Femtosecond Pump-Probe Microscopy Generates Virtual Cross-Sections in Historic Artwork

Femtosecond pump-probe microscopy generates virtual cross-sections in historic artwork

Tana Elizabeth Villafanaa, William P. Brownb, John K. Delaneyc, Michael Palmerc, Warren S. Warrena,1, and Martin C. Fischera

aDepartment of Chemistry, Duke University, Durham, NC 27708; bConservation Department, North Carolina Museum of Art, Raleigh, NC 27607; and cScientific Research Department, National Gallery of Art, Washington, DC 20565

Edited by Michael D. Fayer, Stanford University, Stanford, CA, and approved December 23, 2013 (received for review September 12, 2013) The layering structure of a painting contains a wealth of information generation and two-photon excited fluorescence (13) and the about the artist’s choice of materials and working methods, but mapping of oil and varnish interfaces with third harmonic gen- currently, no 3D noninvasive method exists to replace the taking eration (14). However, most inorganic pigments neither fluoresce of small paint samples in the study of the stratigraphy. Here, we nor generate appreciable harmonic light, leaving these techniques adapt femtosecond pump-probe imaging, previously shown in tis- limited in their scope for cultural heritage. sue, to the case of the color palette in paintings, where chromo- Near-infrared femtosecond pump-probe optical microscopy phores have much greater variety. We show that combining the expands the range of detectable molecular signatures (15) to contrasts of multispectral and multidelay pump-probe spectroscopy include signals from excited state absorption, ground state de- permits nondestructive 3D imaging of paintings with molecular and pletion, and stimulated emission (16). This microscopy technique structural contrast, even for pigments with linear absorption spectra was mainly developed for biomedical imaging and has been used that are broad and relatively featureless. We show virtual cross- to provide high-resolution images for the biological pigments sectioning capabilities in mockup paintings, with pigment separa- hemoglobin (17, 18), eumelanin, and pheomelanin (19, 20) that tion and nondestructive imaging on an intact 14th century painting are present in skin (21) or ocular cancer (22). The Crucifixion ( by Puccio Capanna). Our approach makes it possible Extension of pump-probe microscopy from biological pig- to extract microscopic information for a broad range of applications ments to samples of artist’s pigments has yielded promising SCIENCES to cultural heritage. preliminary results (23). However, achieving pump-probe con- APPLIED PHYSICAL trast in fine art objects is more challenging than skin imaging, nonlinear imaging | pigment spectroscopy because artist colorants range from organic dyes to inorganic minerals, with colors spanning the entire visible spectrum. In ’ dentifying an artist s choice of materials (e.g., support, pigments, contrast, the pigments in a sample of skin tissue are mainly Ibinders, and varnishes in a painting) and working methods can limited to hemoglobin, eumelanin, and pheomelanin, which all lead to greater understanding of past cultures and enhance the provide image contrast with a single pump-probe wavelength ability of conservators to preserve that culture. In a painting, this combination (in this case, 720 and 810 nm, respectively). Here, information is contained in its layered structure, and it is generally we show that an increased spectral range of both the pump and studied by the physical removal of a small paint sample, which the probe beams, from near-IR to visible, and a variable time can be characterized by a plethora of analytical techniques (1). delay of the pump-probe pulses help to address the complexity The sample needs to be representative of the painting but as < introduced by the large range of possible pigments in the paint small as possible (typically 0.5 mm), and only local information layers and allow for in situ 3D imaging of paintings with mo- is obtained. Nondestructive analysis by traditional macroscopic lecular specificity. We first show virtual cross-sectioning capa- methods, such as X-radiography, near-infrared reflectance im- bilities in historically relevant mockup paintings and use specific aging, and UV-visible fluorescence photography, can provide some information about a painting’s support, compositional Significance paint changes, underdrawings, paint and varnish applications, and restorations (1). Materials can be identified in situ on the microscopic scale using Raman (2–4) or the macroscopic scale We show that a nonlinear microscopy technique (femtosecond with reflectance imaging spectroscopy (5, 6) and X-ray fluores- pump-probe microscopy) allows for nondestructive 3D imaging cence intensity mapping (7). Unfortunately, none of these tech- of paintings with molecular and structural contrast. Until now, niques contain quantitative depth-resolved material information. studying the layering structure of a painting has generally re- Methods that could offer 3D information are under active re- quired the physical removal of a cross-section sample. Pump- search, such as confocal X-ray fluorescence, absorption near-edge probe imaging has previously been shown on biological tissue, structure imaging (8), optical coherence tomography (9), and but applications to cultural heritage are more challenging: the ’ terahertz imaging (10), but they are not yet widely used in con- variety of pigments in the artist s palate is enormous compared servation science laboratories because of their limitations: X-ray– with the biological pigments present in skin. Nonetheless, we based techniques have absorption limited depths, whereas optical show virtual cross-sectioning capabilities in mockup paintings coherence tomography and terahertz imaging produce image and nondestructive imaging on an intact 14th century painting. contrast that is largely based on refractive index mismatches and This work represents a comprehensive collaborative effort between laser and biomedical imaging experts and scientists and therefore, only provide structural contrast. conservators in national museums. In general, conventional (linear) optical imaging into the paint layer of a painting is limited in its depth penetration by Author contributions: T.E.V., W.S.W., and M.C.F. designed research; T.E.V., J.K.D., and absorption and scattering from the pigment particles. In biology M.P. performed research; W.P.B., J.K.D., M.P., and W.S.W. contributed new reagents/ and biomedical applications, nonlinear imaging can provide analytic tools; T.E.V., J.K.D., M.P., and M.C.F. analyzed data; and T.E.V., W.S.W., and optical sectioning in highly scattering and absorbing samples M.C.F. wrote the paper. (11, 12). Traditional nonlinear imaging has found a few appli- The authors declare no conflict of interest. cations to cultural heritage; recent research includes the 3D This article is a PNAS Direct Submission. imaging of wood and varnishes in a violin with second harmonic 1To whom correspondence should be addressed. E-mail: [email protected].

www.pnas.org/cgi/doi/10.1073/pnas.1317230111 PNAS Early Edition | 1of6 Downloaded by guest on September 29, 2021 pump-probe signatures to provide pigment separation. We then perform in situ 3D imaging on a 14th century painting (The Crucifixion by Italian artist Puccio Capanna) to highlight our ability to noninvasively image and create virtual cross-sections of complex pigment layers. Although we focus on historic paintings, our approach can be applied to a wide range of cultural heritage objects and provides information extremely relevant to current areas of interest in conservation science. Results Approach. Pump-probe microscopy uses a sequence of ultrafast pulses (typically 0.2 ps in duration) to first electronically excite molecules and then probe their response at a later time (up to about 100 ps). A pump pulse moves a fraction of the ground state population into electronic excited states, creating a corresponding hole in the ground state spectral distribution. In response to the excitation, the population distributions in both ground and excited states rearrange (excited state population tends to eventually relax back to the ground state). The changes in population can be monitored by applying a second delayed (probe) pulse. Different molecular processes have different effects on the probe pulse as a function of pump intensity and pump-probe delay. For example, in sequential two-photon absorption, the probe is absorbed only by molecules in the excited state; hence, the presence of the pump increases the probe absorption (the absorption then diminishes for longer delays). In contrast, for ground state depletion, the probe is absorbed by molecules remaining in the ground state, which has been partially depleted by the pump; hence, the presence of the pump decreases the probe absorption (probe absorption increases back to the equi- librium value for long delays). Pump-probe spectroscopy has been a mainstay of chemical physics for decades using high- powered lasers; however, at the powers that we are willing to use Fig. 1. Schematic of experimental setup. Pump-probe imaging uses an in- on important artwork, the differences in absorption might typi- tensity-modulated pump pulse train and a nonmodulated probe pulse train cally be 1 part in 106 parts or a tiny signal on a large background. separated by a variable time delay. Nonlinear interactions at the sample cause A schematic of our experimental setup (Fig. 1) shows our solu- the pump modulation to transfer to the probe, which is detected by a lock-in tion to this challenge (17, 18). The pump is an intensity-modu- amplifier. The pump is filtered out, and a series of images is collected, each with a different interpulse delay. AOM, acousto-optic modulator. lated, mode-locked pulse train, which is synchronized and combined with an unmodulated probe pulse train and coupled into a laser-scanning microscope. Nonlinear interactions in the been covered with a thin glaze of red pigment (quinacridone red, focal volume within the sample will cause the modulation to a modern transparent, light-stable replacement for the natural transfer from the pump to the probe. The modulation frequency substituted anthraquinone) to create a purple appearance. In is several megahertz, chosen to overcome the noise spectrum of another case, a blue pigment (lapis lazuli) has been mixed with laser fluctuations. Pump-probe microscopy, like other nonlinear the same red pigment to create a similar purple appearance. To imaging methods, is much less affected by light scattering than conventional microscopy; the signal is proportional to the take a virtual cross-section, we first determined the pump-probe product of the intensities of the two lasers, causing scattered light wavelength combinations and interpulse delays that would fully to produce much less signal and giving the method its power in separate ultramarine blue from quinacridone red by imaging 3D imaging. a physical cross-section from the layered mockup at different wavelength combinations. In the future, we can build a pump- Pump-Probe Specificity in Quinacridone Red and Ultramarine Blue. In probe library with cross-section samples from a variety of his- a typical painting, the 3D structure could consist of single to torical artworks that have already been characterized with cur- multiple colorants in layers, mixtures, or a combination of layers rently accepted analytical techniques. The spectroscopy results and mixtures. Given the limited available colorants in Italian are seen in Fig. 2. The ground and excited state dynamics for renaissance paintings compared with contemporary works, pur- each pigment are specific to that pigment, providing structured ples were often made using combinations of red pigments, such and complex pump-probe signatures. At a pump-probe wave- as kermes or red madder (both substituted anthraquinone), length combination of 615/810 nm, the signal in quinacridone red mixed or layered with blue mineral pigments, such as natural ul- is positive (i.e., the amount of detected probe light decreases tramarine or azurite. The combination of kermes and natural ul- when the pump is turned on) and decays in time. In ultramarine tramarine (lapis lazuli, which during the time, was more expensive than gold) gives a rich purple. Such a combination would be blue, the signal is negative, also decaying in time. The combina- suitable for the robes of major characters of a painting, such as the tion of positive and negative pigment-specific transient absorp- Madonna’s robe, whereas a combination with the cheaper azurite tion signals provides an ideal case for creating a virtual cross- can give a danker muted purple useful for less prominent figures. section. Interestingly, shifting to pump-probe wavelengths of 655/ To test our virtual cross-section capabilities to separate mix- 810 nm, the transient absorption amplitudes for these pigments tures vs. layering of pigments, we began by creating a set of are reversed (although much weaker in magnitude for quinacri- mockup paintings that features historically relevant pigment done red), although the linear absorption barely differs at our pairings. In one case, a blue pigment (synthetic ultramarine) has choice of pump wavelengths. The temporal decay characteristics

2of6 | www.pnas.org/cgi/doi/10.1073/pnas.1317230111 Villafana et al. Downloaded by guest on September 29, 2021 A B C Quinacridone Red

Ultramarine Blue

Fig. 2. Multispectral pump-probe investigation of quinacridone red and ultramarine blue. (A) Linear absorption spectra of quinacridone red (red curve) and ultramarine blue (blue curve), with dashed lines indicating the different wavelengths that we used in the pump-probe imaging of the pigments. The pump- probe delay traces for (B) ultramarine blue and (C) quinacridone red at a variety of pump-probe wavelengths (indicated in the legend in nanometers) show vastly different time responses. All wavelength combinations were taken at a total power of 7 mW.

of the pigments also vary with pump-probe wavelengths, pro- structure and the other mock painting features a mixture, both viding yet another method of pigment separation. cases present a purple color; the reflectance spectra acquired Bright-field microscopic images of the physical cross-sections using fiber optic reflectance spectroscopy (FORS) from each

taken from the two mockup paintings (Fig. 3) show that, in the mock-up painting are similar (Fig. 4) and give no indication of SCIENCES layered case, the red glaze layer is ∼5-μm thick and the synthetic either painting’s stratigraphy. However, in both cases, the pump- APPLIED PHYSICAL ultramarine is ∼25-μm thick, whereas the mixed sample has one probe virtual cross-sections not only distinguish between the layer with a varying thickness of 25–90 μm. Pump-probe images layered and mixed stratigraphy but also, highlight variations in (at 615/810 nm) of the physical cross-sections were taken at the paintings caused by the artist’s brushwork. In the virtual different time delays to allow assignment of pigments. The pig- cross-section image in Fig. 4, Right Inset displays data acquired ments are assigned false colors according to their pump-probe with a high-N.A. objective, clearly resolving the thin red glaze response; the red glaze (colored red) has a positive response, and layer. At this particular pump wavelength (615 nm), our imaging the ultramarine blue (colored cyan) has a negative response. depth in lapis lazuli was limited to roughly 10 μm because of There is no signal from the acrylic binder. The pump-probe absorption by ultramarine, but tuning the pump wavelength to images give similar results in terms of both the distribution of 710 nm increased the penetration through this pigment sixfold, pigments and the layer thickness to the structure of the cross- which is discussed below, and let us image through the entire section obtained from the bright-field microscope. ultramarine layer in the mockup (at the expense of a negligible signal from red glaze). Virtual Cross-Sections of the Mock Paintings. Having found appro- Our technique has several advantages over the removal of priate imaging parameters for the two pigments, we created a physical cross-section other than its nondestructive nature. By virtual cross-sections from the mockups by taking a series of en mapping out an entire volume, we can create virtual slices from face images (xy images perpendicular to the beam axis) at dif- the entire field of view in any direction and visualize differences ferent depths (z direction) using the wavelength combination of in brushwork or abrupt changes in layering that may not be ev- 615/810 nm with an interpulse delay of 0.1 ps. We generated ident in a physical cross-section, in which accessible informa- virtual cross-sections from this volume set by selecting a data slice in the xz or yz direction. In Fig. 4, we false-colored the tion is dependent on the sampling orientation. In addition, we images according to the previously discussed methodology; the can sample many areas anywhere in the painting, which is not red glaze has been colored red, and the blue pigment has been possible when acquiring physical cross-sections (generally con- colored cyan. Although one mock painting features a layered servators do not remove samples form pristine areas of the paintings).

Nondestructive Investigation of Intact Artwork. The Crucifixion was Layered Sample Mixed Sample painted by Puccio Capanna in roughly 1330 on a wooden panel using various pigments in egg tempera with gold leaf. We focused on two areas in the painting: the rich blue of the Virgin

Bright Mary’s robe and the light blue robe of a floating angel that is outlined with purple shading. Prior cross-sectional analysis of the Virgin Mary’s robe indi- Probe Pump- cates that the robe has been painted with a thick (up to 60 μm) layer of lapis lazuli. Thus, examination of the robe presented Fig. 3. Pump-probe separation of quinacridone red and ultramarine blue a unique opportunity to test our depth penetration capabilities in using physical cross-section samples. Bright-field and pump-probe images of a real work of art in a relatively uncomplicated setting; previous the physical cross-sections from the layered and mixed mockup painting. The pump-probe images were taken at an interpulse delay of 0.1 ps and a work indicates no pump-probe signal in egg tempera binder or wavelength combination of 615/810 nm with a total power of 5 mW. other binder materials. The results are presented in Fig. 5. Quinacridone red is false-colored red, whereas ultramarine blue is cyan. Pump-probe imaging in the center of the robe gave virtual cross- The pump-probe images are 365 × 90 μminsize. sections consistent with the known thickness of the lapis lazuli,

Villafana et al. PNAS Early Edition | 3of6 Downloaded by guest on September 29, 2021 pump-probe images, we color-coded the positive signals orange, Pump-Probe Images encompassing any of the three materials, and negative signals En-face (xy) Layered En-face (xy) Mixed cyan (lapis lazuli). At the probed location, we found a composition that is slightly different from the physical cross-section. The en face images show positive signal on the surface (most likely from iron oxide), negative signal in the center from lapis lazuli, and positive signal again underneath the lapis lazuli, which is most likely gold with possible contributions from mordant (the gilding in this region is heavily cracked, exposing the mordant z z underneath). This view is supported by virtual cross-sections extracted from this dataset. The virtual xz slice and even more so, x x the maximum intensity projection of the volume along the y di- Fig. 4. Reflectance spectra and virtual cross-sections of mock paintings. rection suggest either a mixture or very thin layers of iron oxide (Left) The linear reflectance spectrum from each painting indicates the with lapis lazuli and gold leaf with mordant underneath it. presence of quinacridone red (peak at 600 nm) and ultramarine blue (increased reflection at 700 nm) but does not indicate which painting is layered Discussion and which is mixed. Inset shows photographs of the (Upper) layered and These results represent a large step in the nondestructive 3D (Lower) mixed mockups. (Right) A volume set of pump-probe images of the analysis of pigments and their composition in historical art intact mock paintings was taken at a wavelength combination of 615/810 works. In an intact 600-y-old painting, we have shown that our nm, fixed interpulse delay of 0.1 ps, and total power of 3 mW with a 20× 0.7 N.A. objective. One image of each set is shown (false-colored red for qui- technique can noninvasively image through a relatively thick nacridone red and cyan for ultramarine blue). Virtual cross-section (xz) layer of paint and map multilayer structures. With our current images immediately reveal the composition difference between the layered microscope design, we can easily image volume data at fixed time and mixed samples. Inset on the virtual cross-section of the layered sample delays or 2D images at varying pump-probe delays. Typical ac- was obtained with a higher-resolution 60× 0.9 N.A. objective. Each en face quisition times of the three-parameter datasets in Fig. 6 were 30– (xy) image is 365 × 365 μm, and the virtual cross-sections are 365 × 90 μm. 60 min. Future improvements in detection sensitivity should decrease imaging time markedly, leading the path to acquisition of 4D datasets (3D space and delay) dense enough for 3D pig- highlighting the ability of this method to noninvasively image ment-specific mapping. The achievable imaging depth depends deeply into pigment layers. on the structure and layering of the artwork. Certain flexibility is To test the ability to obtain a virtual cross-section in a historic — afforded by the choice of pump and probe wavelengths. For each painting, we imaged in an area with known layering specifically, pigment, there will be tradeoffs between signal strength (pumping the outline of the floating angel. Fig. 6 shows a bright-field image close to an absorption line best excites electronic states but results of a physical cross-section taken from this region of the angel’s in the largest linear absorption) and contrast (not all wavelength robe. Bright-field microscopy and scanning electron microscopy combinations yield usable signal). Efforts to establish a pump- with energy-dispersive X-ray spectroscopy (SEM-EDS) analysis probe database for the most common pigment types are currently of this cross-section indicates a very delicate and thin layering of underway. Some materials might not yield distinct pump-probe pigments containing, from top to bottom, a faded red glaze, signals (a fact that we use to our advantage by rendering binders a mixture of lead white and lapis lazuli, iron oxide, an organic invisible), in which case combination of our microscopy technique coating, gold leaf, an iron-rich mordant (a mixture of pigments with other optical contrasts in the same microscope might be and oil used to adhere the gold leaf) (24), and a gypsum ground. beneficial. For example, to study organic glazes, binders, and To image this area, we tuned the pump-probe pulses to a wavelength of 710/810 nm, for which we obtain signals from lapis lazuli, iron oxide, mordant, and gold (gold has a strong two- photon absorption response at this wavelength combination). We classified the temporal dynamics in the pump-probe signals using The Crucifixion Virtual Cross-Section phasor analysis, a method that is commonly used to visualize decay times in fluorescence lifetime measurements (25) and was En-face (xy) recently adapted to pump-probe work (26). We identified three distinct decay behaviors, consistent with previously observed data in lapis lazuli, iron oxide, and gold. Fig. 6 also displays the physical cross-section as a false-colored pump-probe image, cor- relating well with the bright-field image. At the chosen wavelength combination, we do not see a signal in the faded red glaze, lead white, organic coating, or gypsum, except for a few mineral impurities that may be present in those layers. We obtain signal from gold; however, the gilding is very thin and could not be spatially resolved. Also, at this wavelength combination, iron z oxide and mordant showed signals with identical decay behaviors. The optical image was taken after pump-probe imaging, in- x dicating no visible damage. We then imaged an area of the intact painting adjacent to the sample site. At this location, we acquired Fig. 5. Virtual cross-section of the Virgin Mary’sbluerobeinPuccioCappana’s volume data with a fixed pump-probe delay of 0.2 ps, which yields The Crucifixion.(Left) The painting was imaged in an area of Mary’srobe positive pump-probe signals from iron oxide, gold, and mordant containing only a single layer of lapis lazuli with a wavelength combination of and negative signals from lapis lazuli. Because the pump-probe 720/810 nm and an interpulse delay of 0.2 ps with a total power of 2.7 mW. (Upper Right)Theen face image shown (365 × 365 μm) was from roughly dynamics of iron oxide/mordant and gold are very similar, they 30 μm beneath the surface of the robe. Here, the images have been false- can only be cleanly separated by acquiring data at many densely colored cyan for lapis lazuli and magenta for mineral impurities that occur with sampled time delays, which with our current setup, was not fea- natural lapis lazuli. (Lower Right) The virtual xz cross-section (365 × 60 μm) sible during the loan period of the painting. Hence, in these highlights the thickness of the lapis lazuli used to paint Mary’srobe.

4of6 | www.pnas.org/cgi/doi/10.1073/pnas.1317230111 Villafana et al. Downloaded by guest on September 29, 2021 Physical Cross-Section and Pump-Probe Delay Behavior that extensions to 3D brushstroke imaging are possible, pre- senting opportunities and challenges for data mining (38). Iron Oxide/ Mordant Pump-probe microscopy, in conjunction with current techniques Gold in conservation science, could dramatically impact the study and understanding of our cultural heritage. Ultimately, the general Lapis Lazuli application of our technique in situ will require a portable nonlinear microscope, but all of the miniaturization technology needed for Absorption (arb. u.) Pump-Probe Delay (ps) such a device is being developed for biomedical applications (39); En-face Pump-Probe Images at 0.2 ps Delay also, appropriate laser sources have recently dropped drastically in cost (40). We have shown that it is possible to leverage large so- 0 m (Surface) - 5 m- 12 m cietal investment in biomedical and molecular imaging to enable applications with a broader impact. Methods Approach. A Ti:Sapphire mode-locked laser operating at a repetition rate of 80 MHz, with a wavelength in the near-IR and pulse duration of roughly 150 fs, pumps an optical parametric oscillator, with an output in the visible to the near-IR of a similar pulse duration. The pump pulse train is intensity- Virtual Cross-Section Generated from En-face Volume Cube modulated at 2 MHz using an acousto-optic modulator. The probe pulse is unmodulated, and the interpulse delay is controlled by an adjustable optical path length in the probe arm. The two beams are overlapped on a dichroic Cross-Section Maximum Intensity Projection mirror and sent collinearly into a laser-scanning microscope. The pulses are focused onto the sample with a 20× 0.7 N.A. or 60× 0.9 N.A. air objective. Fig. 6. Investigation of the angel’s purple robe in Puccio Cappana’s The Any nonlinear interaction with the sample will transfer the modulation Crucifixion.(Top) A bright-field image of the physical cross-section taken from from the pump to the probe, and changes in the probe intensity are the angel’s robe is shown. A pump-probe delay dataset was acquired (40 detected by a photodiode and a lock-in amplifier with a reference at the interpulse delays, pump-probe = 710/810 nm, total power = 1.5 mW, size is modulation frequency. × μ 545 55 m). From this set, we created a false-colored image according to the SCIENCES pump-probe delay behavior (cyan for lapis lazuli, red for the two iron-rich Mock Paintings of Quinacridone Red and Ultramarine Blue. Layered painting. pigments above and below the gold layer, and yellow for gold). Cumulative Quinacridone red (1310; Golden) and synthetic lapis lazuli (45000; Kremer) APPLIED PHYSICAL pump-probe traces of all identified pigments in the image are shown in Right. were prepared in an acrylic medium. The synthetic lapis lazuli was painted Note that the gold layer is thinner than the resolution of our microscope, and onto a glass slide that had been prepared with a gesso ground. After the layer the gold-labeled trace likely contains some contribution from the adjacent dried, a thin coat of quinacridone red was painted on top and allowed to dry. mordant. (Middle) A pump-probe volume dataset was taken with a fixed A small sample was extracted with a scalpel from the painting and mounted 0.2-ps delay in the angel’srobe(pump-probe= 710/810 nm, total power = in Wards Bio-Plastic for the cross-section. 1.5 mW). The images from this set have been false-colored according to the Mixed painting. Quinacridone red (1310; Golden) and Afghan lapis lazuli signal at this delay: cyan for negative signal (corresponding to lapis lazuli) and (15300; Kremer) were prepared in an acrylic medium, mixed together in orange for positive (iron oxide/mordant and gold; with only a single delay, a roughly 1:1 ratio, and painted onto a glass slide that had been prepared these three materials cannot be separated). Each image is 185 × 185 μm. with a gesso ground. A small sample was extracted with a scalpel from the (Bottom)Anxz slice taken from the volume data shows a positive component painting and mounted in Wards Bio-Plastic for the cross-section. mixed within the lapis lazuli layer (most likely iron oxide) with another posi- FORS analysis. A fiber optic spectroradiometer, FS3 (ASD Inc.), was used to tive component underneath (most likely gold and possibly, underlying mor- obtain FORS spectra from the mock paintings. The spectrometer operates dant that we image through microscopic cracks in the gold layer). This from 350 to 2,500 nm, with a spectral sampling of 1.4 nm from 350 to 1,000 composition is seen more clearly in a maximum intensity projection of the nm. The spectral resolution at 700 nm is 3 nm. The light source of a leaf probe entire volume cube. The virtual cross-section dimension is 185 × 50 μm. head (ASD Inc.) was used at a distance of 20 cm to illuminate the samples (∼400 lx), and the fiber was placed ∼1 cm from the object, giving an ∼3-mm spot size at the painting. We averaged two spectra, with a total acquisition varnishes, it is possible to incorporate nonlinear fluorescence or time of <5 s per point. harmonic generation contrast, which was shown useful in some recent 3D imaging work (13, 14). Nondestructive Investigation of Intact Artwork. The Crucifixion was the cen- Our work with the iron-rich pigments in The Crucifixion fur- tral compartment of one panel of a diptych altarpiece. The pigments— ther suggests interesting applications for earth pigments in a va- typical of an early Renaissance palette—include pure lapis lazuli, azurite, riety of objects from pottery (27) and ancient relics (28, 29) to vermillion, red lake, red lead, terra verte, white lead, black, and earth colors. The medium is estimated to be egg yolk, and the panel support was iden- Greek statuary, which was not white (as believed for centuries) – tified as poplar. The gold was applied to the embroidered decoration by but brightly colored (30 32). Other potential applications in- mordant gilding and the gold field by water gilding onto a red mordant. A clude the in situ mapping of degradation products. For example, small sample was extracted with a scalpel from the angel’s robe and The Joy of Life by Henri Matisse (1905) contains large areas of mounted in Wards Bio-Plastic for the cross-section. cadmium yellow that have degraded to browns and whites (33), an issue that has also affected masterworks by Van Gogh, ACKNOWLEDGMENTS. We acknowledge Prathyush Samineni for his help in Picasso, and others (34). Mapping degradation products could beginning this research with the first paint samples and Jesse Wilson for aid in understanding degradation processes (35, 36). Finally, 2D helpful discussions and providing the schematic in Fig. 1. We thank the North ’ Carolina Museum of Art and the National Gallery of Art for their collabora- wavelet analysis of van Gogh s brushstrokes has been applied to tion and contribution of various art pieces. This material is based on work 101 high-resolution grayscale scans (37), and our work suggests supported by National Science Foundation Grant CHE–1309017.

1. Steward B (2007) Analytical Techniques in Materials Conservation (Wiley, 5. Delaney JK, et al. (2010) Visible and infrared imaging spectroscopy of Picasso’s Harlequin New York). musician: Mapping and identification of artist materials in situ. Appl Spectrosc 64(6):584–594. 2. Clark RJH (2007) Raman microscopy as a structural and analytical tool in the fields of 6. Dooley KA, et al. (2013) Mapping of egg yolk and animal skin glue paint binders in art and archaeology. J Mol Struct 834–836:74–80. Early Renaissance paintings using near infrared reflectance imaging spectroscopy. 3. Vandenabeele P, Edwards HGM, Moens L (2007) A decade of Raman spectroscopy in Analyst (Lond) 138(17):4838–4848. art and archaeology. Chem Rev 107(3):675–686. 7. Dik J, et al. (2008) Visualization of a lost painting by Vincent van Gogh using syn- 4. Brambilla A, et al. (2011) A remote scanning Raman spectrometer for in situ mea- chrotron radiation based X-ray fluorescence elemental mapping. Anal Chem 80(16): surements of works of art. Rev Sci Instrum 82(6):063109. 6436–6442.

Villafana et al. PNAS Early Edition | 5of6 Downloaded by guest on September 29, 2021 8. Kanngießer B, Malzer W, Mantouvalou I, Sokaras D, Karydas AG (2012) A deep view in 24. Gettens RJ, Stout GL (1966) Painting Materials: A Short Encyclopaedia (Dover, New cultural heritage—confocal micro X-ray spectroscopy for depth resolved elemental York). analysis. Appl Phys A Mater Sci Process 106(2):325–338. 25. Redford GI, Clegg RM (2005) Polar plot representation for frequency-domain analysis 9. Targowski P, Iwanicka M (2012) Optical coherence tomography: Its role in the non- of fluorescence lifetimes. J Fluoresc 15(5):805–815. invasive structural examination and conservation of cultural heritage objects—are- 26. Robles FE, Wilson JW, Fischer MC, Warren WS (2012) Phasor analysis for nonlinear view. Appl Phys A Mater Sci Process 106(2):265–277. pump-probe microscopy. Opt Express 20(15):17082–17092. 10. Adam AJL, Planken PCM, Meloni S, Dik J (2009) TeraHertz imaging of hidden paint 27. Sciau P, et al. (2011) Reverse engineering the ancient ceramic technology based on – layers on canvas. Opt Express 17(5):3407–3416. X-ray fluorescence spectromicroscopy. J Anal At Spectrom 26(5):969 976. 11. Helmchen F, Denk W (2005) Deep tissue two-photon microscopy. Nat Methods 2(12): 28. Edwards HGM, Villar SEJ, David AR, de Faria DLA (2004) Nondestructive analysis of 932–940. ancient Egyptian funerary relics by Raman spectroscopic techniques. Anal Chim Acta – 12. Denk W, Strickler JH, Webb WW (1990) Two-photon laser scanning fluorescence 503(2):223 233. microscopy. Science 248(4951):73–76. 29. Colombini MP, et al. (2004) Integrated analytical techniques for the study of ancient – 13. Latour G, Echard J-P, Didier M, Schanne-Klein M-C (2012) In situ 3D characterization Greek polychromy. Talanta 63(4):839 848. of historical coatings and wood using multimodal nonlinear optical microscopy. Opt 30. Gurewitsch M True colors. Smithsonian 2008 July. Available at www.smithsonianmag. com/arts-culture/true-colors.html. Accessed January 4, 2014. Express 20(22):24623–24635. 31. Brinkmann V, Wunsche R (2007) Gods in Color: Painted Sculpture of Classical Antiquity 14. Nevin A, et al. (2010) Multi-photon excitation fluorescence and third-harmonic gen- (Glypotothek, Munich). eration microscopy measurements combined with confocal Raman microscopy for the 32. Panzanelli R, Schmidt ED, Lapatin K (2008) The Color of Life: Polychromy in Sculpture analysis of layered samples of varnished oil films. Appl Phys A Mater Sci Process from Antiquity to the Present (Getty Research Institute, Los Angeles). 100(3):599–606. 33. Mass JL, et al. (2013) The photodegradation of cadmium yellow paints in Henri 15. Warren WS, Fischer MC, Ye T (2007) Cellular imaging—novel nonlinear contrast im- Matisse’s Le Bonheur de vivre (1905-1906). Appl Phys A Mater Sci Process 111(1):59–68. proves deep-tissue microscopy. Laser Focus World 43(6):99–103. 34. Thoury M, et al. (2011) Near-infrared luminescence of cadmium pigments: In situ 16. Ye T, Fu D, Warren WS (2009) Nonlinear absorption microscopy. Photochem Photobiol identification and mapping in paintings. Appl Spectrosc 65(8):939–951. – 85(3):631 645. 35. Van der Snickt G, et al. (2009) Characterization of a degraded cadmium yellow (CdS) 17. Fu D, et al. (2007) High-resolution in vivo imaging of blood vessels without labeling. pigment in an oil painting by means of synchrotron radiation based X-ray techniques. – Opt Lett 32(18):2641 2643. Anal Chem 81(7):2600–2610. 18. Fu D, Matthews TE, Ye T, Piletic IR, Warren WS (2008) Label-free in vivo optical im- 36. Van der Snickt G, et al. (2012) Combined use of synchrotron radiation based micro-X- aging of microvasculature and oxygenation level. J Biomed Opt 13(4):040503. ray fluorescence, micro-X-ray diffraction, micro-X-ray absorption near-edge, and micro- 19. Fu D, Ye T, Matthews TE, Yurtsever G, Warren WS (2007) Two-color, two-photon, and fourier transform infrared spectroscopies for revealing an alternative degradation excited-state absorption microscopy. J Biomed Opt 12(5):054004. pathway of the pigment cadmium yellow in a painting by Van Gogh. Anal Chem 20. Fu D, et al. (2008) Probing skin pigmentation changes with transient absorption im- 84(23):10221–10228. aging of eumelanin and pheomelanin. J Biomed Opt 13(5):054036. 37. Johnson CR, et al. (2008) Image processing for artist identification. IEEE Signal Pro- 21. Matthews TE, Piletic IR, Selim MA, Simpson MJ, Warren WS (2011) Pump-probe im- cessing Mag 25(4):37–48. aging differentiates melanoma from melanocytic nevi. Sci Transl Med 3(71):71ra15. 38. Chen G, Little AV, Maggioni M, Rosasco L (2011) Applied and Numerical Harmonic 22. Wilson JW, et al. (2013) Imaging microscopic pigment chemistry in conjunctival mel- Analysis (Birkhauser, Basel). anocytic lesions using pump-probe laser microscopy. Invest Ophthalmol Vis Sci 54(10): 39. Ghosh KK, et al. (2011) Miniaturized integration of a fluorescence microscope. Nat 6867–6876. Methods 8(10):871–878. 23. Samineni P, deCruz A, Villafaña TE, Warren WS, Fischer MC (2012) Pump-probe im- 40. Durfee CG, et al. (2012) Direct diode-pumped Kerr-lens mode-locked Ti:sapphire laser. aging of historical pigments used in paintings. Opt Lett 37(8):1310–1312. Opt Express 20(13):13677–13683.

6of6 | www.pnas.org/cgi/doi/10.1073/pnas.1317230111 Villafana et al. Downloaded by guest on September 29, 2021