Simultaneous Organic and Inorganic Analysis of Colored Oriental Lacquerware by Pyrolysis-Gas Chromatography / Mass Spectrometry

1Tokyo Metropolitan Industrial Technology Research Institute, Tama Techno Plaza, Textile and Chemical Technology Group, Department of Materials Chemistry Application, 3-6-1 Azuma-cho, Akishima-shi, Tokyo 196-0033, Japan 2Department of Applied Chemistry, Meiji University, 1-1-1 Higashi-Mita, Tama-ku, Kawasaki-shi, Kanagawa 214-8571, Japan 3Rigaku Corporation, X-Ray Research Laboratory, Application Laboratories, XRD Analysis Group, 3-9-12 Matsubara-cho, Akishima-shi, Tokyo 196-8666, Japan


Introduction
Studies of valuable cultural heritage generally call for nondestructive methods of analysis.However, lacquered cultural heritage items are compound objects that generally incorporate a variety of distinct materials, including lacquers, wood, paper, cloth, soil, minerals, and metals.Such studies necessarily involve a combination of multiple analytical methods: observations using various microscopy techniques to assess an artifact's state of deterioration and reveal its internal structure; analysis of organic substances to assess whether or not a given object contains lacquers, oils, or similar ingredients; and elemental analysis to identify species of inorganic substances such as pigments or metals used for decoration.Most of these analytical methods involve destructive analyses, and the precious nature of cultural heritage makes it essential to restrict the quantity of measurement samples subject to destructive analyses to the absolute minimum required.This poses a challenge for methods of destructive analysis, to extract the greatest possible amount of information from the smallest possible volume and quantity of measurement samples.
The lacquer used in lacquerware is an organic substance and a resin.A general-purpose tool that has long been used to analyze organic substances of this type is Fourier-transform infrared spectroscopy (FR-IR).However, when the measurement sample is a mixture, or otherwise exhibits a complicated composition, FT-IR methods not only may fail to identify certain components but may also pose the risk of erroneous analysis.For this reason, the pyrolysis-gas chromatographymass spectrometry (Py-GC/MS) method is increasingly being used to analyze organic substances.The use of Py-GC/MS to analyze lacquer films allows the identification of chemical species in lacquer resins that cannot be identified in analyses  using IR spectra [1].For inorganic analysis of ingredients, such as inorganic pigments blended into paint layers or metallic powders and foils used to decorate lacquerware, many X-ray-based elemental analyses have been performed.However, the analyses of samples with multilayered structures that have been reported to date suffer from one or more of the following drawbacks.A first problem is the depth to which nondestructive X-ray analysis can probe beneath the surface layers.To remedy these difficulties, an analytical method based on scanning electron microscopy augmented by energy dispersive X-ray spectrometry (SEM-EDS) was proposed.This analytical method allows measurements that are pinpointed within individual layers.However, SEM-EDS requires the measurement sample to be rigidly fixed in place, and the carbon tape or other materials used to achieve this purpose may interfere with other analyses.This is the second of the problems mentioned above.
Understanding the material composition of lacquered artifacts of this sort requires the use of multiple analytical methods to gather data; full analyses can require large amounts of time and may be highly inefficient.For this reason, analytical methods that yield multiple types of information easily and rapidly from precisely targeted specific measurement sites are highly desirable.However, organic analysis and inorganic analysis are generally based on different physical principles, and for this reason it is difficult to analyze resins and pigments simultaneously.For these reasons, we have focused on analyses using Py-GC/MS, which allows analysis of tiny samples and offers outstanding ability to separate and identify sample constituents.A team of Italian researchers recently reported that Py-GC/MS could be used to detect the presence of mercury [2] and arsenic sulfide from analysis of analytical reagents [3].According to their results, Py-GC/MS analysis can offer molecular information (information on binding states) on arsenic (As) and sulfur (S).However, although the use of Py-GC/MS to detect mercury in lacquerware has been reported [2], no report has offered evidence to support the use of Py-GC/MS to detect arsenic in lacquerware [4,5].Arsenic sulfide has been used since antiquity as a yellow coloring agent, and it has a history of use in painting and craftwork.In the 1970s, arsenic sulfide was determined to be poisonous for nonpharmaceutical uses; its production as a pigment for lacquerware was terminated in Japan, and arsenic-containing yellow pigments for lacquerware were destroyed.As a result, notwithstanding the use since antiquity of arsenic sulfide as a lacquerware pigment, the detailed composition of yellow pigments containing arsenic sulfide remains unclear.We have obtained yellow lacquerware pigments used in ancient times by craftsmen in the Japanese city of Wajima; after analyzing 7 varieties of these pigments, we concluded that all were sulfur compounds containing no arsenic.Among these was the yellow pigment clearly identified as orpiment, but this was not arsenic sulfide.This demonstrates that information on arsenic sulfide used as a lacquerware pigment was mixed in those days.The continuing ambiguity surrounding the characteristics of the yellow pigments used in the lacquerware production of the past is problematic from the standpoint of the preservation of lacquerware cultural heritage.
For these reasons, we have performed Py-GC/MS measurements of red, yellow, and green-colored lacquer films applied to lacquerware items to assess the feasibility of simultaneously detecting resin ingredients together with certain pigments.We have also compared our findings to the results of SEM-EDS, X-ray fluorescence spectrometry (XRF), and X-ray diffractometry (XRD) measurements.The results we report demonstrate that a single measurement of a small-volume sample extracted from a pinpointed region of an artifact can simultaneously yield information on organic components and some inorganic components.

Materials and Methods
2.1.Samples.The test item was a tray in the Kamakurabori style decorated with a peony pattern colored with red, yellow, and green pigments, and samples were taken of each of the colors.Figure 1 shows the front and back surfaces of the full artifact, as well as the sampling points.The samples taken from the yellow, green, and red regions of the craft work for analysis are hereafter referred to as Samples Y, G, and R, respectively.Kamakura-bori is a style of Japanese craftsmanship in which lacquer coatings were painted in layers onto a design carved from wood; in this lacquerware tradition, ingredients such as makomo (wild rice) were used to accentuate the shading due to surface roughness in carved wood.This craft work uses both the top and the underside surfaces to express a carved image of a peony flower.The pistil at the center of the flower is colored yellow, while the leaves and other regions are colored green.Careful inspection of the leaf regions of the underside surface reveals a deep green color used to indicate the base of each leaf, which evolves gradually into a green color strongly inflected with yellow coloring as one proceeds out to the tip of the leaf; this lends the impression of a three-dimensional depiction of a leaf.

Surface and Cross-Section
Observations.Surface observation was conducted using a digital camera attached to an optical microscope (BX51 TRF-6 [D] and DP73-CU [D], resp., Olympus Co. Ltd.).Cross-section observations were conducted using a digital camera attached to an optical microscope (KH-770 and MX (G)-10C: OL-140II, HIROX Co. Ltd.).The relevant polished cross-sections were characterized by optical microscopy in order to examine all layers comprising the tray.Each sample was embedded in epoxy resin.After the epoxy resin had completely hardened, the surfaces of the embedded samples were ground flat and polished with wet sandpaper.

XRF.
X-ray fluorescence spectrometry (XRF) was performed using a NEX CG (polarized optics energy dispersive XRF, Rigaku Ltd.); Pd radiation was used for the X-ray tube.Secondary targets of two types were used, RX9 for sulfur (S) and molybdenum (Mo) for iron (Fe), arsenic (As), and mercury (Hg).The SDD detector was used.

SEM-EDS.
The inorganic materials in cross-sectional samples of the three colored areas were characterized by SEM-EDS.The SEM-EDS analysis was performed using a SEM (JSM-6610LV, JEOL Ltd.) coupled with an energy dispersive spectrometer (NSS 312E, Thermo Fisher Scientific K.K.).The accelerating voltage was set at 15 kV and the working distance at 10 mm.The range of qualitative analyses included elements in the periodic table ranging from boron to uranium.Analyses of two types of small-volume samples were conducted to ensure a cross-check on the results: (1) the outermost surface and (2) cross-sectional samples for each layer.In both cases, conducting carbon tape was used to fix the sample to the sample stage and analyses were conducted without evaporative coating.

XRD.
XRD was performed using a diffractometer (UltimaIV, Rigaku Ltd.) equipped with a sample vertical goniometer (radius of 285 mm) and Cu K radiation with ).The cup was placed on top of the pyrolyzer at near ambient temperature.The sample cup was introduced into the furnace when the furnace at 500 ∘ C and the temperature of the pyrolyzer interface and injector were set at 280 ∘ C, and then the temperature program of the gas chromatograph oven was started.The gas chromatograph oven was set at 40 ∘ C and held for 2 min, then programmed to provide a constant temperature increase at 12 ∘ C/min from 40 ∘ C to 320 ∘ C, and held for 10 min at 320 ∘ C. The flow rate of the helium gas was 1 mL/min.All pyrolysis products were identified by mass spectrometry.The mass spectrometry ionization energy was 70 eV (EI-mode), the cycle time was 1.94 scans per second, and the mass range was 29-800 /.Collected data were processed and analyzed by MSD ChemStation Software (Agilent Technologies, Ltd.).findings, we surmise that arsenic sulfide was used as a yellow pigment, while cinnabar (mercury sulfide, HgS) was used as a red pigment.However, binding information could not be obtained in cases where sulfur (S) and arsenic (As) formed compounds; thus we are unable to state definitively that the compound was orpiment (As 2 S 3 ).

SEM-EDS.
Table 2 lists the elements detected by SEM-EDS.It was determined with high certainty that Sample Y contained sulfur (S), iron (Fe), and arsenic (As); Sample Y was also believed to contain trace amounts of aluminum (Al), silica (Si), potassium (K), and calcium (Ca).It was determined with high certainty that Sample G contained sulfur (S), iron (Fe), arsenic (As), and silica (Si); Sample G was also believed to contain trace amounts of chlorine (Cl), potassium (K), calcium (Ca), and mercury (Hg).Sample R showed strong spectra of sulfur (S), arsenic (As), and mercury (Hg); also detected in trace amounts in Sample R were aluminum (Al), iron (Fe), and copper (Cu).Figure 3 shows the results of crystal analyses of Samples Y, G, and R. The analysis of Sample R was straightforward, but the analyses of Samples Y and G were considerably more complicated.Indeed, in the crystal analysis of Samples Y and G, it was necessary to inspect the spectra with proper consideration of the possibility of ingredients that might have been used as mineral pigments in the lacquerware tradition; this makes the analysis more difficult.Kitano performed XRD analyses of arsenic sulfide contained in unearthed lacquerware items and showed that the detection rate for arsenic sulfide was lower for green lacquer than for yellow lacquer [6].In the present case, assessing the realgar (As 4 S 4 ) thought to be contained in Samples Y and G amounts to the question of whether or not the sample contains a very small quantity of the substance; we were unable to compare the quantities present in the two samples.A comparison with the results of Kitano's measurements is difficult because  the detailed methods of those measurements are unclear; it is possible that the measurement sample was ground into a powder to enable the measurements.The present study is concerned with nondestructive measurements, and hence we do not grind measurement samples into powders; still, our finding that XRD spectra are more complicated when the lacquer sample contains arsenic sulfide is in good agreement with the findings reported by Kitano.In XRD measurements, determining the presence of arsenic sulfide is more difficult for samples of lacquer mixtures than for individual samples of arsenic sulfide (single-body mineral samples).Thus, although the question is somewhat dependent on the quantity of arsenic sulfide contained in the sample, in general it is best to avoid reliance on XRD measurements alone when determining whether or not oriental lacquer films contain arsenic sulfide.Comparing the two images of cross-sectional samples taken from Sample Y (Figures 4(a) and 4(b)), we see that although the uppermost painted layer in the Sample Y region is yellow in color, some samples taken from this regiondepending on the site from which the sample was collectedcontain red paint films as well.No red paint film lying beneath the yellow paint film was observed for the sample collected from the point closer to the center.We conclude that samples exhibiting red paint films were those taken from regions corresponding to petals of the peony; these regions surround the central region of the flower, which is covered in yellow paint.Presumably the artist was not excessively meticulous when painting the artifact, resulting in a coating of yellow paint.Inspection of the image of Sample G in Figure 4(c) reveals a yellow-painted region lying above a black paint film and marbled with regions of strong green and dark green coloration.This indicates that two shades of green paint were blended on the artifact before the wet paint solidified to form paint films.As seen in Figure 1, the dark green color present at the petiole of the leaves turns gradually into a green color with a strong yellow tint as one traces the leaf to its tip, conveying a sense of the three-dimensional nature of the leaf; we believe that this explains the cross-sectional structure observed in Figure 4(c).The image of Sample R in Figure 4(d) shows a red paint film lying atop a black paint film.The upper paint layer (the red layer) is dotted with yellow particles large in size compared to the red particles, suggesting the possibility that red and yellow pigments were blended.This was most likely done to yield a bright orange color for the petals of the peony that this tray seeks to depict.

SEM-EDS.
Figure 5 shows the results of elemental mappings conducted via SEM-EDS on the cross-sectional samples.Analysis of the red-colored layers observed in Figures 4(a) and 4(d) revealed that the same elements were detected in these two samples.For this reason, Figure 5(a) shows the results of analysis of the cross-sectional sample of Figure 4(a).Figure 5(b) shows the results of analysis of the cross-sectional sample of Figure 4(c).No elements originating from pigments were detected in the black-painted underlayers.The yellow-, green-, and red-painted layers observed in Figure 4 were all observed to contain elements believed to originate from pigments.
Inspecting the element mapping of the Y layer in Figure 5(a), we see that the points at which arsenic (As) is present roughly coincide with the points at which sulfur (S) is present, strongly suggesting the presence of arsenic sulfide.In contrast, in the R layer of Figure 5(a), the points at which arsenic (As) and sulfur (S) are present do not agree to the same extent; on the other hand, the points at which mercury (Hg) and sulfur (S) are present agree quite well.We thus conclude that the R layer of Figure 5(a) contains only small quantities of arsenic sulfide; instead, mercury sulfide constitutes the majority of the pigment in this region.Looking at the element mapping of the G layer in Figure 5(b), we see a lesser presence of arsenic sulfide than in the Y layer of Figure 5(a).Mercury (Hg) is also scattered throughout this region.

Analysis of Resins via Py-GC/MS.
To identify the resins used as paints, we performed Py-GC/MS data analysis.Figure 6 shows an ion chromatogram (/ 108)-extracted from the total ion chromatogram (TIC)-that originates from alkylphenol and is used in the principle component analysis of oriental lacquer.Essentially identical pyrolysis products were observed from all samples (Samples Y, G, and R).Our identification of these pyrolysis products allows us to determine [7,8] the substance in question as Toxicodendron vernicifluum (the academic name for a tree) for which urushiol is a primary component.T. vernicifluum is an oriental lacquer tree cultivated in Japan, China, and Korea.It shows that the paint applied to all colored regions of the Kamakurabori tray used sap tapped from T. vernicifluum.The substance detected with the greatest intensity camphor may be used as a diluting agent when applying coats of oriental lacquer.Because the wood carving of our sample exhibits surface roughness, the application of only an impasto coat of oriental lacquer would create a high risk of paint defects.We surmise  that camphor was used to dilute the oriental lacquer applied to the item to mitigate this risk.

Py-GC/MS Analysis of Pigments.
To investigate the ways in which S, As, and Hg-which may have been used as pigments-were present in our samples, we analyzed the results of Py-GC/MS analyses.
(1) Investigation of Sulfur (S) and Arsenic (As).Samples Y and G gave nearly identical results by Py-GC/MS analysis; the results of Sample R were also similar but of weaker intensity.As 4 (MW 300) and shows a typical fragmentation by sequential loss of As atoms (Figure 7(b)).This agrees with the results reported by Chiavari et al. [3,9].In the mass spectrometry analysis of Sample Y, elemental As was detected at / 74.9 for 1x As, / 149.8 for 2x As, / 224.7 for 3x As, and / 299.6 for 4x As.Chiavari et al. reported that the presence of As 4 indicates the use of orpiment (As 2 S 3 ) as a yellow pigment [3].We performed Py-GC/MS measurements of orpiment (As 2 S 3 ) and realgar (As while the mass spectra reported by Ma et al. [4] were not able to confirm a mass spectrum of As 2 S 3 but nonetheless identified the substance as orpiment (As 2 S 3 ).In the future, further analyses focusing on the binding states of arsenic (As) and sulfur (S) are necessary to investigate the detailed composition of yellow pigments containing arsenic sulfide.
Figure 8 shows ion chromatograms with / 300 (solid curves) and / 427.6 (dashed curves).Results for Sample Y are shown in Figures 8(a) and 8(b), and results for Sample G are shown in Figures 8(c) and 8(d).Samples Y and G both contained As 4 , corroborating the results obtained by both XRD analysis and SEM-EDS mapping analyses, which suggested the presence of isolated arsenic (As).As 4 was also strongly detected in Sample R. While As 4 S 4 was strongly detected in Sample Y, irrespective of the position on the artifact from which the sample was collected for analysis, the intensity of the As 4 S 4 that was detected was variable in Sample G.The reason for this disparity may be due to the varying amounts of differently colored lacquer that were blended together when the artifact was made; for example, in order to elicit stronger coloration in the green-colored region of Sample G, yellow-green lacquer containing a strong yellow tint was blended with a deep green lacquer containing a weak black-inflected yellow tint to produce the two distinct lacquer layers that are visible in the cross-section shown in Figure 4(c).Consequently, the intensity with which As 4 S 4 was detected varied depending on the precise location from which the material was collected for Py-GC/MS analysis.
Most reports of scientific investigations involve the use of X-ray techniques to confirm the presence of the elements arsenic and sulfur-findings which are habitually attributed to As 2 S 3 (orpiment) despite the absence of molecular-level information.Many studies in the field of cultural artifact research have used methods of Raman spectroscopy to identify molecular-level signatures of arsenic sulfide [10].However, in the case of lacquerware artifacts, the existence of light emission from lacquer-coating layers prevents the collection of analyzable spectra, and thus Raman spectroscopy cannot be used for such objects.On the other hand, these findings showed that information on binding states, which is difficult to infer based on XRD analysis, can easily and rapidly be obtained from Py-GC/MS analysis.Indeed, depending on the analytical methods employed, a single measurement can yield a variety of information regarding the use of pigments.
(2) Investigation of Mercury.Many naturally occurring elements are composed of multiple stable isotopes of different mass numbers [11].For mercury, there are 7 stable isotopes, and the use of Py-GC/MS allows the stable isotopes of mercury to be detected with ease [2].The total ion chromatography (TIC) and the / 202 ion chromatograms of standard HgS powder are shown in Figure 9 together with mass spectra and isotope ratios inset.The values listed in the table of isotope ratios were derived as follows.Following the literature by taking the 198 Hg isotope to be 100, the isotope ratios of 196 Hg, 198 Hg, 199 Hg, 200 Hg, 201 Hg, and 204 Hg were each computed from the mass spectra, and the measured values were compared to values available in the literature.From the TIC peak in Figure 9, a total of 6 isotope peaks were detected, with / 202 being the primary peak (Figure 9 inset).The isotope ratios are close to those reported in the literature, whereupon we conclude that these mass spectra originate from mercury [12].
The TIC and the / 202 ion chromatograms of Sample R are shown in Figure 10.Because the mercury isotope peaks  were detected with greatest intensity in Sample R, Figure 10 shows results of measurements on Sample R. Isotope peaks were not detected for Sample Y. On the other hand, for Sample G, isotope peaks similar to those of Sample R but of weaker intensity were detected.The results of these measurements agree well with the results of XRF, SEM-EDS, and XRD analyses.Inspecting the table of isotope ratios in the inset of Figure 10, we see results that lie closer to literature values than the measured values in the inset of Figure 9.However, in Figure 9, which measured powders of a mercury sulfide (HgS) reagent, a strong spectrum was observed at 7 minutes and 30 seconds; in contrast, for Sample R, a strong spectrum was observed at 11 minutes and 30 seconds.Because we did not make any significant modifications to the state of the measurement apparatus, the origin of this discrepancy in the retention time is unclear and remains a question for future work.

Conclusions
Upon comparing the results of the XRF and XRD analyses, SEM-EDS measurements of the outermost layers and the cross-sectional layers, and Py-GC/MS measurements, we arrive at the following conclusions.
(    two elements are present coincide.However, it is difficult to pinpoint the binding configuration of the As and S atoms.The use of XRD measurements allows determination of the binding states of As and S; however, in cases involving mixtures with lacquer, the peaks that allow detailed analysis are difficult to observe, and it is challenging to confirm the presence of substances.In contrast, the use of Py-GC/MS readily allows determination of specific binding states of As and S-such as As 4 S 4 -whereupon the presence of such substances may be rapidly discovered. (2) Observations regarding Substances Other Than Arsenic Sulfide.We were able to verify the presence of mercury via elemental analyses using XRF, SEM-EDS, and XRD, while Py-GC/MS identified mercury by detecting the mass spectra of its isotopes.We were able to verify the fact that mercury was present in the form of cinnabar (mercury sulfide) from the results of XRD analyses.The presence of iron was confirmed by XRF, SEM-EDS, and XRD.The results of XRD analyses indicated the possibility that iron was present in the form of hematite (Fe 2 O 3 ) or magnetite (Fe 3 O 4 ).Therefore, Py-GC/MS analyses simultaneously yielded molecular-level information on arsenic (As) and sulfur (S) together with detection of the primary ingredients of the lacquer.We were also able to confirm the presence of mercury (Hg), in agreement with the results of previous work [2].These findings demonstrate that Py-GC/MS is capable of identifying the presence of arsenic (As), sulfur (S), and mercury (Hg), even when mixed together in a lacquer film; thus Py-GC/MS may be used for simultaneous analysis of organic and inorganic substances.

Figure 1 :
Figure 1: Photographs of a Kamakura-bori tray decorated with a peony design (diameter: 31 cm); (a) top, (b) underside.Locations of collected samples (I) are indicated.

Figure 2 :
Figure 2: Microphotographs showing surface conditions of Samples Y, G, and R.

Figure 4
shows microscope images of cross-sectional samples from the various color regions.Images (a) and (b) are both of cross-sectional samples taken from Sample Y, while images (c) and (d) are cross-sectional samples of Samples G and R, respectively.All of these samples included an underlayer of black paint.

Figure 4 :
Figure 4: Cross-sectional photos of Sample Y (a) and (b), Sample G (c), and Sample R (d).

Figure 7
collects the results for Sample Y.

Figure 6 :
Figure 6: Ion chromatograms of the / 108 fragment ion of the alkylphenol of Sample Y (a), Sample G (b), and Sample R (c).

4 S 4 )
from commercial suppliers.Measured results for the mineral samples of orpiment reveal mass spectra identical to those observed near Peaks B and C (Figures 7(b) and 7(c)).However, the As 3 S 3 and As 4 S 4 features observed near Peak D (Figure 7(b)) were not seen for the orpiment mineral sample.In contrast, the mass spectra detected for the mineral sample of realgar contained the As 3 S 3 and As 4 S 4 features seen near Peak D (Figure 7(b)).On the other hand, none of the samples we measured indicated As 2 S 3 (MW 246) in their mass spectra.Chiavari et al. did not perform measurements with As 2 S 3 as a standard reagent [3],

Figure 7 :
Figure 7: Py-GC/MS analyses of Sample Y. (a) TIC, (b) ion chromatograms of Peak B, (c) ion chromatograms of Peak C, and (d) ion chromatograms of Peak D.

Figure 8 :
Figure 8: Ion chromatograms of the / 300 and / 427.6 fragment ions of Sample Y (a) and (b), Sample G (c) and (d).

Figure 9 :
Figure 9: TIC and ion chromatograms of the / 202 fragment ion of standard HgS powder.

Figure 10 :
Figure 10: TIC and ion chromatograms of the / 202 fragment ion of Sample R.

Table 1 :
Results of surface analysis by XRF.∘ divergence slit, an 8 mm scattering slit, and a 13 mm receiving slit.A one-dimensional silicon strip detector (D/teX Ultra, Rigaku Ltd.) was used.2.2.5.Py-GC/MS.Very small samples were taken from each of the color regions of the craft work and analyzed by Py-GC/MS.Py-GC/MS measurements were carried out using a vertical microfurnace-type pyrolyzer (PY-2020iD, Frontier Lab.Co. Ltd.) with a 6890N/5975 GC/MS system (Agilent Technologies, Ltd.).A stainless steel capillary column (30 m, 0.25 mm I.D.) coated with 0.25 m Ultra Alloy PY-1(MS/HT) (100% methyl silicone) was used for the separation.The sample (0.5 mg) was placed in a stainless steel sample cup (Eco-cup, Frontier Lab.Co. Ltd.

Table 2 :
Results of surface analysis by SEM-EDS.
C, O, Al, Si, S, Ca, K, Fe, As G Green C, O, Si, S, Cl, K, Ca, Fe, As, Hg R Red C, O, Al, S, Fe, Cu, As, Hg

Table 3 :
Results of surface analysis by XRD.