Chemical modification of Norway spruce and Scots pine. A 13 C NMR CP-MAS study

M.G. Boonstra et aL: Wood Components Reactions with Acetic AnhydrideHolzforschung50(1996)215-220215Chemical Modification of Norway Spruce and Scots PineA 13C NMR CP-MAS Study of the Reactivity and Reactions of Polymeric Wood Componentswith Acetic AnhydrideBy M.G. Boonstra1, A. Pizzi2 5, P. Tekely3 and J. Pendlebury41 TNO, Group for Timber Research, Delft, The Netherlands2ENSTIB, Universite de Nancy 1, Epinal, France3Laboratoire de Methodologie RMN, Universite de Nancy 1, Nancy, France4 N. Z. Forest Research Institute, Rotorua, New ZealandKeywordsWoodAcetylationAcetic anhydrideWood protectionWood modificationCross-linkingLigninCelluloseSummaryAnalysis by CP-MAS 13C NMR of solid samples of wood, lignin, cellulose and holocellulose treatedwith acetic anhydride gas showed that in wood both acetylation and substitution of free sites of thearomatic ring occur readily in lignin both when alone and when in presence of carbohydrates such asin the wood matrix; that further rearrangements of the new substituents on the aromatic nuclei of ligninalso occur inferring real or potential network cross-linking; that acetylation on the carbohydrates occursto a considerably lesser extent than on lignin; that acetylation proper occurs on the carbohydrates atC6, and to a greater extent when in the wood then when alone due apparently to the presence of anadditional acetylation site.IntroductionChemical modification of wood by reaction withacetic anhydride has been found to lead to dimen-sional stabilization of wood. Numerous studies onthis treatment have appeared (Arodra et aL 1981;Bekere et al. 1978; Callow 1952; Clermont andBender 1957; Goldstein et al. 1961; Hon 1994; Honand Bangi 1994; Klinga and Tarkow 1966; Matsuda1987; Nikitina 1977; Nishimoto and Imamura 1985;Peterson and Thomas 1979; Placket and Cohen 1994;Risi and Arsenau 1957; Rowell 1982, 1991; Rowellet al. 1981, 1986; Sudo 1979; Tillman 1987; Tillmanet al. 1985; Truskne 1977; Youngquist et al. 1985).Recently two studies have focused on which woodconstituents the reaction of the anhydride is likely tooccur. One of these studies identified the reaction tobe considerable on wood lignin and equally consid-erable, but to a lower extent than in lignin, on thewood holocellulose (Rowell et al. 1994). The tech-niques used in this study were infrared spectroscopyand X-ray dispersive analysis (EDX). The otherstudy (Pizzi et al. 1994) started instead from theconcept that acetylation leads in reality also to someacylation on the aromatic nuclei of lignin and henceto possible cross-linking of the wood network withconsequent improved dimensional stability. In thisstudy acetylation of the phenolic hydroxy groups oflignin was observed, followed in time by a Fries5rearrangement to an acylation and further rearrange-ments suggesting potential cross-linking. The tech-nique used in this latter study was 13C NMR of liquidsamples of lignin treated with acetic anhydride im-mediately after treatment, and 4 days and 18 daysafter treatment.Both the above studies suffer of some disadvantages.The former because techniques of lower reliability inorganic chemistry of natural products were used,leaving open some doubts as to the results and failingto address what reactions are really involved. Thelatter because, while a very reliable technique wasused, it addressed only one of the polymeric con-stituents of wood, namely lignin, without consideringthe possible effect of acetic anhydride treatment onany of the others, and without quantifying the extentof the reactions observed.It is for these reasons that it was decided to combinethe advantages of the two studies and to examine onwood, and on isolated lignin, isolated cellulose andisolated holocellulose what the extent of the differentreactions were and to follow by 13C NMR CP MASthe chemical modifications induced in the polymericconstituents in their solid state.ExperimentalTreatment of samplesNorway spruce (Picea abies (L.) Karst) fibers, isolated alkalilignin (Norway spruce), holocellulose (Norway spruce) and cellu- Author to whom correspondence should be addressedHolzforsehung / Vol. 50 / 1996 / No. 3Brought to you by | Sichuan University (Sichuan University)© Copyright 1996 Walter de Gruyter · Berlin · New YorkAuthenticated | 172.16.1.226

Download Date | 4/19/12 1:19 PM

216M.G. Boonstra et al: Wood Components Reactions with Acetic Anhydridelose (cotton) were treated with vaporized acetic anhydride (120°C,10mm Hg) resulting in a weight percent gain (WPG) of 0.2% tominimum 77% of the WPG of spruce wood is due to43%. The samples were oven dried (104°C, 24 hours) beforereactions of the anhydride with lignin.treatment. Results of the treatments are given in Table 1.This distribution of results is very different fromI3CNMR CP-MAS analysiswhat reported previously by other authors (RowellSolid state of the untreated and acetic anhydride treated lignin,et al. 1994) which have found on southern yellowcellulose, holocellulose and wood samples were examined bypine that the percentage of acetyl groups on holo-CP MAS (Cross-polarisation Magic-angle-spinning) I3C NMR.cellulose is of between 12% and 25% of availableThe spectra were obtained on a Bruker MSL300 FT-NMR spec-holocellulose hydroxyls against 80% to 100% lignintrometer, at a frequency of 74.47 MHz. Chemical shifts werehydroxyls substituted in the range 8.5 to 23.5 WPG.calculated relative to TMS for NMR control. All the spectra wererun overnight. Aquisition time was of 0.026 seconds with numberIn our case the difference observed is consider-of transients of about 10000. All the spectra were run withably greater (Table 1). While this might appearrelaxation delay of 5 seconds and were accurate to Ippm. Theastonishing, it is not really so if one considers thatspectra were run with spinning side bands CP time of 1 ms andRowell et al. (1994) used a liquid treatment systemspectral width of 20 000 Hz. Typical spin lattice relaxation timesfor the types of compounds analyzed were taken from thewhile the present are obtained with a anhydride gasliterature (Breitmaier and Voelter 1987). The results obtained musttreatment system at 120 °C. Gas reactions are well-rather be interpreted as qualitative rather than quantitative.known to be different from liquid ones mainly due tosubstrate swelling, reagent distribution and penetra-Discussiontion of reactants.The results of the reaction of vaporized acetic anhy-While the macroscopic results of weight gain aredride with alkali lignin, cellulose, holocellulose andquantitative, it is also interesting to observe, althoughspruce wood fibres are shown in Table 1. The fourat the qualitative rather than quantitative level, whatmaterials were treated at 120°C from 0 minutes (notappears to occur at the molecular level when thetreated) to 960 minutes for spruce wood fibre and tosamples treated with acetic anhydride are examined960 minutes for alkali lignin, cellulose and holo-by 13CNMR CP-MAS. In Figures la-c 13CNMRcellulose. In Table 1 it is interesting to note thetraces of solid samples of spruce lignin at WPGconsistant increase in weight percent gain (WPG)of 0% (0 minutes), 20% (30 minutes) and 43.2%of the samples on treatment with acetic anhydride.(960 minutes) are reported. The 125ppm band char-While wood fibres and especially lignin presentacteristic of (Ar)C-C increases indicating that sub-noticeably higher WPG as the time of treatmentstitution on the aromatic ring of lignin has occurred,lengthens, this is hardly the case for the carbo-confirming previous results (Pizzi et al. 1994). Ahydrates. Both cellulose and holocellulose showfurther band at 138ppm appears and increases inWPGs of 4.1% and 4.9% respectively after 960intensity at increasing WPGs: this is a different typeminutes treatment, a very low level in relation to theof (Ar)C-C(-O-) which did not exist in the untreated43.2% WPG obtained in lignin. It is then clear that,sample where a similar signal but at 132ppm isat least in spruce, after 960 minutes of treatment thenoted. Thus, the (Ar)C-C(-O-) formed is a different,greater proportion of the anhydride has reacted withadditional substituent on the aromatic ring of lignin.lignin rather than with the carbohydrates and that atIts existence indicates a derivative of the possibleattack of an acylic group on the ring. Furthermore,the appearance of the small, but clear 193ppm bandTable 1. WPG (%) of Norway spruce fibers, isolated lignin,indicates the presence of small amounts of aromatic-isolated holocellulose and isolated cellulose with vapour aceticlinked ketogroups ((Ar)C-C=O), again indicatinganhydridethe presence of acylation as a side reaction of theReaction time (min) WPG (%)treatment. This is consistent with the previous obser-vations (Pizzi et al. 1994).Norway spruce fibers0The decrease of the HOppm band indicates a de-240018.8960crease in the proportion of the free aromatic reactive33.1sites of lignin, again confirming that some reactionAlkaline lignin0300on the aromatic nuclei has occurred. Furthermore, the20.696043.2variations in respective relative intensities of theHolocellulose060o'125ppm, 118ppm and HOppm bands indicate vari-ation in the relative proportions of different types of9600.44.9aromatic sites as the anhydride treatment proceeds.Cellulose0The decrease in intensity of the LlOppm band and3009600.2respective increase of the 125ppni band appears to4.1indicate that some substitution at position 5 of thearomatic ring of lignin has occurred. Of interest areHolzforschung / Vol. 50 / 1996 / No. 3Brought to you by | Sichuan University (Sichuan University)

Authenticated | 172.16.1.226Download Date | 4/19/12 1:19 PM

M.G. Boonstra et al: Wood Components Reactions with Acetic Anhydride217a20B180140-rfr100PPM80604020I . . . ι . . . ι . . ι . . . ι . . . | . . . ι . _. . ι . . . ι . . . ι200 1B0 160 140 120 100 80 60 40 20PPMalso the bands at 87ppm and lOOppm representativeof C=C and C=C-0 groupings, indicating again (Pizziet al. 1994) that some more complex rearrangementsappear to follow after acetylation and acylation in-duced by the anhydride treatment of lignin.Proper acetylation is also evident from the bands at148 and 152ppm, both representing (Ar)C-O-, thefirst initially decreasing and then disappearing atprogressively increasing treatment with anhydride,and the second instead first appearing (20% WPGcase) and then increasing in intensity (43.2% WPGcase). These two bands indicate that acetylation ofthe phenolic hydroxy groups of lignin has indeedoccurred. The disappearance of one band (148ppm,(Ar)C-OH) and the appearance of the other one(152 ppm, (AR)C-OOC-CH3) indicates qualitativelythat acetylation of the aromatic hydroxy group oflignin has been just about total at 43.2% WPG. Asregards acetylation the width of the sharp band at170ppm (CH3-COO- aromatic, Fig. Ic) prevents toascertain if both alyphatic as well as aromatichydroxy groups of lignin have been acetylated. Thesame band width however infers that acetylation atthe aliphatic -OHs of lignin has also occurred. Con-versely, the very small band at 175-177 ppm belongsto the CH3-COO~ residue of the acid, again confirm-ing from its very low intensity that acetylation hasbeen quantitative. The quantitative acetylation of lig-nin finds confirmation in the position of the methylband of the acetyl group being at 25 ppm. The samemethyl group in unreacted acid is at 20-21 ppm andthis is absent in the spectra.The apparent lack of variation of the primary -OCH3band at 5 8 ppm and the noticeably but small decreaseof the secondary -OCH3 band at 39 ppm indicates thatsome aromatic demethylation of some type of ligninmethoxy group occurs, but that it is definetely small.In the spectra of cellulose and holocellulose (compareFigs. 2 to 3a-c) it is evident that in pure cellulosethe proportion of crystalline material is higher thanin holocellulose, as could be expected; this can beseen by the higher intensity of the 90 ppm band(crystalline cellulose) in relation to the 85 ppm band(amorphous cellulose) (Tekely 1994) in Figure 2. Inboth Figures 2 and 3 a-c, a clear carbohydrate patternis observed, namely Cl (105 ppm), C4 (90 ppm and85 ppm), C2 (78ppm), C3, C5 (74; 75ppm) and C6(66-67 ppm).. I . · . I . 1 . . I . . . I . . 1 u^J .—J . . I , . 1 . . . I200 180 168 140 120 10B 80 60 40 20 0Fig. 1. a: CP MAS 13C NMR of untreated Norway spruce (Piceaabies) lignin. b: CP MAS 13C NMR of gas-phase acetic anhydridetreated Norway spruce (Picea abies) solid lignin after 30 minutestreatment and at 20.6% weight percent gain (WPG), c: CPMAS13CNMR of gas phase acetic anhydride treated Norway spruce(Picea abies) solid lignin after 960 minutes treatment and at43.2% WPG.Brought to you by | Sichuan University (Sichuan University)Holzforschung / Vol. 50 / 1996 / No. 3Authenticated | 172.16.1.226Download Date | 4/19/12 1:19 PM

218M.G. Boonstra et al: Wood Components Reactions with Acetic Anhydridea•rfr160 140100PPM6060I ... I ... I ... I ... I .... I . . I ... I ... I188 160' 140 120 100 80 ' 60 40 20Fig. 2. CP MAS 13C NMR of gas phase acetic anhydride treatedsolid cotton cellulose» after 60 Min. treatment at 0.2% WPG.A small amount of acetylation clearly appears to haveoccurred (compare Figs. 3a-c) as shown by theappearance of the 22ppm band characteristic of amethyl group of an aliphatic acetyl group, the de-crease of one of the two C6 bands (that at 64ppm;the one at 66ppm remaining unchanged) indicatingthat the small amount of reaction which has occurredis definetely an acetylation of the hydroxy group atC6. The same is observed for cellulose (Fig. 2) whereacetylation appears to occur at C 6 at earlier stagesthat in the case of holocellulose (64ppm band de-creasing to a shoulder). This indicates that for carbo-hydrates a greater proportion of the very limitedacetylation, at least from a qualitative point of view,appears to have occurred on cellulose rather than onhemicelluloses, and at C6, and inparticular at the C6site of the amorphous portion (64ppm) rather than thecrystalline one (66ppm) of cellulose. The slightlyreduced intensity of Cl and C4 bands after acetyla-tion also indicates that a slight amount of depolymeri-zation of cellulose by some, not extensive cleavageof the ß-glucosidic bond has also occurred. Thismight account for the slight decrease in mechan-ical properties observed for acetylated wood (Rowell1991; Rowell era/. 1994).In spruce wood (Figs. 4a-c) the dominant pattern ofthe 13CNMR spectra is that of the carbohydrates.However, the important patterns of lignin clearlyemerge at the maximum WPG after acetic anhydride,treatment. Thus, before and after treatment there isonly a very small decrease of the 58ppm band,indicating that demethylation of the methoxy groupsdoes occur during treatment although this appears tobe only to a very small extent.Passing from 0% WPG to 33% WPG the ligninbands at 125ppm and particularly at 118ppm increaseHolzforschung / Vol. 50 / 1996 / No. 3100PPM100PPMFig. 3. a: CP MAS 13C NMR of untreated Norway spruce (Piceaabies) holocellulose (0.0% WPG), b: CPMAS 13CNMR of gasphase acetic anhydride treated Norway .spruce (Picea abies)solid holocellulose after 60 minutes treatBrtent and 0.4% WPG. c:CP MAS 13C NMR of gas phase acetic anhydride treated Norwayspruce (Picea abies) solid holocellulose after 960 minutes treat-ment and at 4.9% WPG.Brought to you by | Sichuan University (Sichuan University)

Authenticated | 172.16.1.226Download Date | 4/19/12 1:19 PM

M.G. Boonstra et al\\ Wood Components Reactions with Acetic Anhydride219considerably, confirming that new substituents havebeen introduced in the aromatic ring of lignin, at twodifferent types of site. Acetylation proper hasoccurred on lignin by the predominance in the maxi-mum WPG case of the 152ppm band, although in thiscase this band is not as evident in relation to the125ppm and 118ppm bands as much as in the caseof the pure lignin sample. This might mean thatacetylation proper, when lignin and carbohydrates aretogether as in wood, is much less extensive that onpure lignin samples: an important qualitative obser-vation as it switches even further the reaction onlignin in favour of ring substitution and hence poten-tial cross-linking of the total network. In Figures4a-c the methyl band of the acetyl group at 23ppmand the carboxylic group at 170ppm clearly indicatethe presence of acetyl groups on lignin. Further-more a clear emerging peak (Fig. 4c) at 98-100ppmagain confirms the formation of (Ar)C=C-O- or-O-C(C)2-O- groups which did not exist beforetreatment: again an indication that tridimensionalcross-linking has occurred or might be on the way tooccur (Pizzi et al 1994).The spectra of treated spruce wood (Figs. 4a-c) alsoclearly show that acetylation has also occurred on thecarbohydrates portion of the fibres. The decrease anddisappearance of the C6 band of the amorphousregion confirms that acetylation at C6 has clearlyoccurred. Furthermore, differently from what ob-served in the spectra of the carbohydrates alone, inthe wood spectra there is a clear decrease in theintensity of the C2 band indicating that additionalacetylation at C2 might have occurred. This isimportant because it infers that in the fiber itselfreactions of acetic anhydride on wood carbohydrateswhen in presence of lignin occur, which insteadclearly do not when the treatment is done just onpurely the wood carbohydrates.An unusual effect which is worthwhile noting isthe strengthening of the lignin pattern in the series of13C NMR spectra of wood with increasing amounts ofWPG. This is an unusual effect and no immediateexplanation can be offered for it; increased spacingdue to the bulking action of the chemical and theslight carbohydrates cleavage by the treatment couldcause a more open structure of the network to ensue,and hence the more noticeable NMR signals patternof lignin in relation to the rest of the spectrum.In conclusion, the consequences of all the above arethat in wood both acetylation and substitution of freesites of the aromatic ring occur readily in lignin bothalone and when in presence of carbohydrates; thatfurther rearrangements of the new substituents onthe aromatic nuclei of lignin also occur inferring realor potential network cross-linking; that acetylationproper occurs on the carbohydrates to apparently a180PPM200 1B0 160ι . 1 . ι . I . ι . I . ι140 120 100100PPMFig. 4. a: CP MAS 13C NMR of untreated Norway spruce (Piceaabies) solid wood, b: CP MAS I3C NMR of gas phase aceticanhydride treated Norway spruce (Picea abies) solid wood after240 minutes treatment and at 18.8% WPG. c: CP MAS UC NMRof gas phase acetic anhydride treated Norway spruce (Picea abies)solid wood after 960 minutes treatment and at 33.1 % WPG.Brought to you by | Sichuan University (Sichuan University)

Holzforschung / Vol. 50 / 1996 / No. 3Authenticated | 172.16.1.226Download Date | 4/19/12 1:19 PM

220M.G. Boonstra et al\\ Wood Components Reactions with Acetic AnhydridePizzi, A., A. Stephanou, M.G. Boonstra and J. Pendlebury. 1994.A new concept on the chemical modification of wood byorganic anhydrides. Holzforschung (Suppl.) 48: 91-94.Risi, J. and D.F. Arsenau. 1957. Dimensional stabilization ofwood. For. Prod. J. 7: 210-213.Rowell, R.M. 1982. Distribution of reacted/qhemicals in southernpine modified with acetiy anhydride. Wood Science 15: 172-182.Rowell, R.M. 1991. Chemical Modification of Wood. In: Woodand Cellulose Chemistry. Eds. D.N.S. Hon and N. Shiraishi.Marcel Dekker Inc., New York. Chapter 15, pp. 703-756.Rowell, R.M., W.C.Feist and W.D.Ellis. 1981. Effects ofweathering on chemically modified southern pine. WoodScience 13: 202-208.Rowell, R.M., R. Simonson, S. Hess, D.V. Plackett, D. Cronshowand E. Dunningham. 1994. Acetyl distribution in acetyiatedwhole wood and reactivity of wood cell-wall components toacetic anhydride, Wood and Fiber Sei. 26: 11-18.Rowell, R.M., A.-M. Tillraan and R. Simonson. 1986. A simpli-fied procedure for the acetylation of hardwood and softwoodflakes for flakeboard production. J. Wood Chem. Technol. 6:427-448.Sudo, K. 1979. Character of hardboards from acetylated Aspenpulp. Mokuzai Gakkaishi 25: 203-208.Tekely, P. 1994. Unpublished results.Tillman, A.-M. 1987. Chemical modification of lignocellulosicmaterials. A Comparison of processes. Chalmers Universityof Technology. Department of Engineering Chemistry II,Göteborg.Tillman, A.-M., R. Siraonson and R.M. Rowell. 1985. Dimen-sional stability and biological resistance of particleboard madefrom acetylated pine wood chips. In: Wood Modification.Ed. M. Lawnicear. pp. 436-446. Polish Academy of Science,Poznan.Truskne, D. 1977. Investigation of the content of bound acetylgroups in the main consitutents of acetylated wood. LatvijasLauksaimniecibas Akademija 130: 32-39.Youngquist, J.A., A. Krzysik and R.M. Rowell. 1985. Dimen-sional stability of acetylated aspen flakeboard. Wood Fiber Sei.18: 90-98.Received February 22nd, 1995Dr. M.G. BoonstraTNO Group forTimber ResearchPos Bus 49NL-2600 DelftThe NetherlandsProf. A. PizziENSTIB, Universitede Nancy 1B.P. 1041F-88051 Epinal Cedex 9FranceProf. P. TekelyLaboratoire de Methodologie NMR\"Universite de Nancy 1B.P. 239F-54506 NancyFranceDr. J. PendleburyNew Zealand Forest ResearchInstituteRotoruaNew Zealandslightly greater extent, due to an additional acetyla-tion site, when in the fibre (hence in presence oflignin) than when alone. All this indicates that per-haps a limited amount of cross-linking is possible,and that it is also this rather than just the bulkingaction described previously (Rowell 1991) whichcontributes to the improvement in dimensional sta-bilization of wood.ReferenceArodra, M., J.S. Rajawat and R.C. Gupta. 1981. Effect of acety-lation on properties of particleboards prepared from acetylatedand normal particles of wood. Holzforsch. Holzverwert. 33:8-10.Bekere, M., K. Shvalbe and I. Ozolinya. 1978. Some factorsaffecting the quality of boards made from acetylated woodfibers. Latvijas Lauksaimniecibas Akademija Raksti 163: 31-35.Breitmaier, E. and W. Voelter. 1987. Carbon-13 NMR Spectros-copy. 3rd Edition. VCH, Weinheim, pp. 1-285.Callow, H.J. 1952. Delignification of jute with sodium chlorite. J.Text. Inst. 43: T'243-247.Clermont, L.P. and F. Bender. 1957. The effect of swelling agentsand catalysts on acetylation of wood. Forest Prod. J. 7:167-170.Goldstein, I.S., E.B. Jeroski, A.E.Lund, J.F. Nielson and J.W.Weaver. 1961. Acetylation of wood in lumber thickness. ForestProd. J. 11: 363-370.Hon, D. N. S. 1994. Photoactivity of acetylated wood. Proceed-ings: Second Pacific Rim Bio-Based Composites Symposium.Nov. 6-9, 1994, Vancouver, Canada, pp. 169-180.Hon, D. N. S. and A. P. Bangi. 1994. Chemical modification ofjuvenile wood. Part I. Juvenility and response of southern QSBflakes to acetylation. Proceedings: Second Pacific Rim Bio-Based Composites Symposium. Nov. 6-9, 1994, Vancouver,Canada, pp. 135-143.Klinga, L.O. and H. Tarkow. 1966. Dimensional stabilization ofhardboard by acetylation. Tappi 49: 23-27.Matsuda, H. 1987. Preparation and utilization of esterified woodsbearing carboxyl groups. Wood Sei. Technol. 21: 75-88.Nikitina, N. 1977. Quality control of acetylated wood. LatvijasLauksaimniecibas Akademija 130: 50-53.Nishimoto, K. and Y. Imamura. 1985. Physical and protectiveproperties of particleboards made of mixtures of acetylatedand normal chips. Mokuzai Kogyo 40: 414-418.Peterson, M.D. and R.J. Thomas. 1979. Protection of wood fromdecay fungi by acetylation - An ultrastructural and chemicalstudy. Wood and Fiber 10: 149-163.Plackett, D.V and D.H. Cohen. 1994. Chemically modified woodcomposites: technical issues and market prospects. Proceed-ings: Second Pacific Rim Bio-Based Composites Symposium.Nov. 6-9, 1994, Vancouver, Canada, pp. 38-46.Holzforschung / Vol. 50 / 1996 / No. 3Brought to you by | Sichuan University (Sichuan University)

Authenticated | 172.16.1.226Download Date | 4/19/12 1:19 PM