SEARCH

SEARCH BY CITATION

Keywords:

  • mediaeval murals;
  • crocoite;
  • micro-Raman;
  • micro-XRD;
  • pigment degradation

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results and discussion
  6. Conclusions
  7. Acknowledgements
  8. References
  9. Supporting Information

Mural paintings of exceptional quality, which can be discerned in spite of their extensive mechanical damage and colour fading, have been uncovered in the church of St. Gallus in Kuřívody, Northern Bohemia, dated to the second half of the 13th century. Materials research with particular use of portable X-ray fluorescence, Raman micro-spectroscopy and powder X-ray micro-diffraction revealed the presence of rare pigments. In Kuřívody, it is only a second identification of intentionally used yellow mineral crocoite (PbCrO4) in European art. Its identification is facilitated by providing a very good Raman scattering, even when present in small amounts in fragmentarily preserved colour layers. Light yellow mimetite (Pb5(AsO4)3Cl) was never before mentioned as intentionally used pigment in Europe. Its finding in Kuřívody, however, corresponds more likely with undesirable physical–chemical conditions causing its formation by alteration of orpiment (As2S3) and minium (Pb3O4). Obtained results highlight the importance of Raman spectroscopy for direct identification of mineral pigments in low concentrations, which may be crucial for interpreting cultural heritage objects in historical context. By materials, the almost forgotten paintings in Kuřívody can be seen as outstanding and rare example of ancient artistic tradition that has spread to Europe from Mediterranean in early Middle Ages. After all, mineral crocoite was already used by ancient Egyptians to paint sarcophagi and degraded orpiment decorates the walls of the Nefertari's tomb in Thebes. Copyright © 2014 John Wiley & Sons, Ltd.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results and discussion
  6. Conclusions
  7. Acknowledgements
  8. References
  9. Supporting Information

The church of St. Gallus (formerly dedicated to St. George) was built in the 13th century as a part of the settlement located around the Bezděz Castle in Northern Bohemia. (Fig. S1, Supporting information) The castle was founded by Przemysl Ottokar II (1233–1278), king of Bohemia. Its construction was finished almost simultaneously with the construction of the church, or shortly before – at the end of the 13th century, the church has already been mentioned in local chronicles. The increasing political importance of this domain in Middle Ages has led to a creation of large-scale figural wall paintings in the church's interior. They were executed probably at the beginning of the 14th century and followed the original simple decorations created immediately after the construction of the church. In the following decades, water shortage caused a gradual decline of the town and led to the change of its name from Freistadt (‘free royal town’) to Kuřívody (in Czech) or Hühnerwasser (in German) with literal translation ‘Fowl's water’.[1] The existence of the wall paintings has been completely forgotten. The gradual decrease of the church's state has culminated in the second half of the 20th century, when it reached the verge of total destruction as a result of (i) displacement of the prevailing German population after the World War II; (ii) incorporation of the town into a military area during communist era in the 1950s; and (iii) occupation by Soviet Army in 1968 and the stay of its soldiers in this military area for the next 23 years. The tremendous devastation of the church, resulting from its usage as, e.g. vegetable or ammunition depot represents one of many examples of damage caused by Russian army in occupied areas during the 20th century.

The recent efforts to revive the church led to partial uncovering of the paintings. Although numerous experts predicted the existence of Gothic wall paintings in the church, they were surprised by their exceptional quality that can be discerned in spite of their extensive mechanical damage and colour fading. (Fig. 1 and Fig. S2, Supporting information) In the region of Central Europe, there are only few examples of materials investigation of wall paintings dated around 1300 or earlier, and these remain mostly unpublished. The reasons are varied: either the paintings are poorly preserved, they remain uncovered (hidden under later repaints), or they have not yet been studied by scientific tools. Their existence is, however, mentioned in art-historic literature.[2, 3] In general, from this period of time, one can find more relevant data regarding the material composition of works of art produced in the regions of southern Europe, e.g. Italy, Spain, Greece, Crete, or Turkey (former Byzantium). It is consistent with the spread of the painting technology from the ancient Mediterranean to Europe.

image

Figure 1. Uncovered murals on the southern wall of the presbytery in the St. Gallus Church in Kuřívody, Czech Republic (before 1300), representing the scene with St. Dorothy and Infant Jesus on the left side and only partially preserved scene of sacrifice in the temple on the right side (photo V. Potůček); sampling points are indicated by numbers corresponding to descriptions given in Table 2 (sampling points 9, 11 and 12 are not shown, because they are situated in other parts of walls with ornamental decorations only, and also on the ceiling).

Download figure to PowerPoint

In early mediaeval European art, it is quite rare to identify pigments like natural crocoite (PbCrO4) or chrysocolla ((Cu, Al)2H2Si2O5(OH)4·nH2O), which are known, e.g. from Egyptian funerary artefacts.[4-6] The only evidence comes from the 13th century wall paintings under the Siena Cathedral, Italy, where they were both unambiguously identified by micro-Raman spectroscopy and X-ray diffraction techniques.[7] Tyrian purple and Egyptian blue represent similar examples of pigments related solely to ancient world, which can be found only rarely in the mediaeval European art. The Tyrian purple has been indicated (particularly by the presence of bromine) in mixture with other dyes in the early mediaeval illuminated Byzantine manuscript (created in the 6th century) [8] using non-invasive techniques (e.g. Raman spectroscopy, UV–Vis diffuse reflectance spectrophotometry with optic fibres and X-ray fluorescence spectrometry). Multianalytical approach combining non-invasive as well as non-destructive laboratory tools has been used for identification of Egyptian blue, i.e. cuprorivaite (CaCuSi4O10), in Cretan wall paintings created in Roman and also Early Byzantine period (5th to 6th century AD).[9]

Other traditional ancient pigments have been employed by mediaeval painters to a much larger extent. For illumination of early mediaeval manuscripts, orpiment (As2S3), massicot (PbO), or yellow organic lakes were traditionally used in yellows, cinnabar (HgS), minium (Pb3O4), Fe-based pigments or madder in reds, ultramarine (lazurite, Na8–10Al6Si6O24S2–4), azurite (2CuCO3·Cu(OH)2) or indigo in blues, copper or earth-based greens in greens, white lead (2PbCO3·Pb(OH)2) in whites, and carbon black in blacks. This typical palette has been identified, e.g. in Cistercian 12th–13th century manuscripts in Portugal[10] as well as in the early 12th century manuscript ‘Liber Floridus’ stored in the Ghent University Library, Belgium. [11] The same palette appeared also in Pre-Romanesque, Romanesque, and Early Gothic wall paintings, e.g. in the 12th century monastery of Santa Maria delle Cerrate, Puglia, Italy,[12] or the 11th–13th century painted fragments in San Giovanni Battista Church in Cevio, Switzerland.[13] In addition to pigments used very frequently in Middle Ages, one can find numerous examples of pigments regionally of even locally specific – e.g. bluish clay mineral aerinite, [14] hydrated iron phosphate vivianite[15, 16] and others.

Formal employment of pigments used for manuscripts' illumination in wall paintings while not taking into account physical and chemical conditions of their usage gradually led to numerous displays of degradation – naturally, wall paintings are much more endangered by possible negative effects of lighting, humidity, or dissolved salts than the manuscripts. Colour changes (e.g. darkening of minium caused by its alteration to plattnerite (PbO2),[17] fading of red realgar or yellow orpiment through transition to arsenolite,[18] or blackening of copper-based pigments in alkaline environment by their transformation to tenorite (CuO)[19]) led, eventually, to changes in painting technology. Simple lead oxides (massicot, minium) were partly substituted by mixed oxides (Pb-Sn and Pb-Sb yellows); in fresco-applied paints, they were completely replaced by Fe-based pigments, which are much more stable. In the wall paintings created from the 14th century onward, arsenic sulfides are rare, and Pb-based and Cu-based pigments are applied only in secco technique (usually bound by organic binders). It has been documented in numerous mural works of art: the Cretan 14th century wall paintings in Rethymno,[20] the 15th century murals in the Batalha Monastery, Portugal,[21] and in the Main Town Hall of Gdansk, Poland,[22] in Our Lady's Cathedral in Antwerp, Belgium,[23] in the 16th century murals in Monsaraz, Portugal,[24] or in the churches located in Biañez and Axpe, Biscay, Spain.[25] While the mural painting technique has changed, the usage of chemically unstable pigments for manuscripts' illumination lasted longer and to a much larger extent. Consequently, realgar and/or orpiment can be seen, e.g. in the 16th or even the 18th century East-European manuscripts.[26, 27] The history of As-based pigments use in non-European countries is not fully evidenced; highly degraded orpiment was identified, e.g. in the mid-19th century Ethiopian mural paintings.[28]

To describe the painting technique and stratigraphy of paint layers, it is usually necessary to obtain micro-samples. Nevertheless, non-invasive analytical tools (e.g. portable X-ray fluorescence (XRF) or portable Raman spectrometry) applied in situ can provide crucial information about the employed pigments as well as preliminary description of signs of degradation. For example, the transformation of cinnabar (HgS) to calomel (Hg2Cl2) on degraded polychrome plasterworks in the Alhambra palace, Spain, has been described solely by non-invasive methods.[29] Recently, besides the most common portable XRF and Raman instruments, the use of portable XRF/X-ray diffraction has been on the uptrend.[30]

In this study, an efficient coupling of Raman micro-spectroscopy and powder X-ray micro-diffraction has been applied to properly describe the original composition of mineral pigments and signs of their degradation in fragmentary remnants of colour layers in Kuřívody wall paintings. Materials research in combination with the description of layer stratigraphy led to differentiation of paints based on their relative date of creation.

Materials and methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results and discussion
  6. Conclusions
  7. Acknowledgements
  8. References
  9. Supporting Information

Investigated murals

The materials investigation was carried out in the uncovered and sufficiently preserved parts of figurative paintings, on the southern wall of the church's presbytery, mainly in the scene depicting St. Dorothy and Infant Jesus, and partially also in the neighbouring scene with the Virgin Mary – ‘Sacrifice in the temple’ (Fig. 1 and Fig S2, Supporting information) It is interesting to note that the scenes with St. Dorothy are not very frequent in Bohemia. In Kuřívody, it is probably the earliest depiction of this saint at all (beside later ones in St. Nicolaus church in Kašperské Hory dated to 1330 and in St. John the Baptist church in Jindřichův Hradec dated to around 1350 [2]). Dorothy is a Roman and Middle Eastern saint, born in Ceasarea (Israel) in AD 290. She is usually depicted with flowers, which she, according to the legend, sent to earth from heaven after her martyrdom and death. In Kuřívody, the original colours were best preserved in the floral wreath on the head of St. Dorothy.

Non-invasive X-ray fluorescence (portable XRF)

In the first stage of materials investigation of the Kuřívody wall paintings, non-invasive XRF measurements were carried out in situ using two hand-held energy dispersive XRF spectrometers: the X-MET 3000 TXR by Oxford Instruments, UK (Rh anode, voltage 40 kV, Si-PIN detector) and Delta Premium by Innov-X, USA (Rh anode, voltage 40 kV, silicon drift detector (SDD)). This method has provided information about the distribution of chemical elements heavier than calcium (Z > 20, X-MET) or magnesium (Z > 12, Delta), respectively. The measurements were performed in air; therefore, the identification of lighter elements was prevented.

Sampling and visual observation of the samples

Places suitable for sampling were selected based on the results of non-invasive measurements. The list and the description of the obtained colour layers' micro-samples are given in Table 1. Micro-samples were embedded in polyester resin Neukadur PE 45 and polished in cross-section. Subsequently, layer stratigraphy was described on the basis of materials' colour and luminescence. Observations were carried out using Axio Imager A.2 light microscope (Zeiss) in reflected visible and UV light (365 and 470 nm). Cross-sections were photographed by digital microscope camera Olympus DP 75. Micro-sample of the plaster was taken separately and divided into several fragments. The morphology was studied in cross-section prepared in a similar way as in case of colour layers.

Table 1. Results of non-invasive measurements with portable XRF (selected elements)
No.ColourAlSiSKCaTiCrFeCuZnAsHgPbSn
  1. ■ present as major element (>5 wt.%)

  2. □ present as minor element

  3. - not present

  4. na, not analysed

 Yellow, orange              
5Very light yellow (hair)Nananana------
11Light yellow (clothing)nananana------
4Yellow (nimbus)nananana------
6Yellow (background)nananana-------
13Yellow (clothing, kerchief)nananana------
16Yellow (clothing, sleeve)nananana-----
3Yellow (flower rose)nananana------
36Yellow (clothing, sleeve)----
38Yellow (near window)-------
1Dark yellow (background)nananana------
22Dark yellow (background)nananana-------
35Dark yellow (clothes, sleeve)--------
17Orange yellow (sleeve)nananana-----
2Orange yellow (flower rose)nananana------
40Orange yellow (clothing)-----
32Orange (basket)---------
 Pink, red              
23Pinkish (drapery)nananana------
20Light red – pink (drapery)nananana------
24Light red – pink (ornament)nananana-----
27Light red – pink (ornament)nananana------
31Light pink (flower)--------
37Light pink (flower)-------
33Light red – pink (flower)-------
34Light red – pink (flower)-------
19Red (flower)nananana-----
 Red-brown, brown              
29Red-brown (earliest paint)nananana-------
26Red-brown (ornament)nananana------
9Dark brown (drapery)nananana-------
12Brown (background)nananana-------
18Brown (background)nananana-------
39Brown (clothing)------
21Brown-black (background)nananana-------
10Brown-black (background)nananana-------
 Grey, black              
7Black line (contour)nananana-----
25Grey (ornament)nananana----
28Blue-grey (star)nananana--------
 Green              
8Light green (clothing)nananana----
14Light green (clothing)nananana----
15Light green (clothing)nananana----
 PLASTER              
30Very light – no paintnananana--------

Electron microscopy and micro-analysis (SEM-EDS)

Selected cross-sections were studied by scanning electron microscope Jeol JSM6510 equipped with energy-dispersive spectrometer INCA (Oxford Instruments) with SDD detector allowing detection of elements heavier than Be (Z > 4) at resolution of 125 eV. Measurements were carried out in low vacuum mode under the pressure of 30 Pa and accelerating voltage of 25 kV; backscattered electrons were detected. Low vacuum mode allowed analyses of samples without conductive coating of their surface. Elemental mapping was carried out on the scanning electron microscope Tescan VEGA 3 XM, equipped with energy-dispersive system Quantax Bruker with SDD detector enabling high-speed XFlash 5010th. Working conditions are the following: voltage 20 kV, beam intensity 15, ×1380 magnification, working distance 15 mm, acquisition time being 120 min.

Raman micro-spectroscopy (micro-Raman)

Micro-Raman spectra of the samples were obtained on a Thermo Scientific DXR Raman Microscope with Peltier-cooled charge-coupled device detector interfaced to an Olympus microscope (×10 and ×50 objective lens) in the 50–3400 cm−1 spectral region with 4 cm−1 resolution. The spectrometer was calibrated by software-controlled calibration procedure using multiple neon emission lines (wavelength calibration), multiple polystyrene Raman bands (laser wavenumber calibration) and standardized white light sources (intensity calibration). The spectra were collected using diode-pumped solid state laser (532 nm) with 5 mW laser power. The spectra were not processed and are reported as collected.

Powder X-ray micro-diffraction (micro-XRD)

The identification of crystalline phases present in micro-samples was performed by powder X-ray micro-diffraction. The samples were analysed either as fragments of colour layers or as their cross-sections embedded in polyester resin used previously for microscopic examination. Diffraction patterns were collected using PANalytical X'Pert PRO diffractometer equipped with a conventional X-ray tube (Co Kα radiation, 40 kV, 30 mA, point focus), a glass collimating monocapillary with an exit diameter of 0.1 mm, and a multichannel position sensitive detector (X'Celerator) with an anti-scatter shield. Diffraction patterns were typically recorded in the interval from 4° to 80° 2θ with a step of 0.034° and counting time 2200 s per step, total measurement time per pattern being about 11 h. Qualitative analysis was performed using the HighScore software package (PANalytical, The Netherlands, version 3.0.5), and JCPDS PDF-2 database.[31]

Powder X-ray diffraction in transmission geometry

Fragment of the plaster was ground in an agate mortar in a suspension with cyclohexane. After that, the suspension was transferred to a mylar foil on a transmission sample holder. After evaporation of the solvent, the resulting thin layer of the sample was covered by a second mylar foil for the purpose of quantitative phase analysis by X-ray diffraction in transmission geometry. Diffraction pattern was collected using a PANalytical X'Pert PRO diffractometer equipped with a conventional X-ray tube (CuKα 40 kV, 30 mA, line focus) in transmission mode. An elliptic focusing mirror, a divergence slit 0.5°, an anti-scatter slit 0.5°, and a Soller slit of 0.02 rad were used in the primary beam. A fast linear position sensitive detector PIXcel with an anti-scatter shield and a Soller slit of 0.02 rad were used in the diffracted beam. X-ray patterns were collected in the range of 1° to 88° 2θ with a step of 0.013° and 500 s per step resulting in a scan of about 3.75 h. Diffrac-Plus Topas (Bruker AXS, Germany, version 4.2) with structural models based on Inorganic Crystal Structure Database [32] was used for quantitative analysis of the samples. This program permits to estimate the weight fractions of crystalline phases by the means of Rietveld refinement procedure.

Infrared micro-spectroscopy (micro-FTIR)

Selected fragments or their cross-sections were analysed in reflection mode by FTIR microscope Thermo Scientific Nicolet Continuμm coupled with Nexus spectrometer. The spectra were collected in the range of 4000 to 650 cm−1 with 4 cm−1 spectral resolution, the number of scans being 256. The spectra were processed in OMNIC software package 8.1, Thermo Electron; and when necessary, they were treated using Kramers–Krönig transformation.

Detection of proteinaceous binders (NanoLC-ESI-Q-TOF MS/MS)

Microsamples were mixed with solution of 50 mM ammonium hydrogen carbonate containing 10 mg/ml of trypsin and left digested for 2 h at laboratory temperature. The resulting peptide mixtures were cleaned and concentrated on a reverse phase C18 sorbent in micro-column ZIP TIP. Purified solutions were left to dried up in laboratory conditions and then redissolved in 97:3:0.1% mixture of water : acetonitrile : formic acid. Measurement was carried out using ultrahigh-pressure liquid chromatography Dionex Ultimate3000 RSLC nano (Dionex, Germany) connected with mass spectrometer ESI-Q-TOF Maxis Impact (Bruker, Germany). Raw data were evaluated by Data Analysis (Bruker Daltonics, Germany). Proteins were identified using Mascot version 2.4.1 (Matrix Science, UK) by searching protein database Uniprot version 2011–2012.

Results and discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results and discussion
  6. Conclusions
  7. Acknowledgements
  8. References
  9. Supporting Information

Results of preliminary screening

Results of preliminary screening of the degraded surface of murals by portable XRF are summarized in Table 1. Based on these results, we can expect the presence of very traditional mediaeval pigments such as cinnabar (high contents of Hg in reds), orpiment and/or realgar (high contents of As in some yellows and flesh tones), Pb-based pigments (minium and/or lead white), Cu-based pigments (in greens), and, of course, Fe pigments (earths and ochres). On the other hand, the analyses also yielded some surprising findings. Firstly, arsenic is not accompanied by sufficient amounts of sulfur to form arsenic sulfides. Secondly, tin in association with copper is not always accompanied by lead, therefore, an admixture of (otherwise frequent) lead-tin yellow (Pb2SnO4 or PbSnO3) is questionable. And finally, relatively high amounts of ‘modern’ elements – Ti in red-browns and flesh tones, Zn in greens and particularly Cr in yellows – suggest the presence of over-paints containing pigments of the 19th or even the 20th century (zinc and titanium white, chrome yellow). However, no over-paints were recorded in Kuřívody, and from the historical point of view, they are not even probable. Micro-samples of colour layers were obtained specifically to explain the previously-mentioned discrepancies by the means of micro-Raman and micro-XRD. List of analysed samples is given in Table 2.

Table 2. List of analysed micro-samples, their description, and summary of results
No.Colour and locationLayers observed in the cross-sectionElements in the painting layer (EDS)Phases confirmed in the painting layer (micro-Raman/ micro-XRD)
  1. 4, painting; 3, lime primer; 2, painting (earliest); 1, new white plaster; 0, thermally altered plaster (brownish)

  2. na = not analyzed

 Earliest simple paints
J1223-12Red-brown (ceiling)0, 1, 2Si, Ca, Al, Fe, Ti, S (P, K, Mg)kaolinite Al2Si2O5(OH)4, hematite Fe2O3, anatase TiO2
Figural scenes
J1223-2Green (Virgin Mary – robe)0, 3, 4Ca, Cu, Cl (Si, Al, Zn, Sn, S)atacamite Cu2Cl(OH)3, admixture of cassiterite SnO2
J1223-3Dark red (St. Dorothy – flower rose)3, 4 (double layer red + dark)Pb, Cl, Hg, S, Ca, Cr, (Si, Al, Fe, K, Mg)laurionite PbCl(OH), minium Pb3O4, plattnerite PbO2, cinnabar HgS, crocoite PbCrO4
J1223-4Light yellow, pinkish (Virgin Mary – kerchief on her head)3, 4Pb, Ca, As, Cl, Si, Hg, S, (Cr)mimetite Pb5(AsO4)3Cl, cinnabar HgS, crocoite PbCrO4, calcite CaCO3
J1223-5Dark (St. Dorothy – contour of the rose)3, 4 (With underdrawing)Ca, Pb, As, Si, Al, Cl, (P, K, Fe)na
J1223-14Yellow-orange (St. Dorothy – clothing, shadow)3, 4Pb, Ca, As, Cl, Si, Hg, S, (Cr)mimetite Pb5(AsO4)3Cl, cinnabar HgS, crocoite PbCrO4, carbon black, calcite CaCO3, quartz SiO2, whewellite CaC2O4∙H2O
J1223-15Dark orange (St. Dorothy – flower rose)3, 4 (double layer red + dark)nalaurionite PbCl(OH), minium Pb3O4, plattnerite PbO2, cinnabar HgS, crocoite PbCrO4
J1223-16Yellow-orange (St. Dorothy – clothing)3, 4Pb, Ca, As, Cl, Si, Hg, S, (Cr)mimetite Pb5(AsO4)3Cl, cinnabar HgS, crocoite PbCrO4, calcite CaCO3, gypsum CaSO4·2H2O, cassiterite SnO2
J1223-11Black contour of the grey star0, 1, 3, 4Ca, Si, Fe, Al, Tina
 Ornamental paintings
J1223-6Grey (ornaments)3, 4Si, Al, Ca, Ti, Fe, (P)na
J1223-7Red-brown (ornaments)3, 4Si, Al, Ca, Ti, Fe, (K, P, Mg)na
J1223-9Grey-blue (ceiling)0, 1, 2, 3, 4Si, Al, Ca, (Ti, Fe, K)na

Plaster and the earliest simple paintings

The powder X-ray diffraction data in transmission geometry clearly show that the plaster is composed of calcium carbonate (19 wt.%) and sand, which contains predominantly quartz grains (70 wt.%) but also other silicates such as feldspars (4 wt.%), micas (<1 wt.%), as well as recrystallised opaline silica (3 wt.% of cristobalite and 3 wt.% of tridymite) and sparse small grains of monazite and zircon. The yellow (in the lower part) and red tints (in the upper part) are caused by an admixture of iron-containing minerals, which predominantly form coatings on quartz grains. The gradual colour change toward the surface may be explained by thermal conversion of initially yellow goethite (FeO(OH)) to red hematite (Fe2O3), which normally takes place at about 250 °C and which may be, speculatively, related to a fire in the church shortly after the finishing of the construction. (Fig. S3, Supporting information) Besides the iron oxides, also clay minerals were identified in quartz grains' coatings – particularly kaolinite and mixed-layered clay structure (most probably illite–smectite [33]). (Fig. S4, Supporting information) The thermally induced transition of smectite to illite, accompanied by formation of mixed layered structures refers to increased temperatures (up to 200 °C). However, in this case, it is not possible to relate the formation of illite–smectite mixed layered structure directly to the fire in the church and exclude the possibility of its natural formation at the source locality of the material.

The undesirable change of colour is the probable cause why the original plaster was covered by a new white plaster coating before the painting started. (Fig. S3, Supporting information) In the first stages, primitive simple paintings were created using red clays as pigments. Red clays belong to a large group of ‘earthy pigments’, and differ in composition with respect to the process of their formation in nature.[34, 35] High contents of Ti and low K/Ti ratio together with low content of coarse-grained quartz indicate the usage of high-quality clay (known as ‘bole’) formed usually by intense weathering of basic rocks rich in Ti and Fe.[35, 36] In Kuřívody, Ti-rich red clay was used in simple paints that represent the earliest paintings in the church created evidently soon after the plastering's finishing. Earths of the same composition, exhibiting the same characteristic Si/Al (=1.7) and K/Ti (=0.2) ratios, (Table S1, Supporting information) were again used in later figurative and ornamental paintings. It could indicate that these paintings were created shortly after the primitive ones. According to the EDS analysis, contents of titanium in these materials vary from 4 to 8 at.% and fully explains the increased values of Ti obtained previously by portable XRF.

Crocoite on the authors' palette

Layers with earlier (simple) and later (figurative) paintings are stratigraphically divided by a layer of fine-grained plaster and a subsequent preparatory layer of lime primer. It is estimated (based also on the dendrochronological dating of the fir girder in the presbytery to 1289/90, not published) that the most important figurative paintings were created around 1300. Refining their dating is not an easy task because of the lack of any materials or technological comparison with similar paintings in the region.

In the micro-samples, following pigments were unambiguously identified by Raman micro-spectroscopy and micro-XRD: cinnabar (HgS), minium (Pb3O4), carbon black, calcium carbonate, yellow earths, and completely degraded and transformed Cu-based and As-based pigments. (Table 2, Figs. 2 and 3) However, the most important finding is the presence of yellow to orange mineral crocoite (PbCrO4), intentionally used in small quantities in the painting of red flowers, highlights in the modelling of the figures and in the figures' outlines. Raman micro-spectroscopy has proved to be the most efficient and indispensable method for direct identification of pigments present in low quantities in the mixture thanks to its high spatial resolution allowing analysis of individual grains of only several micrometres in size. (Fig. 4) On the other hand, laboratory micro-XRD provided an average mineralogical composition by integrating an area of 150–200 µm in diameter; the presence of minor crocoite could not be unambiguously confirmed in any of the performed measurements. In the Raman spectra, crocoite was identified by a very strong ν1 symmetric stretching band of CrO42− at approx. 844 cm−1, while ν2 and ν4 bending modes were present only as weak bands – in agreement with already published data for the measurements using 532 nm laser excitation.[37]

image

Figure 2. Fragment of the painting layer in the sample J1223-3 (flower rose) with yellow crocoite grains and partly darkened minium.

Download figure to PowerPoint

image

Figure 3. Fragment of the painting layer in the sample J1223-4 (yellow clothing), rich in arsenic.

Download figure to PowerPoint

image

Figure 4. Raman spectrum of intense yellow grain in the painting of the flower rose (J1223-3) in comparison with reference spectrum of crocoite (www.rruff.info); laser excitation 532 nm.

Download figure to PowerPoint

The grains of crocoite in Kuřívody vary in shape and size, usually in the range from 10 to 30 µm – the larger ones could be, however, aggregates of smaller particles. It is not clear, if crocoite, which only rarely can be found in nature in higher quantities, was somehow modified or enriched in the raw material before use as pigment. There is a complete lack of historical evidence. We can only exclude the use of synthetic lead chromates, because there are no modern over-paints or retouches in Kuřívody.

The finding of crocoite is thus of an exceptional importance. While the synthetic form of this pigment has been frequently applied since the 19th century, the intentional use of natural form of crocoite represents a rarity. In European art, its presence was confirmed only once, also by micro-Raman, in remarkable large-scale wall paintings under the floor of Siena Cathedral in Italy, which represent the earliest period of Sienese painting and whose creation started around 1270.[6] Similarly to Sienese paintings, grains of crocoite in Kuřívody are diffused in the colour layers and applied in very small quantities to modify the tint of the prevailing pigment (e.g. the minium) or to highlight the outlines. The same unique pigment and the same way of its application in similar mixtures of pigments relate to a certain extent the paintings from Kuřívody to those from Siena. Taking into account the fact that the natural crocoite belongs to ancient pigments originally used in Mediterranean [3] and the period of creation of Sienese paintings, the suggested date of origin of the figurative paintings in Kuřívody (around 1300) seems to be highly relevant. This information was obtained despite the considerable mechanical damage of murals, reduced colouring, and only fragmentary preservation of the colour layers.

Mimetite: pigment or alteration product?

Unlike in the case of the monumental paintings in Siena Cathedral, in Kuřívody, no genuine gilding was observed – no metals were applied. To achieve a golden tint and thus imitate gold, orpiment has been traditionally applied as one of very impressive alternatives – in wooden icons, manuscripts, and even wall paintings. In Kuřívody, increased concentrations of arsenic were detected not only in light yellow or almost colourless parts of murals, e.g. in hairs or haloes, but also in pinkish ornaments and draperies. (Tables 1 and 2) Therefore, it has been originally suggested that both the most common As-containing pigments were used – orpiment in yellow and golden colours, realgar in flesh tones, and some other pinks or reds. However, in the micro-samples, no orpiment or realgar was identified, in agreement with the fact that there was no sulfur accompanying the arsenic (as already indicated by portable XRF). Arsenic associated with oxygen, not sulfur, was similarly evidenced, for example, in wall decorations in the Nefertari's tomb in Egypt.[38] The only As-containing phase in Kuřívody, confirmed simultaneously by micro-Raman and micro-XRD, was mimetite – Pb5(AsO4)3Cl. (Figs. 5 and 6) In the Raman spectra, mimetite is indicated by a strong band corresponding to ν1 modes of AsO43− at approximately 810 cm−1; some influence of ν3 mode was also discussed .[39] However, in our case, the band was positioned at approx. 820 cm−1, which is not usual, but it has already been reported, refer to, e.g. the Raman spectrum of mimetite from Vrančice, Czech Republic, published online at www.rruff.info. The shift of this band may correspond to isomorphic substitutions of AsV by PV (in the isostructural series from mimetite to pyromorphite – Pb5(PO4)3Cl) or to isomorphic substitutions of PbII by CaII (as in case of hedyphane – Pb3Ca2(AsO4)3Cl) in the mineral's structure .[39] In the micro-diffraction pattern, the previously mentioned isostructural phases cannot be positively differentiated because of the extent of the overlaps of very closely located diffraction lines. The best fitting, however, was obtained with mimetite. To avoid any misleading interpretation, elemental mapping has been performed in selected areas of colour layers. For example, in the remnant of the yellow-orange paint of the flower, joint occurrence of Pb, As, and Cl was detected in the whole area of the layer together with several small grains of crocoite (Pb, Cr), cinnabar (Hg), and ferric oxides (Fe, not shown). (Fig. 7) No phosphorus was found; therefore, the presence of pyromorphite can be excluded. Contents of calcium are relatively low (in comparison with the underlying lime primer) but sufficient to eventually contaminate (to a certain extent) the mimetite's structure. Therefore, it is not possible to unambiguously determine whether there is, besides mimetite and calcium carbonate, hedyphane present in the layer as well.

image

Figure 5. Raman spectrum of light yellow diffuse grains in the painting of the yellow clothing (J1223-14) in comparison with reference spectrum of mimetite from Vrančice, CZ (www.rruff.info); laser excitation 532 nm.

Download figure to PowerPoint

image

Figure 6. Micro-diffraction pattern of sample J1223-14 in comparison with reference diffraction lines of prevailing mimetite Pb5(AsO4)3Cl (M, blue lines); other identified phases: A calcite, CaCO3, H cinnabar, HgS, Q quartz, SiO2, W whewellite, CaC2O4∙H2O; microphotograph depicts the analysed fragment; the ellipse and the arrow indicate the analysed area and the incident beam direction.

Download figure to PowerPoint

image

Figure 7. X-ray intensity maps of selected elements in the painting layer of the sample J1223-4 (EDS elemental mapping) showing joint occurrence of Pb, As, and Cl in the layer and separate grains containing Cr and Hg, respectively; X-ray intensities correspond to relative concentration of elements within the measured area, but the concentration maxima indicated by the highest intensity are different for each element.

Download figure to PowerPoint

The crucial question is whether the light yellow mimetite represents an intentionally used pigment or rather a secondary phase formed by degradation of other As-containing pigments. Mimetite, as a pigment, is mentioned in the literature only in a marginal way – in the painted surface of ancient stone steles in Greece and in the painted decorations of tombs in Macedonia (dated to 325–275 BC) and Syria (dated to the 2nd century AD).[7, 40] It implies that in Kuřívody, it would eventually be the first evidence of mimetite pigment in mediaeval European art, confirmed simultaneously by micro-Raman and micro-XRD. In the previously mentioned example, in the Tomb of Three Brothers in Palmyra, Syria, mimetite was identified by micro-XRD in small quantities in highlights of the cloak. No other Pb-containing or As-containing pigments were found in other areas of the painting; there was no minium, and no lead white at all. Because of the lack of original pigments containing lead and/or arsenic, the authors did not mention the possibility of mimetite's formation as a result of degradation processes and suggest its intentional usage. However, to this day, natural source of mimetite is not known. Mimetite is formed in oxidation zones of lead ores, but its sparse occurrence reduces the likelihood of its intentional and economically valuable exploitation.

In Kuřívody, minium (Pb3O4) was identified in numerous places but not in association with mimetite in one mixture. Orpiment was not detected anywhere in the samples, but increased concentrations of arsenic were not always accompanied by sufficient amounts of lead, e.g. in the areas representing halos, where one can suggest the use of orpiment as imitation of gold. (Table 1) Therefore, the formation of light yellow mimetite through degradation of orpiment and lead-based pigments seems to be plausible, although this process is not easy to prove. The process of alteration of lead pigments in the presence of orpiment has been already observed in mediaeval illuminated manuscripts and described by model experiments.[10, 41] Both minium and lead white, respectively, tend to darken in aqueous suspensions with orpiment under normal laboratory conditions as a result of the formation of black galena (PbS), while orpiment is transformed to arsenolite (As2O3). The formation of low amounts of arsenates was also corroborated by gradual increase of Raman band at 810 cm−1, but the arsenates have never become dominant, even after 1-year-long experiment. These neoformed arsenates were not further specified by diffraction methods or elemental mapping in contrast to this study. In Kuřívody, no galena or arsenolite was found.

The environment of the wall paintings is alkaline – when the painting is executed on the fresh plaster or after the painting is covered by new lime-based over-paints or new plasters, the pH increases to values higher than 10. In alkaline conditions, the orpiment becomes less stable and tends to dissolve to thioarsenites (AsS3)3− and arsenites (AsO3)3−.[42] To oxidize trivalent arsenic and form thioarsenates (AsS4)3− and arsenates (AsO4)3−, the action of atmospheric oxygen is necessary.

While in the areas where no As was detected, minium (PbII2PbIVO4) decomposed to brown-black plattnerite (PbIVO2) and massicot (PbIIO, which usually immediately carbonizes to PbIICO3), a reaction described in the literature [17], the processes taking place in the As-rich areas are more complicated. The reason why minium tends to darken because of the formation of plattnerite in the sample J1223-3 (PbO2, proved by micro-XRD, Fig. 8), while in the sample J1223-4 (containing large amounts of arsenic), tends to become lighter because of the complete transformation of minium and orpiment to lead arsenates (Figs. 3, 5, and 6), may lie in the influence of orpiment on the reaction of minium, which would be preferentially reduced to Pb2+ ions under simultaneous formation of arsenolite and elemental sulfur. Subsequently, increased activity of dissolved chloride ions resulting from increased humidity, the alkaline environment, and the presence of atmospheric oxygen would finally lead to preferential formation of lead chloroarsenates like mimetite, which is considered to be one of the least soluble phases among the arsenates. [43]

image

Figure 8. Micro-diffraction pattern of sample J1223-15 in comparison with reference diffraction lines of prevailing laurionite, Pb(OH)Cl (L, blue lines); other identified phases: C crocoite, PbCrO4, H cinnabar, HgS, R minium, Pb3O4, P plattnerite, PbO2; microphotograph depicts the analysed cross-section; the ellipse and the arrow indicate the analysed area and the incident beam direction.

Download figure to PowerPoint

Very low concentrations of sulfur in colour layers proved by portable XRF and SEM-EDS analyses may correspond with the fact that under alkaline conditions, the majority of sulfur is released during transformation of thioarsenates to arsenates and may migrate in the form of dissociated sulfides. Subsequent processes, e.g. oxidation and formation of secondary sulfates (e.g. gypsum), are not fully clear because they were not detected in Kuřívody. A certain amount of gypsum was identified as a surface efflorescence (not shown), but the source of sulfur is not known and cannot be formally associated with the decomposed orpiment. Another possibility is that sulfur escaped in the form of gaseous SO2, especially when arsenolite was formed from orpiment in the first stage of the degradation processes. However, also in this case, there is no direct proof of this mechanism. Although the question of release of sulfur is not fully understood, the formation of mimetite by alteration of orpiment and minium looks more probable than its intentional usage as a pigment.

Other signs of degradation

The average elemental composition of light yellow colour obtained by SEM-EDS (sample J1223-4 taken from the clothing, Table S1, Supporting information) and the calculation of the relative amounts of mineral phases according to their ideal formulas show that crocoite, cinnabar, and mimetite are jointly present in one paint layer in a very approximate ratio of 1:2:3. However, the calculation indicates that a certain amount of lead and chlorine remains unassigned, thus suggesting the presence of another phase. Even better evidence of possible Pb–Cl phase(s) was provided by the measurements of red colour of flower where no arsenic (therefore, no mimetite) is present. (Table S1, Supporting information) Finally, in the micro-diffraction pattern, laurionite – PbCl(OH) – was identified as a prevailing phase in the sample together with smaller amounts of minium (Pb3O4) and plattnerite (PbO2) (Fig. 8). Plattnerite was formed by only partial decomposition of minium; in darkened grains, minium is still the predominant phase. (Fig. S5, Supporting information) In the literature, laurionite is described as a typical product of salt corrosion of lead white resulting in its slight yellowing.[17] However, in Kuřívody, intentional use of lead white (cerussite and/or hydrocerussite) has not been confirmed either by micro-XRD or micro-Raman. The only positive identification of hydrocerussite was provided by micro-FTIR analysis. Laurionite can also originate from cerussite created by the decomposition of minium (refer to the previous chapter). Therefore, the presence of lead white in the original mixture of pigments remains questionable, although the contents of lead are increased in numerous places of the painting. (Table 1)

The increased humidity in walls and high activity of chloride ions also caused an almost complete transformation of the original copper carbonates (azurite and/or malachite) to green copper chlorides, particularly atacamite (Cu2Cl(OH)3).[19] In Kuřívody, green atacamite was identified in the paint of the Virgin Mary's robe, which has been according to iconographic tradition always painted in blue colour. (Fig. S6, Supporting information) Therefore, we can suppose that blue azurite (2CuCO3·Cu(OH)2) has been originally applied. The process of salt corrosion of copper pigments is accelerated by oxalic acid produced by microorganisms. [[19]] Simultaneously, calcium oxalates (as for example whewellite CaC2O4∙H2O) are formed by reaction with calcite. (Fig. 6 and Fig. S6, Supporting information) Low amounts of Zn and Sn identified solely in association with copper by portable XRF can be explained as admixtures related to the production of the original pigment. It is a known fact that natural azurites exploited in oxidized zones of ore deposits can be accompanied by zinc and arsenic (usually in the form of zinc arsenates), as well as by barium.[44] The association with Zn and Sn is rather unusual, because copper ore deposits are not often associated with tin mineralisation. As indicated by micro-XRD, in Kuřívody, tin is predominantly present in the form of cassiterite (SnO2), but the source of this natural admixture remains unclear.

Painting technique

Throughout the history, copper-based and lead-based pigments were applied predominantly on dried plaster (secco-application) and rather with organic binders instead of fresh lime (tempera painting) in order to prevent their salt corrosion. For example, in Siena Cathedral, all three most common painting techniques were combined – fresco painting together with secco-applied techniques – lime painting and tempera.[6] In Kuřívody, the information regarding the painting technique is only fragmentary. While in Siena the plaster is applied in monolayer, in Kuřívody, a fine-grained plaster and/or lime primer is used as a preparatory layer for figurative painting. The underdrawing is executed using carbon black, which is another technological difference from Siena, where (as it is frequent in Italy) Fe-based reds and yellows (the so-called ‘sinopia’) were applied. In Kuřívody, it can be clearly seen that the paints applied in fresco (typically ochres and earths) are more adhesive to the preparatory layer; therefore, they are less damaged, and, consequently, they were not sampled. In all micro-samples taken from the mechanically damaged parts, the paint layers are clearly separated from the preparatory layer indicating a secco-application of Cu, Pb, and As containing pigments. (Figs. 2 and 3) The extent of usage of lime-based or organic-based binders still remains questionable, because it was possible to obtain preliminary information only from several fragments. The presence of proteins was excluded by both micro-FTIR and nanoliquid chromatography ESI-Q-TOF MS/MS analytical techniques. On the other hand, micro-FTIR confirmed the presence of polysaccharides in the painting of flowers (not shown). Subsequently, micro-destructive staining tests excluded the presence of starch based on negative reaction with Lugol's iodine. Therefore, the only possible interpretation is the use of gum (e.g. gum arabic), which is a very traditional technique widely applied in ancient Mediterranean.[45]

Conclusions

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results and discussion
  6. Conclusions
  7. Acknowledgements
  8. References
  9. Supporting Information

Natural crocoite (PbCrO4) was detected by Raman micro-spectroscopy in figurative wall paintings in Kuřívody, Northern Bohemia, dated to around 1300, as intentionally used yellow pigment. This is only a second identification of this pigment in traditional European art.

Raman micro-spectroscopy was found to be an indispensable method for identification of mineral pigments present in paint layers in low quantities, which can be, as in this case, degraded and preserved only fragmentarily. If the minor phases are strong Raman scatterers – such as lead chromates – their identification is greatly improved. In that case, Raman micro-spectroscopy is able to answer key questions in technological and art-historical contexts.

On the other hand, effective combination of Raman micro-spectroscopy and X-ray micro-diffraction is always desirable. X-ray micro-diffraction is particularly important in distinguishing clay-based materials as well as products of pigment alteration. For example, plattnerite (PbO2), being a product of decomposition of minium (Pb3O4), was unambiguously identified by micro-XRD in darkened parts of the red paints, but not by micro-Raman.

The presence of mimetite (Pb5(AsO4)3Cl) was confirmed by effective combination of micro-Raman, micro-XRD and elemental mapping in the layer. Although mimetite is mentioned in the literature as a rare pigment and this would potentially be its first documented occurrence in European art, the authors prefer the interpretation that mimetite was formed through degradation of orpiment in mixture with lead-based pigments. This process is theoretically plausible; on the other hand, there is no clear evidence of starting materials or by-products (e.g. sulfates) in degraded parts of the murals.

Finally, attention should be drawn to the risk of simplified interpretation of results of non-invasive measurements. Increased contents of Cr, Ti, and Zn in mediaeval murals in Kuřívody could be misunderstood and interpreted as evidence of modern synthetic pigments such as zinc white (ZnO), titanium white (TiO2), and chrome yellow (PbCrO4). In fact, natural crocoite was used instead of the synthetic one; high-quality red clays are admixed with large quantities of natural anatase (TiO2), and Zn accompanies natural azurite (transformed to atacamite by salt attack under increased humidity conditions).

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results and discussion
  6. Conclusions
  7. Acknowledgements
  8. References
  9. Supporting Information

The authors express their thanks to Václav Potůček and Miloš Krčmář for fruitful cooperation, allowing access to the wall paintings and sampling, as well as providing historical information and yet unpublished data, e.g. from dendrochronological dating. Thanks also belong to Zlata Vrátníčková (Polymer Institute Brno) and Štěpánka Hrdličková Kučková (Institute of Technology in Prague, Czech Republic), who performed the measurements by means of infrared micro-spectroscopy and nanoliquid chromatography MS, respectively. The study has been supported by Czech Science Foundation, project no. 14-22984S, and by institutional budgets of research institutions (RVO 61388980 and 60461446).

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results and discussion
  6. Conclusions
  7. Acknowledgements
  8. References
  9. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results and discussion
  6. Conclusions
  7. Acknowledgements
  8. References
  9. Supporting Information
FilenameFormatSizeDescription
jrs4556-sup-0001-Hradil_FigS1.tifapplication/unknown6165KSupporting info item
jrs4556-sup-0002-Hradil_FigS2.tifapplication/unknown5179KSupporting info item
jrs4556-sup-0003-Hradil_FigS3.tifapplication/unknown6264KSupporting info item
jrs4556-sup-0004-Hradil_FigS4.tifapplication/unknown176KSupporting info item
jrs4556-sup-0005-Hradil_FigS5.tifapplication/unknown363KSupporting info item
jrs4556-sup-0006-Hradil_FigS6.tifapplication/unknown3143KSupporting info item

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.