Analysis of bulk and inorganic degradation products of stones, mortars and wall paintings by portable Raman microprobe spectroscopy
- First Online:
- Cite this article as:
- Pérez-Alonso, M., Castro, K., Martinez-Arkarazo, I. et al. Anal Bioanal Chem (2004) 379: 42. doi:10.1007/s00216-004-2496-2
- 407 Views
This work reports the use of a portable Raman microprobe spectrometer for the analysis of bulk and decaying compounds in carbonaceous materials such as stones, mortars and wall paintings. The analysed stones include limestone, dolomite and carbonaceous sandstone, gypsum and calcium oxalate, both mono- and dihydrated, being the main inorganic degradation products detected. Mortars include bulk phases with pure gypsum, calcite and mixtures of both or with sand, soluble salts being the most important degradation products. The pigments detected in several wall paintings include Prussian blue, iron oxide red, iron oxide yellow, vermilion, carbon black and lead white. Three different decaying processes have been characterised in the mortars of the wall paintings: (a) a massive absorption of nitrates that reacted with calcium carbonate and promoted the unbinding of pigment grains, (b) the formation of black crusts in the vault of the presbytery and (c) the thermodecomposition of pigments due to a fire.
KeywordsPortable Raman spectroscopyBulk and degradation analysisStoneMortarsWall paintings
Air pollution, both outdoor and indoor, has a significant influence on the weathering of monuments and on the deterioration of artworks like paintings, frescoes, sculptures and wallpapers. The most relevant atmospheric pollutants to stone decay and artwork deterioration are sulphur and nitrogen oxides , particulate matter including carbonaceous particles containing metals (V, Fe, Cu, Cr, Pb, etc.) from combustion engines, carbon dioxide, ammonia, ozone, acids like hydrofluoric and hydrochloric, reduced sulphur compounds like H2S or sulphides, mercaptans, carbonyl sulphides and, moreover, sea spray in coastal areas . Also the deterioration caused by microorganisms must be taken into account [3, 4].
The outside stones of facades and decorative artworks, especially if they have some fractions of CaCO3 (marbles, limestones, sandstones, mortars, frescoes, etc.), are among the materials deserving special attention as they are vulnerable to the attack of gaseous acids. The resulting soluble salts, which crystallise in the pores and on the surfaces, can be observed very quickly [5, 6]. However, damage to indoor stones and artworks are observed in the long-term, the effects being quite similar to those observed on the outside, i.e. formation of gypsum, nitrates, chlorides, fluorides, black crusts or blackening effects.
Air pollution is not the only cause of damage to stone monuments. In some cases, colonisation by pioneer organisms, such as lichens, is the most important damage factor . Climatic wetting and drying of the thallus causes it to expand and contract, which in turn mechanically disrupts the substratum . On the other hand, the chemical effects of lichens on rocks and minerals have been intensively investigated [3, 9, 10]. The best known of all these chemical processes is the secretion of oxalic acid (blue-green algae, various fungi and numerous bacteria), which leads to the formation of soluble cation complexes and insoluble calcium oxalate [3, 9, 10].
The bulk and the composition of inorganic deteriorated materials can be determined by many analytical techniques . Among them X-ray fluorescence spectroscopy (XRF), scanning electron microscopy (SEM) [12, 13] and ion chromatography (IC) [14, 15] have been the most used techniques in this field, although practically all the spectroscopies for surface analysis have been tested. Most of these techniques require sampling of the materials (recently portable SEM and XRF instruments are commercially available ) and, depending on the complexity of the buildings and/or artworks to be examined, the number of samples could be high in order to perform an adequate diagnosis.
Among the spectroscopic analytical techniques, Raman spectroscopy has been used in recent years in the field of cultural heritage; the determination of pigments in wall paintings [17, 18, 19, 20], icons , pottery , wallpapers  and bronze  are some examples. The use of portable equipment has been explored recently [25, 26, 27, 28, 29] as soon as these instruments became more commercially available. The advantages of these alternatives are clear because in situ analysis can be performed without the needing of sampling. In addition to this, the number of analyses is not a key factor when portable instruments are used.
This work reports the use of a portable Raman spectrometer to analyse the bulk and inorganic degradation products in stones, mortars and frescoes. A similar instrument has been used by a few authors to analyse pigments on maps , wall paintings  and archaeological material [28, 29].
Description of the samples
Several kinds of carbonaceous samples, subjected to different atmospheric conditions and ages, have been analysed by portable Raman microprobe spectroscopy.
Round tombstones at the San Telmo Museum, San Sebastian, Basque Country
The round tombstones are funeral monuments formed by a disc and a trapezoidal base, which is set up over a tomb. The anthropomorphic menhirs are said to be their immediate predecessors or forerunner and they are some of the most remarkable symbols of the Basque culture. These monuments were quite numerous in the region between the Garone (France) and Ebro (Spain) rivers between the second and seventeenth centuries. A large number of these memorial stones are preserved nowadays in museums.
One of the most valuable collections of round tombstones, regarding archaeological significance, is likely to belong to the San Telmo Museum at San Sebastian (Basque Country). Three of them, a Roman round tombstone (second century), a round tombstone A 113 (ninth century) and a round tombstone A 132 (fourteenth century), were selected for technical examination.
Mortars and wall paintings of Santa María de Hermo church, Asturias
The church of Santa María, located in the Monasterio de Hermo, Asturias, is under restoration. Prior to the curators’ intervention, the church showed serious problems of dampness, as well as clear evidence of flora, fungus and mould presence. The samples of mortars and wall paintings analysed in the present work were hidden bellow a white coating plaster from the eighteenth century.
During its restoration, the restorers found that some plasters showed a black patch of damp and observed that the ones more affected by the dampness exhibited a noticeable increased hardness, being quite difficult to remove. The restorers wanted to know if a film of calcium oxalate had been formed due to biological attack. Moreover, the remains of the wall paintings showed such a high degradation that the pigment grains were removed by simple brushing.
Graveyard near the sea, Getxo, Basque Country
Several samples of white marble and calcite stones pertaining to monuments in the graveyard of Getxo (Basque Country) have been analysed. The cemetery is located very near the sea (100 m). In less than 60 years the marbles have lost their original appearance, showing a blackish aspect (in the less degraded areas) and rough surface (in the most degraded ones), while the calcite stones of the southeastern facades show black crusts.
Church in a non-polluted area, Villalibado, Burgos
A church in a non-polluted area was selected as a pattern of natural decaying. The church is located in Villalibado, a very small village in the north of Burgos (Spain), which is actually made up by not more than ten country houses. In its surroundings there is no industrial activity, agriculture being the main source of economic development in the region. The climatic conditions are completely continental.
The oldest part of the church dates back to the twelfth century. Although its conservation state is thoroughly poor (in fact some structural parts of the building are knocked down), the Romanesque structure still remains. Those stones have been subjected to chemical analysis.
Wall paintings from the Cristina Enea Palace House, San Sebastian, Basque Country
The wall paintings under study belonged to the blue ceiling of a room in the lowest storey of the Palace House. In these blue areas some small yellow grains can be observed with the naked eye. The house was so seriously damaged by a fire that the conservation of the wall paintings was questioned. Unfortunately, its restoration was finally rejected and at present the paintings just do not exist.
The analysis of all samples was performed, before the restoration processes, using a portable Raman microprobe Renishaw RA 100 spectrometer with an excitation wavelength of 785 nm (diode laser) and a CCD detector (Peltier cooled), resulting in non-destructive analysis. The laser power was lower than 50 mW in every case to prevent the thermodecomposition of the samples. The equipment was calibrated using the 520.5 cm−1 silicon line.
The Raman spectra were taken with a resolution of 2 cm−1 in the range between 2500 and 200 cm−1. Several acquisitions of each spectrum were accumulated in order to improve the signal-to-noise ratio. A microscope lens in the microprobe (20×) allowed the laser beam to be focussed on approximately 40-μm spots. A microvideo camera helped the user in the focussing process. The samples were always focussed to obtain the maximum Raman intensity response. Data acquisition was carried out with the Spectracalc software package GRAMS (Galactic Industries, Salem, NH, USA). For the analysis of the results the Omnic software provided by Nicolet was used.
Some samples, those having compounds with a low Raman scattering, were also analysed by FT-IR spectroscopy using a Nexus spectrometer (Nicolet) in the transmittance mode. To obtain a good signal-to-noise ratio, 40 scans were accumulated for each spectrum at a spectral resolution of 4 cm−1, between 4000 and 400 cm−1 (mid-infrared spectral region). The samples were prepared by mixing approx. 0.5 mg of each material with 150 mg of anhydrous KBr in an agate mortar and then pressing under 10 tons to obtain transparent pellets. These pellets, after drying, were mounted in the spectrometer and the spectra were collected. Data acquisition was carried out with the Omnic software (Nicolet).
Spectral analysis was performed by comparing the spectrum of each sample with those in a home-made database of pigments, minerals and other compounds [30, 31] liable to be found in artworks and monuments, even as degradation products. The database was constructed previously using pure materials (standards) supplied by several dealers and it is continuously upgraded with new compounds.
Results and discussion
Summary of the Raman positions in pure compounds analysed by Raman Spectroscopy
Raman bands (cm−1)
1761 vw, 1442 vw, 1098 vs, 882 vw, 725 w, 339 vw, 301 m
1136 m, 1008 vs, 670 m, 619 m, 493 m, 414 m
1748 w, 1436 w, 1086 vs, 1067 w, 712 m, 282 s
1528 m, 1451 w, 1339 w, 1307 vw, 1214 vw, 1184 vw, 1160 vw, 1142 w, 1107 vw, 1081 vw, 953 w, 773 w, 747 m, 681 m, 638 w, 593 w, 464 vs, 395 w, 355 w, 290 w, 261 w, 230 w, 203 m
1727 w, 1629 m, 1490 vs, 1462 vs, 1396 w, 942 w, 896 s, 865 w, 783 vw, 596 m, 521 m, 503 m, 284 sh, 249 m, 222 s, 205 vs
1769 w, 1631 w, 1534 vw, 1476 vs, 1414 w, 910 s, 869 w, 785 vw, 600 w, 506 m, 414 w, 399 w, 263 w(sh), 224 sh
1600 br, 1315 br
551 vs, 478 w, 457 w, 390 s, 314 s, 227 s
Iron oxide red
1536 w, 1336 w, 1215 w, 739 w, 686 w, 610 m, 495 w, 412 s, 294 vs, 244 w, 227 s
Iron oxide yellow
550 m, 472 w, 388 s, 301 s, 246 m, 227 w, 205 w
344 s, 285 m, 254 vs
Sodium nitrate NaNO3
1777 vw, 1665 vw, 1385 w, 1067 vs, 724 s, 535 vw, 521 vw, 416 vw, 190 m
Potassium nitrate KNO3
1776 vw, 1359 w, 1344 w, 1049 vs, 714 w, 536 vw, 517 vw, 416 vw, 198 vw
Ammonium nitrate NH4NO3
1654 vw, 1464 vw, 1412 vw, 1287 vw, 1042 vs, 714 m, 193 m
Calcium nitrate Ca(NO3)·4H2O
1426 vw, 1354 vw, 1119 m, 1050 vs, 743 w, 719 w, 197 w
1523 vs, 1450 m, 1337 s, 1307 w, 1185 w, 1143 m, 1009 w, 954 w, 842 vw, 832 vw, 776 w, 747 s, 680 m, 594 w, 483 w, 257 w
Lead white 2PbCO3.Pb(OH)2
1413 sh, 1363 m, 1051 vs, 974 vw, 865 vw, 680 w, 413 m, 325 m, 271 sh
Prussian blue Fe4[Fe(CN)6]3
2154 s, 2120 w, 2092 m, 1042 vw, 952 w, 588 sh, 535 s, 507 sh, 479 vw, 330 w, 278 m
426 w, 384 w, 290 vs, 217 w
In the other two round tombstones, the results of the analyses are rather similar. The Raman spectra obtained from the Roman round tombstone showed bands of calcite CaCO3 (1086, 712 and 282 cm−1) and quartz (464 and 355 cm−1), leading to the conclusion that it is a carbonaceous sandstone (Fig. 1b does not show the presence of quartz, because the microprobe was focussed in a carbonaceous grain). In round tombstone A 132, all the spectra from the bulk contained the sharp band at 1086 cm−1, in combination with the weaker 712 and 282 cm−1 Raman bands, that are characteristic of the presence of calcite (Fig. 1c).
Despite the different bulk composition due to the carbonaceous nature of the stones, gypsum was the only degradation product determined (see Fig. 1), although in round tombstone A 132 the gypsum was only detected in areas of black crust. Several authors explain that these black crusts are formed by particulate material trapped during the growth of gypsum crystals, and that these black crusts might also contain calcium oxalate crystals (in different hydration states) and metallic oxides, that work as catalysts in the oxidation of SO2 to sulphuric acid [34, 35]. Furthermore, several works have demonstrated that the gypsum in the black crust can play a possible role as a nutritional source for the growth of some microorganisms .
Mortars and wall paintings, Santa María de Hermo, Asturias
The increased hardness of the mortars in black areas of the ceiling in the presbytery, in comparison with other mortars inside the church, made the restorers initially think that a film of calcium oxalate had been formed on the surface of the mortars. After careful study of the components in the three mortar strata, gypsum (1135, 1008, 493 and 414 cm−1) was revealed as the bulk composition in the most external stratum, calcite in the intermediate one and a mixture of calcite and sand (681 and 464 cm−1) in the innermost one.
The bulk of the wall paintings on the vertical walls and plasters is calcite, containing gypsum and silicates in some strata. The chromatic palette for the original paintings consists of the following pigments: minium (551 and 390 cm−1), iron oxide red (610, 412, 294 and 227 cm−1), iron oxide yellow (550, 472, 388, 301, 246 and 205 cm−1), vermilion (344, 285 and 254 cm−1), carbon black of vegetable origin (IR confirmed the absence of phosphates) (1600 and 1315 cm−1) and two organic pigments in a modern repainting (1150–1650 cm−1). The Raman study carried out over the drawings, together with the historical study by the restorers, finally dated the wall paintings to between the fifteenth and seventeenth centuries.
The migration of nitrate from the old cemetery attached to the north wall of the church can explain its presence in the analysed samples, and the nitrate probably does not correspond to a single compound. During corpse decomposition several volatile gases are generated and released in substantial quantities. One of these gases is NH3, and if it is not released to the atmosphere it could react with soil compounds to provide ammonium and nitrate salts. Much of the damage suffered by frescoes on the plaster could originate from the effects of nitrate percolation and upward migration through the stone, as has been detected in other works . The preliminary assumption based on Raman spectroscopy for the existence of nitrates inside the mortars and wall paintings was confirmed by FT-IR spectroscopy and ion chromatography.
Graveyard near the sea
The outcome of this weathering for less than 60 years, which is visible to the naked eye, is an intense breakdown in the homogeneity and clear aspect of these materials. All this noticeable damage suggests a high atmospheric pollution. The graveyard is located on the right side of Bilbao’s estuary and just opposite where an oil refinery is located (left-hand side of the estuary). The predominant winds transport the emissions of the factory to the cemetery. Airborne sulphur dioxide (SO2) interacts with calcite to form gypsum. The location near the Atlantic sea favoured the washing of degradation products by rainfall. In this sense no traces of sulphates or sulphites have been found in samples from facades open to the sea. However, in the protected facades, gypsum is not washed out, promoting the formation of a black crust by a mechanism well described by Del Monte et al. .
Church in a non-polluted area
Apart from the existence of both calcium oxalates, the presence of a mixture of nitrate salts (Raman bands in the region 1067–1040 cm−1) in the stones of the wall next to the graveyard has been determined. Although these salts have Raman spectra, the scattering of these compounds is rather low (in comparison with carbonates, the bulk of the materials), thus IR spectroscopy was used to confirm their presence. The nitrates appearing in the stones might be the result of migration  from the old graveyard close to the church, as has been previously discussed in this work.
Wall paintings from the Cristina Enea Palace in San Sebastian
In other blue areas of the wall paintings, closer to the location of the fire, yellow grains were visually observed. The aesthetical aspect of these blue areas looks like the yellow grains (approximately 20 grains/cm2) were mechanically sprayed. The Raman analysis of these yellow particles revealed the presence of massicot, PbO (384 and 290 cm−1). There was a doubt whether the massicot (spectrum a in Fig. 7) was a degradation product formed in the fire or whether it was formed in the analysis due to the too-high power of the laser . In areas of yellowish appearance the spectrum for massicot was clearly seen even with the lowest laser power, and no other yellow pigment was detected. Because of the failed detection of a yellow pigment it has been concluded that the yellowish aspect of the blue ceiling must be due to the real presence of massicot. In separate experiments it was checked that pure lead white does not burn under the most aggressive measurement conditions of the Raman microprobe equipment, like Burgio and coworkers had shown earlier . Thus, it was finally concluded that the presence of massicot was due to the fire.
The usefulness of the portable micro-Raman equipment has been demonstrated for the determination of bulk and decaying products in several carbonaceous materials such as stones, mortars and wall paintings. The identification of the degradation products has great importance when a restoration process must be executed on an artwork.
The use of the video camera coupled to the microprobe provides really good focussing of the sample point, making it possible to distinguish between different microscopic crystals within the sample. Furthermore, by viewing the sample before and after the collection of the Raman spectra, any thermodecomposition of the sample is easily observed.
One of the most remarkable characteristics of this technique is the non-destruction of the sample analysed. If the equipment is used by trained people, the analyses are really non-destructive, making it possible to perform them in situ without the necessity of sending the artwork to the laboratory. Even the cracks on the surface can be used to analyse inner layers of the artwork .
Better results can be obtained by the combination of Raman microprobe spectroscopy with other analytical techniques like IR spectroscopy and XRF. There are some compounds with low Raman scattering, which are hardly determined by Raman spectroscopy. For those situations the employment of FT-IR spectroscopy could help in the analysis of this kind of material. In this work the presence of nitrate salts suggested by the Raman analysis was further confirmed by the presence of the band at 1384 cm−1 in the FT-IR spectra, and by ion chromatography. Or for situations with multiple Raman signals, an elemental analysis with XRF will help in the selection of candidate materials to be found in the sample.
This work has been financially supported by projects No. UPV 14590/2002 and UE02-A06. M. Pérez-Alonso, I. Martinez-Arkarazo and K. Castro acknowledge the grants from the Basque Government, the University of the Basque Country (UPV/EHU) and the Spanish Government, respectively. The authors acknowledge the referees’ suggestions, especially one of them who helped us in identifying the nature of the biofilm.