THE SCROVEGNI CHAPEL: STUDIES ON STATE OF CONSERVATION AND CLEANING PROCEDURES

E.Borrelli, M.Marabelli, P.Santopadre

1. FOREWORD

Since 1988, the analyses and tests carried out in the Scrovegni Chapel in Padua have had three basic aims: 1 - checking the distribution and the intensity of sulphation of the paint film and the plaster backing by means of analyses with SEM-EDS (scansion electronic microscope using micro-analysis with energy dispersive X-ray spectrometry); 2 - determining the quality and quantity of soluble salts on the painted surfaces by means of extraction with pads made of cellulose pulp under standard conditions; 3 - devising procedures for cleaning and consolidation.

2. ANALYSES WITH SEM OF CALCIUM SULPHATE DISTRIBUTION

Several micro-fragments of the wall paintings were taken, focusing attention on the inner wall of the facade which shows the most serious degree of deterioration

(fig.1).

The samples were analysed without any special preparation, positioning them sideways so as to allow stratigraphic observation and analysis.
X-ray analysis was carried out on the samples in order to highlight the distribution of the component elements and the sulphur (see figs.2 -11).

Sample S1: fragment of black pigment, inner wall of facade, from base (above the figure of Lucifer).
Calcium is present throughout the sample, with significant denser areas in several clearly visible granules, of diameter varying between 100 and 300 microns, made up of calcium carbonate predominantly.
Sulphur, associated with the calcium, is widespread to a depth of about 130 microns, surrounding the granules of inert material (calcite and quartz), occurring not only in the painted layer but also in the plaster backing (fig.2).

Some granules show the presence of potassium, aluminium and silicon.
Magnesium is present in smaller quantities than calcium, associated in particles that are probably of dolomite, dispersed in the plaster backing.

Sample S2 : fragment of red-brown pigment, inner wall of facade, from the rock beneath the first flame on the left.
In this sample the sulphur seems to be concentrated only on the surface of the colour, except for a slight infiltration inside

(up to 17 microns, fig.3);

iron is present in the painted layer.
The granules of inert material are made up of calcite, with a large granule of dolomite (diameter of about 105 microns).

Sample S3: the fragment of black pigment (taken from the inner wall of the facade, from the base, above Lucifer) is made up of a detached scale (thickness 150 microns), showing a marked concentration of sulphur on both surfaces. This may be due to absorption of SO2 both from the external surface, and from the fractured surface; the maximum penetration is about 60 microns

(fig. 4).

Sample S4 : fragment of red-brown pigment, inner wall of facade, from the rock beneath the flames of the Inferno.
Sulphation has affected the whole layer of red-brown pigment (to a depth of 60 microns, fig. 5), per cui la relativa Xgrafia mostra associati zolfo, calcio, ferro.

so the relative X-ray maps show associated sulphur, calcium, and iron.
Magnesium is also very evident in the painted layer, probably by association calcite-dolomite.
Potassium is associated on the surface with aluminium and iron, and in a granule deeper down with aluminium and silicon.

Sample S6 : fragment of green pigment, right wall, from the horizontal band beneath the fourth window (from the entrance).
Sulphur is present only on the surface of the painted layer (up to 10 microns, fig.6) .   In the granules of pigment (green earth), silicon is associated with aluminium, potassium and iron, and with aluminium and potassium in an internal granule; in the plaster backing, there are granules of quartz sand (max. diameter 170 microns).

Sample S7 : fragment of green pigment, left wall, from the band to the right of the detached panel "Disputation amongst the doctors".
The sample shows a fracture parallel to the surface at a depth of about 220 microns.
Sulphur has permeated the sample to the fracture(fig.7).

Potassium and iron are quite clearly evident in the painted layer associated with silicon.

Sample 3: fragment of white pigment from the horizontal decorative band above the penultimate upper panel, left wall, where it meets the vaulting. The penetration of the sulphur, associated with potassium, is to a depth of about 25 microns

(fig.8).

Sample 6: fragment of white pigment from the horizontal decorative band above the last upper panel, right wall, where it meets the vaulting.
The sulphur has penetrated to a depth of about 50 microns; inside, the silicon is associated with aluminium, potassium, and sodium; there are three granules of dolomite (maximum diameter 140 microns, fig.9).

Sample 8: fragment of white pigment from the horizontal decorative band beneath the throne of God, on the main arch.
There is a parallel fracture on the surface to a depth of about 60 microns, the limit of penetration of the sulphur. Inside, silicon is associated with potassium, aluminium, and sodium; there are also small granules of dolomite and two large granules of quartz sand (maximum diameter 280 microns, fig.10).

Sample 11: fragment of white pigment from the left wall, at about eye level, on the main arch.
The sulphur has penetrated to a depth of about 60 microns. Inside, there is a large granule of calcite (diameter of about 120 microns); the silicon is associated with potassium, aluminium and sodium

( fig.11).

These analytical data lead to the following conclusions:
a. Samples S2 and S6 show limited sulphation only on the surface of the painted layer (10-17 microns), while in samples S4, 6, 8 and 11 the attack affects the whole painted layer right through its thickness (max. 60 microns), without penetrating the plaster backing.

b. Sample S7 is somewhat anomalous, but it is evident that sulphation by absorption of sulphur dioxide can affect interrupted parts down to a certain depth, if in communication with external air.
The same can be said of sample S3.

c. Sample S1 certainly represents a case of particular deterioration of the painted layer: not only has the sulphation on the surface changed calcium carbonate into gypsum, but it has also tended to penetrate the plaster backing in depth, following the edges of the granules of silicates and attacking superficially the granules of calcite of larger dimensions (maximum penetration 130 microns).

d. Sample 3 contains potassium sulphate; in fact, aphthitalite (sulphate of sodium ad potassium) together with thenardite (sodium sulphate) has been found, by XRD analysis, in a sample of efflorescence taken from the main arch, the "Gabriel's Mission" scene. Therefore, in order to check whether the presence of salts might be related to the use of cement-based mortar, samples of mortar used in previous restoration were analysed using X-ray diffraction. Results showed the presence of lime and sand-based mortar as well as hydraulic lime mortar with low levels of gypsum, while gypsum was probably used as a consolidating agent for the deeper layers.

e. It is clear that the sampling carried out, though limited with respect to the whole of the chapel, makes it possible to give a sufficient indication of the sulphation degree of the wall paintings.
From the analyses, this phenomenon seems to exist, on average, over a range of thickness varying between 10 and 60 microns, starting from the external surface. This fact shows therefore that the accumulation of calcium sulphate comes mainly from the environment towards the painting.
The XRD analyses, carried out on samples of powder taken from the inner wall of the facade, from the main arch and from the vaulting, show the presence of gypsum on and within the painted layer, associated with calcium oxalate (weddellite).

3. TESTS FOR EXTRACTION OF SOLUBLE SALTS

3.1 EXPERIMENTAL PART

The need to conduct a screening to evaluate the types and the quantities of soluble salts on the surface layers of the frescoes and immediately beneath them, without resorting to harmful removal of painted material, convinced us to use an indirect sampling method.
This method, which is often used as a system for cleaning, consists of applying moistened pads of cellulose paste on the surface and letting them dry out completely (usually about 24 hours).
The special feature of the method is that it has been standardised, and that it can also be used for diagnostic purposes.
In practical terms, we used moistened pads of paper pulp (Arbocell 1000, previously washed with de-ionised water), of known thickness and surface area (15x10 cm).

The thickness of the pads is determined by the experience of the restorer and generally speaking, is never >5 mm. In any case, slight variations in the thickness have no effect on the results. The experimental data refer to an area of 100 sq. cm.
As for the quantity of water absorbed, this too is decided case by case, as long as it provides good adhesion to the surface in question, and as long as it dries out in a period ranging from 24 to 48 hours.
The latter depends on the micro-climatic conditions of the environment in which the work is taking place. In most cases, the drying time is between 24 and 48 hours, and the pads applied in the chapel all came into this category.
For each point being examined, repeated extractions were carried out, checking the effectiveness first by means of measuring the specific conduction, then by dosing anionic and cationic components using ionic chromatography on the extracting water.

In order to measure the conduction and then the amount of extracted salts, the following method was used: the paper pulp corresponding to each 5x10 cm pad was treated with 200 ml of de-ionised water; this extracted water was then used for the analyses; the detection limits of this methodology are shown in table 1. Since the results of the analyses refer to pads of known dimensions, they are expressed in mg/100 sq. cm.
The number of extractions carried out on the same point was determined as a function of the conduction values measured from time to time on site.

3.2 COMMENTS ON RESULTS
The amount of salts extracted using the method of the paper pulp pads, carried out on all the walls of the Chapel (fig.12), made it possible to obtain significant data that was useful for defining the conservation project.
The experimental results are summarised in tables 2-14; a visual representation of the same is shown in graphs 1-3. The sequence of the extractions is indicated by the letters a, b, c, etc.

Graph 1 shows the anionic contents of the extraction pads used on the inner wall of the facade (1,2,3,4,5). For pads 1,2, and 3, figures are given for "normalised" extractions (tables 2-6, graph 1) ignoring the results obtained from the first extractions (one-two) which were conducted using a procedure that was different from the one successively adopted. From the results, one can see that inner wall of the facade shows a concentration of extracted soluble salts with a maximum corresponding to an intermediate extraction; this is explained by the fact that the extraction of soluble salts is a gradual process depending on the spread and concentration of the salts from the inside towards the outside of the wall.
Generally speaking, the inner wall of the facade has the highest concentrations of salts, with a maximum anionic level of 100 mg/100 sq. cm. and with a clear predominance of nitrates, also because they are more soluble than sulphates and therefore more easily extracted with this method.

Moving on to examine the data relating to the pads applied to the left wall (pads 9 and 12, tables 10,13 and graphs 2,3), one immediately notices a variation in scale of the concentration with respect to the inner wall of the facade; in this case, the highest cumulative figure of total anions just exceeds 50 mg/100 sq. cm.
Pad 9 shows the release of salts in a sequence of four successive extractions, while for pad 12 only two extractions were required.
On the right wall, at least in the areas examined, the situation is better: both pads 8 and 11 (tables 9,12 - graphs 2,3) show slightly lower levels of salts extracted. In this case, on the basis of conductivity data of the subsequent extractions and the low concentrations of salts revealed in the first extraction, no further treatment was carried out. The predominance of nitrates is only evident in pad 9.

In the case of the main arch (pads 6 and 10, tables 7 and 11, graphs 2 and 3), nitrates still predominate, while the extractions with pad 6 show a constant decrease of salts following repeated treatments of the same area, probably due to a higher concentration of salts on the surface.
The maximum cumulative value of anions was around 30 mg/100 sq. cm.
Pads 7 and 13 ( tables 8,14 - graphs 2 and 3) relating to the vaulting, do not differ from the others, except for a detail of extraction 13a which represents the only case of significant inversion in the percentage ratio between nitrates and sulphates.
In most cases, these ratios are about 70/20, while in this case the inversion is very evident (NO3-/SO4-- = 20/60).

Examining the overall data obtained, it becomes clear that there is a close connection between the alteration of the painted layer on the inner wall of the facade (in a very poor state), and significant concentrations of soluble salts.
In particular, the figures for anions, expressed in terms of the percentage ratio between them, showed a constant relationship in all cases, i.e. NO3-/SO4--/Cl- average ratio, 70/20/10.
The most interesting figure concerns the presence of nitrates, which is significantly higher than the level of sulphates in every case.
The state of the other walls is similar to that of the inner wall of the facade, though the situation is less serious on these walls (graphs 2-3). This led to the decision to direct activities relating to the quality of the air towards measuring the quantity of nitrous oxides present in the atmosphere as a polluting gas.
Air quality measurements carried out recently have confirmed that the level of sulphur dioxide has diminished compared to the 1970s (though still at harmful levels for the paintings), while there has been an increase in the levels of nitrous oxides.

4. TESTS FOR CLEANING AND CONSOLIDATION

Sulphates are the most harmful compounds because, on average, they are more soluble, and therefore likely to find the conditions for crystallising beneath the painted layer. In order to work out a method for the removal of sulphates, in July 1994 several tests with ion exchange resin of the anionic type (Akeogel-Syremont) able to capture the sulphates, were carried out on the inner wall of the facade, on the detached painting "Ascent to Calvary" and on "Resurrezione". The tests were carried out with a cellulose pulp pad in order to quantify the residual soluble salts. Before and after these tests, checks were carried out with SEM to verify the effective removal of the sulphates.
The application of the resin, mixed with distilled water, and placing a sheet of japanese paper between the pad and the surface of the painting, lasted for about 20 minutes.
Subsequently, between September and October 1994, this method was applied to the lunette on the main arch.
Overall, therefore, the areas of application were as follows:

1 - Inner wall of the facade, the "Inferno", red zone(figg. 13-16)

2 - Inner wall of the facade, the "Inferno", black zone (figs. 13-16)
3 - "Ascent to Calvary", on the left of Christ
4 - "Ascent to Calvary", on the right of Christ
5 - "Resurrection", central area
6 - Lunette of main arch, blue background, on right of the painted panel
7 - Lunette of main arch, upper angel, on left of the painted panel
8 - Lunette of main arch, angel, on right of the painted panel
9 - Lunette of main arch, third from last figure, on right of the painted panel
10 - Lunette of main arch, figure, on left of the painted panel

At points 3 and 10, there were signs of white efflorescence at the end of the treatment.

Analysis with SEM-EDS, carried out on samples 1 and 2, showed the presence of sulphur to a depth of @ 100 microns in sample 1, and to @ 25 microns in sample 2 (figs. 17-18). The same analyses repeated after the resin treatment showed the complete disappearance of sulphur (figs. 19-20).

Table 15 gives the figures for specific conductivity, as well as the quantitative analyses of soluble ions for pads (cm. 5x10) of paper pulp, applied after the resin treatment.
Comparing the data for specific conductivity with those relating to the inner wall of the facade (first extractions) one can conclude that, except for zone 8, all the other pads show an overall content of soluble salts (after the resin treatment) significantly lower than the "risk levels" for the inner wall of the facade, but comparable to the corresponding data relating to the right wall whose state of conservation can be rated "fair". Zone 8 (the angel on the right of the painted panel) showed obvious flaking of the colours, visible to the naked eye, probably due to the larger quantities of salts. 

The same situation can be seen in zone 3 (detached painting "Ascent to Calvary") and 10 (lunette on main arch), where there is an efflorescence, identified by diffraction tests as thenardite (sodium sulphate). In particular, the detached painting "Ascent to Calvary" showed a degree of deterioration that was both evident and progressive, while zone 10 (lunette on main arch, left portion) had already been affected by efflorescence, probably due to an accumulation of salts caused by water infiltration from the vaulting. Close examination of the lunette, which was possible in March 2001, enabled us to see, at a distance of almost seven years, that the 1994 restoration work had been successful on the whole.

The second problem to be dealt with was that of consolidating the plaster. Several tests were carried out in April 1997 employing two materials that are currently much used in the restoration field: Ledan TB1-ICR (produced by Tecnoedile toscana) and PLMa (produced by C.T.S. s.a.s).
These two products are hydraulic mixtures used for consolidating wall paintings which have internal discontinuities of the structure.
Before these two products were used, they were checked analytically to verify their low saline content, according to the indications set out in the Normal regulations 26/87.
Before injecting these two mixtures into the chosen areas (four on the left wall, two on the inner wall of the facade), colour checks were carried out on the areas to be consolidated, using a portable instrument which shows colours in numerical terms (tri-chromatic co-ordinates).
The same checks were repeated at a distance of two months; the results showed that there had been no significant variation in the colour measurements, which in turn means that there had been no efflorescence of soluble salts.
The decision was taken to use Ledan TB1-ICR, because, of the two, this was the material that ICR had wider and more controlled experience with.

4.CONCLUSIONS

The measurements carried out with various methods and at different times, have contributed, when considered all together, to defining a method of intervention.
It was possible to check the effectiveness of the treatment for the extraction of sulphates using an ion exchange resin of the anionic type; this method can be considered as a further aid to restorers, to be used alongside other procedures which have already been checked, such as pads soaked with ammonium carbonate.
Controls of the restoration work will be carried out using non-destructive X-ray fluorescence, a procedure devised recently specifically for the semi-quantitative determination of sulphur and of chlorine; surveys were carried out on sample areas before restoration, and final checking is underway now that the restoration is finished.
However, since the analytical results have shown that in some cases the sulphation has penetrated into the inner layers of the frescoes (mainly in areas where there are surface discontinuities), it is probably true to say that the current restoration work will not achieve the complete elimination of soluble salts. Therefore it is clear that the recently installed micro-climatic control system will play an important role in the long-term conservation of Giotto's paintings.                 

TABLE 1

Measured ions and concentration limits

1) Fluorides F- < 0.2 mg/100 cm2
2) Chlorides Cl- < 0.2 "
3) Nitrites NO2- < 0.4 "
4) Nitrates NO3- < 0.4 "
5) Phosphates PO4--- < 0.4 "
6) Sulphates SO4-- < 0.4 "
7) Oxalates C2O4-- < 0.8 "
8) Sodium Na+ < 0.2 "
9) Potassium K+ < 0.4 "
10) Ammonium NH4+ < 0.4 "
11) Magnesium Mg++ < 0.2 "
12) Calcium Ca++ < 0.8 "

TABLE 2

Data relating to the inner wall of the facade Pad no. 1

TABLE 3

Data relating to the inner wall of the facade Pad no. 2

TABLE 4

Data relating to the inner wall of the facade Pad no. 3

TABLE 5

Data relating to the inner wall of the facade Pad no. 4

TABLE 6

Data relating to the inner wall of the facade Pad no. 5

TABLE 7

Data relating to the main arch Pad no. 6

TABLE 8

Data relating to the vaulting Pad no. 7

TABLE 9

Data relating to the right wall Pad no. 8

TABLE 10

Data relating to the left wall Pad no. 9

TABLE 11

Data relating to the main arch Pad no.10

TABLE 12

Data relating to the right wall Pad no. 11

TABLE 13

Data relating to the left wall Pad no.12

TABLE 14

Data relating to the vaulting Pad no.13

TABLE 15

Data relating to the cleaning tests