The effect of oil binders on paper supports via VOC analysis

G Model

ARTICLE IN PRESS

CULHER-3025; No. of Pages 10

Journal of Cultural Heritage xxx (2016) xxx–xxx

Available online at

ScienceDirect www.sciencedirect.com

Original article

The effect of oil binders on paper supports via VOC analysis Penelope Banou a,∗ , Athena Alexopoulou b , Charikleia Chranioti c , Dimitris Tsimogiannis c , Agni-Vasileia Terlixi d , Spiros Zervos e , Brian W. Singer f a

General State Archives of Greece, Athens, Greece Department of Conservation of Antiquities and Works of Art, TEI of Athens, Egaleo, Greece Laboratory of Food Chemistry and Technology, School of Chemical Engineering, NTUA, Athens, Greece d National Gallery-Alexandros Soutzos Museum, Athens, Greece e Department of Library Science and Information Systems, TEI of Athens, Egaleo, Greece f Faculty of Health and Life Sciences, Northumbria University, Newcastle, UK b

c

a r t i c l e

i n f o

Article history: Received 30 October 2015 Accepted 19 January 2016 Available online xxx Keywords: Paper degradation Drying oils VOC emissions Headspace-SPME GC-MS Art conservation

a b s t r a c t The effect of the presence of drying oils in paper supports on the rate of cellulose degradation is investigated in a novel manner using Solid Phase Micro-extraction (SPME), which is employed to analyse volatile organic compounds (VOCs), emitted from oiled paper. This technique is applied as a non-destructive means of analysing original works of art on paper, in order to detect volatile cellulose degradation products. It is also applied to artificially aged paper samples with and without oil, in order to investigate the extent to which the presence of drying oil accelerates the degradation of cellulose. Furfural and other volatile cellulose degradation products containing a furan ring are selected as representative cellulose degradation products to be measured for the purpose of the investigation. It is demonstrated, by the finding of increased emissions of the selected compounds, that the presence of drying oils accelerates the thermal and oxidative degradation of cellulose in cotton paper and two types of wood pulp based papers. © 2016 Elsevier Masson SAS. All rights reserved.

1. Research aims We have been particularly concerned [1] with a collection of works in the National Gallery of Greece in Athens, which includes oil paintings on paper and oil-based ink prints on paper and contains works by important 19th century and early 20th century Greek artists, such as N. Gysis, N. Lytras, K. Volanakis, K. Parthenis and K. Maleas all of which present particular problems, associated with the absorption of the oil binder by the paper, such as; discoloration, reduction of mechanical strength and embrittlement of the support. We wished to investigate whether the presence of drying oils in paper accelerates cellulose degradation. Our approach in this current work was to use Solid Phase Micro-extraction (SPME), a solvent-free sample preparation technology for analysing volatile organic compounds (VOC). This technology was applied as a non-destructive means of analysing works of art on paper to detect volatile cellulose degradation products and also was applied to mock-ups in order to

∗ Corresponding author. E-mail address: [email protected] (P. Banou).

investigate the extent to which the presence of drying oil accelerates the oxidative thermal degradation of cellulose. Furfural and other volatile cellulose degradation products containing a furan ring were chosen as marker compounds. The immediate aim here was to provide further information regarding paper degradation in the presence of oil, using SPME analysis of VOC and a comparison with changes in pH and mechanical strength of our samples. 2. Introduction 2.1. Drying oils on paper Works of art on paper containing oil-based media degrade more rapidly than works where aqueous paint binders are used [2]. Problems, associated with the oil binder have been attributed to the oxidation of the paper support [3,4] or the oxidation of the oil medium [5]. Some scientific evidence has been added to these claims via pH measurements [6,7], which showed that the supports became notably acidic, even when there was no intense discolouration or embrittlement [6]. Vincent Daniels [8,9] showed by the Russell effect that free radicals were present around linseed oil stains in paper, demonstrating that oxidation was taking place.

http://dx.doi.org/10.1016/j.culher.2016.01.003 1296-2074/© 2016 Elsevier Masson SAS. All rights reserved.

Please cite this article in press as: P. Banou, et al., The effect of oil binders on paper supports via VOC analysis, Journal of Cultural Heritage (2016), http://dx.doi.org/10.1016/j.culher.2016.01.003

G Model CULHER-3025; No. of Pages 10

ARTICLE IN PRESS P. Banou et al. / Journal of Cultural Heritage xxx (2016) xxx–xxx

2

Fig. 1. Structures of furfural (1) and 5-(hydroxymethyl) furfural (2).

Elfecky and Hassan [10] measured; tensile strength, elongation on break, pH and colour by the ‘CIELab’ system, of mock up paintings on Fabriano paper. They compared paper coated with oil paint having a linseed oil coat before the paint and paper with animal glue based ground before the oil paint, the former being stronger at first but both deteriorating on ageing. 2.2. Volatile organic compounds as cellulose degradation products It has been established [11] that oxidative and thermal degradation of cellulose can give furfural (1) and derivatives of furfural such as 5-(hydroxymethyl) furfural (HMF) (2) (Fig. 1). Fagerson [12] also lists other furans as degradation products including 2methyl furfural, 2-methyl furan, 2-ethyl furan and 2-propyl furan, and he reports a possible route from carbohydrates to HMF during thermal degradation. Loss of formaldehyde from HMF then leads to 2-Furfural. More recently, Scheirs et al. [13] have reviewed other possible mechanisms for thermal degradation of cellulose to furfural and HMF, one via levoglucosan. They then propose an alternative mechanism for the acid catalysed hydrolytic thermal degradation of cellulose to furfural derivatives (Fig. 2) after demonstrating that the route via levoglucosan was not the most favoured route. Several researchers have studied the effects of mineral oils on Kraft paper, a combination used as an insulator in electrical transformers [14]. It is reported that furfural and related compounds could be used as indicators to detect the thermal degradation of cellulose and hemicelluloses within the paper. Interestingly, it is noticed that a combination of air, paper and mineral oil gave off the indicators at a greater rate than paper and mineral oil alone [14,15] and that oxygen was absorbed during the reaction, indicating that oxidative as well as thermal degradation was involved [14]. The presence of 4% moisture also increased the rate of degradation [14].

Contact and headspace-SPME has been successfully employed to track VOC emitted from pure cellulosic and ligno-cellulosic papers, both original and artificially aged, book and archival material [16–21]. The degradation products found [20,21] included aromatic hydrocarbons, an homologous series of alkanes, an homologous series of aliphatic aldehydes, an homologous series of aliphatic carboxylic acids, and several furan derivatives including furfural and 5-methyl furfural, the latter being more prevalent with ageing at 90 ◦ C and 100% relative humidity than with dry ageing at 90 ◦ C. Vanillin was also emitted as a lignin degradation product [20,21]. Clark et al. [22] recognised that paper contains some natural lipids. They claimed that the series of aldehydes found were derived from the fatty acids in the lipids. Other researchers have specifically measured furfural emissions [23] or furfural and acetic acid [24,25] emissions to monitor degradation in paper based collections. However, crucially, none of these studies had considered the effects of drying oils on paper. Because of the problems associated with the collection being studied, we were particularly interested to see if the addition of drying oils to paper would accelerate cellulose degradation. Here we chose to study several different papers and to age them at 90 ◦ C and at a 77% humidity level. This humidity level was chosen for experimental reasons since it is given by saturated sodium chloride solution at 90 ◦ C [26] and by dropping the paper samples into such a solution within the headspace vial after ageing we obtained a maximum recovery of VOC’s for analysis. This is explained by the ‘salting out effect’ [27] which helps to expel the volatile organics, out of the aqueous phase, into the vapour phase from which they are then collected by the SPME needle. Also a 77% humidity level is expected to slightly exaggerate the results [28] compared to the standard 50% standard RH maintained in museum storage areas. We also sought to analyse works of art, containing oil on paper, in a non-destructive way for volatile organic compounds by SPME analysis to compare the emissions with that from old book papers [21] and with our own humid ageing tests. 3. Materials and methodology 3.1. Analysis of binding media in works of art The nature of the binding media was established via methylation GC-MS techniques using TFTMTH as described and investigated by others [29,30].

Fig. 2. Proposed route to 5-(hydroxymethyl) furfural and furfural from cellulose. Adapted from Scheirs et al. [13].

Please cite this article in press as: P. Banou, et al., The effect of oil binders on paper supports via VOC analysis, Journal of Cultural Heritage (2016), http://dx.doi.org/10.1016/j.culher.2016.01.003

G Model CULHER-3025; No. of Pages 10

ARTICLE IN PRESS P. Banou et al. / Journal of Cultural Heritage xxx (2016) xxx–xxx

3

• cotton pHotonTM high purity paper by the Munktel paper Mill, 80 gsm (Conservation by Design Limited, UK); • Canson® Montval® watercolour paper, 185 gsm (Art & Hobby, Greece) and; • Kraft paper (Dionisopoulos, paper distributer, Greece), 135 gsm.

Fig. 3. Experimental set up for collection of VOC’s from the verso of a work of art showing the SPME needle close to the work under a petri dish and the whole work of art and needle holder under a second glass encasement.

3.2. Analysis of volatile organic compounds from works of art by SPME-GC-MS An SPME needle cartridge, with a 50/30 ␮m divinylbenzenecarboxen/poly(dimethylsiloxane) fibre (Supelco, Sigma-Aldrich Ltd., Dorset, UK) was preconditioned by heating to 230 ◦ C in the injection port of a GC-MS and retracted. The preconditioned SPME needle was then exposed to within 2 mm of the verso of a work (which was placed face down) for 24 hours. The work and the needle were encased in glass for this period (Fig. 3). The SPME needle was then retracted and reopened in the injection port of the GC-MS and heated to 230 ◦ C for 10 min to release the volatile components and trap them at the beginning of the column. The compounds were then separated and identified by GC-MS analyses carried out in an Agilent Technologies 7890A GC gas chromatograph coupled to an Agilent Technologies 5975 C MSD mass selective triple-axis detector, and an Agilent 1909/S capillary column (30 m × 0.25 mm internal diameter; coating thickness 0.25 ␮m). Carrier gas used was He, with a flow rate of 1 mL/min and at a linear velocity of 40 cm/s; split 1:10; ionization: EI 70 eV. The temperature of the column was held at 40 ◦ C for 10 min and then raised from 40 ◦ C to 250 ◦ C at a rate of 5 ◦ C per min and held at 250 ◦ C for 15 min. Twenty seven target compounds were selected for measurement of peak areas in each run. The selection was based on compounds found by previous workers [20–22]. 3.3. Paper ageing Three types of paper were investigated:

Fibre analysis using optical microscopy, EDX analysis, spot test for lignin (phluoroglucinol test, TAPPI T401) and FTIR analysis were performed so as to provide information about the fibre content and lignin presence. Paper strips 1 × 7 cm were cut and weighed at 23 ◦ C, 55% RH. Half the strips cut were impregnated with cold pressed linseed oil (Windsor & Newton, London) after weighing. Three strips were prepared for analysis for each number of days, with and without oil. Strips were suspended on cotton threads in headspace vials (SU860101 Supelco, with stainless steel screw cap and PTFE/silicone septum, thickness 1.3 mm, Sigma-Aldrich Ltd., Dorset, UK) above 5 mL of 15% sodium chloride for analysis (MERCK, KGaA, Germany) solution. All samples (Tables 1 and 2) were thus aged at 90 ◦ C and at 77% RH for 1, 4, 7, 14, 21 and 28 days and then kept at −20 ◦ C until analysis. 3.4. Standard compounds Five compounds were chosen for monitoring: 2-ethyl furan, furfural, 5-methyl furfural, 5-ethyl furfural and 5-pentyl furanone. These are known volatile cellulose degradation products and had been shown to be produced on ageing of paper, but not during the ageing of oil films [1]. Furfural and 5-pentyl furanone had also been found in emissions from works of art being studied as described below (compounds 4 and 24, Table 2). Weighed samples of the standards: 2-ethyl furan, furfural; 5-methyl furfural and 5-ethyl furfural (all from Sigma-Aldrich Ltd., Dorset, UK) (Table 5) were analysed by the SPME-GC-MS technique. 20 microlitre aliquots of each of the following standard aqueous solutions: one containing 2.20 × 10−4 gmL−1 of 2-furaldeyde, one containing 1.67 × 10−4 gmL−1 of 2-ethyl furan, one containing 0.0219 × 10−4 gmL−1 of 5-ethyl-2-furaldehyde and one containing 2.35 × 10−4 gmL−1 of 5-methyl-2-furaldehyde, were injected into a sealed headspace file with 2.5 mL of salt solution A preconditioned SPME needle was inserted and the set up was incubated at 40o for 40 minutes. The needle was then inserted into the injection port of the GC-MS instrument and a chromatogram of the compounds collected. This was repeated three times. A second set of readings were obtained using 40 microlitre aliquots of each of the following aqueous solutions one containing 2.20 × 10−5 gmL−1 of 2-furaldeyde,

Table 1 List of works studied and results of fibre, binder and other analysis. Artist, Title of work, description, museum object number

Az/P ratio, P/S ratio

Binding medium identified

Fibre content, EDX/spot tests

pH of oiled areas

Nicholas Gyssis, Sewing Room (E␳␥␣␴␶␳␫ ´ о ␳␣␲␶␫␬´ ς , .3434), oil sketch on paper, late 19th century (Fig. 5) K. Fannelis, Figure of Christ (Mо␳␸´ X␳␫␴␶о␷, ´ .2985), oil sketch on paper, mid 19th century (Fig. 5) G. Economides, Occupation (K␣␶о␹, ´ .9812), wood block print, mid 20th century G. Economides, Livadia (␧␫␤␣␦␫˛, ´ .9822), wood block print, mid 20th century (Fig. 6)

1.0, 3.3

Drying oil, beeswax and rosin Drying oil, beeswax and rosin A non-drying oil, and rosin A non-drying oil, possibly with some drying oil mixed in, and rosin A non-drying oil, and rosin Drying oil, possibly walnut oil, possibly mixed with some non-drying oil and rosin

Cotton, linen rag and softwood, alum, lignin Cotton and linen rag and softwood, alum Kozo and softwood

4.70

Linen, cotton

5.20

Linen and cotton

5.10

Cotton rag mixed with softwood

5.53

´ о␯оς , .9823), wood block G. Economides Mykonos (M␷␬ print, mid 20th century G. Economides Album Sachsiche Scheriz, page from album of wood block prints, .9740, mid 20th century

0.32, 3.3 0.01, 3.1 0.21,3.3

0.05, 3,3 0.38, 3.1

4.70 5.88

Please cite this article in press as: P. Banou, et al., The effect of oil binders on paper supports via VOC analysis, Journal of Cultural Heritage (2016), http://dx.doi.org/10.1016/j.culher.2016.01.003

G Model

ARTICLE IN PRESS

CULHER-3025; No. of Pages 10

P. Banou et al. / Journal of Cultural Heritage xxx (2016) xxx–xxx

4

Table 2 Peak areas of target compounds obtained from SPME-GC-MS investigation of 18 year old naturally aged sample of rag paper impregnated with linseed oil and of six museum objects. Target compounds

Typical retention time mins

Test piece: rag paper + linseed oil naturally aged 18 years

Peak area responses as percentage of total for each object (1) ethanoic acid 2.20 0.352 (2) toluene 5.03 0.004 (3) hexanal 6.45 0.007 (4) furfural 8.74 0.003 (5) m-xylene 10.66 0.004 (6) n-nonane 12.95 0.002 14.77 0.000 (7) pentanoic acid 15.65 1.500 (8) verbenene 17.51 0.332 (9) 1,2,4-trimethyl-benzene (10) phenol 17.79 3.052 17.97 0.000 (11) n-decane 18.13 12.483 (12) n-octanal 19.02 0.783 (13) limonene 21.95 9.574 (14) nonanal 22.83 0.000 (15) heptanoic acid (16) trans-verbenol 23.36 14.478 24.16 0.000 (17) menthol 24.27 0.370 (18) napthalene 24.53 1.036 (19) octanoic acid 25.20 3.591 (20) decanal (21) I-verbenone 25.35 11.612 3.376 (1) 26.81 5-butyldihydro-2(3H)-furanone 27.33 0.836 (22) nonanoic acid (23) 5-pentyl-2(5H)-furanone 29.16 9.475 (24) vanillin 30.65 0.715 35.41 (25) 25.836 2,2,4-trimethyl-1,3-pentanediol diisobutyrate (26) isopropyl myristate 40.32 0.580

Object 3434

Object 2985

Object 9812

Object 9822

Object 9823

Object 9740

Background, museum environment

0.000 0.005 0.003 0.001 0.005 0.304 0.109 0.153 12.799 0.302 16.342 2.608 7.481 19.550 0.320 0.391 1.408 0.119 0.197 13.918 0.115 0.139

0.000 0.717 0.006 0.005 0.166 0.162 0.007 0.191 7.575 0.127 3.853 6.722 3.713 25.002 0.478 0.170 1.199 0.114 1.074 15.983 0.135 0.321

0.000 0.538 0.009 0.002 0.213 0.320 0.007 0.079 11.201 0.056 1.606 8.742 2.005 29.290 0.158 0.070 0.608 0.120 0.084 26.092 0.029 0.491

0.000 0.717 0.006 0.005 0.166 0.162 0.007 0.191 7.575 0.127 3.853 6.722 3.713 25.002 0.478 0.170 1.199 0.114 1.074 15.983 0.135 0.321

0.000 0.820 0.012 0.004 0.014 0.909 0.005 0.175 10.852 0.033 4.320 7.548 3.362 26.461 0.087 0.222 0.873 0.158 0.000 16.325 0.082 0.242

0.032 0.002 0.004 0.000 0.013 0.006 0.013 0.061 6.864 0.000 5.628 2.261 9.539 11.582 0.205 0.948 1.173 30.790 0.344 6.344 0.644 0.434

0.000 2.362 0.003 0.001 0.441 0.519 0.005 0.241 9.084 0.038 3.152 4.119 6.595 23.209 0.469 0.167 0.999 0.416 0.792 27.071 0.052 0.034

0.000 0.021 0.232 22.691

0.000 0.005 0.132 31.333

0.000 0.000 0.063 17.763

0.000 0.005 0.132 31.333

0.000 0.006 0.009 26.959

0.000 0.036 0.635 21.890

0.000 0.005 0.066 20.015

0.790

0.812

0.455

0.812

0.523

0.552

0.144

one containing 1.67 × 10−5 gmL−1 of 2-ethyl furan, one containing 0.0219 × 10−5 gmL−1 of 5-ethyl-2-furaldehyde and one containing 2.35 × 10−5 gmL−1 of 5-methyl-2-furaldehyde, in a similar way. The calibration curve is shown (Fig. 4). 5-pentyl furanone was not available for purchase and was not run as a standard but was recognised in the chromatographs by its mass spectrum using the NIST library. In the case of 5-pentyl furanone, an average response per g for the four standard compounds was used to estimate the quantity of this compound from its peak area in each chromatogram, for comparison purposes only. 3.5. Calculation of quantity of volatile target compounds Since the paper samples had been weighed, the above calibration (Fig. 4) was used to calculate the mass of each of these

chosen compounds given off by the paper samples per gram of paper from the peak areas for these compounds in the respective chromatograms (the paper samples with oil impregnation had been weighed before the linseed oil was applied). It should be stressed that the calculated values are only estimates since the concentration of target compounds in the headspace above a solution of the standards may be different to those above a paper sample depending on how the compounds interact with the paper or oil. The use of saturated sodium chloride solution should have minimised these effects. However, the use of standards in this way allowed us to check retention times and give a rough estimate of the mass of VOCs emitted per gram of paper, which allows for consistent comparison of results at least. Third samples of the Montval wood based paper and of the Kraft paper were analysed six months later. Recalibration with standards showed a deterioration in performance of the GC-MS and/or SPME needle over time. This third set of results was therefore calculated from the peak areas using the new calibration.

3.6. SPME-GC-MS analysis

Fig. 4. Calibration of VOC’s peak area v quantity exposed in ng using standards This calibration was used on the first two results but the third results taken at a later time were recalibrated.

An SPME needle was preconditioned by heating to 230 ◦ C in the injection port of a GC-MS and retracted. The preconditioned SPME needle was then inserted into a headspace vial containing the paper sample (or the weighed standard) now immersed in 5 mL of 15% sodium chloride solution and thus exposed to the vapours given off from each sample for 40 minutes at 40 ◦ C. The SPME needle was then retracted and reopened in the injection port of the GC-MS and heated to 230 ◦ C for 10 min to release the volatile components and trap them at the beginning of the column. The compounds were then separated and identified by GC-MS using the conditions

Please cite this article in press as: P. Banou, et al., The effect of oil binders on paper supports via VOC analysis, Journal of Cultural Heritage (2016), http://dx.doi.org/10.1016/j.culher.2016.01.003

G Model CULHER-3025; No. of Pages 10

ARTICLE IN PRESS P. Banou et al. / Journal of Cultural Heritage xxx (2016) xxx–xxx

5

Fig. 5. Two oil sketches on paper: above: N. Gyssis, Sewing Room (E˛ 

´ о␳␣␲␶␫␬´ ς , Inv. No. ˘3434), late 19th century, collection of the National Gallery-Alexandros Soutzos Museum (Athens, Greece), recto (left) and verso (right). Below: K. Fannelis, Figure of Christ (Mоϕ´ X  о␷, ´ Inv. No. ˘2985), mid 19th century, collection of the National Gallery-Alexandros Soutzos Museum (Athens, Greece), recto (left) and verso (right). Copyright National Gallery-Alexandros Soutzos Museum with permission, photographed by Agathi Kaminari.

described above. The peak areas were measured and averaged for each of the strips for each ageing period.

4. Results, data and discussion 4.1. Non-destructive analysis of volatile oxidation products from works of art

3.7. pH changes pH values of test pieces were obtained by cold extraction pH measurements (Tappi T 509). pH values were recorded on original works of art from areas of paper showing discolouration due to oil absorption via surface pH measurements (Tappi T 529).

3.8. Changes of the mechanical properties Paper samples were examined with tear resistant measurements using an Elmendorf type apparatus (ISO 1974, 1990).

Several works of art on paper, which had oil-based printing inks or oil paints in the work (Table 1), presenting problems of discolouration due to oil absorption and diffusion, were selected from the collection (Figs. 5 and 6). The compounds found to be given off by the objects are listed (Table 2). They include: aromatic hydrocarbons (compounds 2, 5, 9 and 18, Table 2); aliphatic hydrocarbons (compounds 6 and 11); straight chain aliphatic aldehydes (compounds 3, 14 and 20, Table 2); straight chained aliphatic carboxylic acids (compounds 1, 7, 15, 19 and 23); esters (compounds 26 and 27, Table 2), all

Please cite this article in press as: P. Banou, et al., The effect of oil binders on paper supports via VOC analysis, Journal of Cultural Heritage (2016), http://dx.doi.org/10.1016/j.culher.2016.01.003

G Model

ARTICLE IN PRESS

CULHER-3025; No. of Pages 10

P. Banou et al. / Journal of Cultural Heritage xxx (2016) xxx–xxx

6

Fig. 6. G. Economides, Livadia ( ˇ˛ıε ˛, ´ Inv. No. ˘9812), b/w woodcut, first half of the 20th century, collection of the National Gallery-Alexandros Soutzos Museum (Athens, Greece), recto (left) and verso (right). Copyright National Gallery-Alexandros Soutzos Museum with permission, photographed by Agathi Kaminari.

of which can be oil [1] or paper degradation products [20,21]; phenol (compound 10, Table 2); mono terpenoids (compounds 8, 13, 16, 17 and 21); furan derivatives (compounds 4, 22 and 24), which are cellulose or hemicellulose degradation products and also vanillin (compound 25, Table 2), which is a known lignin degradation product [20,21]. Three of the mono terpenoids, verbenene (compound 8), trans-verbenol (compound 16) and 1-verbenone (compound 21) are known oxidation products of ␣-pinene [31], from turpentine and therefore may originate; either from rosin size in the paper or turpentine solvent used in the oil paint or oil-based inks. Also a test piece of rag paper, which had been impregnated with linseed oil and then naturally aged for 18 years, was exposed to the needle and showed the presence of 23 of the 27 target compounds (Table 2).

The needle was conditioned between runs by heating to 230 ◦ C for 10 mins. To check that all compounds had been released in this process, 2 chromatograms were collected of such events and the peak areas for the target compounds averaged. Residues of some of the compounds remained but subtraction of these amounts for each target compound from each of the corresponding peak areas obtained from the object made no significant differences, especially from compounds: 5, 6, 8–22, 25–27. Unsurprisingly, the background environment in a paper conservation studio within the museum also yielded a chromatogram after SPME analysis, which contained most of the target compounds in significant quantities (Table 2). However, we were confident that the compounds found from the object had not suffered too much interference from the background environment since the objects were covered in glass during the collection of volatiles over

Table 3 Peak areas of target compounds obtained from SPME-GC-MS investigation of background in museum environment and of six museum objects with the background measurement subtracted. Target compounds

Typical retention time mins

Background, museum environment

Raw Peak area Responses with museum background subtracted 2.20 0.00E + 00 (1) ethanoic acid (2) toluene 5.03 9.07E + 05 6.45 9.86E + 02 (3) hexanal 8.74 4.53E + 02 (4) furfural 10.66 1.69E + 05 (5) n-xylene 12.95 1.99E + 05 (6) n-nonane 14.77 1.80E + 03 (7) pentanoic acid 15.65 9.24E + 04 (8) verbenene 17.51 3.49E + 06 (9) 1,2,4-trimethyl-benzene 17.79 1.47E + 04 (10) phenol 17.97 1.21E + 06 (11) n-decane 18.13 1.58E + 06 (12) n-octanal 19.02 2.53E + 06 (13) limonene 21.95 8.91E + 06 (14) nonanal 22.83 1.80E + 05 (15) heptanoic acid 23.36 6.41E + 04 (16) trans-verbenol (17) menthol 24.16 3.84E + 05 24.27 1.60E + 05 (18) napthalene 24.53 3.04E + 05 (19) octanoic acid 25.20 1.04E + 07 (20) decanal 25.35 2.00E + 04 (21) I-verbenone (22) 5-butyldihydro-2(3H)26.81 1.31E + 04 furanone 27.33 0.00E + 00 (23) nonanoic acid 1.87E + 03 (24) 29.16 5-pentyl-2(5H)-furanone (25) vanillin 30.65 2.53E + 04 7.69E + 06 (26) 2,2,4-trimethyl-1,335.41 pentanediol diisobutyrate 40.32 5.51E + 04 (27) isopropyl myristate

Object 3434 − background

Object 2985 − background

Object 9812 − background

Object 9822 − background

Object 9823 − background

Object 9740 − background

0.00E + 00 −9.05E + 05 4.24E + 02 2.59E + 02 −1.66E + 05 −3.29E + 04 5.79E + 04 −8.63E + 03 3.52E + 06 1.51E + 05 7.74E + 06 −1.53E + 05 1.56E + 06 1.80E + 06 −4.86E + 03 1.50E + 05 3.88E + 05 −9.49E + 04 −1.96E + 05 −2.77E + 06 4.30E + 04 6.30E + 04

4.87E + 03 −9.03E + 05 3.73E + 03 5.95E + 02 −6.66E + 04 7.61E + 03 1.34E + 03 −5.34E + 04 −2.81E + 05 −9.92E + 03 5.34E + 05 −2.86E + 05 −1.47E + 06 −1.59E + 06 −7.00E + 04 −7.57E + 03 −7.66E + 04 9.01E + 05 −3.04E + 05 −5.19E + 06 1.82E + 04 3.15E + 04

0.00E + 00 −6.37E + 05 3.57E + 03 6.09E + 02 −6.24E + 04 −3.88E + 04 1.69E + 03 −5.27E + 04 2.14E + 06 1.34E + 04 −4.04E + 05 2.81E + 06 −1.53E + 06 5.79E + 06 −1.00E + 05 −2.91E + 04 −7.83E + 04 −9.95E + 04 −2.62E + 05 2.70E + 06 −5.50E + 03 2.33E + 05

0.00E + 00 −6.82E + 05 8.14E + 02 9.66E + 02 −1.17E + 05 −1.48E + 05 4.47E + 02 −3.24E + 04 −1.11E + 06 2.52E + 04 5.27E + 02 5.31E + 05 −1.37E + 06 −1.05E + 06 −2.98E + 04 −1.07E + 04 −6.90E + 03 −1.24E + 05 3.35E + 04 −5.37E + 06 2.24E + 04 8.77E + 04

0.00E + 00 −6.83E + 05 2.28E + 03 5.15E + 02 −1.66E + 05 4.90E + 04 −4.76E + 02 −4.46E + 04 −5.23E + 05 −5.74E + 03 −3.02E + 04 4.81E + 05 −1.61E + 06 −1.68E + 06 −1.56E + 05 −3.44E + 03 −1.45E + 05 −1.17E + 05 −3.04E + 05 −5.94E + 06 2.30E + 03 5.32E + 04

1.87E + 04 −9.06E + 05 1.58E + 03 −4.53E + 02 −1.62E + 05 −1.96E + 05 5.67E + 03 −5.62E + 04 5.61E + 05 −1.47E + 04 2.11E + 06 −2.48E + 05 3.10E + 06 −2.08E + 06 −5.88E + 04 4.95E + 05 3.09E + 05 1.80E + 07 −1.01E + 05 −6.65E + 06 3.60E + 05 2.43E + 05

0.00E + 00 9.73E + 03

0.00E + 00 4.60E + 01

0.00E + 00 −1.87E + 03

0.00E + 00 −4.06E + 02

0.00E + 00 −9.20E + 01

0.00E + 00 1.93E + 04

1.02E + 05 4.74E + 06

1.63E + 04 −1.15E + 05

6.14E + 03 1.23E + 06

1.63E + 04 2.16E + 06

−2.28E + 04 −3.20E + 05

3.49E + 05 5.23E + 06

3.78E + 05

9.98E + 04

1.73E + 05

2.00E + 05

8.77E + 04

2.70E + 05

Please cite this article in press as: P. Banou, et al., The effect of oil binders on paper supports via VOC analysis, Journal of Cultural Heritage (2016), http://dx.doi.org/10.1016/j.culher.2016.01.003

G Model

ARTICLE IN PRESS

CULHER-3025; No. of Pages 10

P. Banou et al. / Journal of Cultural Heritage xxx (2016) xxx–xxx

7

Table 4 The average cold extraction pH values of the test pieces at all ageing periods (average measurements out of three test pieces at each ageing period). Days of ageing

Cotton

Cotton + oil

Montval

Montval + oil

Kraft

Kraft + oil

0 1 4 7 14 21 28

7.0 6.3 6.3 6.6 6.6 6.2 6.3

6.7 2.5 2.2 2.3 2.3 2.0 2.3

7.7 7.7 7.7 7.5 7.7 7.2 6.7

6.6 5.4 4.6 4.1 4.2 4.0 4.3

6.7 6.5 6.6 6.5 6.5 6.2 6.2

6.3 2.5 2.4 2.3 2.3 2.3 2.3

a 24 hour period, and also because the quantities of many of the target compounds differed significantly from the environment in most of the objects, as can be seen when the environmental background peak areas are subtracted from the peak areas for each target compound for each of the six objects (Table 3). Indeed, several values were greatly negative indicating that the background had been successfully excluded. Also some were greatly positive, for example object 9740 emitted considerably more vanillin than was in the background, which possibly means that there is more lignin in this paper than in other papers within the studio. Interestingly, there is considerably more 5-butyldihydro-2(3H)-furanone (compound 22) in all six objects and in the order of twice the quantity of furfural (compound 4) in five of the objects (Table 3). This may reflect an increase in the rate of oxidation of the paper in contact with the oil in these objects. This possible phenomenon was investigated further with the ageing and comparison of paper mock-ups containing oil and not containing oil. 4.2. Comparison of artificially aged test pieces of paper with and without oil media 4.2.1. Examination of the test papers The ‘Cotton paper’ is made of 100% pure cotton linters, it is unbuffered, with no fillers or sizing. ‘Montval paper’, is an acid free/buffered wood pulp based artists watercolour paper, and was found to contain soft wood fibres, fillers and additives, and limited lignin content. The ‘Kraft paper’ is an un-bleached, wood based Kraft processed paper, mainly made of soft and hard wood fibres, that gave a strongly positive result for lignin. It is

Fig. 7. Average tear resistant values (10) of the plain and oiled cotton mock-ups before ageing and after 28 days of ageing.

buffered, contains fillers and additives and has a considerable presence of metal contamination from papermaking machinery, as indicated by elemental analysis. This selection represents basic types of paper supports commonly used by artists, but also have fibre content and characteristics similar to some of the works of art from the National Gallery in Athens being investigated in this project. 4.2.2. pH changes Although the plain paper test pieces present minor changes upon the progress of ageing, the application of the oil medium has an immediate effect on the acidity of the paper, since the pH values drop significantly from the first day of ageing (Table 4). Probably both the hydrolysis of oil to free fatty acids and oxidation of the cellulose to give carboxyl groups contribute to the lowering of the pH. Low pH values have also been recorded on the areas of paper showing discolouration due to oil absorption in the original works of art (Table 1). The average decrease of pH value compared with non-discoloured paper within the same work was 1.5. 4.2.3. Changes of the mechanical properties The effect of the linseed oil on the deterioration of the paper was also indicated by the changes recorded in the mechanical changes

Table 5 Emissions from cotton paper, wood based watercolour (Montval) paper and wood based kraft paper with and without a coating of linseed oil, in ng of emitted compound per g of paper as mean of two (cotton paper) or three (Montval and Kraft papers) samples for each period of ageing. Compound emitted→

Furfural

Paper type and Days of ageing↓

With oil

Cotton 1 Cotton 4 Cotton 7 Cotton 14 Cotton 21 Cotton 28 Montval 1 Montval 4 Montval 7 Montval 14 Montval 21 Montval 28 Kraft 1 Kraft 4 Kraft 7 Kraft 14 Kraft 21 Kraft 28 a

4600 6200 7300 5900 6100 620 2200 5700 4200 7700 12000 7600 2200 4100 6300 6000 6400 5900

5-methylfurfural

5-ethylfurfural

5-pentylfuranone

Without oil

2-ethylfuran With oil

Without oil

With oil

Without oil

With oil

Without oil

With oila

Without oila

0 41 170 430 0 120 53 19 54 170 140 820 88 160 300 570 970 810

1300 670 11000 29000 97 96 0 0 0 0 0 0 460 94 60 0 0 0

0 40 11 29 9 8 13 7 11 12 9 0 0 0 0 0 0 0

1600 810 630 490 280 82 1100 0 0 290 300 190 880 210 200 120 110 220

0 0 0 77 0 0 0 16 0 9 5 0 0 0 0 0 0 0

4600 1100 770 400 170 75 2200 1200 430 360 210 74 2300 240 120 47 15 31

0 0 0 0 0 0 0 2 0 0 0 0 0 0 0 0 0 0

14000 4400 3500 2700 630 480 3400 2500 590 160 220 120 0 0 0 0 0 0

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

5-pentyl furanone figures estimated from peak area using average response rate of other four standard compounds.

Please cite this article in press as: P. Banou, et al., The effect of oil binders on paper supports via VOC analysis, Journal of Cultural Heritage (2016), http://dx.doi.org/10.1016/j.culher.2016.01.003

G Model CULHER-3025; No. of Pages 10 8

ARTICLE IN PRESS P. Banou et al. / Journal of Cultural Heritage xxx (2016) xxx–xxx

of the oiled test pieces upon ageing. Tear resistance measurements in plain and oiled cotton paper mock-ups before ageing indicated that paper mock-ups become stronger after the application and drying of the oil for 40 days (Fig. 7). It could be suggested that as the fibres bonded within the dried elastic polymer, extra strength is provided to the support. Tear resistant measurements in plain and oiled cotton paper mock-ups after 28 days of ageing, indicated that mechanical strength of the oiled mock-ups reduces dramatically, while for the plain paper that is limited (Fig. 7). Shrinkage and loss of elasticity of the oil film and its recess in the fibre net of the paper upon ageing could not fully explain the severe effect on the mechanical properties of the paper. The reduction of mechanical strength of the support is better explained by increased degradation of the paper itself, perhaps by hydrolysis of cellulose in the low pH environment provided by the deteriorating oil and paper oxidation products. Breaking supports has also been recorded in heavily discoloured areas of original artworks where there is oil binder absorption. 4.2.4. Results of VOC analysis of aged paper samples without and with oil medium The outcome of the analysis demonstrates the effect of drying oil binders, using cold pressed linseed oil as an example, on the rate of degradation of paper supports used in works of art. The emissions of the five compounds from the cotton paper within 28 days of ageing without oil are extremely small, or none in the cases of 2-ethyl furan and 5-pentyl furanone (Table 5). Small quantities of furfural and 5-methyl furfural were emitted by the wood based papers without oil, while emissions of 2-ethyl furan, 5-pentyl furanone and 5-ethyl furfural within the 28 day period could not be found or were insignificantly low (Table 5). In contrast, emissions of the five target compounds from all three paper types with oil impregnating the paper were mostly significantly larger. One exception was that the emission of 2-ethyl furan from the ‘Montval’ wood based watercolour paper was in insignificant amounts whether the paper was impregnated with oil or not (Table 5). Furfural was a major VOC emission from the humid aged papers impregnated with oil. This was not surprising since furfural has been widely reported as a cellulose degradation product and since we had found it emitted from works of art on paper containing oil-based media. Furfural is emitted fairly steadily over the first 12 days from the cotton paper impregnated with oil, and the emission peaks at this point dropping to a low level (Table 5, Fig. 8). We must assume that the compound is being consumed in another reaction. Furfural is emitted in increasing amounts over the first 21 days from the ‘Montval’ paper impregnated with oil and then drops to a value less than the peak emission level at 28 days (Table 5, Fig. 8). The mean furfural peak emission for the ‘Montval’ paper with oil was 11000 ngg−1 , whereas for the cotton paper the mean peak emission was approximately two thirds of this value, at 7300 ngg−1 , implying that the cellulose in the chemical wood pulp based watercolour paper is degrading at a faster rate than in the almost pure cellulose cotton paper. A similar pattern in furfural emissions was obtained for the Kraft paper but with the highest emission recorded at 21 days of only 6400 ngg-1 (Table 5, Fig. 8), which is lower than the cotton paper. The significance of this difference in results between the three papers is not clear especially as the main difference in the papers is that the two chemical pulp based papers contained at least some lignin, the Kraft paper having indicated a large proportion of lignin when tested with phloroglucinol. Also lignin was indicated in Fourier transform infrared reflectance (FTIR-ATR) spectra of both the wood based papers especially in the Kraft paper. However, it has been suggested that lignin has no effect on cellulose degradation if the paper is well buffered with calcium carbonate [32] or even that lignin content below 28% has no effect on the

Fig. 8. Graph showing emissions of furfural in ng per g of paper over 28 days of humid ageing, from: a: cotton paper impregnated with linseed oil with cotton paper without any oil; b: form wood based artists watercolour paper impregnated with linseed oil compared with Montval, wood based watercolour paper without any oil, and; c: from Kraft paper impregnated with linseed oil compared with Kraft paper without any oil. (The points plotted for: a: are the average of two results and for; b and c: are the average of three results). Trend lines shown are polynomial order 2 and bars show highest and lowest values.

rate of degradation of cellulose in humid ageing tests [33], and that lignin may even protect paper from oxidative degradation of cellulose, at least in light induced reactions, by removing free radicals [34]. The cotton paper is unbuffered, while both wood based papers present Ca and Mg in EDX analysis indicating alkaline additives (buffer). The Montval paper is also described as acid free and the pH is higher than 7. Hence we could speculate that either different concentrations of buffer or different levels of cellulose or hemicelluloses in the cotton and Montval and Kraft papers are affecting the result rather than lignin content. We might, in addition, speculate that the higher lignin content in the Kraft paper is protecting the cellulose from thermal oxidative degradation by scavenging free radicals.

Please cite this article in press as: P. Banou, et al., The effect of oil binders on paper supports via VOC analysis, Journal of Cultural Heritage (2016), http://dx.doi.org/10.1016/j.culher.2016.01.003

G Model CULHER-3025; No. of Pages 10

ARTICLE IN PRESS P. Banou et al. / Journal of Cultural Heritage xxx (2016) xxx–xxx

9

We must assume that these compounds are rapidly consumed after production in the first day or two of ageing as they are converted to another compound. A similar result is shown by 5-pentyl furanone for cotton and the wood based watercolour paper where it seemed to be emitted after one day in considerable quantities (Table 5), [as might be expected since it was found to be emitted from the art objects (compound 24, Table 2)], but the concentration of this compound then declines rapidly on further aging. However, this compound did not seem to be emitted by the Kraft paper in measurable quantities (Table 5). The route from cellulose to furfural is generally thought to involve heterolytic mechanisms involving acid catalysed hydrolysis [13,24] and hence acidity produced by hydrolysis of the oil and free radical oxidation of the cellulose to carboxylic acids probably both contribute, as indicated by our measured pH values (Table 4), which drop as furfural emissions increase in the oiled papers. Free radical reactions in the degrading oil may in turn increase the rate at which oxidation occurs in the paper and hence contribute further to the occurance of acidic paper oxidation products.

5. Conclusions

Fig. 9. Graph showing emissions of 5-ethyl-2-furfural estimated as ng per g of paper over 28 days, from: a: cotton paper impregnated with linseed oil compared with cotton paper without any; b: from Montval wood based watercolour paper impregnated with linseed oil compared with Montval, wood based watercolour paper without any oil; c: from Kraft paper impregnated with linseed oil compared with Kraft paper without any oil. (The points plotted in: a: are the average of two results and in; b and c: the average of three results). The trendline shown is of the form y = kx−n and bars show highest and lowest values.

The results in the graphs for these papers are the mean of two or three results with the error bars showing the highest and lowest value. The errors are quite large but, even so, it can be clearly seen that emission levels of furfural for the papers impregnated with oil (Fig. 8) are significantly raised above the emission levels for the same papers without oil (Fig. 8) within the 28-day humid ageing period. 5-methyl furfural and 5-ethyl furfural present a similar pattern of emissions to each other when emitted from the cotton paper impregnated with oil, and also from the wood based watercolour paper and Kraft papers impregnated with oil, where they peak on the first day of ageing and decline rapidly to only a very low level by 28 days (Table 5). This is illustrated for 5-ethyl furfural (Fig. 9).

The presence of drying oil in paper greatly accelerates the emission of volatile cellulose degradation products both for humid aged cotton based and wood based papers. In particular, the presence of linseed oil in the cotton paper has greatly accelerated the emission of all five target furan derivatives: furfural, 2-ethyl furan, 5-methyl furfural, 5-ethyl furfural and 5pentyl furanone during ageing, while in the chemical wood pulp based watercolour paper has accelerated the emission of all target compounds except 2-ethyl furan and in the Kraft paper it has greatly accelerated the emission of furfural and increased the emission of 2-ethyl furan, 5-methyl furfural and 5-ethyl furfural but not 5-pentyl furanone (Table 5). It is reasonable to assume, that since the chemistry of other drying oils are similar to that of linseed oil, other drying oils may also increase the rate at which cellulose in paper degrades. However, contaminants, additives, artist’s interventions, plant source, oil blends and oxidation/drying rates may well affect the extent of this effect. SPME analysis of original artworks containing oils gave a similar range of volatile oxidation products to the products obtained from the test pieces. These products in common included a large range of aromatic and straight chained aliphatic hydrocarbons, a series of straight-chained aldehydes, some volatile carboxylic acids and esters and furfural and 5-pentyl furanone. It seems reasonable to suggest therefore, that our humid aged test pieces were a good model for studying the effect of drying oil on cellulose oxidation in paper and therefore that the presence of drying oil in paper greatly accelerates the degradation of cellulose in works of art on paper. This was illustrated by the SPME results and backed up by pH and mechanical strength measurements. Since emissions from oil impregnated papers were in the order of 4 to 10 times those from non-oiled paper aged for the same period, in the same conditions, it might be reasonable to suggest that the works in the collections in which oil has penetrated the paper, may have considerably shortened lifetimes. However, it would be difficult to put a figure on this, since the occurrence and extent of damage depend on multiple parameters and interactions. The issue is further complicated by the fact that historic paper can be both a source and a sink of VOCs’ [35]. Our results clearly contribute to the assessment of the conditions of the Artworks in the collection studied here. Current treatments for similar works have been discussed [36]. Possible future treatments will be the focus of our further research.

Please cite this article in press as: P. Banou, et al., The effect of oil binders on paper supports via VOC analysis, Journal of Cultural Heritage (2016), http://dx.doi.org/10.1016/j.culher.2016.01.003

G Model CULHER-3025; No. of Pages 10 10

ARTICLE IN PRESS P. Banou et al. / Journal of Cultural Heritage xxx (2016) xxx–xxx

Acknowledgments The authors wish to thank the Archimedes III Project for funding and also the TEI of Athens, the NTUA of Athens and the National Gallery – Alexandros Soutzos Museum, Athens, Greece for their help and co-operation in this research. References [1] P. Banou, A.A. Kaminari, A. Moutsatsou, A. Alexopoulou, B.W. Singer, Investigation of the conservation problems of oil paintings on paper supports, in: Works of Art & Conservation Science Today November 26–28, 2010, Aristotle University of Thessaloniki/University of Ecclesiastical Academy of Thessaloniki (Greece), contribution 03 in the CD of full papers, 2010. [2] J. Kosek, L.R. Green, A survey of oil paintings and sketches on paper in the collection in the British Museum: an assessment of stability, in: S. Fairbrass (Ed.), The IPC: Conference papers Manchester, 1992, pp. 96–102. [3] J. McAusland, Problems with works of art in oil on paper and their support: a paper conservator’s viewpoint, General Meeting of the Association of British Picture Restorers, UK, 1989, pp. 1–6. [4] C. Engram, Oil painting on paper and board supports, AICCM Bull. 17 (3, 4) (1991) 13–20. [5] C. James, C. Corrigan, M.C. Enshaian, M.R. Greca, Old Master prints and drawings, translated and edited by Marjorie B. Cohn, University Press, Amsterdam, 1997, pp. 85. [6] T. Millard, E. Le Cornu, R. Smith, E. Hasler, H. Cowdy, R. Chisholm, E. King, The conservation of 830 oil paintings on paper by Marianne North, J. Inst. Conserv. 34 (2) (2011) 159–172. [7] F.S. Hanson, Resistance of paper to natural aging, Pap. Ind. Pap. World 20 (1939) 1157–1163. [8] V. Daniels, The discoloration of paper on ageing, Pap. Conserv. 12 (1988) 93–100. [9] V. Daniels, ‘Oxidative damage and the preservation of organic artefacts’, Free Radical Res. Commun. 5 (4–5) (1989) 213–220. [10] O.M. Elfecky, M.L. Hassan, Artificial ageing and deterioration of oil-painted Fabriano paper and cardboard paper supports, J. Appl. Polym. Sci. 109 (2008) 1594–1603. [11] D.K. Shen, S. Gu, The mechanism for thermal decomposition of cellulose and its main products, Bioresour. Technol. 100 (2009) 6496–6504. [12] I.S. Fagerson, Thermal degradation of carbohydrates, a review, J. Agric. Food Chem. 17 (1969) 747–750. [13] J. Scheirs, G. Camino, M. Avidano, W. Tumiatti, Origin of furanic compounds in thermal degradation of cellulosic Insulating Paper, J Appl Polym Sci 69 (1998) 2541–2547. [14] A.M. Emesley, G.C. Stephens, Review of chemical indicators of degradation of cellulosic electrical paper insulation in oil-filled transformers, IEE ProceduresScientific Measurement, Technology 141 (1994) 324–334. [15] D.H. Shroff, Moisture in transformer insulation, in: Proceedings of the Conference on Diagnostic Testing of High Voltage Power Apparatus in Service, Institution of Electrical Engineers, Publication No. 94, Part I, 1973, p. 126. [16] O. Ramalho, A.L. Dupont, C. Egasse, A. Lattuati-Derieux, Emission rates of VOC from paper versus cellulose degradation: an integrated approach to paper characterization, e-Preserv. Sci. 6 (2009) 53–59. ´ A. Levart, L. Gibson, G. de Bruin, K. Ntanos, J. [17] A. Fenech, M. Strliˇc, I. Kralj Cigic, Kolar, M. Cassar, Volatile aldehydes in libraries and archives, Atmos. Environ. 44 (2010) 2067–2073.

[18] L.T. Gibson, A. Ewlad-Ahmed, B. Knight, V. Horie, G. Mitchell, C.J. Robertson, Measurement of volatile organic compounds emitted in libraries and archives: an inferential indicator of paper decay? Chem. Central J. 6 (1) (2012) 42. ´ J. Kolar, M. Cas[19] M. Strliˇc, J. Thomas, T. Trafela, L. Cséfalvayová, I. Kralj Cigic, sar, Material degradomics: on the smell of old books, Anal. Chem. 81 (2009) 8617–8622. [20] A. Lattuati-Derieux, S. Bonnassies-Termes, B. Lavédrine, Identification of volatile organic compounds emitted by a naturally aged book using solid-phase microextraction/gas chromatography/mass spectrometry, J. Chromatogr. A 1026 (2004) 9–18. [21] A. Lattuati-Derieux, S. Bonnassies-Termes, B. Lavédrine, Characterisation of compounds emitted during natural and artificial ageing of a book. Use of headspace-solid-phase microextraction/gas chromatography/mass spectrometry, J. Cultur. Herit. 7 (2006) 123–133. [22] A.J. Clark, J.L. Calvillo, S. Mark, M.S. Roosa, D.B. Green, J.A. Ganske, Degradation product emission from historic and modern books by headspace SPME/GC–MS: evaluation of lipid oxidation and cellulose hydrolysis, Anal. Bioanal. Chem. 399 (2011) 3589–3600. ˛ [23] T. Łojewski, T. Sawoszczuk, J.M. Łagan, K. Zieba, A. Baranski, J. Lojewska, Furfural as a marker of cellulose degradation. A quantitative approach, Appl. Phys. A 100 (2010) 873–884. ´ J. Kolar, G. De Bruin, B. Pihlar, Non-destructive evaluation of [24] M. Strliˇc, I.K. Cigic, historical paper based on pH estimation from VOC emissions, Sensors 7 (2007) 3136–3145. [25] L.T. Gibson, A. Ewlad-Ahmed, B. Knight, V. Horie, G. Mitchell, C.J. Robertson, Measurement of volatile organic compounds emitted in libraries and archives: an inferential indicator of paper decay? Chem. Central J. 6 (2012) 42 (22 pages). [26] L. Greenspan, Humidity fixed points of binary saturated aqueous solutions, Res. National Bureau Stand. 81A (1977) 89–96. [27] D. Sheehan, Physical Biochemistry, Wiley Blackwell, 2009, pp. 285. [28] X. Zou, T. Uesaka, N. Gurnagul, Prediction of paper permanence by accelerated aging, I., Kinetic analysis of the aging process, Cellulose 3 (1996) 243–267. [29] A. Piccirillo, D. Scalarone, O. Chiantore, Comparison between off-line and online derivatisation methods in the characterisation of siccative oils in paint media, J. Anal. Appl. Pyrolysis 74 (2005) 33–38. [30] F.C. Izzo, 20th century artists’ oil paints: a chemical-physical survey (PhD Thesis), Università Ca’ Foscari Venezia, Scuola di dottorato in Scienze e Tecnologie, 2011 (accessed 16.01.16) http://www.dspace.unive.it/handle/10579/1100. [31] T. Vanek, J. Halík, R. Vanková, I. Valterová, Formation of trans-verbenol and verbenone from alpha-pinene catalysed by immobilised Picea abies cells, Biosci. Biotechnol. Biochem. 69 (2) (2005) 321–325. [32] P. Begin, S. Deschatelets, D. Grattan, N. Gurnagul, J. Iraci, E. Kaminska, X. Zou, The impact of lignin on paper permanence: a comprehensive study of the aging behaviour of handsheets and commercial paper samples, Restaurator 19 (1998) 135–154. [33] N. Hon, Formation of free radicals in photoirradiated cellulose. VI. Effect of lignin, J. Polym. Sci. Polym. Chem. 13 (1975) 2641–2652. [34] X. Zou, N. Gurnagul, T. Uesaka, The role of lignin in the mechanical permanence of paper. Part I: effect of lignin content, J. Pulp Pap. Sci. 19 (1993) 235–239. [35] M. Strlic a, I.K. Cigi, A. Mo, Z. Gerrit de Bruin, J. Kolar, M. Cassar, The effect of volatile organic compounds and hypoxia on paper degradation, Polym. Degrad. Stab. 96 (2011) 608–615. [36] P. Banou, A. Alexopoulou, B.W. Singer, The treatment of oil paintings on paper supports – Considerations on the treatment applications used from the past until the present, J. Pap. Conserv. 16 (2015) 29–36.

Please cite this article in press as: P. Banou, et al., The effect of oil binders on paper supports via VOC analysis, Journal of Cultural Heritage (2016), http://dx.doi.org/10.1016/j.culher.2016.01.003