New colourimetric method for determination of hydroxyl radicals during ageing of cellulose
Jana Kolara,*, Matija Strličb, Boris Pihlarb
aPreservation Department, National and University Library, Turjaška 1, SI-1000 Ljubljana, Slovenia
b Faculty of Chemistry and Chemical Technology, University of Ljubljana, Aškerčeva 5, SI-1000 Ljubljana, Slovenia
Received 29 August 2000; accepted 17 November 2000
Abstract
It is believed that oxidative degradation involving free radicals leads to the deterioration of alkaline paper during ageing. However, due to the high detection limit of conventional methods, e.g. ESR, the presence of radical species during ageing of pure cellulose at close to ambient conditions has not yet been established.
In the present work, N,N'-(5-nitro, 1,3-phenylene)bisglutaramide is used as a colour indicator of hydroxyl radical production in cellulosic materials. Quantitative determinations of its hydroxylated products can be performed using colourimetry. The measurements are rapid, non-destructive and demand a small amount of sample, which makes the new method attractive for use in industrial laboratories for a variety of materials. Several sources of systematic errors are identified and evaluated.
Hydroxyl radicals are shown to be important intermediates in the oxidative processes during ageing of alkaline cellulose (pH > 7). Furthermore, the importance of alkalinity on oxidative degradation of bleached chemical pulps is demonstrated. © 2001 Elsevier Science B.V. All rights reserved.
Keywords: Colourimetry; Viscometry; Hydroxyl radicals; Cellulose; Autoxidation; Degradation
1. Introduction
Preservation of decaying historical paper artefacts poses a severe problem for numerous libraries, archives and museums world-wide and a substantial effort is invested into the design of suitable conservation methods. In order to develop much needed stabilising treatments, a superior knowledge of the degradation chemistry of cellulose is imperative. It has been firmly established by now, that acid hydrolysis is responsible for rapid deterioration of acidic paper and mass decidification systems are now in use to adjust the pH of paper. The research dedicated to
* Corresponding author. Fax: +38-611257293. E-mail address: (J. Kolar).
the ageing of alkaline paper, however, remains scarce [1]. The facts that degradation of paper is significantly slower in the absence of oxygen [2] and that macromolecular hydroperoxides are produced during ageing of cellulose [3,4] suggest the importance of oxidative reactions, probably involving radical species [5,6]. Besides, the role of hydroxyl radicals has been postulated and recognised in oxidation reactions in biological systems [7-10], ecology [11-14], food chemistry [15,16], and organic material degradation [1,17,18]. Degradation processes involving free radicals could be investigated by electron spin resonance spectroscopy (ESR) in situ, however, probably due to the low concentration of the radicals, ESR investigations gave no useful data for purified cellulose (Ahlstrom Filtration filter paper) [19]. In the case of bleached chemical pulps, which contain non-cellulosic components, such as hemicelluloses and lignin, the relative radical concentrations under different conditions have been determined, however, radical structures were not identified [20].
Apart from ESR, trapping the radicals with a reactive substance, yielding more stable products which can be determined quantitatively, is a popular indirect method for detection of radical species in many systems. In the present work, the N,N'-(5-nitro, l,3-phenylene)bisglutaramide (NPG) hydroxylation assay for rapid spectrophotometric determination of the rate of hydroxyl radical production in aqueous solutions [21,22], has been adapted for use in solid cellulose. It is shown that the method can be successfully applied to the studies of alkaline cellulosic systems and that it is useful for fast determinations of the initial rates of cellulose degradation thus enabling us to evaluate new antioxidant formulations reliably in a very short time. The hydroxylation assay can also be used to gain a deeper insight into the effect of alkaline pH on ageing behaviour of different cellulose samples.
2. Experimental
The cellulose samples in this work were purified cellulose (sample no. 1, Whatman filter paper no. 1, degree of polymerisation — DP = 2902) and bleached chemical cellulose samples: Riverdale sulphate pulp (sample no. 2, DP = 2535); Royal Kraft sulphate pulp (sample no. 3, DP = 2735) and MoDo Crown sulphite pulp (sample no. 4, DP = 4206).
N,N'-(5-nitro,l ,3-phenylene)bisglutaramide (NPG) was synthesised according to the literature [22]. 1H NMR, melting point (192.7-194.3°C) and Chromatographic analyses indicate satisfactory purity (99%). HPLC was performed using a Hewlett-Packard 1100 Series Chromatographic system equipped with a diode array detector (detection at 222, 232 and 431 nm). A Hypersil ODS Chromatographie column was used (HP, 250 mmx 4 mm, 5 |xm particle size) and 100 u,l of sample was injected. Gradient elution of 1.5 ml min-1 was used and consisted of 3-8% acetonitrile and of 97-92% phosphate buffer (20mmoll-1, pH 7) in 5 min, after which the unreacted NPG was flushed out of the column with a 30% acetonitrile/70% phosphate buffer eluent. The hydroxylated derivates were isolated by preparative HPLC (HP ODS Hyperprep, 8 (Am). The resulting eluates were acidified with sulphuric acid and extracted with ethylacetate (for spectroscopy, Carlo Erba, Miláno). The solvent was removed and the derivates were dried over P4O10. The remaining inorganic impurities (PO43-, SO42-) were determined using a Merck-Hitachi HPLC ion Chromatographie system with a Dionex IonPac AS4A-SC 4 mm column, carbonate buffer eluent, and conduc-tometric detection with anion suppressor and H2SO4 as the régénérant. The unreacted NPG and inorganic impurities amounted to 1% (w/w) of the isolated products.
The cellulose samples were prepared by immersion into aqueous solutions of Ca(OH)2, 0.01 moll-1 or Mg(HCO3)2, 0.04 moll-1. This introduces approximately the same amount of alkali into cellulose samples (~0.03mmolg-1). Half of the dried samples were then immersed into solutions of NPG 0.01 mol 1-1, while pH of these solutions was adjusted to 7 with KOH.
All samples were aged under ISO Standard 5630/3 conditions (80°C at 65% RH) in a Vötsch VC 0020 ageing oven. After a desired period of ageing, ca. 0.07 gof paper sample was extracted into 5 ml of 0.1 mol 1-1phosphate buffer (pH 7) by sonification (3 h) in order to perform Chromatographie determinations of NPG and its derivates.
Colourimetric measurements of solid samples were performed with a Minolta CM-3610d diffuse reflectance UV-VIS spectrophotometer at 460 nm with the specular component excluded. The reflectance was measured in % relative to polymeric Minolta standard. Typical R.S.D. was below 0.7%.
Spectrophotometric measurements of NPG and of the two isolated derivates in buffered solutions were performed with a PE Lambda 2 double beam spectrophotometer with a 1 cm quartz cuvette.
In order to estimate the pH of cellulosic samples, they were extracted 30 min in MilliQ water (2.7 ml water per gram of sample), thermostated at 80°C for 30 min and the pH of the water extract was then measured using a flat combined electrode.
Viscometric determinations of the degree of polymerisation (DP) were performed according to the standard procedure (SCAN-CM 15:88), using lmol-1
Fig. 1. UV-VIS absorption spectra for o-HNPG (c = 0.0284 mmoll-1) and p-HNPG (c = 0.0310mmoll-1) in aqueous solution, buffered at pH 7.
Fig. 2. Absorbances of aqueous solutions of o-HNPG at 422nm (c = 0.0142mmoll-1) and p-HNPG at 444nm (c = 0.0310mmoll-1) buffered to corresponding pH.
cupriethylenediamine solvent (Carlo Erba, Miláno), with the typical R.S.D. below 0.8%.
3. Results and discussion
The reaction of N,N'-(5-nitro,1,3-phenylene)bisglu-taramide (NPG) with hydroxyl radical yields coloured o- and p-hydroxylated derivates (o-HNPG and p-HNPG) which may be determined either spectropho-tometrically or chromatographically. The UV-VIS absorption properties of o-HNPG and p-HNPG (Fig. 1 ) are strongly pH-dependant, though the absorption spectra and the relevant molar absorptivities at 431 nm remain constant at pH > 7 (Fig. 2). NPG and its derivates may thus be determined by the use of spec-trophotometry without adjustments of pH, as long as the samples remain alkaline during the degradation experiments.
3.1. Stability studies
Before an attempt is made to adapt the NPG hy-droxylation assay for determination of the rate of hydroxyl radical production in aqueous solutions for determinations of hydroxyl radical production during accelerated ageing of solid-state cellulose, the stability of NPG and its derivates at elevated temperatures of accelerated ageing (80°C) needs to be addressed due to relatively long experimental times. The possible degradation of NPG hydroxylation products was tested in the following way: a 0.01 moll-1 solution of NPG in phosphate buffer (pH 7) was prepared and some H2O2 was added to provide an initial concentration of the derivates, o-HNPG and p-HNPG. The flask was stoppered, and kept at 82°C in the darkness. From the data presented in Fig. 3, it can be concluded, that after an initial rise in absorbance of visible light at 431 nm, corresponding to the consumption of H2O2, the absorbance does not change during the following month. It is, therefore, very likely that the derivates neither degrade, nor are they formed in the absence of oxidants. Another experi-
Fig. 3.Changes in the absorption of visible light at 431nm of a solution of NPG and its hydroxylated derivates, kept at 82°C(□) and of solutions, prepared from dry NPG, kept at 82°C (O) for a prolonged period of time.
ment was performed to test the stability of dry NPG during a prolonged period of heating. Equal portions (2ml) of a 0.01 moll-1 solution of NPG in phosphate buffer (pH 7) were dried up and after a period of heating at 82°C in darkness, the dry matter was dissolved and absorption of the obtained solution was measured at 431 nm. There is a minimal increase in absorption amounting to 0.22% per day, indicating that in the presence of air, hydroxylation can take place, although to a very limited extent. This was confirmed by HPLC analyses of the same samples, i.e. the changes in absorption intensity are indeed caused by an increased amount of hydroxylated derivates. However, more importantly, no significant changes in the content of unreacted NPG could be determined, indicating that it is stable and does not evaporate. This is not surprising, given that NPG melts at 192-194°C.
Having addressed the stability and volatility of the substrate and of its derivates at the conditions of accelerated ageing, the attempt was made to adapt the NPG hydroxylation assay for determination of the rate of hydroxyl radical production during accelerated ageing of solid-state cellulose.
3.2. Correlation of reflectance with HO' production
Addition of NPG to the cellulose leads to a marked decrease of its diffuse reflectance after a period of accelerated ageing (80°C, 65% RH) with the absorbance maximum at 450-460 nm (Fig. 4).
A comparison of colourimetric and Chromatographic determination of derivates was necessary to establish whether paper colouration is a consequence of NPG hydroxylation. A series of aged samples were extracted in 0.1 moll-1 phosphate buffer (pH 7) and the hydroxylated derivates in the resulting extracts were determined by use of HPLC. A considerable error is associated with the aqueous extraction of paper samples as retention of NPG in the paper samples causes the correlation of the diffuse reflectance at 460 nm (%) versus the sum of p-HNPG and o-HNPG concentration in paper (mmolg-1) to be less than ideal R460= (71.8 ± 0.9) - (52 ± 4)c, R2= 0.8746, N= 22. Since the variable composition of paper samples (especially due to different contents of Fe, Ca and Mg) influences their adsorption properties, evaluation of the systematic error is difficult. Nev-
Fig. 4. Decrease of diffuse reflectance during ageing of cellulose sample no. 1, enriched with calcium carbonate and with an addition of NPG, after 6, 13 and 20 days of accelerated ageing (80°C, 65% RH), relative to the control (unaged) sample.
ertheless, the correlataion confirms that the colour formation may indeed be ascribed to the production of HO●
3.3. Evaluation of errors
3.3.1. NPG consumption
As already shown, the substrate neither evaporates nor degrades at the mild degradation conditions used. From the Chromatographic analysis performed above, it is also established, that there is no significant change in the NPG concentration in the paper during a 20-day ageing experiment and that during this period of time only a small fraction of the original NPG reacts with hydroxyl radicals. The initial amount of NPG introduced into sample no. 1 by immersion is 20.6μmolg-1, R.S.D. = 6%, while the percentage of the consumed substrate in the case of the most degraded sample studied (ΔR460= —50%), amounts to 2% only, which is well within the R.S.D. Thus, we can assume that at least for samples with diffuse reflectance at 460 nm decreased by 50%, the amount of decrease of NPG content due to its hydroxylation during the ageing experiment consumed does not influence the kinetics of formation of coloured hydroxylated derivates. This implies that the concentration of NPG is still high enough to ensure stationary reaction conditions with respect to the steady-state concentration of HO● in the paper samples during ageing.
3.3.2. Product ratio and sample pH
Although the molar absorptivities of the hydroxy-lated derivates at 431 nm remain constant at pH > 7 (Fig. 2), the ratio of p-HNPG and o-HNPG products during hydroxylation of NPG in an aqueous solution was shown to be pH-dependant (Cp-hnpg/co-hnpg = 1.6-2.0) [21]. As the two derivatives have different extinction coefficients (ε422 = 60201 cm-1mol-1for o-HNPG and ε444 = 87101cm-1mol-1 for p-HNPG) [21], a systematic error may be associated with the reflectance measurements when comparing samples of different alkaline pH. In order to establish whether this is the case during ageing of cellulose, various samples were buffered by an addition of calcium or magnesium carbonate, yielding pH values of cellulose extracts 7.9 and 9.0, respectively. After a period of accelerated ageing of the buffered samples, they were subjected to extraction in phosphate buffer (0.1 mol 1-1, pH 7) and the resulting extracts were analysed using HPLC. The ratio of products was calculated from the ratio of peak areas, taking into account the slopes of Chromatographic calibration curves for p-HNPG: peak area431nm= (8490 ± 35)c, R2= 0.9997, N = 10; and o-HNPG: peak area431nm= (2200 ± 60)c, R2= 0.9999, N = 10. The ratio cp-HNPG/co-HNPG was 1.31 ±0.13 (N = 14) for the samples containing calcium carbonate and 1.46 ± 0.12 (N = 16) for those buffered with magnesium carbonate. Taking into account the molar extinction coefficients at the absorbance maxima in phosphate buffer solutions at pH 7 (£422 = 60201cm-1mol-1for o-HNPG and £444= 87101cm-1mol-1for p-HNPG) [21], the difference in product ratio accounts for a systematic error of +1.4% R.S.D. in the colourimetrically determined concentration of the hydroxylated NPG dérivâtes in samples containing MgCO3 if compared to those buffered with CaCO3.
3.3.3. Cellulose chromophore formation
The absorption of light of both derivates is stronger in the UV region which implies an advantageous S/N ratio for colourimetric determinations. However, it turns out that this is not necessary since changes in the visible region are intensive enough to be measured with a typical R.S.D. within 1%. Besides, measurements in the UV region would be rendered considerably more difficult in the frequent cases when organic, highly UV-absorbent antioxidants are studied.
However, even in the visible region the accumulation of cellulose-derived chromophores, which accompanies ageing of cellulosic materials, introduces a typical systematic error of +1% R.S.D. into the measurements of reflectance at 460 nm after 20 days of accelerated ageing. This error is variable and depends on the type of cellulose, but can easily be determined by performing a control experiment. Its contribution is negligible during the initial phases of degradation, for which the presented colourimetric method is most useful.
3.3.4. Colourimetry versus viscometry
It is interesting to compare the typical errors associated with the two methods. While both methods are associated with a relatively small random error (see Section 2), from Fig. 5, we can establish that at least during the initial phase of degradation, the colourimetric determinations are more reliable than the vis-cometric, as the colour changes are more pronounced than the corresponding decrease of DP. This makes the new colourimetric method particularly useful for following the initial period of cellulose degradation.
3.4. Some preliminary studies
Having established that hydroxylation of NPG occurs during accelerated ageing of alkaline cellulosic materials, and that the derivates can be determined either spectrophotometrically or chromatographically, the attempt was made to evaluate the effect of al-
Fig. 5. Decrease of DP (of samples without NPG) vs. decrease of diffuse reflectance (of samples with NPG) of bleached chemical pulp no. 3 after 2, 6, 13 and 20 days of accelerated ageing. C and M denote sample no. 3 with an addition of CaCO3 or MgCO3, respectively, the number indicates the duration of accelerated ageing.
kaline pH on oxidative degradation of different cellulose samples during accelerated ageing. The same samples enriched with CaCO3 or MgCO3 were used as described above (denoted C and M, respectively). In Fig. 5, it can be seen that the decrease of diffuse reflectance of the NPG containing bleached chemical pulp no. 3 after a period of accelerated ageing is related to the decrease of DP of pulp samples, prepared in the same way, but without an addition of NPG and may be described by a regression line ADP = (0.54±0.02)AR460, R2= 0.9797, N = 8. Due to the extensive oxidation of the samples with higher pH (M, containing MgCO3) which resulted in a decrease of pH as the alkaline reserve was depleted, the samples M13 and M20 were excluded from further analysis.
From Fig. 5, it can be concluded that formation of hydroxyl radicals is the driving force behind the chain-scission process and, conversely, ageing of alkaline bleached cellulose pulp. A pH dependence of the process is also suggested.
In another experiment, various samples were aged for a given period of time (8 days) to obtain a comparison between four different samples enriched with CaCO3 and MgCO3(Fig. 6). Three samples are comparable in origin (bleached chemical pulps-samples nos. 2-4), however, purified cotton cellulose (sample no.l) was also included. From the figure, we may conclude that a higher pH of the bleached chemical pulp samples containing MgCO3 leads to a higher production of HO● and a higher rate of bond scission. This can be supported by a reasonable correlation between HO● production and chain scission (ΔDP =
Fig. 6. Decrease of DP (of samples without NPG) vs. decrease of diffuse reflectance (of samples with NPG) after 8 days of accelerated ageing of samples nos. 1-4 with additions of CaCO3 (C) and MgCO3 (M).
(1.89 ± 0.11)ΔR460- (16.8 ± 2.0), R2= 0.9813, N = 8). At the same time, it can be observed that pH does not have such a pronounced effect on ageing of purified cellulose and that the overall rate of chain scission is smaller. This suggests that the pH dependant degradation processes of bleached chemical pulps originate from sample components other than cellulose itself, e.g. lignin residues or non-cellulosic low-molecular weight polysaccharide impurities.
The two experiments clearly show that the quality of experimental data obtained by the new colourimet-ric method is high and that the information obtained can lead to interesting new conclusions regarding the mechanisms of alkaline paper degradation. A full evaluation of similar studies is currently under way in our laboratories.