- Email: [email protected]
Alcoholic deacidification and simultaneous deacidification-reduction of paper evaluated after artificial and natural aging
G Model CULHER-3039; No. of Pages 8
ARTICLE IN PRESS Journal of Cultural Heritage xxx (2016) xxx–xxx
Available online at
ScienceDirect www.sciencedirect.com
Original article
Alcoholic deacidification and simultaneous deacidification-reduction of paper evaluated after artificial and natural aging Marina Bicchieri a,∗ , Armida Sodo a,b a b
Chemistry department, istituto centrale restauro e conservazione patrimonio archivistico e librario, 76, via Milano, 00184 Roma, Italy Dip. Scienze, università Roma Tre, 84, via della Vasca Navale, 00146 Roma, Italy
a r t i c l e
i n f o
Article history: Received 9 November 2015 Accepted 22 February 2016 Available online xxx Keywords: Paper Degradation Deacidification Reduction Conservation Non-aqueous treatments
a b s t r a c t Cellulose oxidative and hydrolytical degradation is one of the greatest problems for the conservation of paper supports. To contrast these degradation processes, both deacidification and reduction of the oxidized functions are needed. Dealing with original documents, it is often impossible to perform the two mentioned treatments in aqueous solutions and in a distinct subsequent way, because of the fragility of the artifacts. After studying, in a separate way, an effective deacidifier (calcium propionate) soluble in ethyl alcohol and many reducers (boron complexes), able to act in different non-aqueous solvents, it was decided to test a simultaneous method of deacidification and reduction in ethanol. This paper presents the chemical-physical results obtained by applying simple deacidification and simultaneous deacidification-reduction on laboratory paper samples that were artificially aged and then re-measured after 10 and 15 years of natural aging. Results show that all alcoholic treatments are very effective: papers are stable also after a long period of both artificial and natural aging. © 2016 Elsevier Masson SAS. All rights reserved.
1. Research aims This paper tries to answer to the needs of conservators of books, archival materials and graphic works of art, who need sometime to use non-aqueous solutions in treating original documents to prevent the dissolution of the graphic media, the alteration of the paper structure and the loss of the printing impression. Moreover, it is often necessary to work on bound books to maintain the original bindings untouched. Also in this case, water solutions could not be used. All the analyses and the used complementary techniques were applied in order to ascertain the effectiveness of a simultaneous non-aqueous deacidification and reduction treatment on artificially and naturally aged papers.
2. Introduction Cellulose oxidative and hydrolytical degradation is one of the greatest problems for the conservation of paper supports such as archival documents, books and graphic works. Degraded papers
∗ Corresponding author. Tel.: +390648291217; fax: +39064814968. E-mail addresses: [email protected] (M. Bicchieri), [email protected], sodo@fis.uniroma3.it (A. Sodo).
show an evident fragility and a yellowing, compromising in some cases the readability of the text or the graphic sign. Many deacidifier can be used [1], but only few reducers can be directly applied on paper documents without damaging the fiber structure and the graphic media. Boron compounds and complexes are very effective and chemoselective in reducing carbonyls that are the most important oxidized groups in the cellulose polymer chain [2–4]. During the reduction reactions, hydrogen is produced and, if the reaction rate is too high, the cellulose fiber wall could be broken [5] by the large amount of hydrogen generated in a short time. This effect has been observed when sodium borohydride, the first reducer used in conservation field, was used as reducing product in the treatment of strongly oxidized papers [6]. Therefore, to minimize this negative effect, mild and “slow” reducers should be chosen, such as the boron-amine complexes. Furthermore, to avoid -alkoxyelimination reaction, which occurs at ambient temperature in high alkaline environment (greater than 10.5) when a polymeric chain contains a large amount of carbonyl groups, the employed reducers should have a pH lower than 10 when applied in water solution [1]. In 1997, we published the first results on the application of borane tert-butylamine complex [(CH3 )3 CNH2 . BH3 , TBAB] for the reduction of carbonyl groups in oxidized paper [7]. Almost in the same period, TBAB was proposed for the bleaching of pulp in the paper industry. Amine boranes appeared to behave
http://dx.doi.org/10.1016/j.culher.2016.02.008 1296-2074/© 2016 Elsevier Masson SAS. All rights reserved.
Please cite this article in press as: M. Bicchieri, A. Sodo, Alcoholic deacidification and simultaneous deacidification-reduction of paper evaluated after artificial and natural aging, Journal of Cultural Heritage (2016), http://dx.doi.org/10.1016/j.culher.2016.02.008
G Model CULHER-3039; No. of Pages 8
ARTICLE IN PRESS M. Bicchieri, A. Sodo / Journal of Cultural Heritage xxx (2016) xxx–xxx
2
better than sodium borohydride and even than sodium hydrosulphite for obtaining a more stable, white and not-degraded cellulose pulp [8]. In a second time, we investigated the possibility of a nonaqueous application of different boron compound for reducing purposes [9] and the applicability of a simultaneous non-aqueous deacidification with ethanol solution of calcium propionate [Ca(CH3 CH2 COO)2 ] and reduction with borane-ammonia complex [NH3 BH3 ] [10]. Non-aqueous treatments are in fact requested in case of restoration of originals containing water-soluble graphic media or when it is mandatory to treat books keeping intact the bookbinding. Due to instability of borane-ammonia complex and the difficulties in its handling by restorers, we decided to perform a new series of studies on TBAB, whose interactions with cellulose are reported in [11], and to use it for a new series of tests on alcoholic deacidification and reduction. This paper presents the comparison between the data obtained on artificially aged samples and those collected on the same samples after natural aging in uncontrolled conditions for 10 and 15 years. Before presenting the results of this work and their discussion, some points need to be underlined. Looking at the specific literature [12 and related references], it seems that conservators are more interested in “bleaching” of paper – that is only a side effect of the reducing treatment – than in the most important result that is the stabilization of the cellulosic support and the increase of its expected life. Despite the researches carried out on different boron complexes, outside Italy only sodium borohydride is used, whose effect on strongly oxidized papers can be dramatic, causing the mechanical breaking of the paper [5]. Another reducer, sodium dithionite [12], has a mild brightening effect on paper, but is very effective in the reduction of iron (III) ions. This makes the dithionite not applicable to all manuscripts containing iron-based inks or pigments, which could disappear after the treatment. Moreover, the articles concerned with chemical methods for paper restoration often report results based on mechanical and optical measurements. Significant variations of mechanical properties need very long time (See supplementary material S1). On the contrary even light chemical variations can cause strong degradation. This implies that potentially harmful treatments – such as reduction with sodium borohydride or deacidification with calcium hydroxide at too high pH – are commonly accepted. For these reasons, we focused our attention on the chemical behavior of the investigated items to perform a rational and scientific acceptance or rejection of the treatment. Therefore, the characterization was carried out by applying chemical methods (pH, carbonyl content and degree of polymerization), spectroscopic techniques (Raman spectroscopy) and optical methods (color coordinates and Optical density measurements).
3. Materials and methods Materials, methods and techniques used in this work are briefly described below.
3.1. Materials • Whatman® cellulose chromatography paper grade 1; 46 cm × 57 cm, a pure cotton and cotton linters cellulose. • hydrochloric acid 37% [HCl] (Aldrich ACS grade) used to simulate “original” acidic samples subjected to hydrolytic degradation.
• 0.015 M aqueous solution of potassium meta-periodate [KIO4 ] (powder, Merck 99.8%) brought to pH 5 with HCl 10−2 M in order to simulate “original” samples subjected to oxidative degradation. • 3.0 g/L solution of calcium propionate [Ca(CH3 CH2 COO)2 ] (Aldrich 95%) in ethyl alcohol (Aldrich 96%). The solution was used to carry out simple non-aqueous deacidification treatments. • 3.0 g/L solution of calcium propionate in ethyl alcohol + borane tert-butylamine complex [(CH3 )3 CNH2 . BH3 ] (Aldrich 97%). The resulting solution is 0.2 M in borane complex. The solution was used to carry out simultaneous deacidification and reduction. • 0.2% w/V water solution of 2,3,5 triphenyl-2H-tetrazolium chloride (TTC, Aldrich 98%). • 0.5 M cupriethylenediamine solution (Carlo Erba for analysis). 3.2. Preparation of paper samples Three sets of laboratory paper samples were considered in the present work; each one was successively divided in three series, as described below: • one set was oxidized by immersion for 15 minutes in a KIO4 0.015 M water solution, at pH = 5.0, and then rinsed by two immersions, 5 minutes each, in distilled water. This treatment was used to simulate “original” papers presenting a strong degradative oxidation. After the treatment, one series was kept as prepared (Ox series in the following); a second one was immersed in the alcoholic deacidification solution of calcium propionate (Ox-D series in the following) as described in a previous work [10] so to have a comparison with established restoration treatments, and the third one underwent a simultaneous non aqueous deacidification and reduction treatment (Ox-DR series in the following); • one set was hydrolyzed by 1 h exposition to fuming HCl vapors, to simulate acidic “original” documents. After hydrolysis, a series was kept as prepared (Hy series in the following), a series was deacidified (Hy-D series in the following) and the third one deacidified and reduced (Hy-DR series in the following); • one set of Whatman N.1 samples (not treated, deacidified and deacidified/reduced) was included in the experiment as control samples (respectively C, C-D and C-DR series in the following). All laboratory samples were artificially aged in climate chamber (Angelantoni Challenge E300) for 35 days, according to International Organization for Standardization ISO 5630-3:1996 at 65% R.H. and 80 ◦ C and characterized each 7 days of ageing with destructive and non-destructive techniques. The chosen temperature is compatible with the stability of TBAB: amine-boranes prepared from primary and secondary amines are stable up to 110 ◦ C [13]. After the artificial ageing experiment, for each set and series, the “zero-time” samples and those artificially aged for 35 days, were left on library shelves for 10 years in order to simulate a non-controlled storage. After their retrieval, the samples were once more characterized and left for other 5 years in uncontrolled ambient for a third characterization. Chemical, optical and spectrometric tests were performed after each cycle of artificial and natural aging. Oxidation and hydrolysis induced by environmental conditions, in fact, cause important chemical and structural modifications with great impact on paper permanence and durability, even if mechanical characteristics are not particularly affected in the short term. Decrease in the degree of polymerization is related to hydrolytic degradation, whereas increase in carbonylic functions is due to oxidative processes. Structural changes produce modification in the Raman spectra of paper.
Please cite this article in press as: M. Bicchieri, A. Sodo, Alcoholic deacidification and simultaneous deacidification-reduction of paper evaluated after artificial and natural aging, Journal of Cultural Heritage (2016), http://dx.doi.org/10.1016/j.culher.2016.02.008
G Model
ARTICLE IN PRESS
CULHER-3039; No. of Pages 8
M. Bicchieri, A. Sodo / Journal of Cultural Heritage xxx (2016) xxx–xxx
3.3. Methods of testing • Acidity measurements were performed according to standard TAPPI T 435 om-11. • Viscometric average degree of polymerization (DP, dimensionless number) was measured following American Society for Testing and Materials (ASTM) according to the standard ASTM D1795-13. • Carbonyl content measures were performed according to a method developed in our Institute and described in a previous paper [7]. • Color coordinates were measured by means of a colorimeter (Minolta Chroma Meter CR22) in the CIE L*a*b* space. • Optical density: the combined processes of absorption, scattering and reflection influence the “transmittance” (T, dimensionless number), which accounts for the residual light transmitted from sample in the forward direction. In order to disentangle the contribution from reflection, absorption, and scattering, it is necessary to measure the “transflectance” (TF, dimensionless number). TF is the fraction of light emerging in all directions from the sample and allows sample’s optical density (OpD, dimensionless number) to be derived directly, following the equations: OpD = −log (T + R) = −log TF, where R is the reflectance of the sample (R, dimensionless number). Measurements were carried out in the 200–2500 nm range, by means of a double-beam spectrophotometer (Varian Cary 5), equipped with calibrated SpectralonTM integrating sphere (LabSphere) and internal sample holder (LabSphere). By positioning the sample at the centre of the sphere, the transflectance is measured. The empty sphere spectrum is assumed as reference, so the experimental optical density is achieved as OpD = −log (TF/TFsphere ). • Raman measurements have been performed by means of a Renishaw In-Via Reflex Raman microscope equipped with a Renishaw diode laser at 785 nm (output power 300 mW). The backscattered light is dispersed by a 1200-line/mm grating and the Raman signal is detected by a Peltier cooled (−70 ◦ C) deep depletion charge-coupled device (CCD RD-VIU, 578 pixel × 384 pixel) optimized for near infrared and ultraviolet. The nominal spectral resolution obtained for the measurements is about 3 cm−1 . The system, equipped with a Leica DM LM microscope to focus the laser on the sample and a color video camera, allows for the positioning of the sample and the selection of a specific region for the investigation. Spectral acquisitions (1–10 accumulations, 50 s each) have been performed with a 50× objective (N.A. 0.75). Under these conditions, the laser spot measures about 20 m2 . 4. Results and discussion 4.1. Acidity measurements The values of pH by hot extraction method (Table 1) show that all the samples subjected to the treatments involving Table 1 pH (±0.05) variations along aging time (average of four measurements on each sample; five samples for each kind of paper and treatment). Sample
0 day
35 days (accelerated aging)
10 years (natural aging on originally 0 day sample)
15 years (natural aging on originally 0 day sample)
C C-D C-DR Ox Ox-D Ox-DR Hy Hy-D Hy-DR
5.37 8.89 9.08 5.52 8.57 8.84 4.90 8.29 8.20
5.64 8.91 8.84 5.40 8.28 8.26 4.96 8.47 8.19
5.60 7.43 8.37 5.81 7.39 8.65 4.40 7.92 8.04
5.55 7.26 8.42 5.63 7.27 8.60 4.01 7.58 8.00
3
deacidification or combined deacidification-reduction remain alkaline, in the range of the maximum stability of the cellulose, also after prolonged accelerated and natural aging. The best stabilization against the pH variation is obtained with the combined treatment of reduction and deacidification, as evidenced by the data of Table 1. 4.2. Viscometric average degree of polymerization The preparation of the cellulose solutions and the measurement of their intrinsic viscosity in cupriethylenediamine were performed in accordance with ASTM D1795-13 with some modifications. The intrinsic viscosity [] of all the samples was not obtained from listed tables (ASTM Method) but by extrapolating at zero concentration the values of Log [(rel -1)/c] as a function of c (rel is the relative viscosity of the solution of paper in cupriethylenediamine hydroxide and c is the concentration of the cellulose solution expressed in g/dL). The straight-line parameters were obtained using a leastsquare-fit method, taking into account the proper error for each measurement (5 concentration for each sample; 4 measurement for each concentration). By multiplying intrinsic viscosity by a specific constant K (cm3 .g−1 ) it is possible to obtain the viscosity average degree of polymerization DP (dimensionless number). Studies on the behavior of cellulose [14,15] have shown that the constant K assumes different values in relation to the degree of polymerization of the cellulose: K = 156 for DP between 3000 and 300 and K = 124 for DP below 300. To obtain statistically valid data, the value of intrinsic viscosity was extrapolated from a series of measurements at different concentration. Errors associated with each individual measurement were obtained by applying the least square method with variable error. The results of the measurements are given in Table 2A. Data of the measures carried out after 15 years of natural aging both for the samples not subjected to the accelerated aging and for those 35 days artificially aged are also shown. The comparison would give indications on the effects and the differences between natural and artificial aging. Fig. 1 reports the results of DP measurements for the natural aging, also on the samples previously artificially aged for 35 days. It can be easily seen that the artificial aging accelerates of a greater extent the hydrolytic degradation, in respect to the natural one. The higher values of all the DR samples in respect to their control samples (i.e. C, Ox and Hy respectively) are due to high pH value of the cupriethylenediamine (pH around 13). At this pH, cupriethylenediamine does not behave as a good solvent: high pH values induces -alkoxyelimination reactions when the samples contain oxidized functions. The maximum difference between the control and its respective DR sample is, in fact, observed for the oxidized series. Moreover, the results underline that the viscometric DP values closest to the “true” DP of any kind of paper can be obtained only by measuring the samples after a reduction treatment that removes the oxidized functions and allows to the cupriethylenediamine to behave as a good solvent without any interference with the samples under analysis. It is noticeable that the DP values of all the samples treated with the combined deacidification and reduction treatment are, also after 15 years of natural aging, comparable with those of the not treated and not aged samples, this meaning that the treatment offered a huge protection against the hydrolytic degradation. For the hydrolyzed series, the initial hydrolytic degradation had been so intense since the beginning that the unique effect of the two treatments – deacidification and in particular combined one – was to obstacle a further degradation, keeping constant the initial DP level.
Please cite this article in press as: M. Bicchieri, A. Sodo, Alcoholic deacidification and simultaneous deacidification-reduction of paper evaluated after artificial and natural aging, Journal of Cultural Heritage (2016), http://dx.doi.org/10.1016/j.culher.2016.02.008
G Model
ARTICLE IN PRESS
CULHER-3039; No. of Pages 8
M. Bicchieri, A. Sodo / Journal of Cultural Heritage xxx (2016) xxx–xxx
4
Table 2 A: DP variations; B: carbonyl content (m.moles/100 g paper) along aging time. Aging Sample
2A: pH C C-D C-DR Ox Ox-D Ox-DR Hy Hy-D Hy-DR 2B: C = O C C-D C-DR Ox Ox-D Ox-DR Hy Hy-D Hy-DR
0 days
35 days (acc. aging)
1189 1429 1397 830 998 1096 238 288 281
± ± ± ± ± ± ± ± ±
10 15 12 13 5 37 9 4 9
0.46 0.20 0.20 0.55 0.20 0.25 4.50 0.65 0.28
± ± ± ± ± ± ± ± ±
0.01 0.01 0.01 0.01 0.01 0.01 0.40 0.07 0.01
442 670 1115 329 828 1006 216 287 319
± ± ± ± ± ± ± ± ±
16 5 18 12 3 16 4 5 12
1.14 0.21 0.24 4.90 1.89 0.23 5.00 0.51 0.19
± ± ± ± ± ± ± ± ±
0.02 0.03 0.01 0.50 0.20 0.01 0.50 0.02 0.01
10 years (natural aging)
15 years (natural aging)
Originally 0 days
Originally 0 days
624 1107 1233 634 940 1130 232 281 301
± ± ± ± ± ± ± ± ±
3 41 15 10 39 13 4 2 15
0.95 0.43 0.30 7.80 1.60 0.45 5.04 0.86 0.39
± ± ± ± ± ± ± ± ±
0.01 0.01 0.04 0.51 0.02 0.01 0.07 0.05 0.01
4.3. Carbonyl content Carbonyl groups can exist in paper both as end-groups in each macromolecule of cellulose and on the anhydrous glucose ring, after oxidation.
Originally 35 days
580 1126 1200 483 867 1100 220 264 280
± ± ± ± ± ± ± ± ±
9 20 10 7 7 20 10 9 7
400 516 1100 300 789 1000 200 231 300
± ± ± ± ± ± ± ± ±
12 16 15 10 9 8 6 7 10
1.02 0.51 0.32 9.17 1.78 0.70 5.86 0.98 0.45
± ± ± ± ± ± ± ± ±
0.03 0.01 0.05 0.12 0.03 0.02 0.15 0.02 0.01
1.08 0.50 0.38 11.63 1.92 0.80 7.01 1.20 0.56
± ± ± ± ± ± ± ± ±
0.08 0.01 0.01 0.16 0.08 0.01 0.20 0.05 0.01
These groups can be measured by redox reaction with triphenyltetrazolium chloride salts. In an alkaline environment, carbonyls are able to reduce 2,3,5-triphenyl-2H-tetrazolium chloride to the corresponding triphenylformazan of a red color. The amount of formazan formed is directly proportional to the amount of carbonyl groups present, and it is spectrophotometrically determined at 480 nm [7,16]. The results obtained for all the different samples are summarized in Table 2B and plotted in Fig. 2, every value being the average of five measurements on each sample. Five samples were used for each kind of paper and treatment. It should be highlighted that the experimental data show that the carbonyl content of all the DR samples, even after 15 years of natural ageing, is about of the same value of that measured on the not aged C samples. This evidences the goodness of the reduction treatment and its effectiveness in the contrast of the oxidative degradation, promoting the stabilization of the paper over time. 4.4. Color coordinates The most representative coordinates related to the modifications induced by aging are L* and b*. A decrease in L* reduces the lightness of the sample, while an increase of b* shifts to yellow the color of the paper. CIE L*a*b* data (See supplementary material S2) present an increase of the b* coordinate for all the samples, in particular for those not subjected to any deacidification or combined deacidification-reduction treatment. At the same time, L* slightly decreases, as predictable after any aging (Figs. 3 and 4). 4.5. Optical density
Fig. 1. Average viscometric degree of polymerization for the natural aging. The 15 years data are related to the natural aging of the originally 0 and 35 days accelerate ageing samples.
Fragmentation and instable oxidation products cause the formation of yellow chromophores and fluorescent states. UV-Vis absorption spectra of paper are affected by these phenomena since both electronic transition and fluorescence bands can be found in the 200–500 nm spectral region and the interpretation must be done cautiously [17,18]. In particular, as it can be seen from Fig. 5, all the three sets of laboratory samples showed after 15 years of natural aging a peak at
Please cite this article in press as: M. Bicchieri, A. Sodo, Alcoholic deacidification and simultaneous deacidification-reduction of paper evaluated after artificial and natural aging, Journal of Cultural Heritage (2016), http://dx.doi.org/10.1016/j.culher.2016.02.008
G Model CULHER-3039; No. of Pages 8
ARTICLE IN PRESS M. Bicchieri, A. Sodo / Journal of Cultural Heritage xxx (2016) xxx–xxx
5
Fig. 2. Carbonyl content of the samples for the natural aging. The 15 years data are related to the natural aging of the originally 0 and 35 days accelerate ageing samples.
around 205 nm with a shoulder at about 260 nm and a tail extending in the blue region till to around 500 nm. This region is of interest, since, according to ISO 2470-1:2009, yellow chromophores content in paper is quantified by an inverse relation in respect to the brightness measured as UV absorbance at 457 nm.
Fig. 3. L* coordinate of all samples before and after 10 and 15 years of natural ageing.
The above-mentioned features are more intense in the Hy sample with respect to the C and Ox samples. Moreover, the Hy sample shows higher absorbance in the blue region, thus confirming the shift to yellow of the color of the paper. Both the deacidification and the simultaneous deacidification and reduction treatments specifically affect the intensity of the shoulder at around 260 nm. The feature becomes less intense in
Fig. 4. b* coordinate of all samples before and after 10 and 15 years of natural ageing.
Please cite this article in press as: M. Bicchieri, A. Sodo, Alcoholic deacidification and simultaneous deacidification-reduction of paper evaluated after artificial and natural aging, Journal of Cultural Heritage (2016), http://dx.doi.org/10.1016/j.culher.2016.02.008
G Model CULHER-3039; No. of Pages 8
ARTICLE IN PRESS M. Bicchieri, A. Sodo / Journal of Cultural Heritage xxx (2016) xxx–xxx
6
Fig. 6. NIR spectra of C, Ox and Hy samples aged 15 years.
Fig. 5. UV-Vis spectra of samples aged 15 years. Each series of control paper is compared with the respective deacidified and deacidified – reduced series.
all D and DR series, and – consequently – the absorbance in the 400–500 nm region (blue) is reduced accordingly, and contrasts the yellowing effect. This spectral evidence is also confirmed by the lower values of the b* color coordinate of D and DR series for the same set of samples (Supplementary materials S2, column 15 years). Taking into account that the cellulose chromophores at around 260 nm are usually interpreted in terms of carbonyl and carboxyl formation [17], it can be inferred that the two conservation treatments positively contrast their formation, contributing to the preservation of paper. The NIR part of the spectrum is dominated by water/OH features but it is possible to observe, between 1700 nm and 2200 nm, some carbonyl/carboxyl stretching overtones [19] (Fig. 6). The variation in the intensity of overtones, due to the increase of the number of oxidized groups, is much less pronounced in NIR than in the midIR region and the analysis of NIR data is typically performed with multivariate methods, using a great variety of data [20]. The NIR analysis along aging time was therefore not considered relevant in this work: an enhancement of the carboxyl C = O and C-O stretching overtones peaks appeared in fact quite well visible only for the hydrolyzed sample Hy after the longer natural aging time (Fig. 6). 4.6. Raman In previous studies, Raman spectroscopy was applied to investigate the state of conservation of differently degraded papers [21]. As all vibrational spectroscopies, Raman is sensible to the structure of the analyzed molecule. Hydrolysis does not change the intrinsic structure of fibers, producing only the breaking of the cellulose chain, so the spectra collected from hydrolyzed samples are not different from those of the untreated. On the contrary, oxidation causes a dramatic change in the structure, with formation of
Fig. 7. Raman spectra of untreated (C), oxidized (OX) and oxidized-deacidifiedreduced (OX-DR) Whatman paper at zero-time. The arrow on OX spectrum points the typical peak due to oxidation.
carbonyls, carboxyls or carbon-carbon double bonds that can be detected by Raman spectroscopy. In the oxidized cellulose spectrum, a band appears at about 1577 cm−1 and can present sub-structures or it appears as a large broad band, depending on the kind of oxidation. In this work, Raman spectroscopy was used mostly to investigate the effectiveness of the combined deacidification-reduction treatment. For a better understanding of the real behavior of the samples, the spectra presented in Fig. 7 (C, OX and OX-DR samples before aging) have been plotted without any removal of the background or smoothing because also the fluorescence-background or its variation, as a function of the ageing, contains information on the investigated sample. The arrow on OX spectrum points the typical feature of oxidized paper. It is easily seen that in OX-DR sample, there are no traces of the oxidation band, thus underlining the effectiveness of the reduction treatment. C sample spectrum is reported for comparison. Fig. 8 presents the spectra of the same samples reported in Fig. 7, collected after 15 years of natural aging. C sample shows the typical oxidation broad band, thus confirming that the oxidative degradation naturally occurs. In OX sample, a much more severe
Please cite this article in press as: M. Bicchieri, A. Sodo, Alcoholic deacidification and simultaneous deacidification-reduction of paper evaluated after artificial and natural aging, Journal of Cultural Heritage (2016), http://dx.doi.org/10.1016/j.culher.2016.02.008
G Model CULHER-3039; No. of Pages 8
ARTICLE IN PRESS M. Bicchieri, A. Sodo / Journal of Cultural Heritage xxx (2016) xxx–xxx
7
• the use of ethyl alcohol solutions makes it possible to restore papers that cannot be immersed or treated with water and documents written in water-soluble inks; • both the treatments, i.e. simple deacidification or combined deacidification-reduction, can be applied in different ways: by immersion, by spraying, or by dispersion in gels, such as gellan gum [22]. The spray method allows for direct intervention over books without dismantling its binding.
Fig. 8. Raman spectra of untreated (C), oxidized (OX) and oxidized-deacidifiedreduced (OX-DR) Whatman paper after 15 years of natural aging. The arrows on C and OX spectra evidence the regions presenting oxidation features.
degradation occurs, as a consequence of the original treatment. The spectrum presents in fact an increase of the fluorescence background, a worst quality of the Raman signal and the modification of the oxidation band that became larger then at zero-time. The OX-DR sample, on the contrary, does not show any occurrence of oxidative processes: the spectrum does not present any band at about 1577 cm−1 , confirming the long-term effectiveness of the reducing treatment. 5. Conclusions The measurements performed on all the samples before aging, after artificial aging and repeated after 10 and 15 years of natural uncontrolled aging show that the products examined are both effective and perfectly compatible with the paper support. The treatment combining deacidification and reduction leads to a greater stabilization of the paper support over time than deacidification alone. The neutralizing action of the calcium propionate slows down the paper degradation by countering the acidic hydrolysis of the cellulose and inhibiting the reaction of oxidation. The reducing agent (TBAB) transforms the groups potentially harmful for the support (carbonyls) into harmless ones (hydroxyls) and is very effective in inhibiting the formation of oxidized groups. Raman spectroscopy confirms the chemical results, highlighting the effectiveness of the deacidification-reduction treatment. The Raman spectra of the DR samples, after the most prolonged aging time, are compatible with those that can be obtained from a good quality paper in the best conservation conditions. The combined deacidification and reduction treatment developed in our Institute presents numerous advantages and original features with respect to the restoration methods currently used in the book sector:
It is important to stress that the use of a non-aqueous liquid medium prevents from the major damages due to water-paper interactions (drop of fold endurance, fibers swelling and structural change due to bonded water) and allows for an optimal delivery of the stabilizing products on the material. In 2008, the three tested alcoholic treatments, i.e. deacidification, reduction and combined method, were inserted in the official conservation protocol of the Italian Cultural Heritage Ministry [23]. From 1995 till now, the laboratories of our Institute tested the effect of the reduction on printed books, manuscripts containing irongall, china or logwood inks, on many different kinds of mineral pigments, on printed works of art (aquatint, mezzotint, etching, wood engraving, chalcography, litho and chromolithography) on watercolors, pastels and tempera, treating around 1000 original documents of different authors [24,25]. Never negative interactions, neither with the support nor with any color, have been recorded. It is also necessary to remember and stress that the reducing treatment is not a bleaching treatment. While from one side, it is true that the reduction, by removing the oxidized carbonyl groups and turning them back into alcohol groups, produces a bleaching, on the other side, its purpose is completely different. The bleaching is performed for purely aesthetic reasons, while the reduction has the purpose to preserve the paper, to increase its average expected life and to strengthen it (See supplementary materials S3). Tensile strength, as well as double bonds, in fact, mainly depends on the number of interfiber bonds present in the cellulose. By recovering OH groups with reduction, the number of H-bonds between fibers increases and this increases the strength of the paper. Only the chemical and the spectroscopic analyses can give information on the conservation state of the analyzed samples. All the data collected in this work evidence the goodness of both the treatments, mostly of the combined deacidification-reduction, and let predict a high increase of the expected life. Another point needs to be taken in account. The reducing agent investigated, as well as the other reducers used in the world, acts on carbonyl groups. It is necessary to be very cautious when dealing with dyes that owe their color to the presence of carbonyls in their molecule: their resistance to the treatment should be tested before the application. Even though in 15 years of use, fading or color changes of graphic works have not been recorded, our Institute is studying the interactions between the reducer and the most common organic dyes.
Acknowledgement • deacidification is effected by means of a fungistatic food preservative, whose effect on paper had not previously been subjected to systematic study over time. The well known toxicological harmlessness of calcium propionate makes it particularly suitable for repeated use by restorers with no risk for the various users of libraries; • application of a single combined process of deacidification and reduction reduces the number of treatments to be performed on the originals;
We would like to acknowledge Michela Monti and Giovanna Piantanida for the help in performing a series of measurements.
Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/ j.culher.2016.02.008.
Please cite this article in press as: M. Bicchieri, A. Sodo, Alcoholic deacidification and simultaneous deacidification-reduction of paper evaluated after artificial and natural aging, Journal of Cultural Heritage (2016), http://dx.doi.org/10.1016/j.culher.2016.02.008
G Model CULHER-3039; No. of Pages 8
ARTICLE IN PRESS M. Bicchieri, A. Sodo / Journal of Cultural Heritage xxx (2016) xxx–xxx
8
References [1] J.W. Baty, C.L. Maitland, W. Minter, M.A. Hubble, S.K. Jordan-Mowery, Deacidification for conservation, BioResources 5 (3) (2010) 1955–2023. [2] G.C. Andrews, Chemoselectivity in the reduction of aldehydes and ketones with amine boranes, Tetrahedron Lett. 21 (1980) 679–700. [3] D.J. Raber, W.C. Guida, D.C. Shoenberger, Reduction of aldehydes and ketones with tetraalkylammonium borohydrides, Tetrahedron Lett. 22 (1981) 5107–5110. [4] E.R.H. Walker, The functional group selectivity of complex hydride reducing agents, Chem. Soc. Rev. 5 (1) (1976) 23–50. [5] D. Clement, The blistering of paper during hydrogen peroxide bleaching, J. Am. Inst. Conserv. 23 (1) (1983) 47–62. [6] H.D. Burgess, The stabilization of cellulosic fibres by borohydride derivatives, ICOM Comm. Conserv. 2 (1990) 447–452. [7] M. Bicchieri, M. Bella, F. Sementilli, A quantitative measure of borane tertbutylamine complex effectiveness in carbonyl reduction of aged papers, Restaurator 20 (1) (1999) 22–29. [8] C. Pedneault, S. Robert, C. Pellerin, Bleaching with new reductive chemicals: replacement of hydrosulfite, Pulp Pap. Can. 98 (3) (1997) 51–55. [9] M. Bicchieri, F.M. Sementilli, A. Sodo, Application of seven borane complexes in paper conservation, Restaurator 21 (4) (2000) 213–228. [10] M. Bicchieri, M. Monti, M.L. Antonelli, A new low-cost and complete restoration method: a simultaneous non-aqueous treatment of deacidification and reduction, in: J. Apuente (Ed.), Proceedings of the 3rd International Conference on Science and Technology for the safeguard of cultural heritage in the Mediterranean basin, I, Universitad de Alcalà, Alcalà, 2001, pp. 276–280. [11] C. Sanna, A. Sodo, G. Laguzzi, G. Mancini, M. Bicchieri, Tert-butyl amine borane complex: an unusual application of a reducing agent on model molecules of cellulose based materials, J. Cult. Herit. 10 (3) (2009) 356–361. ˇ [12] J. Maleˇsiˇc, M. Kojc, Selih, Assessment of the effect of various bleaching agents on papers with foxing stains, Restaurator 29 (3) (2008) 142–154. [13] C.F. Lane, Ammonia-borane and related N-B-H compounds and materials: safety aspects, properties and applications, N-B-H Survey (2006) 1–33. [14] E.H. Immergut, B.G. Ranby, H.F. Mark, Recent work on molecular weight of cellulose, Ind Eng Chem 45 (11) (1953) 2483–2490. [15] J. Lauriol, J. Comtat, P. Froment, F. Pla, A. Robert, Molecular weight distribution of cellulose by on-line size exclusion chromatography - Low angle laser light scattering, Holzforschung 41 (3) (1987) 165–169. [16] A.M. Mattson, C.O. Jensen, Colorimetric determination of reducing sugars with triphenyltetrazolium chloride, Anal Chem 22 (1950) 182–185. [17] J. Silvy, J.F. Le Nest, A photochromic effect shown during the aging of a cellulose material by heat and light, in: F. Bolam (Ed.), The fundamental properties of
[18]
[19] [20]
[21]
[22]
[23] [24]
[25]
paper related to its uses: transactions of the symposium held at Cambridge, September 1973, The British Paper and Board Industry Federation, London, 1976, pp. 755–760. U. Henniges, M. Schwanninger, A. Potthast, Non-destructive determination of cellulose functional groups and molecular weight in pulp hand sheets and historic papers by NIR-PLS-R, Carbohydr. Polym. 76 (3) (2009) 374–380. M. Ali, A.M. Emsley, H. Herman, R.J. Heywood, Spectroscopic studies of the aging of cellulosic paper, Polymer. 42 (7) (2001) 2893–2900. D. Lichtblau, M. Strliˇc, T. Trafela, J. Kolar, M. Anders, Determination of mechanical properties of historical paper based on NIR spectroscopy and chemometrics - a new instrument, Appl. Phys. A 92 (1) (2008) 191–195. M. Bicchieri, A. Sodo, G. Piantanida, C. Coluzza, Analysis of degraded papers by non-destructive spectroscopic techniques, J. Raman Spectrosc. 37 (10) (2006) 1186–1192. S. Iannuccelli, S. Sotgiu, A new methodology for wet conservation treatments of graphic art on paper with a rigid polysaccharide gel of Gellan gum, in: L. Watteuw (Ed.), GDWG ICOM-CC Interim Meeting, The Royal Library of Denmark, Copenhagen, 2010, pp. 47–51. http://www.icpal.beniculturali.it/allegati/Capitolato Speciale Tecnico Tipo.pdf. G. Pace, Giuseppe De Nitti’s graphical works in the Museo Civico di Barletta – Restoration, in: F. Fiorani, R. Di Noia (Eds.), De Nittis incisore, Artemide, Roma, 1999, pp. 281–287. M. Bicchieri, M. Monti, G. Piantanida, F. Pinzari, S. Iannuccelli, S. Sotgiu, et al., The Indian drawings of the poet Cesare Pascarella. Non destructive analyses and conservation treatments, Anal. Bioanal. Chem. 402 (4) (2012) 1517– 1528.
Further reading ASTM D1795-13 Standard test method for intrinsic viscosity of cellulose. ISO 5630-3:1996 (Rev. 2012) Paper and board – Accelerated ageing – Part 3: moist heat treatment at 80◦ C and 65% relative humidity. ISO 2470-1:2009 Paper, board and pulps – Measurement of diffuse blue reflectance factor – Part 1: indoor daylight conditions (ISO brightness). TAPPI T 435 om-11. Hydrogen ion concentration (pH) of paper extracts (hot extraction method).
Please cite this article in press as: M. Bicchieri, A. Sodo, Alcoholic deacidification and simultaneous deacidification-reduction of paper evaluated after artificial and natural aging, Journal of Cultural Heritage (2016), http://dx.doi.org/10.1016/j.culher.2016.02.008
ARTICLE IN PRESS Journal of Cultural Heritage xxx (2016) xxx–xxx
Available online at
ScienceDirect www.sciencedirect.com
Original article
Alcoholic deacidification and simultaneous deacidification-reduction of paper evaluated after artificial and natural aging Marina Bicchieri a,∗ , Armida Sodo a,b a b
Chemistry department, istituto centrale restauro e conservazione patrimonio archivistico e librario, 76, via Milano, 00184 Roma, Italy Dip. Scienze, università Roma Tre, 84, via della Vasca Navale, 00146 Roma, Italy
a r t i c l e
i n f o
Article history: Received 9 November 2015 Accepted 22 February 2016 Available online xxx Keywords: Paper Degradation Deacidification Reduction Conservation Non-aqueous treatments
a b s t r a c t Cellulose oxidative and hydrolytical degradation is one of the greatest problems for the conservation of paper supports. To contrast these degradation processes, both deacidification and reduction of the oxidized functions are needed. Dealing with original documents, it is often impossible to perform the two mentioned treatments in aqueous solutions and in a distinct subsequent way, because of the fragility of the artifacts. After studying, in a separate way, an effective deacidifier (calcium propionate) soluble in ethyl alcohol and many reducers (boron complexes), able to act in different non-aqueous solvents, it was decided to test a simultaneous method of deacidification and reduction in ethanol. This paper presents the chemical-physical results obtained by applying simple deacidification and simultaneous deacidification-reduction on laboratory paper samples that were artificially aged and then re-measured after 10 and 15 years of natural aging. Results show that all alcoholic treatments are very effective: papers are stable also after a long period of both artificial and natural aging. © 2016 Elsevier Masson SAS. All rights reserved.
1. Research aims This paper tries to answer to the needs of conservators of books, archival materials and graphic works of art, who need sometime to use non-aqueous solutions in treating original documents to prevent the dissolution of the graphic media, the alteration of the paper structure and the loss of the printing impression. Moreover, it is often necessary to work on bound books to maintain the original bindings untouched. Also in this case, water solutions could not be used. All the analyses and the used complementary techniques were applied in order to ascertain the effectiveness of a simultaneous non-aqueous deacidification and reduction treatment on artificially and naturally aged papers.
2. Introduction Cellulose oxidative and hydrolytical degradation is one of the greatest problems for the conservation of paper supports such as archival documents, books and graphic works. Degraded papers
∗ Corresponding author. Tel.: +390648291217; fax: +39064814968. E-mail addresses: [email protected] (M. Bicchieri), [email protected], sodo@fis.uniroma3.it (A. Sodo).
show an evident fragility and a yellowing, compromising in some cases the readability of the text or the graphic sign. Many deacidifier can be used [1], but only few reducers can be directly applied on paper documents without damaging the fiber structure and the graphic media. Boron compounds and complexes are very effective and chemoselective in reducing carbonyls that are the most important oxidized groups in the cellulose polymer chain [2–4]. During the reduction reactions, hydrogen is produced and, if the reaction rate is too high, the cellulose fiber wall could be broken [5] by the large amount of hydrogen generated in a short time. This effect has been observed when sodium borohydride, the first reducer used in conservation field, was used as reducing product in the treatment of strongly oxidized papers [6]. Therefore, to minimize this negative effect, mild and “slow” reducers should be chosen, such as the boron-amine complexes. Furthermore, to avoid -alkoxyelimination reaction, which occurs at ambient temperature in high alkaline environment (greater than 10.5) when a polymeric chain contains a large amount of carbonyl groups, the employed reducers should have a pH lower than 10 when applied in water solution [1]. In 1997, we published the first results on the application of borane tert-butylamine complex [(CH3 )3 CNH2 . BH3 , TBAB] for the reduction of carbonyl groups in oxidized paper [7]. Almost in the same period, TBAB was proposed for the bleaching of pulp in the paper industry. Amine boranes appeared to behave
http://dx.doi.org/10.1016/j.culher.2016.02.008 1296-2074/© 2016 Elsevier Masson SAS. All rights reserved.
Please cite this article in press as: M. Bicchieri, A. Sodo, Alcoholic deacidification and simultaneous deacidification-reduction of paper evaluated after artificial and natural aging, Journal of Cultural Heritage (2016), http://dx.doi.org/10.1016/j.culher.2016.02.008
G Model CULHER-3039; No. of Pages 8
ARTICLE IN PRESS M. Bicchieri, A. Sodo / Journal of Cultural Heritage xxx (2016) xxx–xxx
2
better than sodium borohydride and even than sodium hydrosulphite for obtaining a more stable, white and not-degraded cellulose pulp [8]. In a second time, we investigated the possibility of a nonaqueous application of different boron compound for reducing purposes [9] and the applicability of a simultaneous non-aqueous deacidification with ethanol solution of calcium propionate [Ca(CH3 CH2 COO)2 ] and reduction with borane-ammonia complex [NH3 BH3 ] [10]. Non-aqueous treatments are in fact requested in case of restoration of originals containing water-soluble graphic media or when it is mandatory to treat books keeping intact the bookbinding. Due to instability of borane-ammonia complex and the difficulties in its handling by restorers, we decided to perform a new series of studies on TBAB, whose interactions with cellulose are reported in [11], and to use it for a new series of tests on alcoholic deacidification and reduction. This paper presents the comparison between the data obtained on artificially aged samples and those collected on the same samples after natural aging in uncontrolled conditions for 10 and 15 years. Before presenting the results of this work and their discussion, some points need to be underlined. Looking at the specific literature [12 and related references], it seems that conservators are more interested in “bleaching” of paper – that is only a side effect of the reducing treatment – than in the most important result that is the stabilization of the cellulosic support and the increase of its expected life. Despite the researches carried out on different boron complexes, outside Italy only sodium borohydride is used, whose effect on strongly oxidized papers can be dramatic, causing the mechanical breaking of the paper [5]. Another reducer, sodium dithionite [12], has a mild brightening effect on paper, but is very effective in the reduction of iron (III) ions. This makes the dithionite not applicable to all manuscripts containing iron-based inks or pigments, which could disappear after the treatment. Moreover, the articles concerned with chemical methods for paper restoration often report results based on mechanical and optical measurements. Significant variations of mechanical properties need very long time (See supplementary material S1). On the contrary even light chemical variations can cause strong degradation. This implies that potentially harmful treatments – such as reduction with sodium borohydride or deacidification with calcium hydroxide at too high pH – are commonly accepted. For these reasons, we focused our attention on the chemical behavior of the investigated items to perform a rational and scientific acceptance or rejection of the treatment. Therefore, the characterization was carried out by applying chemical methods (pH, carbonyl content and degree of polymerization), spectroscopic techniques (Raman spectroscopy) and optical methods (color coordinates and Optical density measurements).
3. Materials and methods Materials, methods and techniques used in this work are briefly described below.
3.1. Materials • Whatman® cellulose chromatography paper grade 1; 46 cm × 57 cm, a pure cotton and cotton linters cellulose. • hydrochloric acid 37% [HCl] (Aldrich ACS grade) used to simulate “original” acidic samples subjected to hydrolytic degradation.
• 0.015 M aqueous solution of potassium meta-periodate [KIO4 ] (powder, Merck 99.8%) brought to pH 5 with HCl 10−2 M in order to simulate “original” samples subjected to oxidative degradation. • 3.0 g/L solution of calcium propionate [Ca(CH3 CH2 COO)2 ] (Aldrich 95%) in ethyl alcohol (Aldrich 96%). The solution was used to carry out simple non-aqueous deacidification treatments. • 3.0 g/L solution of calcium propionate in ethyl alcohol + borane tert-butylamine complex [(CH3 )3 CNH2 . BH3 ] (Aldrich 97%). The resulting solution is 0.2 M in borane complex. The solution was used to carry out simultaneous deacidification and reduction. • 0.2% w/V water solution of 2,3,5 triphenyl-2H-tetrazolium chloride (TTC, Aldrich 98%). • 0.5 M cupriethylenediamine solution (Carlo Erba for analysis). 3.2. Preparation of paper samples Three sets of laboratory paper samples were considered in the present work; each one was successively divided in three series, as described below: • one set was oxidized by immersion for 15 minutes in a KIO4 0.015 M water solution, at pH = 5.0, and then rinsed by two immersions, 5 minutes each, in distilled water. This treatment was used to simulate “original” papers presenting a strong degradative oxidation. After the treatment, one series was kept as prepared (Ox series in the following); a second one was immersed in the alcoholic deacidification solution of calcium propionate (Ox-D series in the following) as described in a previous work [10] so to have a comparison with established restoration treatments, and the third one underwent a simultaneous non aqueous deacidification and reduction treatment (Ox-DR series in the following); • one set was hydrolyzed by 1 h exposition to fuming HCl vapors, to simulate acidic “original” documents. After hydrolysis, a series was kept as prepared (Hy series in the following), a series was deacidified (Hy-D series in the following) and the third one deacidified and reduced (Hy-DR series in the following); • one set of Whatman N.1 samples (not treated, deacidified and deacidified/reduced) was included in the experiment as control samples (respectively C, C-D and C-DR series in the following). All laboratory samples were artificially aged in climate chamber (Angelantoni Challenge E300) for 35 days, according to International Organization for Standardization ISO 5630-3:1996 at 65% R.H. and 80 ◦ C and characterized each 7 days of ageing with destructive and non-destructive techniques. The chosen temperature is compatible with the stability of TBAB: amine-boranes prepared from primary and secondary amines are stable up to 110 ◦ C [13]. After the artificial ageing experiment, for each set and series, the “zero-time” samples and those artificially aged for 35 days, were left on library shelves for 10 years in order to simulate a non-controlled storage. After their retrieval, the samples were once more characterized and left for other 5 years in uncontrolled ambient for a third characterization. Chemical, optical and spectrometric tests were performed after each cycle of artificial and natural aging. Oxidation and hydrolysis induced by environmental conditions, in fact, cause important chemical and structural modifications with great impact on paper permanence and durability, even if mechanical characteristics are not particularly affected in the short term. Decrease in the degree of polymerization is related to hydrolytic degradation, whereas increase in carbonylic functions is due to oxidative processes. Structural changes produce modification in the Raman spectra of paper.
Please cite this article in press as: M. Bicchieri, A. Sodo, Alcoholic deacidification and simultaneous deacidification-reduction of paper evaluated after artificial and natural aging, Journal of Cultural Heritage (2016), http://dx.doi.org/10.1016/j.culher.2016.02.008
G Model
ARTICLE IN PRESS
CULHER-3039; No. of Pages 8
M. Bicchieri, A. Sodo / Journal of Cultural Heritage xxx (2016) xxx–xxx
3.3. Methods of testing • Acidity measurements were performed according to standard TAPPI T 435 om-11. • Viscometric average degree of polymerization (DP, dimensionless number) was measured following American Society for Testing and Materials (ASTM) according to the standard ASTM D1795-13. • Carbonyl content measures were performed according to a method developed in our Institute and described in a previous paper [7]. • Color coordinates were measured by means of a colorimeter (Minolta Chroma Meter CR22) in the CIE L*a*b* space. • Optical density: the combined processes of absorption, scattering and reflection influence the “transmittance” (T, dimensionless number), which accounts for the residual light transmitted from sample in the forward direction. In order to disentangle the contribution from reflection, absorption, and scattering, it is necessary to measure the “transflectance” (TF, dimensionless number). TF is the fraction of light emerging in all directions from the sample and allows sample’s optical density (OpD, dimensionless number) to be derived directly, following the equations: OpD = −log (T + R) = −log TF, where R is the reflectance of the sample (R, dimensionless number). Measurements were carried out in the 200–2500 nm range, by means of a double-beam spectrophotometer (Varian Cary 5), equipped with calibrated SpectralonTM integrating sphere (LabSphere) and internal sample holder (LabSphere). By positioning the sample at the centre of the sphere, the transflectance is measured. The empty sphere spectrum is assumed as reference, so the experimental optical density is achieved as OpD = −log (TF/TFsphere ). • Raman measurements have been performed by means of a Renishaw In-Via Reflex Raman microscope equipped with a Renishaw diode laser at 785 nm (output power 300 mW). The backscattered light is dispersed by a 1200-line/mm grating and the Raman signal is detected by a Peltier cooled (−70 ◦ C) deep depletion charge-coupled device (CCD RD-VIU, 578 pixel × 384 pixel) optimized for near infrared and ultraviolet. The nominal spectral resolution obtained for the measurements is about 3 cm−1 . The system, equipped with a Leica DM LM microscope to focus the laser on the sample and a color video camera, allows for the positioning of the sample and the selection of a specific region for the investigation. Spectral acquisitions (1–10 accumulations, 50 s each) have been performed with a 50× objective (N.A. 0.75). Under these conditions, the laser spot measures about 20 m2 . 4. Results and discussion 4.1. Acidity measurements The values of pH by hot extraction method (Table 1) show that all the samples subjected to the treatments involving Table 1 pH (±0.05) variations along aging time (average of four measurements on each sample; five samples for each kind of paper and treatment). Sample
0 day
35 days (accelerated aging)
10 years (natural aging on originally 0 day sample)
15 years (natural aging on originally 0 day sample)
C C-D C-DR Ox Ox-D Ox-DR Hy Hy-D Hy-DR
5.37 8.89 9.08 5.52 8.57 8.84 4.90 8.29 8.20
5.64 8.91 8.84 5.40 8.28 8.26 4.96 8.47 8.19
5.60 7.43 8.37 5.81 7.39 8.65 4.40 7.92 8.04
5.55 7.26 8.42 5.63 7.27 8.60 4.01 7.58 8.00
3
deacidification or combined deacidification-reduction remain alkaline, in the range of the maximum stability of the cellulose, also after prolonged accelerated and natural aging. The best stabilization against the pH variation is obtained with the combined treatment of reduction and deacidification, as evidenced by the data of Table 1. 4.2. Viscometric average degree of polymerization The preparation of the cellulose solutions and the measurement of their intrinsic viscosity in cupriethylenediamine were performed in accordance with ASTM D1795-13 with some modifications. The intrinsic viscosity [] of all the samples was not obtained from listed tables (ASTM Method) but by extrapolating at zero concentration the values of Log [(rel -1)/c] as a function of c (rel is the relative viscosity of the solution of paper in cupriethylenediamine hydroxide and c is the concentration of the cellulose solution expressed in g/dL). The straight-line parameters were obtained using a leastsquare-fit method, taking into account the proper error for each measurement (5 concentration for each sample; 4 measurement for each concentration). By multiplying intrinsic viscosity by a specific constant K (cm3 .g−1 ) it is possible to obtain the viscosity average degree of polymerization DP (dimensionless number). Studies on the behavior of cellulose [14,15] have shown that the constant K assumes different values in relation to the degree of polymerization of the cellulose: K = 156 for DP between 3000 and 300 and K = 124 for DP below 300. To obtain statistically valid data, the value of intrinsic viscosity was extrapolated from a series of measurements at different concentration. Errors associated with each individual measurement were obtained by applying the least square method with variable error. The results of the measurements are given in Table 2A. Data of the measures carried out after 15 years of natural aging both for the samples not subjected to the accelerated aging and for those 35 days artificially aged are also shown. The comparison would give indications on the effects and the differences between natural and artificial aging. Fig. 1 reports the results of DP measurements for the natural aging, also on the samples previously artificially aged for 35 days. It can be easily seen that the artificial aging accelerates of a greater extent the hydrolytic degradation, in respect to the natural one. The higher values of all the DR samples in respect to their control samples (i.e. C, Ox and Hy respectively) are due to high pH value of the cupriethylenediamine (pH around 13). At this pH, cupriethylenediamine does not behave as a good solvent: high pH values induces -alkoxyelimination reactions when the samples contain oxidized functions. The maximum difference between the control and its respective DR sample is, in fact, observed for the oxidized series. Moreover, the results underline that the viscometric DP values closest to the “true” DP of any kind of paper can be obtained only by measuring the samples after a reduction treatment that removes the oxidized functions and allows to the cupriethylenediamine to behave as a good solvent without any interference with the samples under analysis. It is noticeable that the DP values of all the samples treated with the combined deacidification and reduction treatment are, also after 15 years of natural aging, comparable with those of the not treated and not aged samples, this meaning that the treatment offered a huge protection against the hydrolytic degradation. For the hydrolyzed series, the initial hydrolytic degradation had been so intense since the beginning that the unique effect of the two treatments – deacidification and in particular combined one – was to obstacle a further degradation, keeping constant the initial DP level.
Please cite this article in press as: M. Bicchieri, A. Sodo, Alcoholic deacidification and simultaneous deacidification-reduction of paper evaluated after artificial and natural aging, Journal of Cultural Heritage (2016), http://dx.doi.org/10.1016/j.culher.2016.02.008
G Model
ARTICLE IN PRESS
CULHER-3039; No. of Pages 8
M. Bicchieri, A. Sodo / Journal of Cultural Heritage xxx (2016) xxx–xxx
4
Table 2 A: DP variations; B: carbonyl content (m.moles/100 g paper) along aging time. Aging Sample
2A: pH C C-D C-DR Ox Ox-D Ox-DR Hy Hy-D Hy-DR 2B: C = O C C-D C-DR Ox Ox-D Ox-DR Hy Hy-D Hy-DR
0 days
35 days (acc. aging)
1189 1429 1397 830 998 1096 238 288 281
± ± ± ± ± ± ± ± ±
10 15 12 13 5 37 9 4 9
0.46 0.20 0.20 0.55 0.20 0.25 4.50 0.65 0.28
± ± ± ± ± ± ± ± ±
0.01 0.01 0.01 0.01 0.01 0.01 0.40 0.07 0.01
442 670 1115 329 828 1006 216 287 319
± ± ± ± ± ± ± ± ±
16 5 18 12 3 16 4 5 12
1.14 0.21 0.24 4.90 1.89 0.23 5.00 0.51 0.19
± ± ± ± ± ± ± ± ±
0.02 0.03 0.01 0.50 0.20 0.01 0.50 0.02 0.01
10 years (natural aging)
15 years (natural aging)
Originally 0 days
Originally 0 days
624 1107 1233 634 940 1130 232 281 301
± ± ± ± ± ± ± ± ±
3 41 15 10 39 13 4 2 15
0.95 0.43 0.30 7.80 1.60 0.45 5.04 0.86 0.39
± ± ± ± ± ± ± ± ±
0.01 0.01 0.04 0.51 0.02 0.01 0.07 0.05 0.01
4.3. Carbonyl content Carbonyl groups can exist in paper both as end-groups in each macromolecule of cellulose and on the anhydrous glucose ring, after oxidation.
Originally 35 days
580 1126 1200 483 867 1100 220 264 280
± ± ± ± ± ± ± ± ±
9 20 10 7 7 20 10 9 7
400 516 1100 300 789 1000 200 231 300
± ± ± ± ± ± ± ± ±
12 16 15 10 9 8 6 7 10
1.02 0.51 0.32 9.17 1.78 0.70 5.86 0.98 0.45
± ± ± ± ± ± ± ± ±
0.03 0.01 0.05 0.12 0.03 0.02 0.15 0.02 0.01
1.08 0.50 0.38 11.63 1.92 0.80 7.01 1.20 0.56
± ± ± ± ± ± ± ± ±
0.08 0.01 0.01 0.16 0.08 0.01 0.20 0.05 0.01
These groups can be measured by redox reaction with triphenyltetrazolium chloride salts. In an alkaline environment, carbonyls are able to reduce 2,3,5-triphenyl-2H-tetrazolium chloride to the corresponding triphenylformazan of a red color. The amount of formazan formed is directly proportional to the amount of carbonyl groups present, and it is spectrophotometrically determined at 480 nm [7,16]. The results obtained for all the different samples are summarized in Table 2B and plotted in Fig. 2, every value being the average of five measurements on each sample. Five samples were used for each kind of paper and treatment. It should be highlighted that the experimental data show that the carbonyl content of all the DR samples, even after 15 years of natural ageing, is about of the same value of that measured on the not aged C samples. This evidences the goodness of the reduction treatment and its effectiveness in the contrast of the oxidative degradation, promoting the stabilization of the paper over time. 4.4. Color coordinates The most representative coordinates related to the modifications induced by aging are L* and b*. A decrease in L* reduces the lightness of the sample, while an increase of b* shifts to yellow the color of the paper. CIE L*a*b* data (See supplementary material S2) present an increase of the b* coordinate for all the samples, in particular for those not subjected to any deacidification or combined deacidification-reduction treatment. At the same time, L* slightly decreases, as predictable after any aging (Figs. 3 and 4). 4.5. Optical density
Fig. 1. Average viscometric degree of polymerization for the natural aging. The 15 years data are related to the natural aging of the originally 0 and 35 days accelerate ageing samples.
Fragmentation and instable oxidation products cause the formation of yellow chromophores and fluorescent states. UV-Vis absorption spectra of paper are affected by these phenomena since both electronic transition and fluorescence bands can be found in the 200–500 nm spectral region and the interpretation must be done cautiously [17,18]. In particular, as it can be seen from Fig. 5, all the three sets of laboratory samples showed after 15 years of natural aging a peak at
Please cite this article in press as: M. Bicchieri, A. Sodo, Alcoholic deacidification and simultaneous deacidification-reduction of paper evaluated after artificial and natural aging, Journal of Cultural Heritage (2016), http://dx.doi.org/10.1016/j.culher.2016.02.008
G Model CULHER-3039; No. of Pages 8
ARTICLE IN PRESS M. Bicchieri, A. Sodo / Journal of Cultural Heritage xxx (2016) xxx–xxx
5
Fig. 2. Carbonyl content of the samples for the natural aging. The 15 years data are related to the natural aging of the originally 0 and 35 days accelerate ageing samples.
around 205 nm with a shoulder at about 260 nm and a tail extending in the blue region till to around 500 nm. This region is of interest, since, according to ISO 2470-1:2009, yellow chromophores content in paper is quantified by an inverse relation in respect to the brightness measured as UV absorbance at 457 nm.
Fig. 3. L* coordinate of all samples before and after 10 and 15 years of natural ageing.
The above-mentioned features are more intense in the Hy sample with respect to the C and Ox samples. Moreover, the Hy sample shows higher absorbance in the blue region, thus confirming the shift to yellow of the color of the paper. Both the deacidification and the simultaneous deacidification and reduction treatments specifically affect the intensity of the shoulder at around 260 nm. The feature becomes less intense in
Fig. 4. b* coordinate of all samples before and after 10 and 15 years of natural ageing.
Please cite this article in press as: M. Bicchieri, A. Sodo, Alcoholic deacidification and simultaneous deacidification-reduction of paper evaluated after artificial and natural aging, Journal of Cultural Heritage (2016), http://dx.doi.org/10.1016/j.culher.2016.02.008
G Model CULHER-3039; No. of Pages 8
ARTICLE IN PRESS M. Bicchieri, A. Sodo / Journal of Cultural Heritage xxx (2016) xxx–xxx
6
Fig. 6. NIR spectra of C, Ox and Hy samples aged 15 years.
Fig. 5. UV-Vis spectra of samples aged 15 years. Each series of control paper is compared with the respective deacidified and deacidified – reduced series.
all D and DR series, and – consequently – the absorbance in the 400–500 nm region (blue) is reduced accordingly, and contrasts the yellowing effect. This spectral evidence is also confirmed by the lower values of the b* color coordinate of D and DR series for the same set of samples (Supplementary materials S2, column 15 years). Taking into account that the cellulose chromophores at around 260 nm are usually interpreted in terms of carbonyl and carboxyl formation [17], it can be inferred that the two conservation treatments positively contrast their formation, contributing to the preservation of paper. The NIR part of the spectrum is dominated by water/OH features but it is possible to observe, between 1700 nm and 2200 nm, some carbonyl/carboxyl stretching overtones [19] (Fig. 6). The variation in the intensity of overtones, due to the increase of the number of oxidized groups, is much less pronounced in NIR than in the midIR region and the analysis of NIR data is typically performed with multivariate methods, using a great variety of data [20]. The NIR analysis along aging time was therefore not considered relevant in this work: an enhancement of the carboxyl C = O and C-O stretching overtones peaks appeared in fact quite well visible only for the hydrolyzed sample Hy after the longer natural aging time (Fig. 6). 4.6. Raman In previous studies, Raman spectroscopy was applied to investigate the state of conservation of differently degraded papers [21]. As all vibrational spectroscopies, Raman is sensible to the structure of the analyzed molecule. Hydrolysis does not change the intrinsic structure of fibers, producing only the breaking of the cellulose chain, so the spectra collected from hydrolyzed samples are not different from those of the untreated. On the contrary, oxidation causes a dramatic change in the structure, with formation of
Fig. 7. Raman spectra of untreated (C), oxidized (OX) and oxidized-deacidifiedreduced (OX-DR) Whatman paper at zero-time. The arrow on OX spectrum points the typical peak due to oxidation.
carbonyls, carboxyls or carbon-carbon double bonds that can be detected by Raman spectroscopy. In the oxidized cellulose spectrum, a band appears at about 1577 cm−1 and can present sub-structures or it appears as a large broad band, depending on the kind of oxidation. In this work, Raman spectroscopy was used mostly to investigate the effectiveness of the combined deacidification-reduction treatment. For a better understanding of the real behavior of the samples, the spectra presented in Fig. 7 (C, OX and OX-DR samples before aging) have been plotted without any removal of the background or smoothing because also the fluorescence-background or its variation, as a function of the ageing, contains information on the investigated sample. The arrow on OX spectrum points the typical feature of oxidized paper. It is easily seen that in OX-DR sample, there are no traces of the oxidation band, thus underlining the effectiveness of the reduction treatment. C sample spectrum is reported for comparison. Fig. 8 presents the spectra of the same samples reported in Fig. 7, collected after 15 years of natural aging. C sample shows the typical oxidation broad band, thus confirming that the oxidative degradation naturally occurs. In OX sample, a much more severe
Please cite this article in press as: M. Bicchieri, A. Sodo, Alcoholic deacidification and simultaneous deacidification-reduction of paper evaluated after artificial and natural aging, Journal of Cultural Heritage (2016), http://dx.doi.org/10.1016/j.culher.2016.02.008
G Model CULHER-3039; No. of Pages 8
ARTICLE IN PRESS M. Bicchieri, A. Sodo / Journal of Cultural Heritage xxx (2016) xxx–xxx
7
• the use of ethyl alcohol solutions makes it possible to restore papers that cannot be immersed or treated with water and documents written in water-soluble inks; • both the treatments, i.e. simple deacidification or combined deacidification-reduction, can be applied in different ways: by immersion, by spraying, or by dispersion in gels, such as gellan gum [22]. The spray method allows for direct intervention over books without dismantling its binding.
Fig. 8. Raman spectra of untreated (C), oxidized (OX) and oxidized-deacidifiedreduced (OX-DR) Whatman paper after 15 years of natural aging. The arrows on C and OX spectra evidence the regions presenting oxidation features.
degradation occurs, as a consequence of the original treatment. The spectrum presents in fact an increase of the fluorescence background, a worst quality of the Raman signal and the modification of the oxidation band that became larger then at zero-time. The OX-DR sample, on the contrary, does not show any occurrence of oxidative processes: the spectrum does not present any band at about 1577 cm−1 , confirming the long-term effectiveness of the reducing treatment. 5. Conclusions The measurements performed on all the samples before aging, after artificial aging and repeated after 10 and 15 years of natural uncontrolled aging show that the products examined are both effective and perfectly compatible with the paper support. The treatment combining deacidification and reduction leads to a greater stabilization of the paper support over time than deacidification alone. The neutralizing action of the calcium propionate slows down the paper degradation by countering the acidic hydrolysis of the cellulose and inhibiting the reaction of oxidation. The reducing agent (TBAB) transforms the groups potentially harmful for the support (carbonyls) into harmless ones (hydroxyls) and is very effective in inhibiting the formation of oxidized groups. Raman spectroscopy confirms the chemical results, highlighting the effectiveness of the deacidification-reduction treatment. The Raman spectra of the DR samples, after the most prolonged aging time, are compatible with those that can be obtained from a good quality paper in the best conservation conditions. The combined deacidification and reduction treatment developed in our Institute presents numerous advantages and original features with respect to the restoration methods currently used in the book sector:
It is important to stress that the use of a non-aqueous liquid medium prevents from the major damages due to water-paper interactions (drop of fold endurance, fibers swelling and structural change due to bonded water) and allows for an optimal delivery of the stabilizing products on the material. In 2008, the three tested alcoholic treatments, i.e. deacidification, reduction and combined method, were inserted in the official conservation protocol of the Italian Cultural Heritage Ministry [23]. From 1995 till now, the laboratories of our Institute tested the effect of the reduction on printed books, manuscripts containing irongall, china or logwood inks, on many different kinds of mineral pigments, on printed works of art (aquatint, mezzotint, etching, wood engraving, chalcography, litho and chromolithography) on watercolors, pastels and tempera, treating around 1000 original documents of different authors [24,25]. Never negative interactions, neither with the support nor with any color, have been recorded. It is also necessary to remember and stress that the reducing treatment is not a bleaching treatment. While from one side, it is true that the reduction, by removing the oxidized carbonyl groups and turning them back into alcohol groups, produces a bleaching, on the other side, its purpose is completely different. The bleaching is performed for purely aesthetic reasons, while the reduction has the purpose to preserve the paper, to increase its average expected life and to strengthen it (See supplementary materials S3). Tensile strength, as well as double bonds, in fact, mainly depends on the number of interfiber bonds present in the cellulose. By recovering OH groups with reduction, the number of H-bonds between fibers increases and this increases the strength of the paper. Only the chemical and the spectroscopic analyses can give information on the conservation state of the analyzed samples. All the data collected in this work evidence the goodness of both the treatments, mostly of the combined deacidification-reduction, and let predict a high increase of the expected life. Another point needs to be taken in account. The reducing agent investigated, as well as the other reducers used in the world, acts on carbonyl groups. It is necessary to be very cautious when dealing with dyes that owe their color to the presence of carbonyls in their molecule: their resistance to the treatment should be tested before the application. Even though in 15 years of use, fading or color changes of graphic works have not been recorded, our Institute is studying the interactions between the reducer and the most common organic dyes.
Acknowledgement • deacidification is effected by means of a fungistatic food preservative, whose effect on paper had not previously been subjected to systematic study over time. The well known toxicological harmlessness of calcium propionate makes it particularly suitable for repeated use by restorers with no risk for the various users of libraries; • application of a single combined process of deacidification and reduction reduces the number of treatments to be performed on the originals;
We would like to acknowledge Michela Monti and Giovanna Piantanida for the help in performing a series of measurements.
Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/ j.culher.2016.02.008.
Please cite this article in press as: M. Bicchieri, A. Sodo, Alcoholic deacidification and simultaneous deacidification-reduction of paper evaluated after artificial and natural aging, Journal of Cultural Heritage (2016), http://dx.doi.org/10.1016/j.culher.2016.02.008
G Model CULHER-3039; No. of Pages 8
ARTICLE IN PRESS M. Bicchieri, A. Sodo / Journal of Cultural Heritage xxx (2016) xxx–xxx
8
References [1] J.W. Baty, C.L. Maitland, W. Minter, M.A. Hubble, S.K. Jordan-Mowery, Deacidification for conservation, BioResources 5 (3) (2010) 1955–2023. [2] G.C. Andrews, Chemoselectivity in the reduction of aldehydes and ketones with amine boranes, Tetrahedron Lett. 21 (1980) 679–700. [3] D.J. Raber, W.C. Guida, D.C. Shoenberger, Reduction of aldehydes and ketones with tetraalkylammonium borohydrides, Tetrahedron Lett. 22 (1981) 5107–5110. [4] E.R.H. Walker, The functional group selectivity of complex hydride reducing agents, Chem. Soc. Rev. 5 (1) (1976) 23–50. [5] D. Clement, The blistering of paper during hydrogen peroxide bleaching, J. Am. Inst. Conserv. 23 (1) (1983) 47–62. [6] H.D. Burgess, The stabilization of cellulosic fibres by borohydride derivatives, ICOM Comm. Conserv. 2 (1990) 447–452. [7] M. Bicchieri, M. Bella, F. Sementilli, A quantitative measure of borane tertbutylamine complex effectiveness in carbonyl reduction of aged papers, Restaurator 20 (1) (1999) 22–29. [8] C. Pedneault, S. Robert, C. Pellerin, Bleaching with new reductive chemicals: replacement of hydrosulfite, Pulp Pap. Can. 98 (3) (1997) 51–55. [9] M. Bicchieri, F.M. Sementilli, A. Sodo, Application of seven borane complexes in paper conservation, Restaurator 21 (4) (2000) 213–228. [10] M. Bicchieri, M. Monti, M.L. Antonelli, A new low-cost and complete restoration method: a simultaneous non-aqueous treatment of deacidification and reduction, in: J. Apuente (Ed.), Proceedings of the 3rd International Conference on Science and Technology for the safeguard of cultural heritage in the Mediterranean basin, I, Universitad de Alcalà, Alcalà, 2001, pp. 276–280. [11] C. Sanna, A. Sodo, G. Laguzzi, G. Mancini, M. Bicchieri, Tert-butyl amine borane complex: an unusual application of a reducing agent on model molecules of cellulose based materials, J. Cult. Herit. 10 (3) (2009) 356–361. ˇ [12] J. Maleˇsiˇc, M. Kojc, Selih, Assessment of the effect of various bleaching agents on papers with foxing stains, Restaurator 29 (3) (2008) 142–154. [13] C.F. Lane, Ammonia-borane and related N-B-H compounds and materials: safety aspects, properties and applications, N-B-H Survey (2006) 1–33. [14] E.H. Immergut, B.G. Ranby, H.F. Mark, Recent work on molecular weight of cellulose, Ind Eng Chem 45 (11) (1953) 2483–2490. [15] J. Lauriol, J. Comtat, P. Froment, F. Pla, A. Robert, Molecular weight distribution of cellulose by on-line size exclusion chromatography - Low angle laser light scattering, Holzforschung 41 (3) (1987) 165–169. [16] A.M. Mattson, C.O. Jensen, Colorimetric determination of reducing sugars with triphenyltetrazolium chloride, Anal Chem 22 (1950) 182–185. [17] J. Silvy, J.F. Le Nest, A photochromic effect shown during the aging of a cellulose material by heat and light, in: F. Bolam (Ed.), The fundamental properties of
[18]
[19] [20]
[21]
[22]
[23] [24]
[25]
paper related to its uses: transactions of the symposium held at Cambridge, September 1973, The British Paper and Board Industry Federation, London, 1976, pp. 755–760. U. Henniges, M. Schwanninger, A. Potthast, Non-destructive determination of cellulose functional groups and molecular weight in pulp hand sheets and historic papers by NIR-PLS-R, Carbohydr. Polym. 76 (3) (2009) 374–380. M. Ali, A.M. Emsley, H. Herman, R.J. Heywood, Spectroscopic studies of the aging of cellulosic paper, Polymer. 42 (7) (2001) 2893–2900. D. Lichtblau, M. Strliˇc, T. Trafela, J. Kolar, M. Anders, Determination of mechanical properties of historical paper based on NIR spectroscopy and chemometrics - a new instrument, Appl. Phys. A 92 (1) (2008) 191–195. M. Bicchieri, A. Sodo, G. Piantanida, C. Coluzza, Analysis of degraded papers by non-destructive spectroscopic techniques, J. Raman Spectrosc. 37 (10) (2006) 1186–1192. S. Iannuccelli, S. Sotgiu, A new methodology for wet conservation treatments of graphic art on paper with a rigid polysaccharide gel of Gellan gum, in: L. Watteuw (Ed.), GDWG ICOM-CC Interim Meeting, The Royal Library of Denmark, Copenhagen, 2010, pp. 47–51. http://www.icpal.beniculturali.it/allegati/Capitolato Speciale Tecnico Tipo.pdf. G. Pace, Giuseppe De Nitti’s graphical works in the Museo Civico di Barletta – Restoration, in: F. Fiorani, R. Di Noia (Eds.), De Nittis incisore, Artemide, Roma, 1999, pp. 281–287. M. Bicchieri, M. Monti, G. Piantanida, F. Pinzari, S. Iannuccelli, S. Sotgiu, et al., The Indian drawings of the poet Cesare Pascarella. Non destructive analyses and conservation treatments, Anal. Bioanal. Chem. 402 (4) (2012) 1517– 1528.
Further reading ASTM D1795-13 Standard test method for intrinsic viscosity of cellulose. ISO 5630-3:1996 (Rev. 2012) Paper and board – Accelerated ageing – Part 3: moist heat treatment at 80◦ C and 65% relative humidity. ISO 2470-1:2009 Paper, board and pulps – Measurement of diffuse blue reflectance factor – Part 1: indoor daylight conditions (ISO brightness). TAPPI T 435 om-11. Hydrogen ion concentration (pH) of paper extracts (hot extraction method).
Please cite this article in press as: M. Bicchieri, A. Sodo, Alcoholic deacidification and simultaneous deacidification-reduction of paper evaluated after artificial and natural aging, Journal of Cultural Heritage (2016), http://dx.doi.org/10.1016/j.culher.2016.02.008