Revealing the Nature and Distribution of Metal Carboxylates in

Dec 19, 2016 - Burnstock , A. and van der Berg , K. J. In Issues in Contemporary Oil Paint; van den Berg , K. J. , Burnstock , A. , de Keijzer , M. , ...
0 downloads 20 Views 2MB Size
Download PDF
Article pubs.acs.org/ac

Revealing the Nature and Distribution of Metal Carboxylates in Jackson Pollock’s Alchemy (1947) by Micro-Attenuated Total Reflection FT-IR Spectroscopic Imaging Francesca Gabrieli,† Francesca Rosi,*,† Alessandra Vichi,‡ Laura Cartechini,† Luciano Pensabene Buemi,§ Sergei G. Kazarian,*,‡ and Costanza Miliani† †

CNR-ISTM, Istituto di Science e Tecnologie Molecolari, Perugia, Italy Department of Chemical Engineering, Imperial College London, South Kensington Campus, London, United Kingdom § Peggy Guggenheim Collection, Venice, Italy ‡

S Supporting Information *

ABSTRACT: Protrusions, efflorescence, delamination, and opacity decreasing are severe degradation phenomena affecting oil paints with zinc oxide, one of the most common white pigments of the 20th century. Responsible for these dramatic alterations are the Zn carboxylates (also known as Zn soaps) originated by the interaction of the pigment and the fatty acids resulting from the hydrolysis of glycerides in the oil binding medium. Despite their widespread occurrence in paintings and the growing interest of the scientific community, the process of formation and evolution of Zn soaps is not yet fully understood. In this study micro-attenuated total reflection (ATR)−FT-IR spectroscopic imaging was required for the investigation at the microscale level of the nature and distribution of Zn soaps in the painting Alchemy by J. Pollock (1947, Peggy Guggenheim Collection, Venice) and for comparison with artificially aged model samples. For both actual samples and models, the role of AlSt(OH)2, a jellifying agent commonly added in 20th century paint tube formulations, proved decisive for the formation of zinc stearate-like (ZnSt2) soaps. It was observed that ZnSt2-like soaps first form around the added AlSt(OH)2 particles and then eventually grow within the whole painting stratigraphy as irregularly shaped particles. In some of the Alchemy samples, and diversely from the models, a peculiar distribution of ZnSt2 aggregates arranged as rounded and larger particles was also documented. Notably, in one of these samples, larger agglomerates of ZnSt2 expanding toward the support of the painting were observed and interpreted as the early stage of the formation of internal protrusions. Micro-ATR−FT-IR spectroscopic imaging, thanks to a very high chemical specificity combined with high spatial resolution, was proved to give valuable information for assessing the conservation state of irreplaceable 20th century oil paintings, revealing the chemical distribution of Zn soaps within the paint stratigraphy before their effect becomes disruptive.

A

their cohesion and (ii) appear on the surface of the painting as whitish protrusions, thus altering the chromatic properties of the surface.2−4 Modern 20th century oil formulations typically contain traditional linseed oil (eventually mixed with untraditional less yellowing oils such as sunflower or safflower) and a variety of newly introduced additives such as driers and dispersion agents.1 As for the color palette, the 20th century art has been characterized by the use of a number of synthetic inorganic pigments whose superior optical properties attracted artists since the end of the previous century (e.g., lead chromate, chromium oxides, cobalt oxides, cadmium sulfides). As a result,

lthough a large choice of new synthetic polymers was available as art materials, many 20th century artists continued to adopt drying oils as binding media for their artworks, mainly attracted by their good working and optical properties.1 Nowadays in the field of painting conservation, concerns exist regarding if and eventually to what extent modern oil formulations could be affected by saponification, a severe degradation process which has been detected in traditional oil paints containing lead- and copper-based pigments (e.g., PbCO3Pb(OH)2, CuCO3Cu(OH)2).2,3 The saponification process takes place during the aging of the lipid binder through the reaction of the free fatty acids and the metal cation of the pigment leading to the formation of metal carboxylates. The occurrence of metal soaps can affect both the mechanical and the aesthetical properties of a painting since they can (i) aggregate inside/between paint layers affecting © XXXX American Chemical Society

Received: October 17, 2016 Accepted: December 19, 2016 Published: December 19, 2016 A

DOI: 10.1021/acs.analchem.6b04065 Anal. Chem. XXXX, XXX, XXX−XXX

Article

Analytical Chemistry

Pollock or are they rather the result of an alteration process? (ii) Which is their spatial distribution within the paint stratigraphy? (iii) Is there any tendency to form protrusions and/or detachment of the paint layers? In this work we exploit the high spatial resolution, molecular specificity, and sampling versatility of micro-attenuated total reflection (ATR)−FT-IR spectroscopic imaging to answer these questions.8 The advantages of this technique for investigating the multilayered structure of painting cross sections have been reported only in a few number of papers.9,10 Additional literature concerning the detection of metal carboxylates by micro-FT-IR includes the use of ATR mapping,11,12 micro-transmission measurements performed on thin sections, as well as external reflection FT-IR spectroscopy on cross sections.13,14 Micro-ATR−FT-IR imaging is the method of choice for facing the relevant questions previously mentioned for several reasons: (i) higher spatial resolution achievable with the same conventional IR source with respect to transmission and external reflection modes, (ii) faster acquisition time reachable by the use of a 2D infrared array detector compared with a single element detector, (iii) better signal-to-noise ratio as well as easier spectral interpretation with respect to micro-FT-IR in external reflection mode, and (iv) easier sample preparation steps as the necessity of preparing thin sections to work in transmission mode. Prior to the micro-ATR−FT-IR imaging investigation of the samples from Alchemy, some paint models were artificially aged with the aim of reproducing over a shorter time scale the alterations encountered in the modern painting and for acquiring information about their molecular composition. The aged and unaged paint models made of ZnO mixed with linseed oil and with the addition of aluminum stearate were analyzed in order to simulate the effect of a jellifying agent commonly added in 20th century paint tube formulations, which has been proved to play a fundamental role in the formation of zinc soaps.16

because of the compositional similarities with traditional oil paints, these paint formulations have been expected to give saponification as well but not necessarily with the same mechanism. The substitution of toxic lead white (PbCO3Pb(OH)2) with zinc white (ZnO) first and titanium white (TiO2) later, which began in the early 20th century, is of particular relevance. The behavior of ZnO in oil-based paints gives a particular cause for concern since in a number of modern paintings containing zinc white (used both as pigment and as filler in paints and preparation layers) protrusions, detaching, delamination, and opacity-decreasing phenomena have been documented linked to the saponification process.5 Recently, within a noninvasive analytical study of Pollock’s masterpiece Alchemy (1947, Peggy Guggenheim Collection), the presence of zinc carboxylates has been pointed out by external reflection FT-IR spectroscopy in some of the drying oil-based paints (Figure 1).6 Complementary investigation by

■

EXPERIMENTAL SECTION Methods. Micro-ATR−FT-IR Imaging. Micro-ATR−FT-IR imaging measurements were performed using a 670 FT-IR spectrometer operating in continuous scan mode attached to a UMA 600 IR microscope (Agilent) equipped with a 2D focal plane array (FPA) detector (Santa Barbara, CA, U.S.A.). A 64 pixel × 64 pixel array of the FPA detector was selected as field of view (FOV) allowing 4096 FT-IR spectra to be simultaneously collected. The IR microscope is fitted with a 15× Cassegrain objective and a hemispherical-shaped germanium slide-on crystal to perform micro-ATR measurements. The Ge crystal (refractive index n ∼ 4.00) enables a spatial resolution of about 4 μm in the fingerprint region. All spectra were collected in the mid-IR range of 4000−900 cm−1 by coadding 256 or 128 scans and with a spectral resolution of 8 cm−1. The size of the chemical images obtained with this setup is estimated to be of about 70 μm × 70 μm. Each chemical image is obtained by integrating the FT-IR band of interest in all 4096 spectra; the resulting values are then plotted simultaneously for each pixel of the FPA detector, thus representing the band spatial distribution as a false color scale (high intensity red, low intensity blue). All ATR−FT-IR spectra obtained in imaging measurements were extracted from a single pixel of the FPA detector. Conventional ATR, Transmission, and External Reflection FT-IR. Conventional ATR, transmission, and external reflection

Figure 1. (a) Visible image of Alchemy (1947, Peggy Guggenheim Collection, Venice). (b) Details of ZnO/TiO2 white (I), phthalocyanine blue (II), and cobalt phosphate violet (III) paints analyzed by mid-FT-IR reflection spectroscopy. (c) Spectra acquired on the paints shown in (b); sharp inverted signals at 1540, 1460, and 1400 cm−1 indicated the presence of zinc soaps of saturated long chain fatty acids (i.e., stearic acid and/or palmitic acid7). Spectrum acquired in transmission mode from a reference sample of Zn stearate (ZnSt2) is also reported. Asterisk (∗) indicates signals of Zn oxalates.

XRF and UV−vis spectroscopy revealed that these Zn soaps are strictly correlated to the zinc white pigment employed in mixtures with phthalocyanine and cobalt phosphate in blue and violet paints, respectively, as well as with anatase in white paints. Although at visual examination the painting appears in a rather good conservation state, the detection of zinc carboxylates have aroused some questions ultimately for improving the guidelines for its conservation treatment: (i) Are these Zn soaps already present in the artists’ tubes used by B

DOI: 10.1021/acs.analchem.6b04065 Anal. Chem. XXXX, XXX, XXX−XXX

Article

Analytical Chemistry

undefined zinc complex structure. In fact, it does not match for spectral position and shape any of the Zn soaps possibly generated by the interaction of Zn2+ with the free fatty acids released during the aging (namely, stearate/palmitate, azelate, oleate7,17). In the recent literature, two different interpretations of the origin of this anomalous broad and up-shifted carboxylate band have been reported, namely, (i) the functionalization of the ZnO surface by carboxylic acids18 and (ii) the formation of zinc ionomers.17,19 This band was notably the prevailing feature of the aged ZnO-oil paint model, and no other types of the most common Zn carboxylates (characterized by sharper and downshifted νas COO−) were observed (Figure 2a, black lines). Diversely, in the presence of AlSt(OH)2 (red lines, Figure 2a), a sharp band at 1538 cm−1 assigned to the νas mode of the group COO− in Zn stearate/palmitate (they have the same FT-IR profile)7 appeared already at ta = 0. With the aging time, this sharp band strongly increased, and at ta = 30 days, the corresponding symmetric stretching mode (νs) at about 1400 cm−1 became also visible. Only the 30 days aged ZnO-oil paint model, enriched with AlSt(OH)2, reproduced the reactivity observed in Alchemy evidenced by sharp-featured Zn carboxylate bands.6 These findings are also in agreement with previous studies15 in which it has been demonstrated that there is a strong correlation between the presence of Zn soaps and the use of AlSt(OH)2 in historical paint formulations naturally aged for a few decades. Surprisingly, we observed rather strong carboxylates features at 1538 cm−1 (νas ZnSt2) and 1400 cm−1 (νs ZnSt2) also on the unaged sample after merely one month from the preparation (hence, in a relatively young oil paint). With the aim of better investigation of the molecular modification happening before the film is touch dry, the early stages of the oil polymerization (polymerization time, tp) for ZnO-oil-AlSt(OH)2 paint were also considered. The spectral modifications starting at tp = 0, which correspond to the fresh paint (right after the mixing of the compounds in oil) up until 8 days of reaction are reported in Figure 2b. Many of the spectral modifications correspond to the oil polymerization process in agreement with the literature.20,21 Additionally, at tp = 0, the band assigned to the COO− antisymmetric stretching of the added AlSt(OH)2 at 1588 cm−1 was observed. After just 24 h, the νas at about 1540 cm−1 of the ZnSt2-like structure appeared. The presence of the two carboxylate species (Al and Zn) persisted for about the first 2 days with no evident variation in intensity until after 3 days when they became less visible because of the increasing broad band at about 1590 cm−1, as observed in the ZnO−oil system analyzed in the same way (Figure S-1, SI). The formation of ZnSt2 at the expense of ZnO during this stage of the polymerization and in the presence of AlSt(OH)2 may be explained by considering the reaction with the stearic acid present in commercial grades of Al stearate15 or available from its hydrolysis. With the temperature/moisture aging, the strong increase in the absorbance of spectral bands of ZnSt2 (Figure 2a) clearly demonstrates the central role of AlSt(OH)2 in promoting the soap formation in a quantity far above the one foreseen by stoichiometry of the hypothesized reactions.15,16 In order to provide a deeper insight into the behavior of the different carboxylate forms and the dynamics of formation of the new ones, the same models were studied as cross sections by micro-ATR−FT-IR spectroscopic imaging. The measured

FT-IR measurements were carried out by an ALPHA spectrometer (Bruker). A spectral resolution of 4 cm−1 was selected for the ATR and external reflection modes, while transmission mode spectra were recorded at 2 cm−1. SEM-EDS. SEM-EDS compositional images were recorded with a FEG-SEM microscope LEO 1525 (ZEISS) fitted with a third generation GEMINI column. Materials. The following chemicals were employed for the preparation of the paint models: ZnO (Erba, 99%, lead white (Sigma-Aldrich 99%), aluminum stearate (AlSt(OH)2, SigmaAldrich, 75%), and linseed oil (Zecchi). The paint models were prepared on a lead white grounded canvas, applying a mixture of ZnO linseed oil (1:1 in weight) with and without the addition of AlSt(OH)2 (5% by wt of oil). The models, named ZnO-oil and ZnO-oil-AlSt(OH)2, were left to dry for about one month and then exposed to accelerate aging at 45 °C and RH values of 95%−98%. A parallel set of the same models was left to age naturally (unaged). The same mixtures of ZnO/oil with and without AlSt(OH)2 were also applied as a thin layer on a KBr pellet and analyzed by FT-IR spectroscopy in transmission mode for investigating the pigment−oil interaction at the early stage of the polymerization. Cross sections of the samples coming from both paint models and Alchemy were embedded in epoxy resin, and the final polishing was done with micromesh cloths (from 6000 to 12000 mesh).

■

RESULTS AND DISCUSSION Model Samples. The pigment−oil interaction process was followed in the paint models as a function of the accelerated aging time, ta, (T = 45 °C, RH = 95%−98%) by ATR−FT-IR spectroscopy. The spectra corresponding to aging steps ta = 0, 1, 30 days for ZnO-oil models with and without AlSt(OH)2 are reported in Figure 2a (red and black lines, respectively). All the FT-IR spectra showed a broad band positioned at about 1600 cm−1 already present before the accelerated aging process started (at ta = 0). This band slightly shifted to lower wavenumber with the aging time (its lowest value was observed at 1585 cm−1). It can be assigned to the COO− antisymmetric stretching (νas) of the carboxylate moiety related to an

Figure 2. (a) ATR−FT-IR spectra of the paint models ZnO-oil (black line) and ZnO-oil-AlSt(OH)2 (red line) at ta= 0, 1, and 30 days of aging. (b) FT-IR spectra evolution of the paint mixture with ZnO-oilAlSt(OH)2 (applied as thin layer on a KBr pellet) during the early stage of the oil polymerization (tp= 0, 1, 2, 3, and 8 days of reaction). C

DOI: 10.1021/acs.analchem.6b04065 Anal. Chem. XXXX, XXX, XXX−XXX

Article

Analytical Chemistry

Figure 3. (a) Micro-ATR−FT-IR spectra extracted from single pixels of the FPA detector corresponding to the locations in presented chemical images by numeric labels; the integrated area of the νas COO− bands of the three types of carboxylates are plotted for the reconstruction of the corresponding chemical images. (b) Chemical images of the unaged model in two different areas of the stratigraphy obtained from the integration of (I) broad band at 1590 cm−1 and (II) band at 1588 cm−1 of AlSt(OH)2 for the same area, (III) band at 1588 cm−1 of AlSt(OH)2, and (IV) band at 1540 cm−1 of ZnSt2 for the same area. (c) Chemical images of the 30 days-aged model in three different areas of the stratigraphy obtained from the integration of (I) broad band at 1590 cm−1 and (II) band at 1540 cm−1 of ZnSt2 in the same area and (III) and (IV) band at 1540 cm−1 of ZnSt2 in two additional areas. In all the chemical images, the scale bar corresponds to 10 μm.

Figure 4. (a) Micro-ATR−FT-IR spectra extracted from single pixels of the FPA detector corresponding to the locations in presented chemical images of cross section A1. Numeric labels refer to their spatial localization within the corresponding chemical images reported on (b). (b) Chemical images resulting from the integration of (I) broad band at 1590 cm−1, (II) band at 1540 cm−1 of ZnSt2, and (III) band at 1588 cm−1 of AlSt(OH)2. (c) BSE-SEM and elemental images (Al, Zn, and Ti) in the same area of the cross section. In the micro-ATR−FT-IR spectroscopic images, the scale bar corresponds to 10 μm.

the sharp band of AlSt(OH)2 at 1588 cm−1 overlapping with the broad ZnO-oil band. As a result, in the chemical images plotted by integrating the broad ZnO-oil band, the simultaneous presence of AlSt(OH)2 produced an overestimation of the broad band intensity in correspondence of the areas of coexistence. This can be appreciated in Figure 3bI and II where both the broad FT-IR band at 1590 cm−1 and the AlSt(OH)2 band at 1588 cm−1 are plotted for the same area.

spectra and the corresponding chemical images of the unaged and 4 weeks-aged models are shown in Figure 3. The three types of carboxylate features were spatially located in the stratigraphy by integrating the νas COO− band (Figure 3a): (i) at 1588 cm−1 for AlSt(OH)2, (ii) at 1540 cm−1 for ZnSt2, and (iii) in the range 1510−1650 cm−1 for the broad band. Generally, the broad band was widely detected in the whole stratigraphy of both unaged and aged models (Figure 3bI and cI). Here, band deconvolution was not performed to separate D

DOI: 10.1021/acs.analchem.6b04065 Anal. Chem. XXXX, XXX, XXX−XXX

Article

Analytical Chemistry

Figure 5. Visible microscopy images (90×) and micro-ATR−FT-IR spectroscopic images representing the distribution of ZnSt2 (1540 cm−1) in the Alchemy cross sections A1, A2, and A3 (painting) and B1, B2, and B3 (frame). In the micro-ATR−FT-IR spectroscopic images, the bar corresponds to a distance of 10 μm.

observed, indicating that it had almost completely reacted within the paint layer. In a previous paper,15 FT-IR microspectroscopy, in transmission mapping mode using a synchrotron IR source to analyze historical model samples naturally aged, has revealed a rather different distribution. A higher concentration of Al stearate has been detected in correspondence with the zinc carboxylate represented by the broad band at 1590 cm−1, and ZnSt2 has been found mainly distributed at the bottom of the cross section. This discrepancy might be connected, as suggested also by the authors of the cited article, to the possible influence of the PET (polyethylene terephthalate) substrate on which their paints were prepared. The high extent of stearate-like Zn soap formation and the fact that AlSt(OH)2 was introduced in our paint models at 5 wt % (vs oil) suggests that AlSt(OH)2 can promote oil hydrolysis, increasing the relative quantity of free fatty acids available to react with the pigment ZnO to give sharp ZnSt2-like carboxylates.15

Additionally, the chemical images of the unaged sample (Figure 3bII and III) show the presence of both elongated and rounded agglomerates of AlSt(OH)2 having dimensions generally greater than 10 μm. Most interestingly some of these agglomerates are surrounded by newly formed aggregated particles of ZnSt2-like structure (Figure 3bIII and IV, the two chemical images correspond to the same investigated area). This evidence strongly supports the process of facilitated formation of Zn carboxylates occurring in the presence of AlSt(OH)2 proposed in the literature16 and observed in this work with conventional FT-IR spectroscopy at the first stage of oil polymerization. After 30 days of temperature/moisture accelerated aging, the sample stratigraphy is mainly composed of ZnSt 2 -like carboxylates (Figure 3cII, III, and IV), which distribution results spatially complementary to one of the Zn carboxylates featuring the broad band at 1590 cm−1 (Figure 3cI and II correspond to the same area of the cross section). Furthermore, no spectral bands related to aluminum stearate have been E

DOI: 10.1021/acs.analchem.6b04065 Anal. Chem. XXXX, XXX, XXX−XXX

Article

Analytical Chemistry Paint Microsamples from Jackson Pollock’s Alchemy. The understanding acquired with the study of the model paints discussed above has been very helpful for the interpretation of the data obtained by micro-ATR−FT-IR spectroscopic imaging of the cross sections from the painting Alchemy by Jackson Pollock, where noninvasive reflection FT-IR had evidenced a wide distribution of sharp features related to Zn soaps. Three microsamples (A1−3) from white thick paints of Alchemy were available for the study, along with three more samples (B1−3) taken from the frame, which presents several drops and lines of the paints as Pollock used it to stretch the canvas. Visible microscopy and scanning electron microscopy images of the related cross sections are reported in Figure S-2 of the Supporting Information (SI). The micro-ATR−FT-IR imaging of the cross sections revealed the presence within the stratigraphy of Alchemy of all the carboxylate species detected in the aged model paint containing AlSt(OH)2. In Figure 4a, the spectral profiles of ZnSt2 and AlSt(OH)2 and the broad band at 1590 cm−1 recorded in sample A1 are reported as representative of the whole set of samples. In all the microsample stratigraphies from Pollock’s painting, the chemical images showed features similar to the artificially aged models: a wide and homogeneous distribution of the broad-featured Zn soap and a close distribution of ZnSt2 and AlSt(OH)2 (Figure 4bI, II, and III). In particular, we found a reaction shell of ZnSt2 surrounding the particles of the Al stearate used as additive (Figure 4bII and III). Figure 4c depicts the corresponding BSE-SEM images and elemental maps of Zn, Ti, and Al in the same area highlighting the higher concentration of Al in the aggregate surrounded by Ti and Zn. An overall examination of the chemical images acquired for all the six cross sections (Figure 5) revealed at least two different morphologies of the ZnSt2 aggregates. Samples A1, A2 (from the painting), and B2 (from the frame) showed clusters of ZnSt2 with irregular shapes and variable dimensions (3−10 μm). The smaller ones are randomly scattered, while the larger ones are mainly localized around the AlSt(OH)2 particles. This distribution is very similar to what was observed in the aged model discussed in the previous section (Figure 3bIV and cII, cIII, and cIV). The other samples (Figure 5A3, B1, and B3) showed a very different morphology of ZnSt2 aggregates with rounded shape particles generally of dimensions larger than 10 μm that in some portions coalesced to larger clusters. The different distributions of ZnSt2 could represent a different stage of the saponification process. Considering the high similarity of the first group (A1, A2, and B2, Figure 5) with the model samples aged for just a month, the second type of ZnSt2 distribution (A3, B1, and B3, Figure 5) might correspond to a more advanced stage of saponification which foresees the migration and aggregation of metal soaps in analogy to what have been already observed for lead soaps.22 In particular, we noticed in sample B1 (from the frame) the occurrence of some compositional heterogeneities in the inner layers of the paint in contact with the wooden support (Figure 6 and Figure S-2, SI), which appear in the BSE-SEM images as a dark gray region since they are depleted in zinc. These heterogeneities are visible also by visible microscopy as a more transparent paint with yellowish rounded agglomerates recalling the chemical images of the ZnSt2 particles. The results of the ATR-FT-IR imaging experiments carried out toward the inner edge of the section effectively highlighted a higher concentration of ZnSt2 agglomerates in the form of

Figure 6. (a) BSE-SEM and (b) visible microscopy image of sample B1 corresponding to the inner part of the painting stratigraphy in contact with the wooden support. (c) Micro-ATR−FT-IR spectroscopic images representing the distribution of ZnSt2 (1540 cm−1) performed in succeeding imaging experiments toward the lower portion of the stratigraphy.

bigger particles (>10 μm) that evidently coalesced to larger clusters prone to form inner protrusions. The two different distributions of the ZnSt2 particles observed in the six cross sections are not correlated to the different conservation history undergone by the frame and the painting. Instead, our findings strongly support the hypothesis that they are more related to the migration of zinc soaps through the paint stratigraphy most probably driven by the different thickness and type of support in contact with the paint. In fact, it should be mentioned that in Alchemy the preparation layer (made of lead white) was applied by Pollock in a highly inhomogeneous fashion;6 hence, the paint could have been dropped directly on the canvas, as well as on the preparation layer. Additionally, the paint application on both the painting and the frame is highly variable in terms of thickness. These facts could have locally created similar environments for samples A3 and B1−3 with respect to A1, A2, and B2, thus explaining their different behavior.

■

CONCLUSION Because of its high specificity in chemical speciation combined with high spatial resolution capabilities, micro-ATR-FT-IR spectroscopic imaging was successfully applied for a detailed account of the pigment−oil interaction in the complex and heterogeneous microstructure of Alchemy. The study, profitably informed by a preliminary investigation of aged and unaged paint models, allowed us to answer relevant conservation questions. (i) Are these Zn soaps already present in the paint tubes used by Pollock or rather the result of an alteration process? The micro-ATR−FT-IR imaging results of the model paint samples proved that the ZnSt2-like particles aggregate around the F

DOI: 10.1021/acs.analchem.6b04065 Anal. Chem. XXXX, XXX, XXX−XXX

Article

Analytical Chemistry

FIRB (RBFR12PHL4) projects. Philip Rylands, director of the Peggy Guggenheim Collection (PGC, Venice), is acknowledged for giving access to the painting. The authors thank Carol Stringari of the Peggy Guggenheim Collection (PGC, NY).

agglomerates of the added AlSt(OH)2, showing the ring of reaction. With the aging of the paint, the Zn carboxylates grow and distribute as irregular clusters increasing in number until coalescing to larger agglomerates through the whole painting stratigraphy. (ii) Which is their spatial distribution within the paint stratigraphy? The micro-ATR−FT-IR spectroscopic imaging study revealed at least two different morphological distributions of ZnSt2. Scattered, irregular, and small particles (