Journal of Cultural Heritage 10 (2009) 403–409

Original article

Consolidation of paint on stained glass windows: Comparative study and new approaches Noemi Carmona a,b,∗ , Katrin Wittstadt a , Hannelore Römich a,c a

Fraunhofer-Institut für Silicatforschung (ISC), Bronnbach 28, D-97877 Wertheim-Bronnbach, Germany b Materials Physic Department, Complutense University at Madrid, Spain c Conservation Center, Institute of Fine Arts, New York University, USA Received 8 April 2008; accepted 10 December 2008

Abstract Stained glass windows belong to the most precious pieces of art in many European countries. Examples of heavily endangered paint on glass are reported in the literature and mainly related to condensation effects and air pollution, as stained glass windows preferably remain in their original architectural surrounding. Several surface coatings and paint treatments have been proposed to consolidate and protect degraded paint. Very often, the selection of the materials is based more on practical aspects than on scientific research. This study concerns the comparison of some traditional, modern and newly developed consolidants for the preservation of historic glass paintings. Experiments have been carried out with model painted glass samples simulating weathering phenomena of originals. Traditional materials like Paraloid B72, modern ones like SZA (proposed by the Fraunhofer-Institut für Silicatforschung, ISC), and three new consolidants prepared by the sol–gel method and based on different hybrid organic–inorganic alkyl-alkoxysilane systems have been considered. The adhesion, penetration, stability, hydrophobicity, mechanical and chemical resistance are properties and requirements tested to prove their effectiveness and range of use. The three new materials developed in this study for the consolidation of paint on glass have the potential to offer alternatives to existing materials. Nevertheless, further research is necessary before their application in restoration workshops can be recommended. A strategic approach is requested to avoid risks for these valuable historical originals and to contribute to the long-term preservation of the paint on stained glass windows in their original sites. © 2009 Elsevier Masson SAS. All rights reserved. Keywords: Paint; Consolidation; Stained glass window; Sol–gel technique; Conservation

1. Research aim

2. Introduction

The consolidation of unstable paint is one of the major problems during the restoration of stained glass windows. Some of the early treatments for loose paint layers have failed and caused even more damage in some cases. Due to this reason, the aims of the present study were: (1) the development of new sol–gel based consolidants for paint on stained glasses, with improved properties and considering recent developments in sol–gel chemistry; and (2) the comparison of the new systems with the traditional products currently available and with the modern consolidant SZA.

Stained glass windows form an eminent part of the European Cultural Heritage. The earliest and most precious pieces date to the 12th century (gothic period) (e.g. in York, Canterbury, Chartres, Reims, Cologne) [1–4]. A change of style in the 15th up to the 17th century implied variations in manufacturing techniques (important examples in Cologne, Antwerp, León, Gouda and Edam) [5,6]. Nineteenth century stained glass windows stand out by introducing new glass and paint types and different ways of production, although sometimes old religious themes were copied. Stained glass windows are formed by small coloured glass pieces joined with a lead frame, which not only holds the glass segments together but are also part of the artistic composition. The surfaces of some glass pieces are decorated by a layer

∗

Corresponding author. E-mail address: [email protected] (N. Carmona).

1296-2074/$ – see front matter © 2009 Elsevier Masson SAS. All rights reserved. doi:10.1016/j.culher.2008.12.004

404

N. Carmona et al. / Journal of Cultural Heritage 10 (2009) 403–409

Fig. 1. a: SEM cross-section of a painted glass from León cathedral (interior side), Renaissance period. The bulk glass is the area at the bottom and the paint or grisaille is the heterogeneous area in the middle of the picture; b: detail of a glass piece with altered paint from the Crodel panel II, cathedral of Erfurt (transmitting light).

of paint (mostly black trace lines or half-tones) that has been attached by a heat treatment [7]. Glass paint or grisaille is composed of a mixture of metal oxides (PbO, CuO, MnO, CoO, Fe2 O3 ), with silicon oxide or sand (SiO2 ) in different proportions. Normally, they are mixed at ambient temperature with a bonding medium like arabic gum or sugar, and dispersed in water, acetic acid or other solvents. This mixture is finely ground and applied on the glass surface with a brush. Afterwards, a thermal treatment up to 500–600 ◦ C is necessary to adhere the paint on the glass surface [8,9]. The micrograph, Fig. 1a, shows the transversal section of an original painted glass from León cathedral. The heterogeneous band in the middle of the picture corresponds to the paint (average composition given in Table 1). In this example, the paint is well conserved. The heating process was probably carried out under moderate temperature, because the line between the glass (area on the bottom of the picture) and the paint is rather sharp (no broad transition between the paint and the substrate). Since their creation, stained glass windows are subjected to deterioration due to environmental conditions [10-13]. In some cases, the composition of the glass pieces makes them highly sensitive to atmospheric weathering and thus causes the alteration of the paint adhered. In other cases the paint itself deteriorates due to an incomplete firing process and/or a bad adaptation of the paint to the glass surface [14,15]. Independent from the origin of bad adhesion of paint, two types of deterioration phenomena of glass paint can be distinguished: it can disintegrate and appear like powdery paint, or areas of paint may keep internal cohesion but loose adhesion to the substrate, called flaky paint [16,17]. These powdery or flaky paint fragments are partially detached from the glass, leaving behind only negative images as shadows after removal or loss. In addition to any degradation due to material properties or manufacturing, paint on glass is heavily endangered by condensation effects, as stained glass windows mostly remain in their original architec-

tural surrounding (Fig. 1b). Since preservation strategies aim at keeping stained glass windows placed in their original settings (historic churches, cathedrals), it is necessary to protect the paint on the glass from further decay [18]. One of the major problems during the restoration of stained glass windows is how to achieve the consolidation of unstable paint [19,20]. Many surface coatings and paint treatments for stained glasses have included a variety of organic polymers (epoxies, polyvinylacetates, acrylates) originally developed for industrial purposes [21–23]. Many of them have failed and caused even more damage by pealing off the surface and thus damaging the paint. Historically, natural wax has been used as consolidant for the paint in some cathedrals [24,25]. Nowadays, Paraloid B72 is the most commonly used consolidant for this application [26–29]. Its reversibility is considered as an advantage, but the long-term stability is limited by the fact that as an organic material it will be subject to the environmental impact with time. Sol–gel technology shows important advantages concerned the preparation of glassy and glasslike matrices at low temperature with functional properties, like improved chemical resistance, low viscosity or good adhesion to glass substrates [30–33]. Some inorganic sol–gel matrices have been proposed for the consolidation of degraded stones with controversial results [34–36]. A new approach to paint consolidation was proposed by the Fraunhofer-Institut für Silicalforschung (ISC), based on inorganic sol–gel systems (SZA, silicium-zirconiumalkoxides). As laboratory results were promising, pilot studies were performed in 1991/1992 and since then, treated originals have been exposed on site. The main requirements for a good paint consolidant are the following: it needs to penetrate well into the porous structure of the paint; the material should improve the cohesion of the paint itself and the adhesion to the glass substrate; a reasonable stability of the solution is necessary before application (storage

Table 1 EDX analyses of one selected example from Leon Cathedral: paint and glass (Fig. 1 a); numbers given in weight-% of oxides.

Paint Bulk glass

CO2

SiO2

Na2 O

K2 O

CaO

MgO

MnO

Al2 O3

Fe2 O3

P2 O5

PbO

12.9 -

22.5 55.1

3.9

16.6

9.5

7.7

1.0

1.5 1.7

37.4 -

4.5

25.7 -

N. Carmona et al. / Journal of Cultural Heritage 10 (2009) 403–409

405

Table 2 Preparation process and types of simulated damaged paint. Paint type

Powdery paint Flaky paint

Paint characteristics Paint/Arabic gum [weight %]

Mixture/Solvent [weight %]

Heat treatment

1:2.5 1:3

1:10 (distilled water) 1: 10 (acetic acid solution 5 weight-%)

300 ◦ C 10 min. 520 ◦ C 10 min.

Remark: the temperature for heat treatment has to be selected below the recommended temperature for firing paint on glass, in order to produce fragile paint.

conditions) and a long-term resistance to degradation is essential after application of the material; the application should be easy and without damage to the original; the treatment should have a minimal visual impact on the aesthetic value of the paint or the glass around or below the paint. Some limitations about reversibility on porous paint have been accepted by experts in this field. The consolidation process implies the penetration through the paint, and the hardening to hold the porous paint structure together. In this situation, the removal of the consolidant will harm the paint, no matter what material is used. For the consolidation of porous paint, a treatment has to be considered irreversible; instead, the issue of re-treatability should be discussed. All of the commercial products used until now as consolidants have some drawbacks, at least in one of the requirements listed above. SZA has received attention by the expert community, but was never marketed by ISC. It is prepared by the sol–gel process and thus contains inorganic components, triggered for long term stability. Since the development 20 years ago, sol–gel chemistry became more and more popular in materials science, leading to many industrial products on the market based on new approaches in the recipes and new starting compounds. 3. Experimental 3.1. Preparation of damaged paint layers Commercially available black glass paint was employed for the preparation of model samples simulating damaged paint layers (chemical composition determined: PbO 58.8, CoO 18.8 and SiO2 16.8 weight-% as main components). Simulation of powdery and flaky paint types were made as it is explained in Table 2. Three glass types were used as substrates for the paint application: microscopic glass slides (M), with a chemical composition typical for commercially available modern glasses (SiO2 ∼ 72.2, Na2 O ∼ 14.3; CaO ∼ 6.4 and MgO ∼ 4.3 weight-% as main components); model glass MI (MI), a potash-lime glass similar to medieval glasses (SiO2 ∼ 48.0, K2 O ∼ 25.5; CaO ∼ 15.0 and P2 O5 ∼ 4.0 weight-% as main components); and antique glass (A), commercially available for new manufactures of stained glass windows (SiO2 ∼ 72.3, Na2 O ∼ 16.3; CaO ∼ 8.7 and Al2 O3 ∼ 1.2 weight-% as main components).

for this study have been prepared containing an inorganic matrix as main component to obtain a good chemical resistance and stability (tetraethylortosilane, TEOS, as starting compound). Different alkyl alkoxydes have been added to avoid crack formation and improve the consolidants elasticity (3-trimethoxysilyl propyl methacrylate, MEMO; polydimethyl siloxane vinyl terminated, PDMSv and 3-glycidyloxypropyl trimethoxysilane, GLYMO). All of them purchased from Fluka. Absolute ethanol (EtOH), 2-propanol (IPA) and tetrahydrofurane (THF) were employed as solvents. Hydrochloric acid (HCl) was used as catalyst. Among all the recipes prepared in the screening period, the molar ratios of the selected ones are shown in Table 3. Once prepared, the consolidants were applied carefully by leaving some drops of consolidant flow over the paint on the model glasses painted with grisaille and they were dried 50 h at ambient conditions (Laboratory conditions: temperature ∼ 20 ◦ C, relative humidity ∼ 40 %). Further thermal densification treatments were avoided, since future application on originals would not allow thermal stress on the glass. 3.3. Accelerated tests of adhesion and chemical durability In order to prove the suitability of the selected consolidants as compared to the requirements previously mentioned, accelerated tests were performed to simulate extreme stresses. Mechanical resistance and adhesion properties of the consolidants were studied by application of the selected ones on painted glasses in direct comparison with untreated paint (only half of the sample was treated). The samples were submitted to an ultrasonic bath for two minutes, which represents a severe adhesion test. Chemical resistance of the consolidants was explored by immersion of samples in acid and alkaline solutions for four months. Both experiments are extremely aggressive and not directly comparable to atmospheric weathering tests. Nevertheless, they give an indication for the stability of materials to acid environments (acid attack) and to the aggressive treatments with cleaning products, like ammonia or its solutions (alkaline attack). Half of the painted glass samples were consolidated and immersed afterwards into Teflon vessels containing a Table 3 Composition of the selected consolidants (molar ratios of starting compounds). Consolidant M

3.2. Preparation of the new sol–gel consolidants Consolidant S

The sol–gel method allows the introduction of organic molecules inside an inorganic network, forming “hybrids” with a broad variety in their composition and properties. All recipes

Consolidant G

0.5 TEOS: 0.5 MEMO: 4 EtOH: 4 H2 O: 0.05 HCl 0.4 TEOS: 0.1 PDMSv: 0.5 MEMO: 4 IPA: 1 THF: 2 H2 O: 0.05 HCl 0.4 TEOS: 0.1 GLYMO: 0.5 MEMO: 4 EtOH: 4 H2 O: 0.05 HCl

406

N. Carmona et al. / Journal of Cultural Heritage 10 (2009) 403–409

0.1N NaOH solution (pH = 13) or a 0.1N HCl solution (pH = 1) respectively. As previously mentioned, samples remained in the solutions for four months. 3.4. Characterisation methods The morphology of the surfaces was examined by optical microscopy (OM), with a Leika Litz DMRxE microscope. For the SEM pictures and EDX analyses, a Philips scanning electron microscope (XL30 DX4) coupled to an EDAX CDU Leap detector was employed. Metallization of the samples were realised with a Sputter Coater Cressington 108 Auto. Consolidant solution viscosities were measured just after preparation with an Ubbelohde viscosimeter Scott CK100 at 20.0 ± 0.020 ◦ C. Hydrophobicity was investigated by recording the contact angles of thin layers of consolidant applied on glass substrates after curing/densification and evaporation of the solvents. An automatic contact angle system (Dataphysics mod. ACA 50) was employed for the measurements. Drops of 5 ␮l of distilled water were deposited on the coatings surface and the angle formed between the solid surface and the line tangent to the droplet radius from the point of contact with the solid was measured. The quantity of consolidant entering and remaining in the paint has been determined by weight measurements. The weight loss of just consolidated painted glasses has been recorded each 24 h until 300 h after treatment. 4. Results and discussion Two different types of paint on three model glass substrates were employed to simulate all possibilities of stained glass windows from medieval to modern times. Results after accelerated tests showed slight differences for the flaky paint type, which seems to protect the glass underneath against corrosion, even for the potash-lime glass (similar to medieval compositions). Apart from this, there was no noticeable influence between the observed type of decay and the paint or glass substrate employed in the test. Visual appearance, optical transparency, colour, presence of phase separation, etc. were relevant characteristics for the selection of the new consolidants. The final recipes tested in this study (Consolidant M, Consolidant S and Consolidant G, see Table 3) are transparent, colourless and the solutions prepared for application consist of only one component (no primers or hardeners are necessary).

Some other practical aspects and physical properties of the selected consolidants have been summarised in Table 4. The degree of penetration into the porous structure of the paint is a very important property of a consolidant but it is difficult to evaluate. The viscosity of the solution is easy to measure and gives at least an estimate value for the flow into the porous structure. Similar values were obtained for the three new sol–gel recipes. These values are lower as compared with Paraloid B72 (10 weight-% solution in acetylacetone) and SZA. These values indicate that the new consolidants are sufficiently fluent for being easily applied on the paint and to penetrate through it. Wettability of the consolidants is interesting for practical applications, in particular, since hydrophobic materials have the potential to prevent the formation of water droplets in high humid environments. Table 4 also compares the contact angle measurements. The consolidants M, S, G and SZA form a surface where the water drops slide. The most hydrophilic consolidant is Paraloid B 72 with the lowest value. By comparing the solid content after curing, it is possible to estimate the amount of consolidant still present in the paint after drying, keeping the structure together. Some examples have shown that after evaporation of the solvent, the solid amount of consolidant is low, in such a way, that consolidation of the paint itself is physically not possible. The consolidant M lead to a solid content of 30 weight-% which remains in the paint after drying. The highest solid content was registered for the consolidant G (34 weight-%). The chosen concentration for the Paraloid and SZA leads to a 10 weight-% solid deposition. Adhesion between the paint itself and to the glass underneath has been determined by the immersion of the consolidated painted samples in an ultrasound bath. Fig. 2 shows the behaviour of the consolidants applied on the right half-side of the painted model glasses after two minutes treatment. Results show the unconsolidated side (left half-side) with the paint completely lost, and good adhesion for the right half-side for all the consolidants tested. The pot life (the time during which the consolidant is still usable) is important for application. The consolidant G is liquid for only 24 h after preparation, which is a limitation to take into account (Table 4). The consolidant S pot life is about one month, and the consolidant M pot life is more than six months. Results of the chemical durability of the consolidants applied on painted glass samples showed that the pH of the solution was an important factor for the degradation of the paint. Acid conditions have led to the dissolution of the paint, macroscopically

Table 4 Practical aspects and physical properties of the selected consolidants. Viscosity [mm2 /s] Paraloid B72 5.280 SZA 4.350 Consolidant M 3.030 Consolidant S 3.002 Consolidant G 3.266 √ : Yes / good; X: No / bad.

Contact angle [◦ ± 2]

Solid content [weight-%]

Pot life

Resistance to degradation

51.6 73.8 88.8 70.8 71.8

10 10 30 10 34

Several months ∼ 1 month ∼ 6 months ∼ 1 month ∼ 24 hours

X √ √ √ √

Reversibility on non porous substrates √ X X X X

N. Carmona et al. / Journal of Cultural Heritage 10 (2009) 403–409

407

Fig. 2. OM pictures of model glass MI with powdery paint and consolidated (only the right side) with consolidant M (a) initially and (b), after 2 min in an ultrasonic bath; (c) consolidated with consolidant G, (d) with consolidant S, (e) with SZA and (f) with Paraloid B72, each after after 2 minutes adhesion test in an ultrasonic bath.

visible by a white-grey colour of the surface. At microscopic scale, the losses of the paint are evident (Fig. 3a), where only some black points of paint remain still visible in relation with the reference sample not immersed (Fig. 3b). Under alkaline conditions, the surface seems not to be deteriorated (Fig. 3c). From some EDX analyses performed, it was noticed that the PbO has been completely lixiviated from the paint during the acid attack. Degradation mechanism for samples immersed into alkaline solutions showed a lighter decrease of the lead content, while the rest of the elements are relative increased proportionally to the initial content. New sol–gel recipes applied upon painted samples behaved good in both alkaline and acid conditions. Although some paint flakes came off after four months attack in acid conditions, in general, the deterioration of the consolidated areas (Fig. 4a, right half-side) is lower than in the unconsolidated ones (Fig. 4a, left half-side), where the paint is completely lost. Painted samples

showed no visual change or corrosion after four months weathering in alkaline solution. Fig. 4b shows both areas, unconsolidated (left) and consolidated (right), both visually well preserved. The extreme conditions of exposure to acid solution for four months have reduced the adhesion of paint consolidated with Paraloid. This indicates that this material may lose its efficiency under high humid conditions. For SZA treated samples, some cracks were noted after acid exposure, but the adhesion of the paint still remained satisfactory. On more general terms, it shall be noted that the SZA efficiency is dependent on the humidity of the atmosphere during treatment and curing. Furthermore, the solutions have to be prepared freshly and they remain stable for less than one month, which is considered a practical disadvantage. The three new consolidants penetrate well into the porous paint structure, whereas the cohesion of the paint was better preserved after artificial attack in consolidants G and M. They both behaved good, but

408

N. Carmona et al. / Journal of Cultural Heritage 10 (2009) 403–409

Fig. 3. Detail of a flaky paint type glass sample after four months immersion into (a) acid solution, (b) reference sample not immersed and (c) alkaline solution.

Fig. 4. Painted microscopic glass slide consolidated with consolidant G (right side), and painted area unconsolidated (left side), where the paint has flaked off, after four months immersion into (a) acid, and (b) alkaline solution.

consolidant G showed a pot life of only 24 h. The storage time of around six months of consolidant M, and the fact that it has good mechanical and chemical resistance, good penetration, etc. makes it an alternative to the treatment of degraded paint layers on glasses. 5. Conclusions The comparison of traditional (Paraloid B 72), modern (SZA) and newly developed consolidants (consolidant M, S and G), needs to include practical aspects, such as the pot life of the solution as well as material properties, such as the adhesion to paint and the substrate. Three new materials based on sol–gel chemistry have been tested for the consolidation of degraded

paint on historical glasses. The three new consolidants fulfil the main requirements demanded for a paint consolidant. They are transparent and colourless, causing no visual impact on the paint or the glass. Their low viscosities allow a good penetration in the porous structure of the paint. They should be applied and dried at ambient conditions. They improve the adhesion of the paint itself and with the glass substrate. The resistance to extreme conditions during exposure in acid solutions was better for consolidants G and M, both rated excellent compared with other materials. However, consolidant G has some practical limitations (pot life of only 24 h and storage time of around 6 months). Consolidate M is superior in practical considerations (pot life of 6 months) and is rated excellent concerning penetration, adhesion and chemical resistance. Considering the results

N. Carmona et al. / Journal of Cultural Heritage 10 (2009) 403–409

from the three new consolidants described in this study, further research may concentrate on consolidant M, prepared from TEOS, MEMO in Ethanol and water (for details see Table 2). Finally, the new consolidants are not reversible, but re-treatments are not restricted, if necessary, during a following restoration campaign. This study can be considered as the first step in the development of new sol–gel based products for the consolidation of degraded paint on historical stained glass windows. Pilot studies on originals and exposures in natural environments are necessary as next stages to evaluate their potential use in restoration workshops. Acknowledgements The authors want to acknowledge the financial support from the Marie Curie Individual fellowship (EVK4-CT200250007). All experiments described here were carried out at Fraunhofer ISC in Bronnbach. Further research at ISC is conducted under the supervision of Dr. Peter Mottner ([email protected]). References [1] M. Perez y Jorba, G. Tilloca, D. Michel, J.P. Dallas, Quelques aspects du phénomène de corrosion des vitraux anciens des églises franc¸aises, Verres Réfract. 29 (1975) 53–63. [2] R.G. Newton, D. Fuchs, Chemical compositions and weathering of some medieval glasses from York Minster, part 1, Glass Technol. 29 (1988) 43–48. [3] K.J.S. Gillies, A. Cox, Decay of medieval stained glass at York, Canterbury and Carlise. Part 1. Composition of the glass and its weathering products, Glastech. Ber. 61 (1988) 75–84. [4] K.J.S. Gillies, A. Cox, Decay of medieval stained glass at York, Canterbury and Carlise. Part 2. Relationship between the composition of the glass, its durability and the weathering products, Glastech. Ber. 61 (1988) 101–107. [5] R. Jacobi, Die Konservierung alter Glasmalereien des des Kölner Doms, Glastech. Ber. 30–12 (1957) 509–514. [6] M. Gómez-Rascón, Catedral de León, las vidrieras. I Análisis temático, Edilesa, León, 2000. [7] Theophilus, The Diversis Artibus (English traduction), edited by Robert Hendrie, London, 1847. [8] J.M. Bettembourg, Composition et durabilité des grisailles, Sci. Techn. Conserv. Restaur. 2 (1991) 47–54. [9] H. Debitus, Recherche pour une formulation nouvelle de grisailles, Sci. Techn. Conserv. Restaur. 2 (1991) 24–28. [10] H. Römich, D.R. Fuchs, A new comprehensive concept for the conservation of stained glass windows, Bol. Soc. Esp. Ceram. 31-C (1992) 137–141. [11] J.M. Fernández Navarro, Procesos de alteración de vidrieras medievales. Estudio y tratamientos de protección, Mat. Constr. 242–243 (1996) 5–25. [12] H. Römich, Historic glass and its interaction with the environment / Laboratory experiments to simulate corrosion on stained glass windows, in: N. Tennent (Ed.), The Conservation of Glass and Ceramics, 5-14, James & James, London, 1999, pp. 57–65. [13] N. Carmona, M.A. Villegas, J.M. Fernández Navarro, Characterisation of an intermediate decay phenomenon on historical glasses, J. Mat. Sci. 41 (2006) 2339–2346. [14] M. Schreiner, El deterioro de los vidrios pintados medievales. Caracterización analítica del proceso de corrosión y sus consecuencias para los tratamientos de prevención, in: Proceedings of the historical stained glass windows congress. The Getty Conservation Institute, Los Angeles, 1994.

409

[15] R. Newton, Unusual effects of the weathering of ancient glass, in: Grown in Glory, Jarrold & Sons, Norwich, 1982, pp. 73-80. [16] O. Schalm, K. Janssens, J. Caen, Charaterization of the main causes of deterioration of grisaille paint in the 19th century stained-glass windows by J.-B. Capronnier, Spectr. Acta B 58 (2003) 589–607. [17] N. Carmona, M.A. Villegas, J.M. Fernández Navarro, Study of glasses with grisailles from historic stained glass windows of the cathedral of León (Spain), Appl. Surface Sci. 252 (2006) 5936–5945. [18] I. Pallot-Frossard, La conservation et la restauration des grisailles, émaux et jaune d’argent, in: Grisaille, jaune dárgent, sanguine, émail et peinture à froid, Forum pour la Conservation et la Restauration des vitraux, Commission Royale des Monuments, Sites et Fouilles de la Région Wallone, Liège, 1996, pp. 43–51. [19] W. Müller, H. Pouillon, G. Bochynek, H. Mehner, Extreme Dunkelung von Glasmalereien, Glastech. Ber. 59 (1986) 96–102. [20] F. Bornschein, M. Jähn, T. Gla␤, Die Glasmalereien von Charles Crodel im Dom zu Erfurt, Leipzig, Erfurt, 1999. [21] J.M. Bettembourg, Protection des verres de vitraux contre les agents atmosphériques. Étude des films de résines synthétiques, Verres Refract. 30 (1976) 87–91. [22] O. Chiantore, M. Lazzari, Characterization of acrylic resins, Int. Polymer Anal. Charact. 2 (1996) 395–408. [23] J.C. Ferrazzini, Vorteile und Technik der Anwendung von CyanoacrylatMonomer-Klebstoffen zur Sicherung der Malerei von Glasgemälden, Glastechn. Ber. 49 (1976) 264–268. [24] F. Bornschein, H. Forberg, Vitraux médiévaux de la cathédrale d’Erfurt. Détériorations et solutions pour la conservation du Patrimoine, in : Grisaille, jaune dárgent, sanguine, émail et peinture à froid, Forum pour la Conservation et la Restauration des vitraux, Commission Royale des Monuments, Sites et Fouilles de la Région Wallone, Liège, 1996, pp. 169–173. [25] J.R. Carroll, R.M. Bradley, A.I. Kalmikoff, Using waxes and polymers to improve coatings properties, Surf. Coatings Int. 10 (1994) 435–441. [26] R. Bertelmann, H. Marschner, Alternatives au Paraloid B72 pour la fixation des peintures sur verre, in: Grisaille, jaune dárgent, sanguine, émail et peinture à froid, Forum pour la Conservation et la Restauration des vitraux, Commission Royale des Monuments, Sites et Fouilles de la Région Wallone, Liège, 1996, pp. 79–89. [27] S. Davison, A review of adhesives and consolidants used on glass antiquities, in: Adhesives and consolidants, Paris Congress, Int. Inst. Conserv. Hist. Artist. Works (1984) 191–194. [28] S. Davison, Conservation and Restoration of Glass, 2nd edition, Butterworth-Heinemann, Oxford, 2003. [29] H. Römich, D.R. Fuchs, Stained glass conservation: Innovative research and new materials, in: Proceedings of Congreso Internacional de Rehabilitacion del Patrimonio Arquitectonico y Edificacion, Islas Canarias, 1992, pp. 277–280. [30] S.Y. Chang, T.A. Ring, Silica sol-gels from tetraetoxysilane, ethanol and water mixtures, in: Int. congress: The Physics of Non-Crystalline Solids, Ed Pye, Lacourse, Stevens, Taylor & Francis, London, 1992, pp. 228–230. [31] A. Yuasa, H. Inaba, K. Tadanaga, M. Tatsumisago, T. Minami, Preparation of water-repellent coating films by the sol-gel method, in: Proceedings of XVII International Congress on Glass, vol. 4, Glass Surface Sol-Gel glasses, Chinese Ceramic Society, Beijing, 1995, pp. p445–449. [32] M.A. Villegas, Chemical and microstructural characterisation of sol-gel coatings in circonia-silica system, Thin Solid Films 382 (2001) 124–132. [33] M. Pilz, H. Römich, Sol-gel derived coatings for outdoor bronze conservation, J. Sol-Gel Sci. Techn. 8 (1997) 1071–1075. [34] C.A. Grissom, A.E. Charola, A. Boulton, M.F. Mecklenburg, Evaluation over time of an ethyl silicate consolidant applied to ancient lime plaster, Stud. Conserv. 44 (1999) 113–120. [35] M.J. Mosquera, J. Pozo, L. Esquivias, T. Rivas, B. Silva, Application of mercury porosimetry to the study of xerogels used as stone consolidants, J. Non-Crystalline Solids 311 (2002) 185–194. [36] C. Miliani, M.L. Velo-Simpson, G.W. Scherer, Particle-modified consolidants: A study on the effect of particles on sol-gel properties and consolidation effectiveness, J. Cult. Herit. 8 (2007) 1–6.