The author would like to thank PhD Ivica Ljubenkov, Head of the Conservation Research Laboratory of the Conservation-Restoration Department of the Arts Academy in Split, for the mentorship of the technical study of St. Michael's antependium, as well as for numerous helpful comments and suggestions.
She would also like to thank Jurica Matijević, Head of the Easel Paintings Conservation of the Conservation-Restoration Department, for the mentorship of the conservation-restoration treatment of St. Michael's antependium.
The author is thankful to MSc Katarina Hraste for the proofreading of the paper and to MSc Goran Nikšić for the French translation of the summary.
Finally, the author is very grateful to Sagita Mirjam Sunara, the Head of the Conservation-Restoration Department, for editing this paper and for the support and incentive of her work and advancement.
1At the end of the 17th century, Dalmatian Zagora was an area of restlessness and instability, war and override between the Ottoman Empire and Republic of Venice. In this seemingly remote, backwater place, the little village of Tugare had – already during the war – equipped its middleage Medieval parish church with works of the artist Matej Otoni. Otoni worked in the broader area of Split, Omiš, Makarska and the island of Brač, making wooden polychromed altarpieces and paintings. He repeatedly used simple motives from earlier graphic templates. Multicolor floral ornaments evocative of ornaments of the folklore are characteristic of his work.
2The wooden polychrome antependium from the church of Saint Catherine in Osič, Tugare, is a part of the altarpiece dedicated to st. Michael. Today, only the antependium and the altarpiece "frame" are preserved.
3The antependium is simply painted. St. Michael is represented in the centre, treading upon a dragon. To the left and right of the central medallion, there are two angel heads with wings. Free space is filled with stylized floral ornaments and rosettes. The antependium is framed with an outer decorative frame.
Fig. 1 Wooden polychrome antependium of St. Michael's altar
Antependium from the church of Saint Catherine in Osič, Croatia.
Credits: Aida Grga
4Otoni first applied a coat of white paint. He then painted using red, blue, green, ochre, brown and shades of pink. Black lines were painted rather freely. The battens of the two frames were overpainted two times. After the application of the first overpainting, the front side of the antependium was varnished. This probably happened in the 19th century.
5Due to poor storage (fluctuations of relative humidity in the church, insect infestation), the object was in a bad condition. The varnish had discoloured and the surface was covered with dirt and candle wax. The conservation-restoration treatment is being carried out at the Department for Conservation-Restoration of the Arts Academy in Split. A part of the treatment and the technical study of the painting materials were the topic of author's Master's thesis.
6When painting, Otoni used primary colours. This, along with the fact that he combined pure pigment with the binding medium and applied the paint directly to the wooden support, facilitated the identification of the materials used on the antependium of St. Michael's altar. The dating of the object to the 17th century narrowed down the scope of the reference materials. The varnish, which is of a later date, was also analyzed.
7The binding medium was analyzed using thin layer chromatography (TLC). Micro-chemical analysis and Fourier transform infrared spectroscopy(FTIR) were used to identify the pigments, while TLC and FTIR were used to determine the varnish. To confirm the results, additional techniques were used.
8In order to investigate the binding medium, a small piece of paint was taken from the edge of a wood crack.
9To determine the pigments (white, red, ochre, green, blue), samples of paint were collected from the edges of the antependium and the edges of the lacunae. To collect the samples, surgical scalpel was used. The samples were scraped out from the surface.
Fig. 2 Sampling scheme
(1) binding medium, (2) white paint, (3) red paint, (4) ochre paint, (5) green paint and (6) blue paint.
Credits: Tome Bakić
10As a part of the restoration treatment, the varnish was removed using cotton swabs wetted with acetone. A certain number of cotton swabs were preserved and dried. The residues of the resin on the cotton were then analyzed.
- 1 Because of the appearance and the characteristics of the paint layer on the antependium of St. Mic (...)
11Reference samples included egg yolk, egg white, mixture of egg yolk and egg white, amino acid standards, casein and thickened linseed oil.1
- 2 Ochre and umber are compounds of different metal oxides, primarily iron oxides and hydroxides. To (...)
- 3 To determine the pigment in the green paint, green earth and malachite were used. Verdigris was no (...)
12Kremer pigments, selected according to the 17th century pigment palette, were used as reference samples. Lead white ((PbCO3)2·Pb(OH)2) was the most widely used white pigment in the 17th century. The most widely used inorganic red pigments were vermilion (HgS) and minium (Pb3O4). Natural earth pigments such as ochre and umber have been used since prehistoric times.2 The most commonly used green pigments were green earth (mixture of silicon, aluminium, iron, magnesium, calcium and other oxides), malachite (CuCO3×Cu(OH)2 – alkaline copper carbonate) and verdigris (Cu(OH)2×(CH3COO)2×5H2O – alkaline copper acetate).3 Blue was assumed to be blue smalt, a pigment that Otoni was known to have used.
13Sandarac, Manila copal, colophony, mastic and dammar were used, as they could have been available in Dalmatian Zagora in the 19th century.
- 4 The time needed for proteins to hydrolyze completely.
14All samples, apart from the amino acid standards, were first hydrolyzed (the amino acid standards were directly spotted on the TLC plate). 50 mg of each sample and 1 mL HCl : H2O (1 : 1) was placed in a test tube. The test tube was kept for 24 h at 105 °C4.
- 5 The plates were sprayed with ninhydrin solution so that the spots of the separated components woul (...)
15After cooling at a room temperature the samples were spotted on TLC plate using capillary tube. Different mixtures of solvents were used as mobile phase. The type of sample analyzed on each plate was taken into account. When the development was completed, the TLC plates were dried at room temperature and then put into a dryer at 105 °C (10 to 15 minutes). After cooling down, the plates were sprayed with ninhydrin solution.5
16TLC separation of the hydrolyzed sample from the antependium and the sample of egg yolk were also carried out. The idea was to detect the presence of phospholipid lecithin, which can be found in egg yolk, but it was not present in either of the reference samples. A mixture of chloroform, methanol and water (65 : 25 : 4) was used as the mobile phase. As a spot indicator, Hanes reagent for esters of phosphoric acid was used. TLC plate was developed and put into the dryer at 105 °C. The plate was then examined under longer wavelengths of UV light.
- 6 Several samples were measured in a wider range od IR spectrum (400 - 4400 cm-1).
17To measure the spectrums of the reference samples (pure pigments) and the pigment samples from St. Michael’s antependium, a small amount of sample was mixed with KBr. The mixture was ground in the mortar and put into the mould for making the pellet for FTIR-spectroscopy. The ground powder was pressed into a transparent pellet, using a hydraulic press (~ 300 atp). The pellet was then carefully put into a holder and measured with IR-spectrometer (FTIR – 8400S Furier Transform Infrared Spectrophotometer SHIMADZU). The data was processed using the computer. The samples were measured in the range from 500 to 4200 cm-1 spectral range of IR-spectrum.6
- 7 Plumbtesmo paper (Machinery-Nagel) is used to determine the presence of lead in metal objects, pig (...)
18Two drops of de-ionized water were spotted with pipette on the Plumbtesmo paper.7 A sample of white paint from the antependium was placed on the wet test paper.
- 8 Plumbtesmo paper can be used to determine minium, because minium is lead oxide (Pb3O4). If the red (...)
19Red paint was also analyzed.8
20The test reveals the presence of mercury in the red pigment, or mineral, and can identify cinnabar (vermilion, HgS).
21A small amount of red colour sample material (tip of a spatula) was dissolved in a watch glass with a few drops of aqua regia. The sample was then evaporated in the dryer. The crystalline residue left after evaporation was dissolved in a drop of distilled water. A small spot was abraded with emery cloth on a piece of aluminium foil. One drop of 10 % NaOH was placed on the spot. Due to the creation of H2 gas, the drop became milky white. The excess alkali was removed with blotting paper. Next, a drop of the sample solution was deposited on the prepared surface.
22The test is used to determine the presence of iron (III) ions in corrosion products, stains, or pigments. Iron (III) ions dissolve in hydrochloric acid and react with potassium ferrocyanide (K4Fe(CN)6) producing ferric ferrocyanide, a bright blue complex called Prussian blue.
23Before the test was carried out, two reagents were prepared:
(1) 3 M hydrochloric acid (HCl) solution (1 : 3) : 1 mL concentrated HCl was added to 3 mL distilled water,
(2) Potassium ferrocyanide solution: 1 g K4Fe(CN)6×3H2O was added to 25 mL water.
24The yellow ochre paint sample almost completely dissolved on the spot-test plate with concentrated HCl. The solution was then dried in the laboratory dryer. The sample was cooled down to room temperature and a drop of potassium ferrocyanide was added to it.
25This test was also used to determine the green pigment. Four samples were tested: the sample of green paint from St. Michael’s antependium, two different green earths and malachite.
26Dried cotton swabs with varnish residues were placed on the bottom of a Petri dish. One Petri dish was used to dissolve the varnish residues in ethanol, another to dissolve them in toluene. The solvents were carefully poured over the swabs, so that the resin would dissolve better. After a couple of days, the solvents with extracted resins were filtered through filter paper. The swabs were drained and rinsed with the appropriate solvents.
27Reference samples were put into test tubes with an impermeable cap and allowed to dissolve in ethanol and toluene.
28The solutions of the varnish samples and reference resins were spotted, applying the capillary tube to selected places on the TLC plate. The concentration of the resins in the spots was checked using UV light. The solvent mixture for the mobile phase was prepared using 95 ml benzene and 5 ml methanol. The TLC plate was placed in the solvent tank. When the plate was removed from the tank, the solvent front was marked. After the solvents of the mobile phase had evaporated from the TLC plate, the plate was sprayed with the resin indicator antimony (III) chloride in chloroform. After spraying, the plates were first dried in the fume hood and then heated at about 110 °C 5 for 10 minutes.
29A part of the filtrate sample with the solution of the resinous varnish was dried in the laboratory dryer. The firm residue of the resin was then mixed with KBr to make a pellet for the FTIR measuring using the same procedure as it was described for the pigment samples.
30Spectrums of all reference resins used in thin layer chromatography - dammar, colofonium, sandarac, Manila copal, mastic, and the varnish sample from the St. Michael antependium - were measured. The samples were measured at 500 to 4200 cm-1 spectral range of IR-spectrum.
31None of the spots of the components of the reference samples completely matched those of the sample from the antependium. By examining the developed plates with the samples of different binding mediums, it was noticed that thickened linseed oil shows results similar (not identical!) to those of the sample analyzed.
Fig. 3 TLC plate with reference samples
Different binding mediums and the sample of the binding medium from St. Michael's antependium.
Credits: Aida Grga
32By measuring the sample of lead white, the absorption bands characteristic for lead carbonate appeared the most intense at 1406,15 cm-1, 1045,45 cm-1 and at ~ 680 cm-1. The measuring of IR-spectrum of the sample from St. Michael’s antependium showed the same result. By joining the spectrums, it was concluded that the white pigment used at the antependium was lead white.
Fig. 4 IR-spectrum
IR-spectrum of lead white and IR-spectrum of the sample of the white paint from St. Michael's antependium.
Credits: Aida Grga
33The presence of lead, i.e. lead white, was detected in the white paint of the antependium. When the drops of de-ionized water were placed on the Plumbtesmo paper, it turned orange. A minute later, the orange colour disappeared and paper turned pink (Fig. 5). This was due to the presence of lead in the sample. As a very small sample was used, the colour change was more easily observed under the microscope.
Fig. 5 Observation under microscope
Due to the presence of lead in the white paint from the antependium, pink colour appeared on the Plumbtesmo paper. This was observed under the microscope.
Credits: Aida Grga
34The results of Plumbtesmo test of the reference minium and of the red paint sample were compared. It was concluded that the red paint sample did not give positive reaction, i.e. it did not contain lead compounds. This excluded the presence of minium in the red paint on St. Michael's antepenidum.
- 9 This did not occur on the blank determination.
35When a drop of the sample solution was deposited on the prepared surface, the aluminium foil quickly corroded and grey-black colouring appeared. This indicated the presence of mercury in the sample.9(Fig. 6) The pigment in the sample is probably a mercury compound, i.e. vermilion.
Fig. 6 Dark corrosion of aluminium foil caused by mercury
Testing of the red paint from St. Michael's antependium.
Credits: Aida Grga
36When potassium ferrocyanide was added, blue colour (Prussian blue) appeared along the edge of the solution of the sample (Fig. 7). This proved the presence of iron ions in the sample of ochre paint from St. Michael’s antependium.
Fig. 7 Test for iron (copper and aluminium) using potassium ferrocyanide
On the left is the ochre used as a reference sample, and on the right is the sample of the ochre paint from St. Michael's antependium. Blue colour proves the presence of iron ions.
Credits: Aida Grga
37When measuring the IR-spectrum of the reference sample of ochre, its characteristic absorbance bands appeared at 1622,19 cm-1 and the two most intense at 1139,97 cm-1 and 1114,89 cm-1. The ochre paint sample from the antependium showed absorbance bands at 1620,26 cm-1, 1078,24 cm-1 and 1043,52 cm-1. By joining the spectrums, the results of the micro-chemical analysis, i.e. the presence of the ochre pigment in the ochre paint from St. Michael's antependium, were confirmed.
38When the drops of potassium ferocyanide were added to reference samples with green earth, they turned blue. This was due to the presence of iron ions. The sample of malachite and the sample of green paint taken from the antependium turned reddish-brown (Fig. 8). This colouring proved the presence of copper ions.
Fig. 8 Formation of the reddish brown colour
Formation due to the presence of copper ions in the green paint from St. Michael's antependium
Credits: Aida Grga
39The presence of copper was revealed in the sample of green paint from St. Michael's antependium. The ensuing question was whether the pigment in the paint was malachite (copper carbonate) or verdigris (alkaline copper acetate). To answer this question, the presence of acetate or carbonate group had to be determined.
40IR-spectrum of smalt showed the most intense absorption band at 1031,95 cm-1 and a smaller one at 769,62 cm-1. In the measuring of the blue paint sample from St. Michael's antependium, absorption bands characteristic for smalt appeared at 1078,24 cm-1 and 783,13 cm-1. Also, lead white values appeared (1045,45 cm-1, 1417,73 cm-1, ~ 680 cm-1), indicating mixing of the pigments of the blue colour sample with other white pigment from the paint.
41By comparing the IR-spectrums, it was confirmed that the blue paint on St. Michael's antependium contains smalt.
Fig. 9 Comparison of IR-spectrum
IR-spectrum of smalt and IR-spectrum of the sample of the blue paint from St. Michael's antependium
Credits: Aida Grga
42Different concentrations of the resin solutions were analyzed, which did not bring about satisfactory results. TLC had to be repeated several times in order to get a satisfactory imaging of the components, which was important for the precise identification of the type of resins used in the varnish of the antependium. The resins used as varnish were not pure compounds. Instead, they were made up of several components. TLC method was used to separate these components. In this way, spots arrangement characteristic for individual resins were obtained. This was the basis for the identification of the varnish. The arrangement of the separated components of the resin sample was compared with the separated components of the reference sample. Based on the similarity of patterns, the conclusion was reached as to the composition of the sample taken from St. Michael's antependium.
- 10 Mastic was common resin in Dalmatia in the 19th century.
43The indicator was used to make the spots of the resins visible under the UV light (365 nm). The arrangement of the spots in the sample of the varnish from St. Michael's antependium matched the arrangement of spots/components of mastic. It was thus concluded that mastic was present in the varnish.10
44IR-spectrum of mastic showed absorption bands at 2947,33 cm-1 (the most intense), 2874,03 cm-1, 1707,06 cm-1, 1456,30 cm-1 and 1384,94 cm-1. As for the sample of the varnish from St. Michael's antependium, following values for the absorption bands were recorded: 2926,11 cm-1 (the most intense), 2852,81 cm-1, 1712,85 cm-1, 1460,16 cm-1 and 1384,94 cm-1.
45By joining/comparing the measured spectrums, i.e. the wavelengths of the absorption bands of the resins, it was concluded that IR-spectrum of the varnish sample from St. Michael’s antependium matched the IR-spectrum of the sampled mastic (Fig. 10). That confirmed the results of the varnish analysis, carried out using TLC.
Fig. 10 IR-spectrum of mastic and IR-spectrum of the sample of the varnish from St. Michael's antependium
Credits: Aida Grga
46When painting the wooden antependium of St. Michael's altar, Matej Otoni used materials that were common in the 17th century and available in Dalmatian Zagora. To identify these materials, as well as the varnish that was applied at a later date, a number of analytical techniques were used.
- 11 It should be noted that IR-spectrums of pigments from St. Michael's antependium have different int (...)
- 12 This was the case even with very small samples.
47TLC was a helpful tool, especially when determining the composition of the varnish, which contains mastic resin. This technique, however, did not give satisfying results when analyzing the binding medium in the paint layer. FTIR confirmed the presence of mastic in the varnish and determined the type of the pigments analyzed (lead white, ochre and smalt).11 Micro-chemical analysis proved to be useful in determining the metals present in some pigments.12 The study showed that Otoni used lead white, vermilion, ochre and green copper pigment to paint the antependium of St. Michael's altar.