Novel applications of the Er:YAG laser cleaning of old paintings Alessia Andreotti1, Paola Bracco2, Maria Perla Colombini1, Adele deCruz2, Giancarlo Lanterna2, Kyoko Nakahara2, Francesca Penaglia1 1

Dipartimento di Chimica e Chimica Industriale, Università di Pisa, via Risorgimento 35, 56126 Pisa, Italy 2 Opificio delle Pietre Dure, Fortezza da Basso, Viale Filippo Strozzi 1, Firenze, Italy 3 Duke University, Department of Chemistry, Durham, NC, USA Abstract. The present contribution focuses on the use of Er:YAG laser cleaning technique for the removal of unwanted and/or degraded materials both from a large series of reference standards (overpainting, varnishes, patinas and restoration materials) which simulate the layering of old paintings, and also examples from old paintings. A series of diagnostic controls (optical microscopy, SEM, FT-IR, GC-MS, and topographic techniques) was designed to study the effects of the laser radiation on the surface components, including morphological, optical and chemical examination. The most significant results show that an effective thin-layer-removal of about 90% is obtained by submitting the painted surfaces to the laser exposure, while the rest of cleaning is rapidly accomplished in safety by applying mild solvents or aqueous methods. Consequently, possible interference with the original substrate can be noticeably minimized. No degradation compound induced by laser energy was formed. The laser cleaning procedure applied on an oil painting canvas “Morte di Adone” (17th century), and on a panel tempera painting “San Nicola e San Giusto” of Domenico di Michelino (15th century) shows that the surfaces cleaned by this system exhibit a morphology quite similar to that obtained by traditional cleaning methods.

1.

Introduction

The cleaning of painted surface is one of the most critical operation in conservation. Contemporary criteria for the cleaning procedures requires a selective and progressive removal of materials, which can enable an expert conservator to preserve any external thin original glaze or even old varnishes or patinas. Moreover, for environmental needs and safety reasons it is required to abandon toxic solvents cleaning and to use alternative methods, such as aqueous methods or less-toxic solvent cleaning. Er:YAG laser ablation, in fact, meets these requirements. Actually, laser techniques have demonstrated very promising applications for diagnostic and conservation purposes in art conservation [1]. As reported in the literature[2], Er:YAG laser exposure at 2.94 micrometers on a surface dampened with a liquid containing –OH groups, effectively removes old varnish and other encrustations without inducing unwanted chemical or physical changes on the original painted surfaces on canvas and wood panels. On the basis of the previous results [3,4], this paper presents an exhaustive protocol for the use of Er:YAG laser technique in the removal of superficial layers from paintings to obtain a comprehensive knowledge on laser efficiency, fluence,

Andreotti A. 1

threshold limits, and its practical cleaning approach combined with other ancillary methods. These parameters are tests first on the laboratory models and afterwards on old paintings. The study and control of the polychrome surfaces before and after removal of the coatings by Er:YAG laser were performed by a series of diagnostic analyses which provides fundamental information on the morphology and chemical composition of the treated surfaces.

2.

Experimental Methods

Experiments of Er:YAG laser cleaning were conducted firstly on the laboratory tempera or oil paint models with top layers varying from natural resin varnishes, oil-resin varnishes, synthetic varnishes/fixatives/adhesives, artificial patinas, to overpaintings, which were previously prepared in 1999 [3]. The top-layers examined by laser ablation were: shellac (natural resin varnish); ketone and vinylic resin (synthetic varnish); mastic and walnut oil mixture varnish, boiled linseed oil varnish (oil based varnish); burnt umber-casein overpainting, Naples yellow-linseed oil overpainting, Naples yellow-casein overpainting, burnt umberlinseed oil overpainting on gypsum/rabbit glue (dark-colour/light-colour overpainting and overpainting on new ground); EVA based resin, acrylic resin (BMA and EA/MMA). In order to study the behaviour of the paint layer (tempera/oil) exposed directly to laser, uncoated surfaces were also examined. On the paint model simulating multiple layers (burnt umber-linseed oil overpainting, stucco with gypsum/rabbit glue, mastic varnish, yellow ochre/lead white in egg tempera, ground of gypsum/rabbit glue), laser ablation was tested to prove its gradual thinning action. All the laser tests have been executed under the control of stereo microscope by painting conservators in order to study the action of laser ablation and to determine threshold for each material examined. The laser ablation testing has been proceeded with application of progressively increasing energy to the surface, firstly starting with dry methods and secondary with wet methods. To study the overall effects of laser, the energy levels beyond the threshold limit were also examined. The auxiliary wetting liquids, applied with small cotton swab, were:OH containing substances (distilled water, ethanol, water-ethanol mixture (1:1 v/v), distilled water with 1% surfactant [Tween20]); non O-H containing liquid (light aliphatic hydrocarbons [Ligroine]); non O-H containing liquid added with O-H containing substances [White Spirit with 15% di-ethylene-glycol]). The range of energies used was between 3mJ and 200mJ, with a laser beam diameter of 1 mm at 10 and 15 Hz macropulse frequencies. The Er:YAG lasers used were ‘CrystaLase 2940’ and Light Scalpel© of Mona Laser Inc., U.S.A. A previously published procedure [3,4] was employed for the operative laser conditions, and the collection of ablated materials. These materials were analyzed by PY-GC-MS (Hewlett Packard, Palo Alto, CA, USA) and GC-MS (Thermo Electron Corporation, USA) methods for the analysis of synthetic polymers, amino acids, fatty acids, and terpenoids [5,6]. OM with VIS and UV sources (Zeiss Axioplan) were used to observe the painted specimen surfaces and their modification after the laser pulses. SEM (Leica Cambridge) and EDS (Link-Oxford) were used to

Andreotti A. 2

study the complex morphology changes of the surfaces both in flat samples and in cross sections. FTIR (Thermo-Quest) was used to compare the materials left on the surfaces and the ablated materials collect on the cover-slip glass. μProfilometry INOA prototype (Florence, Italy) was used to measure micrometers differences in depth after the different laser pulses.

3.

Results and Discussion

In general, the optimal energy thresholds for thin top-layers (thickness ≈ 15 µm), of natural or synthetic resin, is between 8 - 13 mJ with auxiliary liquid containing O-H bond. In particular, synthetic polymers such as Plexisol (n-butyl methacrylate), or Plextol (methyl methacrylate/ethyl acrylate) were successfully removed, while BEVA (ethylene vinyl acetate) was not ablated: however, a deep surface modification was provoked by the laser exposure which allowed the cleaning by a simple swabbing technique. Oil based-varnish models resulted quite resistant to laser ablation and showed a morphology (Figure 1) with an increased roughness of the surface after laser ablation. SAMP VIS

SAMP SEM/SE

SEZ VIS.

A1 YOET 11-1-3 Shellac 13 mJ Dry SEZ UV

SEZ SEM/SE

SEZ SEM/BSE

Figure 1. Microscope images of the sample YOET 1-1-3 (shellac varnish on lead white/yellow ochre dispersed in egg yolk tempera; gypsum/glue ground). All the images are referred to the same sample observed under Optical Microscopy and SEM (gold sputtering was used). The fragment was then embedded in resin and observed again in OM and SEM. All the images show on the left side the unexposed surface

Andreotti A. 3

This increase, together with the introduction of chemical agents, permits to achieve a cleaned surface in a shorter time using a low concentration of chemicals. Thick oil based overpaintings (about 30-50 µm) can be ablated at an energy level between 30 and 100 mJ with auxiliary hydroxyl liquids. Repeated exposure may be performed according to the thickness and nature of the superficial layers, without exceeding established energy threshold. This allows a gradual removal of the overpainting without causing damage or discoloration of the substrate. Experiments on the multi-layers model demonstrated a gradual and very thin laser ablation as shown in Figure 2, where the energy of each passage and the wetting agent are reported. 1°

2°

3°

10080mJ 10080mJ 80mJ 80mJ 105mJ 105 mJ

H2O + Ethanol

4°

80mJ 80mJ

5°

50mJ 50mJ

6° 6°

150mJ 150mJ

White Spirit

7° 7°

8° 8°

150mJ 80mJ 150mJ 80mJ

9° 9°

50mJ 50mJ

10° 10°

20mJ 20mJ

H2O + Ethanol

Figure 2. The gradual action of Er:YAG laser is shown on a laboratory multi-layers sample: burnt umber-linseed oil overpainting on stucco (gypsum/rabbit glue) which covers mastic varnish on yellow ochre/lead white egg tempera.

The thick overpainting (burnt umber in boiled linseed oil) was gradually thinned by five laser applications (100mJ, 80mJ, and 50mJ) using distilled water/ethanol mixture (1:1) except for the last treatment, where white spirit added with 15% di-ethylene-glycol was used. Then the thick stucco with gypsum and rabbit glue was thinned by five laser applications (150mJ, 80mJ, 50mJ, and 20mJ) using distilled water/ethanol mixture (1:1). The thin layer of mastic varnish as well as the egg tempera with yellow ochre/lead white underneath were well preserved. The laser cleaning was then applied on a 15th century tempera painting on panel, “S. Nicola e S. Giusto”, attributed to Domenico di Michelino, from S. Giusto in Piazzanese church in Prato (Italy), under restoration at the Opificio delle Pietre Dure (Laboratories of restoration in Florence, Italy). The laser ablation was executed on the degraded greyish thin and compact film tightly bound to the original layer (a very fragile white tempera). This layer locally contains discontinuous old brown varnish and some residues of grey-coloured overpainting in the area of the mitre of Saint Nicolas. (Figure 3)

Andreotti A. 4

5-6 mJ, White Spirit + solvent 1. laser

2. 3.

Fatty emulsion added with artificial saliva and coccocollagene Fatty emulsion pH 7 Ligroin

5-6 mJ, White Spirit + solvent 1. 2. 3. 4. 5. 6.

Fatty emulsion added of resin Soap made with DCA-TEA Fatty emulsion pH 7 Ligroin Fatty emulsion added of Citric acid Fatty emulsion pH 7 Ligroin

Old tested area with traditional solvent

Figure 3. 15th century tempera painting on panel, “S. Nicola e S. Giusto”, attributed to Domenico di Michelino. Cleaning tests with combined Er:YAG laser and emulsions (top and middle) compared with an old test area (bottom) cleaned with mechanical tools and traditional solvents

The chemical identification of the grey patina was performed by analysing the organic material with the GC-MS analysis: the amino acids percentage data showed that the patina was mainly constituted by egg with small traces of animal glue. Moreover, the FTIR analysis demonstrated the presence of calcium oxalate on sample surfaces. This area was considered to be suitable for laser test, as it was extremely difficult to execute a selective and safe cleaning even by using sophisticated aqueous methods based on emulsions containing surfactants, resin soaps, or artificial saliva [7,8]. After preliminary tests using the energy between 3 mJ and 9 mJ at the frequency of 15Hz, and wetting agents white spirit or distilled water/ethanol mixture (1:1 v/v), a combined method with laser ablation and aqueous and mechanical method was considered to be appropriate. Initially the energy level of 8mJ with white spirit as wetting agent and a clearing agent as water/ethanol mixture (1:1 v/v) after laser ablation, seemed to offer a good result as far as removal itself concerned. The combined method with laser at 5 – 6mJ with white spirit as wetting agent and aqueous methods using an emulsion added with resin soap, followed by a milder pH 7 emulsion and by Ligroin one, seemed to be also rather efficient. However, the first method did not seem to guarantee a selective and gradual cleaning, judging from observation under stereoscopic microscope especially due to extreme sensibility of the original colour to aqueous solution; the second combined method either was considered inappropriate, as it removed also the antique inner varnish which was considered to be preserved. Finally, the appropriate method consists of:

Andreotti A. 5

• a laser ablation at 5 - 6mJ with wetting agent (white spirit), which permits homogenous superficial desegregation of the grey-coloured layer; • a following cleaning with a fatty emulsion [7] (Brij 35 2g, artificial saliva (Mucin 0,25g, Tribasic Ammonium Citrate 0,25g, deionized water 100ml) 20ml, coccocollagen 2,5ml, Ligroin 80ml); • a treatment with a clearing agent, a pH 7 emulsion (Brij 35 2g, deionized water 10ml, Tween 20 (non-ionic surfactant) 2ml, Ligroin 90ml) [8]; • a final treatment with Ligroin alone. By this way, it was possible to achieve a real selective removal of the degraded materials without removing the old varnish underneath. This approach is in agreement with the latest criteria of contemporary conservation, oriented towards the minimum intervention, and the reduction of toxic chemicals. The laser cleaning approach was also used for removing the insoluble and hard overpaintings on the canvas painting “Morte di Adone” (17th century). Either using various cleaning approaches based on polar or basic solvents, acid solution, enzymes, enzyme-soaps, chelating agents, and surfactants, any efficient and safely removal of the unwanted materials was observed; nor the scalpel was successful due to the hardness of the overpaintings and their strong adhesion on the original substrate. The investigation on the grey overpainting in the area of the Venere’s shoulder by the GC-MS analysis, highlighted the presence of egg as proteinaceous material and a small amount of beeswax. The analysis of the organic material, laser sampled in different areas from a tick overpainting on the Venere’s face, showed that it was mainly constituted by animal glue (Figure 4) with a small trace of a pinacae resin. PC2 3

eg 1

0 -

-

animal glue

-

-

0

2

4

-

-

Database Pos. 3 2° passage, 15mJ 2 Pos. 3 1° passage, 19mJ 3 Pos. 2 1° passage, 18mJ 4 Pos. 1 1° passage, 19mJ 5 Pos. 1 1° passage, 15mJ 6 Pos. 2 2° passage, 17mJ 7 Pos. 3-4 1° passage, 10mJ 1

2

casein

isopropylic alcohol + Ligroine (1:1) as wetting agent

-

PC1

Figure 4. Principal Component Analysis of samples from 17th century oil painting on canvas “Morte di Adone”: the relative amino acid percentage contents of the ablate material from different area of Venere’s face reveal the presence of animal glue

Andreotti A. 6

Laser pulse was repeated several times on the surface, in decreasing energy levels, (18-20mJ, 15mJ, and 10mJ) as shown in Figure 5.

N°1

N°2

N°3 N°4

Figure 5. 17th century oil painting on canvas “Morte di Adone”: four overpainting removal tests realised by combined Er:YAG laser and emulsions actions. Left: tested area; right: close details

This procedure permits the removal of the most of the overpainting and allows a final cleaning with application of a mild solvent mixture and a soft employment of the scalpel without removing the old varnish. A non-invasive analysis, μprofilometry, executed before, during, and after laser cleaning allows to quantify the depth of laser ablation as 3-5 µm (Figure 6). This avoids the need to take samples for cross-sections.

Figure 6. Laser micro-profilometry results of first and second exposure of Er:YAG laser and emulsion on area N°1. Left: level difference between unexposed surface (orange-red) and first passage ablation level (yellow-green); centre: difference between surface and second passage level (cyan); right: relative level difference between first and second passage (5 μm on average)

The operative conditions adopted for the removal of the overpainting were 1518 mJ energy, wetting liquid iso-propanol/ligroine 50% and iso-propanol 27%, followed by a clearance with the same liquid, repeated several times. Results show a very homogeneous and superficial removal of irregular top-layers after each laser application. However, small changes in laser exposure (about 10mJ) and in the composition of the solvent mixture (iso-propanol/ligroine 1:1), may also give successful results as shown for the different areas in Figure 5.

Andreotti A. 7

4.

Conclusions

The Er:YAG laser ablation offers a gradual, homogeneous, and OH-selective cleaning with its thinning action of a few micron’s depth. The general mechanism of ablation consists in modifying the surface by breaking up of a few micron’s depth, which enables the conservator to complete the cleaning using less invasive methods which are not efficient if used by themselves. The preliminary tests under microscope are indispensable to obtain the optimal threshold for each specific material, and to avoid possible surface modification or chromatic alteration of the original substrate which may not be directly observed. The macropulse frequency of 15 Hz was confirmed as the optimum to obtain a more homogeneous ablation. No chromatic alteration was observed below the energy threshold, except for certain colours such as yellow ochre, some mineral iron based blue or Naples yellow in casein. The use of hydroxyl-wetting agents is appropriate, as it increases the efficiency of laser ablation as well as safety of the original substrate. It is important to note that chemical analyses show that ablation under the threshold limits provokes no significant variation in chemical composition of both ablated materials and substrate. Analyses also show that the energy levels used below the safety threshold, combined with the conservators’ skills, permit gradual and progressive cleaning. Particularly, repeated laser applications at lower fluencies in respect to the optimal energy thresholds allows the cleaning of degraded. To conclude, we can assert that the combination of Er:YAG laser, used within the energy thresholds of each material, together with the ultimate chemical and biochemical systems, allows the cleaning of a broad variety of unwanted layers. With the suggested cleaning procedure, conservators/restorers have a further and alternative chance to properly solve difficult restorations.

Acknowledgements Ente Cassa di Risparmio of Florence is gratefully acknowledged for the large contribute given for these researches. The authors are very gratefully to Raffaella Fontana, Enrico Pampaloni and Maria Chiara Gambino (INOA, UNIFI) for the μprofilometry measurements and interpretations, to Fabrizio Cinotti for the macroimages on old paintings, and to Annette Keller for the graphic elaborations.

References [1] De Cruz, A. Hauger, S. Wolbarsht, M. L. “ The role of lasers in fine arts conservation and restoration. Opt. Photon. News 10 (1999) 36-40 [2] De Cruz, A. Hauger, S. Wolbarsht, M. L. “Laser removal of contaminants from painted surfaces” J. Cult. heritage 1 (2000) 173-180. [3] E. Adamkiewitz, P. Bracco, M. P. Colombini, A. De Cruz, G. Lanterna, M. Matteini, K. Nakahara, O. Sartiani, M. L. Wolbarsht, “Er:YAG laser, an innovative tool for controlled cleaning of old paintings: testing and evaluation” J. Cult. Heritage, 4 (2003) 202-208.

Andreotti A. 8

[4] A. Andreotti. M.P. Colombini, G. Lanterna, M. Rizzi: “A novel approach for high selective micro-sampling of organic painting materials by Er:YAG laser ablation” J. Cult: Heritage, 4 (2003) 355-361. [5] I. Bonaduce and M.P. Colombini, “The Characterization of Beewax in works of Art by Gas Chromatography-Mass Spectrometry and Pyrolysis-Gas Chromatography-Mass Spectrometry Procedures” Journal of Chromatography, 1028 (2004) 297-306 [6] I. Bonaduce and M.P. Colombini, “Gas Chromatography/Mass spectrometry for the chracterization of organic materials in frescoes of the Monumental Cemetery of Pisa (Italy)” Rapid communication in Mass Spectrometry, 17 (2003) 2523-2527 [7] P. Cremonesi, “L’uso degli enzimi nella pulitura di opera policrome”, Padova 1999 [8] P. Cremonesi, “L’uso di tensioattivi e chelante nella pulitura di opere policrome”, Padova 2001

Andreotti A. 9