Skip to navigation – Site map

Aiming at reproducibility in lining canvas paintings

A research on cold lining methods
Antonio Iaccarino Idelson and Valerio Garofalo

Abstracts

This paper describes a process started in 2011 with the use of the “mist lining” technique for a large painting by Titian, conserved in Venice uncontrolled environment. After several experiences with a specific version of the method, Valerio Garofalo’s research thesis helped the standardisation of lining procedures aiming to obtain predictable, reproducible results. This was achieved through the production of a large quantity of mock-up painting linings and samples for peel tests. The bond created with a viscoelastic adhesive appears stronger with high speed testing, if compared with the same at a lower speed, but literature shows high variability of peel speeds. Preliminary tests with different speeds for Plextol B500 allowed to compare results with literature data. The study gives a deeper understanding of the mechanisms involved in adhesion when lining with this method.

Top of page

Full text

The authors wish to thank the editorial board for choosing experienced and competent peer reviewers and the latter for their thorough and careful revision.

General introduction

Traditional practices, underlying criteria and evolution

1Early lining techniques were based on protein adhesives and required heat to extract moisture and complete the setting process. A simple, and widely documented, example is the sturgeon glue lining from the Russian tradition, in which honey is the only additive to the protein adhesive. This is used to modify the stiffness of the protein and affect its relationship with water. According to this method, both original and lining canvas are impregnated and allowed to dry separately. A fresh layer reactivates the adhesive, and setting is achieved with repeated ironing until complete evaporation of water occurs.

2Most techniques in Italian and French tradition (since the 17th century) are based on a different concept: a thickened protein adhesive is applied through an open weave lining canvas, and adhesion is obtained with no need for reactivation. The adhesive is pressed with a wooden tool until an even layer saturates the new canvas. Subsequent ironing removes moisture from the adhesive mixture to complete the adhesion between the two canvases. Thickening animal glues allows a faster and more efficient process, also because it reduces drying shrinkage of the adhesive. Starch or flour mixtures are used as thickeners. Other additives, such as molasses, alum, etc. are also frequently included. The exact recipe and additives used depended on local tradition, that was often passed down from generation to generation.

3The method introduced in 20th century Italy by the Istituto Centrale per il Restauro (ICR), aimed at the definition of a standardised national lining method. Though the task was not achieved, a reduction of the variability of countless local traditions was obtained. It sees the use of a rather thick paste, a gel that rapidly ensures enough adhesion without ironing. Open air drying is enough for complete setting of the adhesive, though the method is seldom used as a cold lining because ironing the moist, thus temporarily plasticized, oil paintings helps reducing cupping and distortions: all traditional lining techniques tend to solve a whole lot of problems within a single process.

  • 1 In the Netherlands, by Nicolas Hopman (1794-1870).

4The use of moisture and heat may cause problems restorers and conservators have always been aware of, such as delamination and distortion of paint layers, or shrinkage of the canvas. Moreover, glue-paste adhesives favour biological growth when 75-80% relative humidity (RH) is exceeded for more than a few days. Wax-resin lining methods were introduced in the early 19th century1 (van Oudheusden, 2014) to avoid such threats and biological growth in unstable environments. Though, the side effects are not negligible, as impregnation with wax and resin alters the refractive index of the canvas, ground and oil paint layers (Froment, E. M., 2019). Additionally, wax-resin lining irrevocably alters the mechanical-physical properties of the painting. The solubility parameters of the individual components are affected and some wax-resin lined paintings are found to become sensitive to fibres swelling, causing shrinking at high relative humidity. Furthermore, the weight of the painting increases more than with other alternative techniques.

5In traditional lining, a degree of reproducibility can be found in the standardised methods but not in the quantification of the materials or of their results. The goal has always been that of “reinforcement of the painting and sufficient adhesion of the new canvas”, and nowadays it is still difficult to find a better definition for lining in general.

Changes in perspective at the time of the 1974 Greenwich Conference on Comparative Lining Techniques

6Awareness of the inevitable risks and damages connected to all traditional lining techniques saw its peak in the Greenwich Conference on Comparative Lining Techniques in 1974 (Villers, 2003). A seminal meeting, in which lining techniques and alternatives to lining were described in deep detail. Problems related to lining were described with awareness and knowledge, and new lining techniques were proposed. Approaches to quantitative description of the materials and their behaviour were introduced. Regarding reproducibility, the most innovative contribution of the conference was probably in the introduction of an engineering approach in the field of structural conservation of canvas paintings (papers from Berger, Boissonians, Fieux, Tassinari). Still, the largest impact on practice was in the new lining techniques proposed by Gustav Berger and Vishva Raj Mehra.

  • 2 As an example, see US patent n. 3,232,895. Feb. 1, 1966: “Adhesive compositions comprising ethylene (...)

7Berger proposed a new adhesive, whose composition is bridging traditional wax-resin and that of formulations from industrial research2, a patented mixture of synthetic materials, BEVA 371.

  • 3 Still, its composition was recently changed, its formulation in EU differs from that in the USA, it (...)

8Data on related research on paintings’ mechanics was widely published, and the fixed ratio of the components increased the standardization3. The material proved to be efficient and allows the solution of different classes of structural problems, such as lining, edge lining, consolidation. BEVA’s great success is in the versatile use of a new wax-resin formulation that reduces the risk of discoloration of paint layers. Moreover, it also allowed to keep working with well-established traditional methods. Though a mixture of materials that are seldom chosen for conservation purposes, it is currently considered as a “new” traditional technique, after more than half a century probably the most used worldwide.

  • 4 Mehra’s cold ling method synthetic materials is derived from the ICR glue paste. It is based on a t (...)

9Mehra, wishing to avoid the use of heat and to reduce weight gain due to lining, proposed a cold lining method derived from a deep rethinking of traditional Italian glue-paste lining4. Using the acrylic emulsion Plextol B500 as adhesive in aqueous dispersion applied only on the lining canvas, he also strongly reduced the amount of moisture reaching the painting. The introduction of the low-pressure table further increased this achievement and allowed a whole range of new treatment possibilities.

10The reduction of moisture and the elimination of heat were considerable innovations, but he also made a point on the idea of looking for separate solutions for each conservation problem, allowing for freer specific thought and innovation. His lining method allows reducing the diffusion of the adhesive into the painting’s canvas, aiming at an adhesion obtained with surface contact: a concept defined “nap-bond lining”. Solutions are proposed to adapt the force of adhesion to the needs of the painting (weight, stiffness, specific mechanical problems), and lining is performed after treating all other issues, as the mere addition of a mechanic support.

  • 5 A problem avoided in heat-based methods, that allow easy local treatment of the adhesive, in which (...)

11On the technical side, elimination of heat implied the problem of a short “open time” of the wet adhesive and the need of finding methods for prolonging it.5 When treating larger paintings, he proposed to reactivate the adhesive after its complete drying on the lining canvas, by spraying liquid solvent (Mehra, 1981). Still, achieving homogeneous adhesion on medium to large surfaces with this method is difficult because solvent volatility competes with the time needed for spraying. Thus, the area that was reactivated first may become dry when spraying is completed, or may achieve a lesser bond. Using high retention or slowly evaporating solvents helps reducing the magnitude of this problem, but the use of low retention solvent with fast evaporation is required for the sake of the painting. Fast setting is useful for handling the process, as pressure is needed to keep the canvases in position until adhesion is completed. Mechanical research on the adhesion obtained was performed and published (Mehra, 1975, pages 23-26).

12Mehra’s methods became less popular than BEVA because they required a change in attitude and more complex approach.

Solvent vapor reactivation in the “mist lining” at the Stichting Restauratie Atelier Limburg (SRAL) Maastricht

13During the 1990s and 2000s, Jos Van Och and his colleagues at the SRAL Maastricht, started the systematic use of a reactivation method based on solvent vapours within an enclosed environment, an important innovation that opened new possibilities. The reactivation process takes place on a low-pressure table or in a low-pressure envelope, used to keep the canvases in close contact while the solvent vapours are extracted. Solvent vapours are generated with an additional canvas, used as a carrier for the solvent. The latter is dampened with a measured quantity of solvent and rapidly unrolled on the reverse of the lining canvas before closing the suction system, left in the envelope for a predetermined time and removed before bond formation.

14Lining canvas is chosen with spun-yarn, moderately open weave so as to favour vapor diffusion, and the adhesive is applied with a spray gun. This contributes to naming the method “mist lining”, because the sprayed adhesive takes the form of an airy deposition of white acrylic adhesive on lining canvas filaments (fig. 1). The effect is enhanced with a surface abrasion of the canvas releasing partially loosened fibers, nap, raised up from the canvas with the help of a vacuum cleaner, that become coated with adhesive and offer an extension of the surface available for adhesion (Van Och, 2003; Seymour, 2005). Such an open and porous structure of the adhesive and lining canvas implies a fast diffusion of solvent vapours.

Fig. 1. Adhesive distribution in “mist lining”

Fig. 1. Adhesive distribution in “mist lining”

The sprayed adhesive takes the form of an airy deposition of white acrylic adhesive on lining canvas filaments.

Credits Van Och, 2003.

  • 6 Both D 360 and D 540 are no longer commercially available. Currently D 512 is under testing for use (...)
  • 7 According to Lascaux datasheet and (HORIE, 2010), page 196, Plextol D360 (or K360) is -8°C.
  • 8 Several mixtures of solvents are used, also depending to the size and other qualities of the painti (...)

15The closed environment in mist lining allows to use a much smaller quantity of solvent if compared to Mehra’s spray methods. In (Phenix, 1984) up to 10 times more solvent is required for the reactivation of a Plextol B500 dry film, mostly because of solvent dispersion in the environment. The reason why the quantity of solvent can be so drastically reduced is not only in the system allowing to use the vapours efficiently, but also in the nature of the adhesive. Instead of B500 alone, a mixture of two acrylic emulsions is used, 70% Plextol K360 (previously D360) and 30% Plextol D541 (previously Plextol D540),6 in which K360 provides adhesion at room temperature, even without reactivation. It’s very low glass transition temperature (Tg) (Tg -8°C),7 is very useful for solvent reduction in the system. On the contrary, Plextol D540 (or D541) is a much stronger molecule that has no tack at room temperature (Tg 29°C), ensuring the stability of the bond. The solvent used is a mixture of ethanol (or isopropanol) and xylene (or Shellsol A) (Seymour, 2012). Aromatic solvents are used to increase adhesion, especially on surfaces that are contaminated with previous wax-resin lining treatments,8 but also because they provide a very efficient reactivation of the adhesive.

  • 9 Authors are authorized to state that Seymour and van Och are currently working on this aspect of th (...)

16Mist lining technique also introduced a higher level of reproducibility, by providing quantitative data on the adhesive and on the solvent mixtures. The exact communication of quantitative information on the treatment allows taking decisions in the awareness of the results described in previous experience. At least until early 2019, no data was published on the strength of the bond obtained nor on the values that would be considered acceptable, because of the need of finding specific solutions for each painting9. As it is normal in fields dominated by a great variability, like that of a paintings’ needs, adjustments in the formulations are based on experience that is hard to quantify.

17Mist lining’s reactivation method allowed the use of Mehra’s principles on medium to large size paintings, implying a major change in perspective for the reduction of thermal stresses.

Adapting the “mist lining” technique for the high temperatures of Mediterranean climate

  • 10 All choices were shared with Gloria Tranquilli, Director of the Conservation Laboratories of Venice (...)
  • 11 Sakase industry, SK802, vacuum sized in epoxy for dimensional stability and adherence between the p (...)
  • 12 Thanks to Matteo Rossi Doria and Olivier Verheyden who made the contact.

18In August 2011 the water used to extinguish a fire damaged the large ceiling painting by Tiziano, “David and Goliath”, in the Basilica della Salute in Venice. The painting had become extremely fragile and needed a new lining. The choice was made10 to use an inextensible support made of a lightweight, high modulus, carbon fiber canvas11 on an elastic stretcher, so as to minimize gravity-related sagging (the painting is displayed horizontally at 13 m from the ground) without stressing the painting. Thanks to a conversation with Jos Van Och at SRAL Maastricht12, the decision was made to try and use the mist lining technique.

  • 13 Plextol B500, Rhoplex AC 33 and AC 234, Lascaux 360 HV and 498 HV.

19The hot summer environment in Venice easily reaches more than 40° C under a ceiling, and this is the temperature used for thermal reactivation of K360. The low glass transition temperature (Tg -8°C) of K360 and its elongation at break of 1000% (Horie, 2010), describing a very yielding polymer, imply that this material should not be used in warm environments. The Plextol B500 used in Mehra’s techniques since the late 1960s, has a Tg of 9°C, a reactivation temperature around 80°C and 600% elongation at break (Horie, 2010), typical of a polymer of medium tenacity. Michael Duffy’s comparative study on the stability of peel strength and yellowing after artificial ageing (Duffy, 1989) showed that, of the five adhesives tested13, the Plextol B500 appears to be the most resistant to peel strength changes over time. Yellowing is a minor concern, since lining adhesives are protected from UV (the major cross-linking agent for all acrylics). The choice to use only B500 alone for lining the David and Goliath seemed therefore to offer sufficient guarantees (Costantini, 2013).

  • 14 With 0,1% w/w Rohagit.

20The thickened14 emulsion was sprayed on the very even surface of the open weave carbon canvas, obtaining a different result from that of the “mist” learned at the SRAL. The absence of raised fibres on the surface produced the deposition of small droplets, that developed a lenticular shape upon drying, finely distributed on the surface and strongly adhering to the canvas. Their small size (less than 0.4 mm in diameter, almost a dust) still allows free space for the passage of the solvent vapours (fig. 2).

Fig. 2. Adhesive distribution on the carbon fibre canvas

Fig. 2. Adhesive distribution on the carbon fibre canvas

The absence of raised fibres on the surface produced the deposition of small droplets, that developed a lenticular shape upon drying, finely distributed on the surface and strongly adhering to the canvas.

Credits: A. Iaccarino Idelson and C. Serino, 2012.

  • 15 According to the availability, a dry “rotary vane pump” or a regular oil pump have been used for th (...)
  • 16 For the specific painting 0,6 bar. It should be noted that the well-known surface texture changes d (...)

21The adhesive’s different chemical nature and shape, required rethinking also for the solvent. The goal being an efficient reactivation with a low retention and adequate evaporation speed. Methyl Ethyl Ketone (MEK) is a good solvent for acrylics and has a relatively slower evaporation if compared to Acetone, what seemed to offer more guarantees regarding homogeneous reactivation on a 9 m2 painting. A vacuum envelope was used instead of the low-pressure system, measuring the force with a vacuum meter on the inlet of the vacuum pump15 and setting it on a value deemed to be safe for the painting.16 A completely sealed bag was preferred, because it offered hydrostatic (thus very even) distribution of the pressure on a large painting. It also seemed to allow keeping the solvent vapours in place for a longer time.

22Preliminary tests were done to obtain a reliable bond that could still be reversed by hand-peeling lining canvas without solvent reactivation. The main variables were the quantities of adhesive and solvent, as the setting time was determined by that of solvent evaporation. The choice was for 110 g/m2 of dry adhesive and 180 g/m2 of solvent. The method seemed therefore to require more materials if compared with mist lining, in which typical quantities are described as approx. 40-70 g/m2 of adhesive (Van Och, 2003) and 60 g/m2 of solvent (Seymour 2012). This seems to be due to the absence of the room-temperature adhesion provided by the K360 and to the higher efficacy of the solvent mixtures for reactivation. Nevertheless, the horizontal position of the painting in a high and inaccessible ceiling also suggested that a relatively strong bond was required (Iaccarino Idelson, 2012).

  • 17 A temporary opening of the bag is not as easy as for a low-pressure envelope.
  • 18 Evaporation is forced since the beginning because of the reduction of atmospheric pressure by 0.6 b (...)

23After this challenging first experience, the method was used again for lining paintings that did not have such a strong need for an inextensible carbon fiber canvas, or could not afford to have one. As lining canvas, the preference has been for a linen open weave called “pattina” (9/9 threads/cm; 150 g/m2), a robust and relatively inextensible material. But finer and tighter weave canvases were also used under specific requirements. The linen canvas is sized with diluted B500 (1:3 w/w in water), in order to improve adhesion of the sprayed glue to the lining canvas, and also to increase canvas cohesion, dimensional stability and elastic modulus. Reactivation time is 30 minutes, followed by a variable drying -setting- time in the vacuum envelope, depending on the quantity of solvent used for the specific case. Pressure is kept during the entire process, at around 0.6 bar. The reactivation fabric is not removed as it is in the mist lining, but kept in the vacuum bag until the end of the treatment17. After 30 mins pressure is decreased, allowing clean air to flow into the vacuum bag from holes opposite to the suction pipe18. As air circulates through the “bleeder”, the process allows almost complete removal of the solvent from the system during the following 30-40 minutes.

24Table 1 compares the procedures successfully used after the David and Goliath. As previously said, no quantification of the effects of the procedures is possible on real paintings, unless partial de-lining is carried out for testing purposes. In such a context, the only information available is if the bond appears to be strong enough for future conservation of the painting. Still, this does not allow learning from mistakes as it is rather easy to provide more adhesion than necessary.

25The need for deeper understanding became urgent, in order to obtain some kind of prevision on the bond obtained using different combinations of the variables.

Table 1. Examples of variables used lining paintings differing for age and dimensions

Table 1. Examples of variables used lining paintings differing for age and dimensions

Credits: Antonio Iaccarino Idelson

Aiming at reproducibility

26Valerio Garofalo’s research work for the five-year master’s thesis in Paintings Conservation was centred on the effect of the single variables in this specific variant of cold lining, aiming at the description of the influence of the quantities of adhesive and solvent on the force of adhesion. Research was designed for the treatment of a large format canvas (38 m2), the “Last supper” by G. C. Procaccini (from the church of the Santissima Annunziata del Vastato in Genoa, Italy). The experimental part was undertaken within the course “Conservazione e Restauro dei Beni Culturali” at the “Centro Conservazione e Restauro La Venaria Reale” in collaboration with the “Laboratorio Alta Qualità del Dipartimento di Meccanica Aeronautica e Spaziale” of the Polytechnic University of Torino, under the direction of Antonio Iaccarino Idelson.

  • 19 Another relevant literature reference for peel tests on Plextol B500 is Alain Roche’s work (Roche, (...)

27In the early 1970s, V. R. Mehra performed comparative peel tests on several lining methods, stating that “quantifying such aspects will enhance further improvement in the practice of lining and the approach to it, which should become even more gentle” (Mehra, 1975). Though, literature sees as the main reference19 the work of Gerry Hedley and Alan Phenix (Phenix,1984) who investigated the relationship between the quantity of solvent used for reactivation and the force of adhesion in a dry-film lining, proposing a tentative definition of correct minimum value of force in 200 N/m for peel strength. Mehra’s method, used by Phenix and Hedley for their tests, required either brushing or spraying solvent for the reactivation of the dry-film. Both methods implied high dispersion in the environment of the -relatively fast evaporating- solvent, thus a rough and approximated quantification was inherent and inevitable. Their results (see graph in fig. 3) indicate that a large amount of solvent is needed to reactivate B500, a continuous film that could only be reached from the painting’s side.

28Peel tests were therefore chosen, as the test that allowed comparison with previous research and also for the practical reason that it describes the action performed for de-lining a painting.

29Solvent interaction with paint films is undesired and dangerous, as it may cause swelling, solubilisation and plasticisation (Phenix, 1998). Upon drying, the effects are reversed, but leaching and extraction of volatile parts from the binder are a well-known risk (Tumosa, 1999; Sutherland, 2010), especially with liquid solvent. Finally, the dispersion of solvent in the working environment implies health hazards and safety issues for the conservators. As seen above, the sealed environment reduces the solvent needed, as evaporation is part of the reactivation process, and allows a more precise quantification.

Fig. 3. Weight of liquid solvent needed for reactivation vs. force of adhesion

Fig. 3. Weight of liquid solvent needed for reactivation vs. force of adhesion

Credits: Valerio Garofalo

Experimental

Definition of the lining procedures

30The following variables influence the force of adhesion during lining:

  • quantity of adhesive

  • quantity of solvent

  • reactivation time (time of permanence of the solvent vapours)

  • setting time (time for solvent removal from the adhesive)

  • pressure applied during the process of adhesion.

31In order to reduce the number of variables and obtain an in-depth study of the ratio between the quantities of adhesive and solvent, a “standardised” method was adopted from the previous practice. The reactivation time was fixed in 25 minutes; setting time in 45 minutes (with an exception for the samples treated with a very limited amount of solvent, as complete drying occurred already after 20 minutes); pressure in 0.3 bars.

  • 20 The quantity of dry residue is calculated on the base of the solids in the dispersion, as the weigh (...)

32The quantities of pulverised adhesive (dry residue on the lining canvas20) were limited to 60 g/m2 or 100 g/m2 and the quantities of solvent to 50 g/m2, 100 g/m2 or 150 g/m2.

33In order to create an easy and univocal naming of the samples, the weight of Plextol B500 was indicated as “P” and the weight of solvent as “S” in the following abbreviations. Therefore, P100S150 stands for: 100 g/m2 dry residue of Plextol B500 sprayed on the lining canvas, reactivated with 150 g/m2 of Methyl Ethyl Ketone.

34Volatility of the solvent, and its use in relatively small quantities, suggested to produce large mock-ups (100x100 cm) in order to reduce the influence of accidental differences on the results. Each class was produced in 3 specimens (18 m2 in total), out of which 5 samples were extracted, for a total of 15 samples for each of the 6 classes (90 samples).

35Another important variable is the presence of substances, on the original canvas’ surface, that are also reactivated with the solvent vapours. Among others, acrylic or PVA adhesives, used for consolidation purposes, will increase the adhesion force.

36As the large (38 m2) and heavy painting studied for the thesis, later described in this paper, had been impregnated with a layer of PVA emulsion, samples were prepared to investigate its influence on the lining force.

Table 2. The lining conditions for the mock-ups

Table 2. The lining conditions for the mock-ups

Credits: Valerio Garofalo

The making of the mock-ups

  • 21 Manifatture tessili Sironi, Gallarate, Italy. Canvas named DM 18.

37The painting mock-ups were produced using a linen canvas21 with a thread count of 12x12/cm2. The canvas was washed, stretched on a strainer, wet and restretched three times before the application of a rabbit skin glue sizing (1/14 in weight in water, with an average dry residue of 33 g/m2), which was also used for binding the gesso ground (1/1,2 in weight -glue/gesso- with an average dry residue of 118 g/m2).

  • 22 For their preparation, the canvas was chosen for being as similar as possible to the original. A 30 (...)

38The mock-ups prepared for testing the influence of the PVA on the painting’s canvas22 were lined with only one quantity of adhesive (100 g/m2 - P100-). The quantities of solvent were 70-100-120 g/m2; the goal was set on the finer investigation of the medium range of solvent.

  • 23 Rohagit S hv, Evonik GmbH, Darmstad, DE. An acrylic acid, soluble in alcaline environment, that all (...)

39As lining canvas, the previously mentioned pure-linen “pattina” (9/9 threads/cm; 150 g/m2) was used, with the B500 sizing layer (1:3 w/w in water). The low density of the canvas contributes to its dimensional stability in case of extreme fluctuations of the environment (Rossi Doria, 2003), therefore allowing to avoid washing and repeated wetting on strainer. This was regarded as particularly useful in consideration of the large dimensions of the painting object of the dissertation, but also for the paintings in Tab.1. Each piece of lining canvas was stretched and sized with diluted B500 (1:3 w/w in water, approx. 20 g/m2 of dry residue). The lining adhesive was thickened with 0,1% in weight of Rohagit S HV23 and sprayed on the canvas surface using a 2.4 mm nozzle and 2.5 bar air pressure. The even distribution of the adhesive was assured with continuous movements on the entire surface of the canvas, always requiring several passages to obtain the desired deposition (fig. 4). The quantity of adhesive is measured weighing the adhesive poured into the spray gun’s tank, and also counterchecking with the difference of weight of the tank before and after the application.

40Just like in previous experiences (see fig. 2) the deposition of the adhesive builds up a dense network of droplets, closely adhering to the canvas surface and its few raised fibres, but leaving most of the space in the weave open for the passage of solvent vapours. Comparison between the pictures in fig. 1 and 4 shows the considerable difference in pattern for the adhesive distribution.

Fig. 4. Adhesive distribution

Fig. 4. Adhesive distribution

The deposition of the adhesive builds up a dense network of droplets, closely adhering to the canvas surface and its few raised fibres, but leaving most of the space in the weave open for the passage of solvent vapours.

Credits: A. Iaccarino Idelson

The lining protocols

  • 24 The adhesive let dry in the open air for a minimum of 24 hours.
  • 25 “Peel ply” 17x17 threads/cm, normally used as a separation layer in epoxy resin laminations.
  • 26 Three different cotton fabrics were used for the purpose, each thick enough to carry the desired am (...)
  • 27 Polyester non-woven “bleeder”, 150 g/m2.

41The painting mock-up and its lining canvas are aligned so as to perform the peel tests always in the weft direction for both. On the lining table, protected with silicon mylar, the painting mock-up was placed face down. The lining canvas24 was placed on top of it. On the back of the lining canvas, a thin nylon fabric25 avoids direct contact with the reactivation fabric, letting solvent vapours get thorough. The reactivation fabric26, imbibed with the desired amount of solvent until it is still unable to release it in liquid form, is unrolled on the nylon fabric just like in the “mist lining”. The thick polyester non-woven bleeder27 completes the sandwich and favours air flow in the vacuum envelope and the even distribution of pressure (fig. 5). The suction pipe from the vacuum pump is placed on the perimeter of the envelope. Two pressure gauges are placed on the opposite side. These are used only after the completion of the reactivation process, in order to allow access of clean air during the last phase, to remove the solvent with a constant value of pressure (fig.6).

Fig. 5. Layer build-up for lining

Fig. 5. Layer build-up for lining

From the painting mock-up on the bottom, to the white polyester bleeder

Credits: Valerio Garofalo

Fig. 6. Lining the mock-ups

Fig. 6. Lining the mock-ups

The suction pipe from the vacuum pump is placed on the perimeter of the envelope. Two pressure gauges are placed on the opposite side. These are used only after the completion of the reactivation process, in order to allow access of clean air during the last phase, to remove the solvent with a constant value of pressure.

Credits: Valerio Garofalo

The influence of speed in peel tests on the viscoelastic adhesive

  • 28 The graph is characterised by high and low peaks.

42Peel tests measure the force needed to separate two laminated materials, pulling them apart with an angle of 180°, under a constant rate of travel of the power-actuated grip of the testing machine. The “peel strength” is defined by the average force recorded28 during the test (force measured in N) related with the width of the specimen in order to obtain the value for a chosen unit of length (N/m).

43Testing high modulus adhesives, the rate of travel (speed) does not affect significantly the results, but with medium to low modulus adhesives (like acrylics and BEVA), speed involves an interaction with their viscoelastic behaviour. As a matter of fact, slow motion causes their plastic deformation before they reach the ultimate strength, and this requires less force for separation than a fast movement, which causes instead a sudden failure of the polymer.

  • 29 Gustav Berger’s tests on rabbit skin glue showed almost identical behaviour with different speeds ( (...)

44Data available in literature often describe de-lining with peel tests based on ASTM D903-49, but with a relatively free choice of testing speeds.29 Differences in delamination speed for viscoelastic adhesives make direct comparison of the information. Mehra (Mehra, 1975) mentions peel tests for B500 lining with the speed of 305 mm/min. Later authors, Kennet Katz and Alan Phenix, tested B500 delamination at 2 mm/min; Juliet Hawker used both 2mm/min and 100 mm/min; Alain Roche used 100 mm/min (Katz 1985, Phenix 1985, Hawker 1987, Roche 2003).

45In order to compare test results with the previous work done in the field, it seemed necessary to better investigate the influence of speed on the peel force and determine the possibility of finding a linear conversion factor.

Preliminary testing for peel speed

46Preliminary tests were performed on sets of identical mock-up samples (lined with P100S150), with four different speeds (2, 30, 100, 3000 mm/min) in order to obtain a broad description of the correlation. Tests were performed according to the same ASTM D903-49, describing 180° delamination protocol, with a total run of 150 mm. Only the middle 130 mm of the run were used for the interpretation of data, cutting the head and the tail of the test in order to reduce experimental error as these are often disturbed areas in the graph (see the graph in fig. 7). The testing machine is an INSTRON (Model 8516) with a load cell measuring up to 2 KN, with an accuracy of 2mV/V, 25 Hz sampling rate.

47Measurements allowed to draw the results (average force and standard deviation) summarized in table n.3. Experimental data show the relevant increase of the force needed for delamination associated with speed. Reduction of the standard deviation with high speeds suggests that the phenomena connected with plastic deformation of the adhesive are less predictable.

48Speeds below 100 mm/min were deemed to be too slow for the description of the actual process of de-lining a painting. Such value corresponds to approx. 1,7 mm/sec, which is a speed that can be considered as describing a gentle manual separation of the lining canvas from the original.

49In order to allow comparison with the peel test data for Plextol B500 performed at the other two speeds from literature (that is, mainly 2 and 30 mm/min), a tentative linear conversion factor was calculated with a simple ratio as in table 4.

Fig. 7. The force needed for delamination with different peel speeds

Fig. 7. The force needed for delamination with different peel speeds

The force needed for delamination with different peel speeds.

Credits: Valerio Garofalo

Table 3. Results of the preliminary tests for the influence of speed on the force needed for delamination

Table 3. Results of the preliminary tests for the influence of speed on the force needed for delamination

Credits: Valerio Garofalo

Table 4. Ratio between the force obtained at 100 mm/min vs. 2 mm/min and 30 mm/min

Table 4. Ratio between the force obtained at 100 mm/min vs. 2 mm/min and 30 mm/min

Credits: Valerio Garofalo

Test results on the lined samples

  • 30 The relatively small difference (approx. 10%) from the results obtained with 30 mm/min, also allowe (...)

50Tests on lined mock-ups were therefore performed at the speed of 100 mm/min30. Five samples were extracted from the central area of each of the three mock-up paintings lined with each of the six combinations of adhesive and solvent, for a total of 90 samples. They all had both canvases in weft direction, and measure 25 mm in width. The mock-ups for the influence of the PVA were lined with a single value of adhesive (M_P100), resulting in three series of five samples. For each test the average force was calculated as described for the speed tests. Data can be read in table 5.

Table 5. Average values of peel strength and standard deviation for all test

Table 5. Average values of peel strength and standard deviation for all test

Credits: Valerio Garofalo

Interpretation of experimental data

51Peel strength increases, as expected, when larger quantities of adhesive and solvent are used: the graph of the data obtained with the 6 series is in fig. 8.

52An unexpected behaviour was found for the minimum value of solvent, giving almost identical force with the different quantities of adhesive. Statistical evidence given by the high number of tests provides reliability to the information, therefore it can be stated that with a low quantity of solvent peel strength is less dependent on the quantity of adhesive. This suggests that adhesion happens mostly on the resin’s surface, reducing the influence of its bottom layers. Under these conditions, changing the adhesive weight (i.e. reducing it) may therefore not mean a change of the force of adhesion.

53The standard deviation values are lower in the P60- series than in the P100- suggesting that with less adhesive interactions are less complex. On the other side, in the P100- series (richer in adhesive), the peel strength increases with a linear behaviour, allowing an interesting predictability of intermediate combinations of solvent and adhesive quantities.

54Most probably, more adhesive provides a linear behaviour through a redundancy of connections. Reactivation with larger quantities of solvent vapour increases adhesion through a stronger cohesion between the particles of the acrylic emulsion pulverised on the surface of the lining canvas, and not only because of higher superficial adhesion.

Fig. 8. Peel strength with increasing quantities of solvent

Fig. 8. Peel strength with increasing quantities of solvent

Peel strength with increasing quantities of solvent.

Credits: Valerio Garofalo

  • 31 The canvas was impregnated with two brush applications of a 5% w/w dilution of the mixture of PVA e (...)

55The values obtained with PVA emulsion on the painting’s canvas are considerably higher, despite its quantity31: as expected, PVA was also reactivated, providing a stronger bond. See the graph in fig. 9 for the 6 series with 100 g/m2 of adhesive on the lining canvas.

Fig. 9. Peel strength with and without PVA on the painting’s canvas

Fig. 9. Peel strength with and without PVA on the painting’s canvas

Peel strength with and without PVA on the painting’s canvas.

Credits: Valerio Garofalo

56Further investigation will hopefully allow deeper understanding of this complex subject, but this research seems to add reliable data for conservation choices, as we will see.

Correlation between peel strength and conservation needs

57When coming to the evaluation of peel strength, connection with previous research becomes even more important, as the question is very much one of experience and shared knowledge. Within this framework, the definition of a force of adhesion as “correct”, or at least adequate for a specific painting, implies considerations on its condition and on the cohesion of the original canvas, in relation with the stress due to mechanical removal of the lining. Still, the information is hardly relevant for the real practice because very seldom de-lining is performed without some kind of softening of the lining adhesive: BEVA and wax-resin linings are removed with the use of heat, those with paste-glue after humidification, acrylics are at least partially reactivated with solvent. The peel strength is therefore mostly an indicator of the reliability of the adhesion in carrying the weight of the painting. The threshold value should be considered as the description of the ability of the joined areas to avoid spontaneous delamination due to the painting’s weight and mechanical or environmental stresses.

58Test results were compared with manual de-lining trials, so as to obtain a hands-on understanding of the information from the point of view of the conservator. The descriptions in table 6 are for the hand peel off of the lining canvas with no reactivation. It is important to remember that a relatively short time (approx. 5 mins) reactivation of the adhesive with solvent vapours allows de-lining with hardly any mechanical stress on the painting.

Table 6. Hand de-lining with no reactivation of the adhesive

Table 6. Hand de-lining with no reactivation of the adhesive

Credits: Valerio Garofalo

  • 32 In the paper the data appears as 300 g for 2,5 cm wide strips, with the peel speed of 2 mm/min. The (...)
  • 33 Authors propose the forces of 800 and 1000 g for 2,5cm wide strips, at the speed of 2mm/min.
  • 34 The difference between the forces obtained with increasing speeds has a logarithmic pattern, theref (...)

Using the linear conversion factor calculated with the preliminary investigation on the influence of speed, it is possible to compare experimental data with that found in literature. Phenix and Hedley (Phenix,1984) suggest that below 202 N/m32, the adhesion is such that the painting may spontaneously delaminate from the lining canvas because of in-plane movements connected with environmental changes. They consider values between 543 and 678 N/m33 as reliable but still not too strong for manual de-lining. The speed in their peel test was set at 2 mm/min, therefore the conversion factor used is 1,73. Daly Hartin (Daly, 1993) considers 108 N/m unreliable, defining a force around 540 N/m strong enough and 972 N/m too strong. The speed in her peel test was set at 40 mm/min, therefore the conversion factor used is 1,08 (as found for 30 mm/min34). Alain Roche defines a force 500 N/m as the low limit value for a reliable adhesion (Roche, 2003), at the speed of 100 mm/min. The judgement of these previous authors on the values obtained in the present research is reported in the graph in fig. 10.

Fig. 10. Peel strength in the judgement of previous authors

Fig. 10. Peel strength in the judgement of previous authors

Credits: Valerio Garofalo

Description of the residues on the original canvas after mechanical separation

59Some residues of B500 are left on the reverse of the original canvas after mechanical delamination. Digital microscopy shows that those particles have no penetration in the original canvas. Adhesion happens on the surface. It also shows that adhesive failure (in the form of the transfer of poorly bound particles from the lining canvas to the original) is the most frequent with 50 and 100 g/m2 of solvent. With 150 g/m2 of solvent, cohesive failure (failure of the adhesive when strongly bound to both canvases) reaches a relevant distribution because of a stronger bond on the original canvas.

  • 35 The granules needed to be free of dust, as small particles may interact with the canvas and stay in (...)

60The quantity of residues is directly correlated with the quantity of solvent used for reactivation during the lining process, but naked eye evaluation of their quantity and distribution on the surface is difficult, as they are almost completely transparent. Colouring or fluorescing additives may migrate during the lining process giving controversial results. An easy solution was found in dispersing dark granules on the canvas after a new reactivation of the acrylic residues with solvent, what makes granules stick to their surface35 (see fig. 11).

Fig. 11 Residues of adhesive

Fig. 11 Residues of adhesive

Residues on the painting canvas after the peel tests made visible with dark granules

Credits: Valerio Garofalo

61Residues are described as follows:

  • 50 g/m2: in both series, almost no residues.

  • 100 g/m2: in both series, quantity of residues starts being relevant; even though experimental data shows a difference in the peel force, the quantity of residues is similar.

  • 150 g/m2: in both series, considerable quantity of residues.

Overall description of the results

62The two series provide a minimum peel force value that is considered to be above the safety minimum but below the value judged as optimal, when 50 g/m2 of solvent are used for reactivation. Manual delamination of the samples of both series is easy and not dangerous for fragile paintings, residues on the original canvas are negligible.

63With the intermediate quantity of solvent (100 g/m2), P60 and P100 show a different behaviour, as the first is just above the value considered as a minimum suggested by previous literature (500 N/m), and the second reaches what is described as the maximum desirable (972 N/m). Residues start being clearly perceivable in both cases.

64With the maximum quantity of solvent, for P60S150 the force needed for peeling is very similar to that needed for P100S100 (the maximum desirable in previous research: 972 N/m), while P100S150 exceeds that value by 40%. A considerable quantity of residues is left on the painting.

65Interpolation allows to determine with an apparently reasonable approximation the quantities of adhesive and solvent for intermediate values. For example, with 100 g/m2 adhesive and 70 g/m2 solvent the recommended minimum value (500 N/m) could be attained.

66Similar interpolations may be done for the quantity of residues, but this would be impossible in the presence of a PVA, acrylic or similar adhesives on the original canvas. As a matter of fact, any solvent-reactivated substance impregnating the painting’s canvas changes the pattern and mechanisms of adhesion. Forces become higher, recommending in all cases the use of solvent vapours, so as lining canvas can be removed when the adhesive is softened.

Conclusions

67Most conservation treatments require, and are based on, personal experience, what often makes choices difficult to share. Lining canvas paintings in particular, could be defined using quantifiable information. Since the late 1960es (Urbani, 1969), it seems possible to use such information as the base of shared, comprehensible, choices. The work in this paper shows that reproducibility of the results becomes an attainable goal, when working within a relatively standardised procedure.

68Peel tests allow to describe in a quantitative way the effect of the ratio of the materials used. Of course, the mock-ups are not real paintings and their rear surface is described in a very constant way by clean linen threads with specific patterns. Still, a general understanding requires a certain degree of simplification and data should never be taken as a reliable raw material, but as a reference information that needs to be adapted to a specific context, if possible.

69Nevertheless, graphs allow a reliable degree of prevision of the pattern force of adhesion obtained with untested combinations of adhesive and solvent quantities thorough extrapolation between tested couples.

70Very small quantities of solvent allow only a superficial adhesion. Increasing solvent causes larger interaction at the interface, but also a stronger bond between the particles that constitute the mass of the adhesive on the lining canvas.

71Permanence of residues on the original canvas after mechanical removal is connected to the quantity of solvent used for reactivation. It seems hard to avoid residues when a particularly strong adhesion is necessary. For the paintings’ safety, the limiting factor is the quantity of solvent, which may interact with varnish or paint layers. Tests suggest that using more adhesive may help reduce the solvent, as the two factors affect adhesion independently. Still, further research is being carried out on the use of different solvents that may have less interaction with specific paintings’ materials.

  • 36 The work was carried out in spring 2018 by Equilibrarte srl, for the conservation company Recoop, i (...)

72The research and experience related with Valerio Garofalo’s experimental work (concluded in 2016) could be used shortly after, for the lining of a large painting of great relevance: The Martyrdom of St. Laurent in the church of St. Laurent in Birgu, Malta, (31 m2) by Mattia Preti36. Its large dimensions, 6,7 x 4,6 m, its weight, the complexity of access to the painting and of maintenance, in a highly unstable environment, suggested to stay on the safe side. A combination of 120 g/m2 of adhesive and 180 g/m2 of solvent privileged reliability of adhesion over mechanical reversibility, implying the need of solvent vapours on the lining canvas in order to obtain a gentle de-lining process in future treatments.

73Research is currently being carried out on a different method of application of the adhesive. A layer of the same thickened Plextol B500 can easily be applied on the lining canvas with a paint roll, obtaining a homogeneous deposition and leaving the spaces between the canvas threads open for the passage of solvent vapours. Comparative peel tests have been performed and will soon be made available.

Top of page

Bibliography

Berger, G. A. 1966. Weave Interference in Vacuum Lining of Pictures. In: Studies in Conservation, 11 (4): 170.

Berger, G. A. 2003. “Lining a Torn Painting with BEVA 371”. In: Lining Paintings, paper from Greenwich conference on comparative lining techniques (Villers C., ed), Archetype Publication, London :49-62.

Boissonians, P. 2003. “Comparison of dimensional stability between woven glass fibre fabric and conventional linen canvas as supports for paintings”. In: Lining Paintings, paper from Greenwich conference on comparative lining techniques (Villers C., ed) Archetype Publication, London: 37-37.

Daly Hartin, D., et al. 1993. “Ongoing Research in the CCI Lining Project: Peel Testing of BEVA 371 and Wax-Resin Adhesives with Different Lining support”. In: Preprints of the 10th Triennial Meeting ICOM - Committee for Conservation, Washington. International Council of Museum: 128-134.

De Ruggeri, M. B., Cardinali, M., Ghia, G. S., Iaccarino Idelson, A., Leone, G., Serino, C. 2013. “Carlo Saraceni e la tela di san Carlo Borromeo in San Lorenzo in Lucina, analisi e recupero di un testo pittorico”. In: Kermes, luglio-settembre, Nardini Editore, Firenze: 47-64.

Duffy, M. 1989. “A study of acrylic dispersions used in the treatment of paintings”. In: Journal of the American Institute for Conservation, 28 (2): 67-77.

Fieux, R.E. 2003. “Consolidation and Lining Adhesive Compared”. In: Lining Paintings, paper from Greenwich conference on comparative lining techniques (Villers C., ed) Archetype Publication, London: 37-37.

Froment, E.M., 2019. The Consequences of Wax-Resin Linings for the Present Appearance and Conservation of Seventeenth Century Netherlandish Paintings On Canvas. PhD thesis Faculty of Humanities, Amsterdam School for Heritage and Memory Studies, University of Amsterdam.

Hawker, J. J. 1987. “The bond strength of two hot table lining adhesive: BEVA 371 and Plextol D360”. In: Preprints of the 8th Triennial Meeting ICOM - Committee for Conservation, Sydney. International Council of Museum, 1:161-168.

Horie, V. 2010. Materials for conservation, second edition, Elsevier Ltd, Oxford UK.

Iaccarino Idelson, A., Serino, C. 2012. “L’intervento strutturale”. In: La sfida di Davide e Golia, un capolavoro di Tiziano restaurato. Marcianum press, Venezia.

Iaccarino Idelson, A., Serino, C. 2014. “L’intervento strutturale”. In: Lungo la via degli Abruzzi, un restauro per Pestocostanzo, Massimo Stanzione e Giambattista Gamba nella Chiesa di Gesù e Maria. (Verderame Progetto Cultura), Etgraphiae: 64-72.

Katz, K. 1985. “The quantitative testing and comparisons of peel and lap/shear for Lascaux 360 H.V. and Beva 371”. In: Journal of the American Institute for Conservation, 24 (2): 60-68.

Mehra, V. R. 1975. “Further developments in cold lining (nap bond system)”. In: Preprints of the 4th Triennial Meeting ICOM - Committee for Conservation, Venice. International Council of Museum.

Mehra, V. R. 1981. “The cold lining of paintings”. In: The Conservator, 5: 12-14.

Phenix, A., Hedley, G. 1984. “Lining without heat or moisture”. In: Preprints of the 7th Triennial Meeting ICOM - committee for conservation, Copenhagen 10-14 September: 40-44.

Phenix, A,1998. “Solvent-inducted swelling of paint films: some preliminary results”. In: WAAC Newsletter, September 1998, Western Association for Art Conservation, 20 (3).

Roche, A. 2003. « Approche du principe de réversibilité des doublages des peintures sur toile ». In: Studies in Conservation, 48 ( 2): 83-94.

Rossi Doria, M. 2003. « Diversificazione delle metodologie nel trattamento dei grandi formati ». In: Atti del congresso IGIIC Lo Stato dell’Arte, Torino.

Seymour, K., Van Och, J. 2005. “A cold lining technique for large-scale paintings”. In: Big Picture, Archetype Publication Ltd, London: 96-104.

Seymour, K., Van Och, J. 2012. “De mystifying Mist Lining an introduction”. In: The Picture Restorer, Spring 2012, British Association of Paintings Conservator-Restorer (BAPCR).

Sutherland, K. 2013. “Solvent leaching effects on aged oil paint film”. In: New Insights into the Cleaning of Paintings, Proceedings from the Cleaning 2010 International Conference Universidad Politécnica de Valencia and Museum Conservation. Istitute (Mecklemburg M.F, Charola E.A, Koestler R.J, edit), Washington D.C., Smithsonian Istitution Scholarty Press.

Tassinari, E. 2003. “Caracterisation of lining canvases”. In: Lining Paintings, paper from Greenwich conference on comparative lining techniques (Villers C., ed) Archetype Publication, London: 96-100.

Tumosa, C. S., Millard, J., Erhardt, D., Mecklenburg, M. F. 1999. “Effects of solvents on the physical properties of paint films”. In: 12th Triennial Meeting Lyon 29 August- 3 September 1999 ICOM Committee for conservation, James & James, London, I : 347-352,.

Urbani, G. 1969. « Propositions pour un programme de recherche sur la conservation des peintures sur toile ». In: Plenary meeting Amsterdam 14-19 September ICOM - committee for conservation. International Council of Museum, Paris.

Van Och, J., Hoppenbrouwers, R., 2003. “Mist lining and low-pressure envelope an alternative lining method for the reinforcement of canvas paintings”. In: Zeitschrift für Kunsttechnologie und Konservierung: ZKK, Jahrgang 17, Heft 1: 116-128.

Villers C. ed. 2003. Lining Paintings, paper from Greenwich conference on comparative lining techniques, Archetype Publications, London.

Electronic reference

Costantini, D. “Cold Lining and Mist Lining: insights and possibilities of adaptation to the Mediterranean climate”. CeROArt [Online] EGG 3, 2013, connection on 05 November 2014, http://journals.openedition.org/ceroart/3090

van Oudheusden, S., “The procedure of wax-resin linings by the painting restorers Johannes Albertus Hesterman (1848-1916) and sons” , CeROArt [Online], EGG 4 | 2014, Online since 18 March 2014, connection on 10 July 2019. URL : http://journals.openedition.org/ceroart/4081

Top of page

Notes

1 In the Netherlands, by Nicolas Hopman (1794-1870).

2 As an example, see US patent n. 3,232,895. Feb. 1, 1966: “Adhesive compositions comprising ethylene/vinyl acetate, chlorinated paraffin, and rosin”.

3 Still, its composition was recently changed, its formulation in EU differs from that in the USA, its reversibility and stability are often questioned.

4 Mehra’s cold ling method synthetic materials is derived from the ICR glue paste. It is based on a thickened acrylic emulsion (Plextol B500) evenly distributed on the lining canvas through a mesh. The mesh plays the role of the traditional open weave canvases but it has only a temporary use, allowing the regulation of the quantity of the adhesive. Vishwa Raj Mehra recalled the influence of ICR and that of Giovannni Urbani on his approach to structural conservation of canvas paintings when he started a research on cold lining methods at the Central Research Laboratory of Art, Object and Science in Amsterdam in 1968. At an international level, the role played by ICR during those years was very relevant. The work done in collaboration with Conti and Tassinari, dealing with fabric mechanics at ENI (Ente Nazionale Idrocarburi), had a seminal effect when presented at the Greenwich Conference on comparative lining methods in1974.

5 A problem avoided in heat-based methods, that allow easy local treatment of the adhesive, in which lining is the accumulation of the treatment of separate sectors.

6 Both D 360 and D 540 are no longer commercially available. Currently D 512 is under testing for use in such lining as a replacement for D 540. K 360 has pH 2,5-3 so it has to be buffered to neutral.

7 According to Lascaux datasheet and (HORIE, 2010), page 196, Plextol D360 (or K360) is -8°C.

8 Several mixtures of solvents are used, also depending to the size and other qualities of the painting. In some cases (small paintings) ethanol / isopropanol alone.

9 Authors are authorized to state that Seymour and van Och are currently working on this aspect of the method.

10 All choices were shared with Gloria Tranquilli, Director of the Conservation Laboratories of Venice Soprintendenza per i Beni Culturali e Ambientali. Marion Mecklenburg was kindly available for discussion of the structural subjects.

11 Sakase industry, SK802, vacuum sized in epoxy for dimensional stability and adherence between the pieces of fabric.

12 Thanks to Matteo Rossi Doria and Olivier Verheyden who made the contact.

13 Plextol B500, Rhoplex AC 33 and AC 234, Lascaux 360 HV and 498 HV.

14 With 0,1% w/w Rohagit.

15 According to the availability, a dry “rotary vane pump” or a regular oil pump have been used for the task.

16 For the specific painting 0,6 bar. It should be noted that the well-known surface texture changes discussed in 19th century literature, as in (Berger, 1966), were connected to the use of vacuum in systems that were also plasticizing the painting, such as humidity and heath.

17 A temporary opening of the bag is not as easy as for a low-pressure envelope.

18 Evaporation is forced since the beginning because of the reduction of atmospheric pressure by 0.6 bar, therefore some solvent starts being removed from the system from the beginning of the treatment. A vacuum pump does not imply passage of the solvent filled air close to electrical connections, as it uses a sealed chamber for suction.

19 Another relevant literature reference for peel tests on Plextol B500 is Alain Roche’s work (Roche, 2003), but his adhesive mixtures are not solvent-reactivated.

20 The quantity of dry residue is calculated on the base of the solids in the dispersion, as the weight of the dispersion sprayed on a square meter of lining canvas is easily measurable.

21 Manifatture tessili Sironi, Gallarate, Italy. Canvas named DM 18.

22 For their preparation, the canvas was chosen for being as similar as possible to the original. A 300g/m2 linen twill was used, with a thread count of 12/cm in weft and 20/cm in warp, prepared with identical procedures and quantities to all other mock-ups. The canvas was then impregnated with two brush applications of a 5% w/w dilution of the mixture of PVA emulsions: Mowilth DM5 (1 part) and Mowilth DMC 2 (5 parts). Lining was done after a minimum of two weeks open air drying.

23 Rohagit S hv, Evonik GmbH, Darmstad, DE. An acrylic acid, soluble in alcaline environment, that allows obtaining high viscosity disperions. A concentration of 0,1% w/w, as used, maintains a constant viscosity for more than 24 hours.

24 The adhesive let dry in the open air for a minimum of 24 hours.

25 “Peel ply” 17x17 threads/cm, normally used as a separation layer in epoxy resin laminations.

26 Three different cotton fabrics were used for the purpose, each thick enough to carry the desired amount of solvent.

27 Polyester non-woven “bleeder”, 150 g/m2.

28 The graph is characterised by high and low peaks.

29 Gustav Berger’s tests on rabbit skin glue showed almost identical behaviour with different speeds (1 mm/min and 100 mm/min) as this is a very rigid adhesive at low to medium RH (Berger, 2000). He preferred to use the slow speed of 1mm/min to simulate long term delamination of a BEVA lining.

30 The relatively small difference (approx. 10%) from the results obtained with 30 mm/min, also allowed to consider it as describing a wide range of speeds.

31 The canvas was impregnated with two brush applications of a 5% w/w dilution of the mixture of PVA emulsions: Mowilth DM5 (1 part) and Mowilth DMC 2 (5 parts).

32 In the paper the data appears as 300 g for 2,5 cm wide strips, with the peel speed of 2 mm/min. The conversion is done as follows: 300/2,5*100/1000*9,81= 117 N/m (speed 2mm/min); 117 N/m*1,73=202 N/m (theoretical value at the speed of 100 mm/min).

33 Authors propose the forces of 800 and 1000 g for 2,5cm wide strips, at the speed of 2mm/min.

34 The difference between the forces obtained with increasing speeds has a logarithmic pattern, therefore the difference between the speeds of 30 mm/min and 40 mm/min can be considered close to 0.

35 The granules needed to be free of dust, as small particles may interact with the canvas and stay in contact even with no adhesion on the residues. A very easily available material proved to be perfect for the task: grounded coffee.

36 The work was carried out in spring 2018 by Equilibrarte srl, for the conservation company Recoop, in Malta, sponsored by the Bank of Valletta.

Top of page

List of illustrations

Title Fig. 1. Adhesive distribution in “mist lining”
Caption The sprayed adhesive takes the form of an airy deposition of white acrylic adhesive on lining canvas filaments.
Credits Credits Van Och, 2003.
URL http://journals.openedition.org/ceroart/docannexe/image/6488/img-1.jpg
File image/jpeg, 1.9M
Title Fig. 2. Adhesive distribution on the carbon fibre canvas
Caption The absence of raised fibres on the surface produced the deposition of small droplets, that developed a lenticular shape upon drying, finely distributed on the surface and strongly adhering to the canvas.
Credits Credits: A. Iaccarino Idelson and C. Serino, 2012.
URL http://journals.openedition.org/ceroart/docannexe/image/6488/img-2.jpg
File image/jpeg, 952k
Title Table 1. Examples of variables used lining paintings differing for age and dimensions
Credits Credits: Antonio Iaccarino Idelson
URL http://journals.openedition.org/ceroart/docannexe/image/6488/img-3.jpg
File image/jpeg, 1.4M
Title Fig. 3. Weight of liquid solvent needed for reactivation vs. force of adhesion
Credits Credits: Valerio Garofalo
URL http://journals.openedition.org/ceroart/docannexe/image/6488/img-4.jpg
File image/jpeg, 408k
Title Table 2. The lining conditions for the mock-ups
Credits Credits: Valerio Garofalo
URL http://journals.openedition.org/ceroart/docannexe/image/6488/img-5.jpg
File image/jpeg, 1.2M
Title Fig. 4. Adhesive distribution
Caption The deposition of the adhesive builds up a dense network of droplets, closely adhering to the canvas surface and its few raised fibres, but leaving most of the space in the weave open for the passage of solvent vapours.
Credits Credits: A. Iaccarino Idelson
URL http://journals.openedition.org/ceroart/docannexe/image/6488/img-6.jpg
File image/jpeg, 712k
Title Fig. 5. Layer build-up for lining
Caption From the painting mock-up on the bottom, to the white polyester bleeder
Credits Credits: Valerio Garofalo
URL http://journals.openedition.org/ceroart/docannexe/image/6488/img-7.jpg
File image/jpeg, 596k
Title Fig. 6. Lining the mock-ups
Caption The suction pipe from the vacuum pump is placed on the perimeter of the envelope. Two pressure gauges are placed on the opposite side. These are used only after the completion of the reactivation process, in order to allow access of clean air during the last phase, to remove the solvent with a constant value of pressure.
Credits Credits: Valerio Garofalo
URL http://journals.openedition.org/ceroart/docannexe/image/6488/img-8.jpg
File image/jpeg, 2.0M
Title Fig. 7. The force needed for delamination with different peel speeds
Caption The force needed for delamination with different peel speeds.
Credits Credits: Valerio Garofalo
URL http://journals.openedition.org/ceroart/docannexe/image/6488/img-9.jpg
File image/jpeg, 1.2M
Title Table 3. Results of the preliminary tests for the influence of speed on the force needed for delamination
Credits Credits: Valerio Garofalo
URL http://journals.openedition.org/ceroart/docannexe/image/6488/img-10.jpg
File image/jpeg, 96k
Title Table 4. Ratio between the force obtained at 100 mm/min vs. 2 mm/min and 30 mm/min
Credits Credits: Valerio Garofalo
URL http://journals.openedition.org/ceroart/docannexe/image/6488/img-11.jpg
File image/jpeg, 472k
Title Table 5. Average values of peel strength and standard deviation for all test
Credits Credits: Valerio Garofalo
URL http://journals.openedition.org/ceroart/docannexe/image/6488/img-12.jpg
File image/jpeg, 1.2M
Title Fig. 8. Peel strength with increasing quantities of solvent
Caption Peel strength with increasing quantities of solvent.
Credits Credits: Valerio Garofalo
URL http://journals.openedition.org/ceroart/docannexe/image/6488/img-13.jpg
File image/jpeg, 404k
Title Fig. 9. Peel strength with and without PVA on the painting’s canvas
Caption Peel strength with and without PVA on the painting’s canvas.
Credits Credits: Valerio Garofalo
URL http://journals.openedition.org/ceroart/docannexe/image/6488/img-14.jpg
File image/jpeg, 440k
Title Table 6. Hand de-lining with no reactivation of the adhesive
Credits Credits: Valerio Garofalo
URL http://journals.openedition.org/ceroart/docannexe/image/6488/img-15.jpg
File image/jpeg, 312k
Title Fig. 10. Peel strength in the judgement of previous authors
Credits Credits: Valerio Garofalo
URL http://journals.openedition.org/ceroart/docannexe/image/6488/img-16.jpg
File image/jpeg, 424k
Title Fig. 11 Residues of adhesive
Caption Residues on the painting canvas after the peel tests made visible with dark granules
Credits Credits: Valerio Garofalo
URL http://journals.openedition.org/ceroart/docannexe/image/6488/img-17.jpg
File image/jpeg, 2.0M
Top of page

References

Electronic reference

Antonio Iaccarino Idelson and Valerio Garofalo, « Aiming at reproducibility in lining canvas paintings  », CeROArt [Online], 11 | 2019, Online since 30 September 2019, connection on 14 November 2019. URL : http://journals.openedition.org/ceroart/6488 ; DOI : 10.4000/ceroart.6488

Top of page

About the authors

Antonio Iaccarino Idelson

Antonio Iaccarino Idelson studied “Conservation of paintings and architectural surfaces” at Istituto Centrale per il Restauro in Rome (1989-1993) and “Conservation of wooden artifacts” (1986-1988) in Florence. Since 2002 is CEO and Technical Director of Equilibrarte srl, a conservation company in Rome, with important commissions in Italy, Europe and the USA. He has been teaching subjects related to conservation since 2001, in Italy and abroad. Has published the book “Il tensionamento di dipinti su tela” in 2004 and actively promotes and participates in research projects on conservation subjects. iaccarino.a@gmail.com

By this author

Valerio Garofalo

Valerio Garofalo is a paintings and sculpture conservator, graduated with maximum grades at the University of Torino, CCR La Venaria Reale in 2016. He started his career in 2007 with collaborations in important public and private conservation projects, since 2010 with a part-time collaboration for the conservator Delfina Fagnani, in Bergamo, Italy. After the completion of his studies in 2016, he started working as a freelance conservator, with multiple collaborations. valeriogarofalo728@gmail.com

Top of page

Copyright

Licence Creative Commons
CeROArt – Conservation, exposition, restauration d'objets d'arts est mis à disposition selon les termes de la licence Creative Commons Attribution - Pas d'Utilisation Commerciale - Pas de Modification 4.0 International.

Top of page
  • Logo Fondation Périer-d’Ieteren
  • Logo DOAJ – Directory of Open Access Journals
  • OpenEdition Journals