlack’ (II), red area of Edith Campendonk-van Leckwyck's painting ‘Lofoten’. ■ = binder, + = barium sulphate.
enzyme-linked immunosorbent assay (ELISA) e.g. [65,64] or proteomic techniques [66] can be employed. The DRIFT spectrum of linseed oil (Fig. 9 III) reveals inverted bands with minima at 718 (CH2 rocking), 1240 (C\\O stretching), 1695 (C_O stretching), 2845 and 2914 cm−1 (CH2 stretching). The bands at 1481 and ~2960 cm−1 can be ascribed to CH2 scissoring and CH3 stretching movements, respectively (24 and references therein). The negative bands were corrected after successfully applying the KK transformation (Fig. 9 II). The resulting bands at 731 (CH2 rocking), 1470 (CH2 scissoring), 1740 (C_O stretching), 2852 & 2925 (CH2 stretching) and 2955 cm−1 (CH3 stretching) are in good agreement with the transmission spectrum (I). However, the KK transform may give inconclusive results for the 1000–1400 cm−1 region. The proteinaceous binder (whole egg) yields inverted bands with minima at 2847, 2914 cm−1 (C\\H stretching) and 1631, 1531 cm−1 (amide I, amide II), respectively (Fig. 10 II). Additionally, signals at 1752 cm−1 (C_O stretching) and 1465 cm−1 (amide III) are visible. Mastic is the natural resin of Pistacia lentiscus and has been widely used as an additive or varnish material. The spectrum (Fig. 10 III) shows intense bands at 1390, 1477, 1720 (C_O stretching), 2985
Fig. 9. Raw DRIFT spectrum of linseed oil (III). KK-corrected spectrum (II), transmission spectrum of linseed oil (I). RSB = reststrahlen band.
Fig. 10. DRIFT spectra of various binders: polyvinyl acetate (I), egg (II), mastic (III), bees wax (IV). RSB = reststrahlen band.
cm−1 (C\\H stretching) and derivative shaped bands at 1277 cm−1 and 2886 (C\\H stretching). Bees wax (Fig. 10 IV) yields sharp but inverted bands with minima at 715 (CH2 rocking), 1159, 1461 (CH2 bending), 1730 (C_O stretching), 2843 and 2912 cm−1 (CH2 stretching) [50]. The spectrum of polyvinyl acetate (Fig. 10 I) shows intense inverted bands with minima at 1015 (C\\C stretching), 1233 (C\\O\\C stretching), 1360 (C\\H bending) and 1730 cm−1 (C_O asymmetric stretching) [42]. Additionally, there are weaker signals at 950 (C\\O stretching), 1446 (C\\H bending), 2910–3030 (C\\H stretching), ~3450, 3540 and 3610 cm−1. Gum Arabic (Fig. 11 III) yields an intense inverted band at 1020 cm− 1 (C\\O stretching) with shoulders at 982 cm−1 and 1055 cm−1. Bands at 1458, 1660 (C_O stretching), 2948 and 3550 cm−1 (C\\H stretching) are visible [62]. However, the spectrum deviates from the data in the literature [25], and its interpretation isn't fully clear. The inverted signals with minima at 1592 and ~3200 cm−1 could also be ascribed to the carbonyl stretching and the C\\H stretching band, respectively. The KK transformation properly corrects the bands at 1605 and 3430 cm−1 (Fig. 10 II), but it yields inconclusive results for the C\\O
Fig. 11. DRIFT spectrum of gum Arabic from Marianne Uhlenhuth's painting ‘Ohne Titel’ (IV). Raw (III) and KK-corrected DRIFT (II) spectrum of gum Arabic in comparison with a transmission reference spectrum (I). RSB = reststrahlen band.
S. Steger et al. / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 195 (2018) 103–112
stretching band (resulting in a doublet at 1107 and 1151 cm−1). Earlier studies had shown that structured reststrahlen bands (i.e. band splitting) cannot be corrected by KK transform [41,63]. Fig. 11 II shows the DRIFT spectrum from a violet area of Marianne Uhlenhuth's painting ‘Ohne Titel’. The structured reststrahlen band with the minimum at 1020 cm−1 (shoulders at 982 and 1053 cm−1), the broad bands at 1654 and 3540 cm−1 and the spectral feature at 2945 cm−1 correspond very well with the gum Arabic reference. No signal from the pigments can be seen in the spectrum, although PG7 and pigment red 81 were detected by means of Raman spectroscopy [59]. The transmission spectrum (I) reveals bands at 1030, 1073, 1427, 1630, 2924 and 3435 cm−1. 4. Conclusions The results show the strong influence of the surface on the spectral appearance of gypsum and lead white. Powders yield almost pure diffuse-reflection spectra when the grain size is in the range of the IR wavelength (2,5–15 μm). In powder/binder mock-ups, derivativeshaped features and inverted bands occur for both pigment and binder, owing to the strong influence of the specular-reflected light. However, knowing both cases can aid in correctly ascribing the band in a sample spectrum. Inverted bands can be corrected with the KK transformation when surface reflection is dominant. However, this yields inconclusive results when both specular and diffuse reflections are active. The evaluation of the analytical potential of the mobile DRIFTS device for synthetic organic pigments yields mainly positive results. We managed to get good reference spectra of PG7, PY1 and PR83 from powders and from the historical colour charts. Moreover, using DRIFTS, we successfully identified PY1 in Marianne Uhlenhuth's painting ‘Ohne Titel’ and PR83 in Edith Campendonk-van Leckwyck's painting ‘Lofoten’. However, spectral contributions of fillers and extenders are the main causes of band overlapping and may hamper correct interpretation. The results show the high potential of mobile DRIFTS for the analysis of colourants and binders in 20th-century reverse glass paintings. Unexpected oddities like undistorted spectra and a pseudo-transmission spectrum of gypsum were reported. Binding media can be classified in the best cases, when band overlapping and spectral distortions do not hamper the spectrum. Transporting precious and fragile objects to the lab is often not feasible. Therefore, in-situ, non-invasive methods like the handheld DRIFTS device are necessary to analyse colourants and binders. However, this is limited to surface analysis. The device cannot measure through glass and when the paint layer is covered by paper or metal foils, the painting cannot be used for reflectance FTIR studies. Due to the reverse paint stratigraphy in reverse glass paintings, the measured layer is always the backmost one. In case of multi-layered colour systems, it is possible that the areas visible from the front are not accessible with this technique. This becomes crucial when the painting was recently restored. Sometimes the painted areas need to be conserved, using synthetic waxes and resins, because delamination of the paint layers is a common disfigurement. These areas can be still measured with DRIFTS, which, however, would provide information only on the fixatives. Therefore, micro-invasive methods need to be used as a supplementary approach to obtain more specific information. Conflict of Interest The authors declare that they have no competing interest. Acknowledgements We would like to thank Ira Rabin, Olivier Bonnerot, Gisela Geiger, Diana Oesterle, Christoph Steuer and Emmanuel Kindzorra. We are grateful to Clarimma Sessa and Patrick Dietemann for their valuable comments on the manuscript. We thank the Museum Penzberg — Sammlung Campendonk for the friendly cooperation. The project
111
“Hinterglasmalerei als Technik der Klassischen Moderne 1905–1955” is funded by the Volkswagen-Stiftung, “Forschung in Museen” reference 89921. References [1] G. Geiger, S. Bretz, Heinrich Campendonk, Die Hinterglasbilder, Wienand Verlag, Cologne, 2017. [2] E. Wessels, Die Hinterglasmalerei: Anleitungen zur Herstellung von Malereien hinter ober unter Glas, sowie Glasmalerei-Imitation, Glas-Vergoldung und dergleichenin German. Esslingen 1913. [3] K. Wehlte, Werkstoffe und Techniken der Malereiin German. Ravensburg 1967. [4] C.J. Stahl, Dekorative Glasmalerei (Unterglasmalerei und Malen auf Glas) in ihrem Gesamtumfange dargestelltin German. Wien/Leipzig 1915 (= A. Hartlebens chemisch-technische Bibliothek, Bd. 354). [5] O. Freytag, Hinterglasmalerei. Ihre künstlerische Eigenart und Arbeitsweise in Vergangenheit und Gegenwartin German. Ravensburg 1937. [6] S. Bretz, Hinterglasmalerei, …die Farben leuchten so klar und rein. Maltechnik, Geschichte, Restaurierungin German. München 2013. [7] S. Bretz, U. Baumer, H. Stege, J. von Miller, D. von Kerssenbrock-Krosig, A German house altar from the sixteenth century: conservation and research of reverse paintings on glass, Stud. Conserv. 53 (4) (2009) 209–224. [8] U. Baumer, P. Dietemann, J. Koller, Identification of resinous materials on 16th and 17th century reverse-glass objects by gas chromatography/mass spectrometry, Int. J. Mass Spectrom. 284 (2009) 131–141. [9] U. Baumer, P. Dietemann, Identification and differentiation of dragon's blood in works of art using gas chromatography/mass spectrometry, Anal. Bioanal. Chem. 397 (2010) 1363–1376. [10] U. Baumer, I. Fiedler, S. Bretz, H.J. Ranz, P. Dietemann, Decorative reverse-painted glass objects from the fourteenth to twentieth centuries: an overview of the binding media, IIC Preprints Vienna Congress, Studies in Conservation, Supplement 1, 57, 2012, pp. 9–18. [11] O. Hahn, S. Bretz, C. Hagnau, H.-J. Ranz, T. Wolff, Pigments, dyes, and black enamel — the colorants of reverse paintings on glass, Archaeol. Anthropol. Sci. 1 (2009) 263–271. [12] C. Miliani, F. Rosi, F. Daveri, B.G. Brunetti, Reflection infrared spectroscopy for the non-invasive in-situ study of artists' pigments, Appl. Phys. A Mater. Sci. Process. 106 (2012) 295–307. [13] F. Rosi, A. Federici, B.G. Brunetti, A. Sgamellotti, S. Clementi, C. Miliani, Multivariate chemical mapping of pigments and binders in easel painting cross-sections by micro IR reflection spectroscopy, Anal. Bioanal. Chem. 399 (2011) 3133–3145. [14] M. Manfredi, E. Barberis, A. Rava, E. Robotti, F. Gosetti, E. Marengo, Portable diffuse reflectance infrared Fourier transform (DRIFT) technique for the non-invasive identification of canvas ground: IR spectra reference collection, Anal. Methods 7 (2015) 2313–2322. [15] B.G. Brunetti, C. Miliani, F. Rosi, B. Doherty, L. Monico, A. Romani, A. Sgamellotti, Non-invasive investigations of paintings by portable instrumentation: the MOLAB experience, Top. Curr. Chem. 374 (10) (2016) 1–35. [16] L. Rampazzi, V. Brunello, C. Corti, E. Lissoni, Non-invasive techniques for revealing the palette of the Romantic painter Francesco Hayez, Spectrochim. Acta A 176 (2017) 142–154. [17] C. Miliani, F. Rosi, I. Borgia, P. Benedetti, B.G. Brunetti, A. Sgamellotti, Fiber-optic Fourier Transform mid-infrared reflectance spectroscopy: a suitable technique for insitu studies of mural paintings, Appl. Spectrosc. 61 (2007) 293–299. [18] M. Irazola, M. Olivares, K. Castro, M. Maguregui, I. Martinez-Arkarazo, J.M. Madariaga, In-situ Raman spectroscopy analysis combined with Raman and SEMEDS imaging to assess the conservation state of 16th century wall paintings, J. Raman Spectrosc. 43 (2012) 1676–1684. [19] J.M. Madariaga, M. Maguregui, K. Castro, U. Knuutinen, I. Martínez-Arkarazo, Portable Raman, DRIFTS, and XRF analysis to diagnose the conservation state of two wall painting panels from Pompeii deposited in the Naples National Archaeological Museum (Italy), Appl. Spectrosc. 70 (2016) 137–146. [20] B. Doherty, A. Daveri, C. Clementi, A. Romani, S. Bioletti, B. Brunetti, A. Sgamellotti, C. Miliani, The Book of Kells: a non-invasive MOLAB investigation by complementary spectroscopic techniques, Spectrochim. Acta A 115 (2013) 330–336. [21] C. Zaffino, V. Guglielmi, S. Faraone, A. Vinaccia, S. Bruni, Exploiting external reflection FTIR spectroscopy for the in-situ identification of pigments and binders in illuminated manuscripts. Brochantite and posnjakite as a case study, Spectrochim. Acta A 136 ( (2015) 1076–1085. [22] M. Vagnini, F. Gabrieli, A. Daveri, D. Sali, Handheld new technology Raman and portable FT-IR spectrometers as complementary tools for the in-situ identification of organic materials in modern art, Spectrochim. Acta A 176 (2017) 174–182. [23] C. Invernizzi, A. Daveri, T. Rovetta, M. Vagnini, M. Licchelli, F. Cacciatori, M. Malagodi, A multi-analytical non-invasive approach to violin materials: the case of Antonio Stradivari “Hellier” (1679), Microchem. J. 124 (2016) 743–750. [24] C. Invernizzi, A. Daveri, M. Vagnini, M. Malagodi, Non-invasive identification of organic materials in historical stringed musical instruments by reflection infrared spectroscopy: a methodological approach, Anal. Bioanal. Chem. (2017)https://doi. org/10.1007/s00216-017-0296-8. [25] I. Arrizabalaga, O. Gomez-Laserna, J. Aramendia, G. Arana, J.M. Madariaga, Applicability of a diffuse reflectance infrared Fourier transform handheld spectrometer to perform in analyses on cultural heritage materials, Spectrochim. Acta A 129 (2014) 259–267. [26] I. Arrizabalaga, O. Gomez-Laserna, J.A. Carrero, J. Bustamante, A. Rodriguez, G. Arana, J.M. Madariaga, Diffuse reflectance FTIR database for the interpretation of the
112
[27]
[28] [29] [30] [31]
[32]
[33]
[34] [35] [36]
[37] [38]
[39]
[40] [41]
[42]
[43]
[44]
[45]
[46]
[47]
S. Steger et al. / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 195 (2018) 103–112 spectra obtained with a handheld device on built heritage materials, Anal. Methods 7 (2015) 1061–1070. O. Gomez-Laserna, I. Arrizabalaga, N. Prieto-Taboada, M.A. Olazabal, G. Arana, J.M. Madariaga, In-situ DRIFT, Raman, and XRF implementation in a multianalytical methodology to diagnose the impact suffered by built heritage in urban atmospheres, Anal. Bioanal. Chem. 407 (2015) 5635–5647. P. Griffiths, J.A. De Haseth, Fourier Transform Infrared Spectrometry, 2nd ed. Wiley, New York, 2007. E.H. Korte, A. Röseler, Infrared reststrahlen revisited: commonly disregarded optical details related to n b1, Anal. Bioanal. Chem. 382 (2005) 1987–1992. A. Röseler, E.H. Korte, Reflection anomalies related to n = 1, Vib. Spectrosc. 43 (2007) 111–115. S. Bruni, F. Cariati, F. Casadio, L. Toniolo, Spectrochemical characterization by microFTIR spectroscopy of blue pigments in different polychrome works of art, Vib. Spectrosc. 20 (1999) 15–25. W. Vetter, M. Schreiner, Pigment-binding media systems – comparison of non-invasive in-situ reflection FTIR with transmission FTIR microscopy, e-Preservation Sci. 8 (2011) 10–22. L. Monico, F. Rosi, C. Miliani, A. Daveri, B.G. Brunetti, Non-invasive identification of metal-oxalate complexes on polychrome artwork surfaces by reflection mid-infrared spectroscopy, Spectrochim. Acta A 116 (2013) 270–280. D. Buti, F. Rosi, B.G. Brunetti, C. Miliani, In-situ identification of copper-based green pigments, Anal. Bioanal. Chem. 405 (2013) 2699–2711. M.R. Derrick, D. Stulik, J.M. Landry, Infrared Spectroscopy, Getty Conservation Institute, Los Angeles, 1999. E. Van Nimmen, K. De Clerck, I. Verschuren, K. Gellynck, T. Gheysens, J. Mertens, L. Van Langenhove, FT-IR spectroscopy of spider and silkworm silks Part I. Different sampling techniques, Vib. Spectrosc. 46 (2008) 63–68. M. Milosevic, S.L. Berets, A review of FT-IR diffuse reflection sampling considerations, Appl. Spectrosc. Rev. 37 (2002) 347–364. M. Fabbri, M. Picollo, S. Porcinai, M. Bacci, Mid-infrared fiber-optics reflectance spectroscopy: a non-invasive technique for remote analysis of painted layers. Part I: technical setup, Appl. Spectrosc. 55 (2001) 420–427. M. Fabbri, M. Picollo, S. Porcinai, M. Bacci, Mid-infrared fiber-optics reflectance spectroscopy: a non-invasive technique for remote analysis of painted layers. Part II: statistical analysis of spectra, Appl. Spectrosc. 55 (2001) 428–433. C. Miliani, F. Rosi, B.G. Brunetti, A. Sgamellotti, In-situ non-invasive study of artworks: the MOLAB multi-technique approach, Acc. Chem. Res. 43 (2010) 728–738. C. Ricci, C. Miliani, B.G. Brunetti, A. Sgamellotti, Non-invasive identification of surface materials on marble artifacts with fiber optic mid-FTIR reflectance spectroscopy, Talanta 69 (2006) 1221–1226. F. Rosi, A. Daveri, P. Moretti, B.G. Brunetti, C. Miliani, Interpretation of mid and nearinfrared reflection properties of synthetic polymer paints for the non-invasive assessment of binding media in twentieth-century pictorial artworks, Microchem. J. 124 (2016) 898–908. B. Doherty, M. Vagnini, K. Dufourmantelle, A. Sgamellotti, B. Brunetti, C. Miliani, A vibrational spectroscopic and principal component analysis of triarylmethane dyes by comparative laboratory and portable instrumentation, Spectrochim. Acta A 121 (2014) 292–305. J. Aramendia, L. Gomez-Nubla, I. Arrizabalaga, N. Prieto-Taboada, K. Castro, J.M. Madariaga, Multianalytical approach to study the dissolution process of weathering steel: the role of urban pollution, Corros. Sci. 76 (2013) 154–162. M. Manfredi, E. Barberis, M. Aceto, E. Marengo, Non-invasive characterization of colorants by portable diffuse reflectance infrared Fourier transform (DRIFT) spectroscopy and chemometrics, Spectrochim. Acta A 181 (2017) 171–179. F. Rosi, A. Daveri, B. Doherty, S. Nazzareni, B.G. Brunetti, A. Sgamellotti, C. Miliani, On the use of overtone and combination bands for the analysis of the CaSO4–H2O system by mid-infrared reflection spectroscopy, Appl. Spectrosc. 64 (2010) 956–963. F. Rosi, A. Daveri, C. Miliani, G. Verri, P. Benedetti, F. Piqué, B.G. Brunetti, A. Sgamellotti, Non-invasive identification of organic materials in wall paintings by fiber optic reflectance infrared spectroscopy: a statistical multivariate approach, Anal. Bioanal. Chem. 395 (2009) 2097–2106.
[48] R. Ploeger, D. Scalarone, O. Chiantore, Noninvasive characterization of binding media on painted glass magic lantern plates using mid-infrared fibre-optic reflectance spectroscopy, J. Cult. Herit. 11 (2010) 35–41. [49] R. Ploeger, O. Chiantore, D. Scalarone, T. Poli, Mid-infrared fiber-optic reflection spectroscopy (FORS) analysis of artists' alkyd paints on different supports, Appl. Spectrosc. 65 (2011) 429–435. [50] T. Poli, O. Chiantore, M. Nervo, A. Piccirillo, Mid-IR fiber-optic reflectance spectroscopy for identifying the finish on wooden furniture, Anal. Bioanal. Chem. 400 (2011) 1161–1171. [51] C.E. Silva, L.P. Silva, H.G.M. Edwards, L.F. De Oliveira, Diffuse reflection FTIR spectral database of dyes and pigments, Anal. Bioanal. Chem. 386 (2006) 2183–2191. [52] F. Rosi, C. Grazie, R. Fontana, F. Gabrieli, L. Pensabene Buemi, E. Pamapaloni, A. Romani, C. Stringari, C. Miliani, Disclosing Jackson Pollock's palette in Alchemy (1947) by non-invasive spectroscopies, Herit. Sci. 4 (2016) 18–30. [53] C. Sessa, H. Bagam, J.F. Garcia, Evaluation of mid-IR fibre optic reflectance: detection limit, reproducibility and binary mixture discrimination, Spectrochim. Acta A Mol. Biomol. Spectrosc. 115 (2013) 617–628. [54] C. Sessa, H. Bagam, J.F. Garcia, Influence of composition and roughness on the pigment mapping of paintings using mid-infrared fiber optics reflectance spectroscopy (mid-IR FORS) and multivariate calibration, Anal. Bioanal. Chem. 406 (2014) 6735–6747. [55] M.H. Brooker, S. Sunder, P. Taylor, Infrared and Raman spectra and X-ray diffraction studies of solid lead(1I) carbonates, Can. J. Chem. 61 (1983) 494. [56] T.L. Hughes, C.M. Methven, T.G.J. Jones, S.E. Pelham, P. Fletcher, C. Hall, Determining cement composition by Fourier transform infrared spectroscopy, Adv. Cem. Based Mater. 2 (1995) 91–104. [57] B. Barszcz, A. Bogucki, A. Biadasz, B. Bursa, D. Wróbel, A. Graja, Molecular orientation and spectral investigations of Langmuir–Blodgett films of selected copper phthalocyanines, J. Photochem. Photobiol. A Chem. 218 (2011) 48–55. [58] D. Li, Z. Peng, L. Deng, Y. Shen, Y. Zhao, Theoretical studies on molecular structure and vibrational spectra of copper phthalocyanine, Vib. Spectrosc. 39 (2005) 191–199. [59] S. Steger, O. Hahn, I. Rabin, H. Stege, Non-invasive spectroscopic study on a modern reverse glass painting: Marianne Uhlenhuth's “Ohne Titel, 1954”, Colloquium Spectroscopicum Internationale XL – IX Euro-Mediterranean Symposium on LIBS, Pisa, Italy 2017, p. 416. [60] T.J. Learner, Analysis of Modern Paints, Research in Conservation, Getty Conservation Institute, Los Angeles, 2004. [61] J. Peris-Vicente, U. Baumer, H. Stege, K. Lutzenberger, J.V. Gimeno Adelantado, Characterization of commercial synthetic resins by pyrolysis–gas chromatography/mass spectrometry: application to modern art and conservation, Anal. Chem. 81 (2009) 3180–3187. [62] D.N. Williams, K. Gold, T.R.P. Holoman, S.H. Ehrman, O.C. Wilson, Surface modification of magnetic nanoparticles using gum Arabic, J. Nanopart. Res. 8 (2006) 749–753. [63] T. Poli, A. Elia, O. Chiantore, Surface finishes and materials: fiber-optic reflectance spectroscopy (FORS) problems in cultural heritage diagnostics, e-Preservation Sci. 6 (2009) 174–179. [64] M. Palmieri, M. Vagnini, L. Pitzurra, P. Rocchi, B.G. Brunetti, A. Sgamellotti, L. Cartechini, Development of an analytical protocol for a fast, sensitive and specific protein recognition in paintings by enzyme-linked immunosorbent assay (ELISA), Anal. Bioanal. Chem. 399 (2011) 3011–3023. [65] M. Palmieri, M. Vagnini, L. Pitzurra, B.G. Brunetti, L. Cartechini, Identification of animal glue and hen-egg yolk in paintings by use of enzyme-linked immunosorbent assay (ELISA), Anal. Bioanal. Chem. 405 (2013) 6365–6371. [66] S. Dallongeville, N. Garnier, C. Rolando, C. Tokarski, Proteins in art, archaeology, and paleontology: from detection to identification, Chem. Rev. 116 (2016) 2–79. [67] W. Fremout, S. Saverwyns, Identification of synthetic organic pigments: the role of a comprehensive digital Raman spectral library, J. Raman Spectrosc. 43 (2012) 1536–1544.
