Comparative Evaluation of Three Lignin Isolation Protocols for Various Wood Species

Comparative Evaluation of Three Lignin Isolation Protocols forVarious Wood Species

ANDERSON GUERRA, ILARI FILPPONEN, LUCIAN A. LUCIA, AND

DIMITRIS S. ARGYROPOULOS*

Organic Chemistry of Wood Components Laboratory, Department of Forest Biomaterials Science &Engineering, North Carolina State University, Raleigh, North Carolina 27695-8005

Milled wood lignin (MWL), cellulolytic enzyme lignin (CEL), and enzymatic mild acidolysis lignin (EMAL)were isolated from different wood species and characterized by various techniques. The EMAL protocoloffered gravimetric lignin yields 2-5 times greater than those of the corresponding MWL and CEL.The purities of the EMALs were 3.75-10.6% higher than those of their corresponding CELs, dependingupon the wood species from which they were isolated. Molecular weight analyses showed that theEMAL protocol isolates lignin fractions that are not accessed by the other procedures evaluated,while 31P NMR spectroscopy revealed that MWL is more condensed and bears more phenolic hydroxylgroups than EMAL and CEL. The yields and purities of EMAL, MWL, and CEL from hardwood weregreater than those obtained for the examined softwoods. Structural details obtained by DFRC(derivatization followed by reductive cleavage)/31P NMR revealed different contents of condensedand uncondensed â-O-aryl ether structures, dibenzodioxocins, and condensed and uncondensedphenolic hydroxyl and carboxylic acid groups within lignins isolated from different wood species.

KEYWORDS: EMAL; MWL; CEL; DFRC; 31P NMR; ball milling; lignin; southern pine; Douglas fir; white

fir; redwood; Eucalyptus globulus ; compression wood

INTRODUCTION

Lignin is a complex natural polymer resulting from oxidativecoupling primarily of (4-hydroxyphenyl)propanoids (1). Thecurrently accepted theory is that the lignin polymer is formedby combinatorial-like phenolic coupling reactions, via a radicalgenerated by peroxidase-H2O2, where monolignols react end-wise with the growing polymer (2). Such “random” dehydro-genative reactions produce a heterogeneous and highly cross-linked macromolecule, built up of different interunit linkagessuch as â-O-4, â-â, â-5, â-1, 5-5, 4-O-5, etc. (1).Furthermore, lignin is covalently linked to carbohydrates (3,4), forming a lignin-carbohydrate network made up of benzyl-ether (3, 5), benzyl-ester (3, 6, 7), and phenyl-glycoside (8-10) bonds.

Although lignin has been studied for more than 100 years,its structural details continue to emerge (11). One of the mostimportant problems in elucidating the lignin structure has beenthe isolation of the total lignin from wood in a chemicallyunaltered form (11-14). Early lignin preparation techniquesused strong mineral acids to reach high lignin yields (15). Suchdrastic conditions, however, were found to cause irreversiblereactions that severely alter the structure of the isolated material.Currently, the most used techniques aimed at isolating ligninfrom wood in a chemically unaltered form are based on the

extraction of ball-milled wood by neutral solvents (12, 16).While milled wood lignin (MWL) is extracted from finely milledwood without any previous treatment (16, 17), cellulolyticenzyme lignin (CEL) utilizes cellulolytic enzymes to removemost of the carbohydrate fractions prior to aqueous dioxaneextraction of ball-milled wood meal (12, 14). Recent comparisonof the chemical structures of MWL and CEL using wetchemistry and modern NMR spectroscopy has revealed thatMWL is slightly more condensed than CEL, suggesting thatMWL may contain a higher proportion of lignin from the middlelamellae (14, 18). Albeit being extensively used to isolate ligninfrom different sources, such lignin preparations offer moderateyields, which depend on the severity of the wood pulverization.In general, the more severe the milling conditions, the higherthe yields achieved by such isolation processes. However, asteady decrease inâ-O-4 linkages with increasing ball-millingintensity has been observed (19-21), showing that substantiallignin depolymerizationVia the cleavage of uncondensedâ-arylether linkages may take place under severe mechanical action(21). In this light, intensive milling protocols offered byvibratory or orbital milling devices should be considered withcaution since they provide higher lignin yields within relativelyshort milling intervals at the expense of the integrity of the ligninmacromolecule and associated condensation and oxidationreactions (19-21).

Recent progress toward isolating lignin from wood has shownthat a novel procedure using the combination of enzymatic and

* To whom correspondence should be addressed. Phone: (919) 515-7708. Fax: (919) 515-6302. E-mail: [email protected].

9696 J. Agric. Food Chem. 2006, 54, 9696−9705

10.1021/jf062433c CCC: $33.50 © 2006 American Chemical SocietyPublished on Web 12/20/2006

mild acidolysis (EMAL, enzymatic mild acidolysis lignin)isolates lignin that may be more representative of the total ligninpresent in milled wood (11, 21). Because a mild acid hydrolysiscan liberate lignin from lignin-carbohydrate complexes, knownto preclude lignin isolation in high yields, it can be combinedwith low severity of milling, facilitating the isolation of lessmodified lignin in high yields from milled wood (21). We haverecently shown that the yield of EMAL from Norway spruce isabout 4 times greater than that of the corresponding MWL andabout 2 times greater compared to that of CEL isolated fromthe same batch of milled wood (21). Comparison of the chemicalstructures of EMAL, MWL, and CEL revealed only subtledifferences, showing that EMAL is released by cleaving lignin-carbohydrate bonds rather than other linkages within the ligninmacromolecule. Molecular weight distribution analyses alsopointed out that the EMAL protocol allows the isolation of ligninfractions from spruce that are not accessed by any other ligninisolation procedures (21). In this study we have further exploredthe extent to which EMAL may be more representative of thetotal lignin in wood than CEL and MWL. To do this, EMAL,MWL, and CEL were isolated from four different species ofsoftwoods and one hardwood and characterized by wet chem-istry, quantitative31P NMR spectroscopy, and derivatizationfollowed by reductive cleavage (DFRC) coupled with31P NMR.Moreover, we used the virtues of such lignin isolation protocolsto better understand how the wood species affects the ligninyield, purity, structure, and molecular weight distribution whenisolated with the same method.

MATERIALS AND METHODS

Isolation of EMALs, MWLs, and CELs . EMAL, CEL, and MWLwere isolated form Douglas fir (Pseudotsuga menziessi), white fir (Abiesconcolor), redwood (Sequoia semperVirens), eucalyptus (Eucalyptusglobulus), and normal and compression wood from southern pine (Pinuspalustris). The wood chips from each different wood species wereground to pass a 20-mesh screen in a Wiley mill and Soxhlet extractedwith acetone for 48 h. The resulting Wiley-milled wood powder wasair-dried and stored in a desiccator under vacuum. TheE. globuluswood powder was submitted to an alkaline extraction with (0.075 mol/L) NaOH for 1 h (liquid-to-wood ratio 50:1) to remove tannins beforeuse (22). Rotary ball milling was performed in a 5.5 L porcelain jar inthe presence of 474 porcelain balls (9.4 mm diameter), which occupied18% of the active jar volume. A 100 g portion of extractive-free woodpowder was loaded into the jar, creating a porcelain ball/wood weightratio of 16.6. The milling process was conducted at room temperaturefor up to 28 days with a rotation speed of 60 rpm (21). EMALs wereisolated from ball-milled wood as previously reported (21). MWL wasisolated from the extractive-free wood according to the methoddescribed by Bjo¨rkman (16, 17). CEL was isolated from the insolublematerial obtained after isolation of MWL according to the method ofChang et al. (12) modified by Ikeda et al. (14) and Holtman et al. (18).Both preparations were purified as described elsewhere (16).

Determination of Lignin Purity . The purities of EMAL and CELwere calculated by summing the acid-insoluble (Klason lignin) and acid-soluble lignin contents, measured according to the method reported byYeh et al. (23).

Acetobromination Derivatization Procedure. Acetobrominationwas used as the derivatization method of choice for all samples priorto size exclusion measurements (21). Approximately 2.5 mL of amixture composed of 8 parts of acetyl bromide and 92 parts (v/v) ofglacial acetic acid was added to about 10 mg of a lignin sample (EMAL,MWL, or CEL) in a 15 mL round-bottom flask. The flask was sealedand placed in a water bath set at 50°C for 2 h with continuous magneticstirring. The solvent was rapidly evaporated at 25-28 °C in a rotaryevaporator connected to a high-vacuum pump and a cold trap. Theresidue was immediately dissolved in THF (5 mL) and subjected tosize exclusion analyses.

Size Exclusion Chromatography (SEC). SEC of EMAL, MWL,and CEL samples was performed on a size exclusion chromatographicsystem (Waters system) equipped with a UV detector set at 280 nm.The analyses were carried out at 40°C using THF as the eluent at aflow rate of 0.44 mL/min. A 120µL volume of the sample dissolvedin THF (2 mg/mL) was injected into HR5E and HR 1 columns (Waters)connected in series. The HR5E column specifications allow formolecular weights up to 4× 106 g/mol to be reliably detected. TheSEC system was calibrated with polystyrene standards in the molecularweight range of (890-1.86) × 106 g/mol, and Millenium 32 GPCsoftware (Waters) was used for data processing.

Quantitative 31P Nuclear Magnetic Resonance. Quantitative31PNMR spectra of all lignin preparations were obtained using publishedprocedures (21, 24, 25). To improve resolution, a delay time of 5 swas used and a total of 256 scans were acquired.

DFRC/31P NMR. DFRC was performed as described by Lu andRalph (26). The precise amounts of the lignin and precautions due tothe ensuing31P NMR steps were nearly identical to those reportedelsewhere (21).

RESULTS AND DISCUSSION

Our continuing efforts to better understand the lignin isolationprocess from wood have prompted us to examine various salientfeatures of lignin isolation variables. In this study, therefore,we evaluated the yield, purity, molecular weight, and structureof lignin samples isolated from different wood species usingdifferent isolation protocols. To ensure that the effects of woodpulverization on the lignin structure would not lead to amisinterpretation of our data, MWL, CEL, and EMAL wereisolated from the same batch of ball-milled wood (rotary ballmilling). Douglas fir, white fir, and redwood were used in thiswork to evaluate the aforementioned effects on lignin samplesfrom different species of softwood, while eucalyptus was chosenas a source of lignin from hardwood. To better understand thestructural differences between lignin from regular and reactionwood, the three lignin preparations were also isolated fromnormal and compression wood of southern pine.

Lignin Yield and Purity . The yields of MWL, CEL, andEMAL isolated from the different wood species are shown inFigure 1. As previously observed, the yields of EMAL werefound to be greater than those of the corresponding MWL andCEL, regardless of the wood species from which they wereisolated. The yields of MWL (w/w, based on the amount ofKlason lignin of the starting wood and the isolated lignin) variedfrom 1.34% to 34%, with eucalyptus giving a higher yield ascompared to the yields of MWL from softwoods. Overall, theyields of MWL were found to be lower than 16% for allevaluated softwoods. Such low yields are not totally surprising,however, when viewed in light of the recent conclusions of Huet al. (20) and Fujimoto et al. (19), where the extent ofextractable MWL was shown to be dependent upon millingseverity. To isolate MWL in higher yields, more extensivemilling is required (14). Intensive milling protocols offered byvibratory- or orbital-milling devices should be, however,considered with caution since they provide higher lignin yieldswithin relatively short milling intervals at the expense of theintegrity of the lignin macromolecule and associated condensa-tion and oxidation reactions (14, 19-21).

To improve yields while minimizing the extent of mechanicalaction, the insoluble material from aqueous dioxane extractionof MWL can be treated with cellulolytic enzymes (14, 18). Thisenzymatic treatment removes the majority of the carbohydrates,and a subsequent aqueous dioxane extraction solubilizes anotherlignin portion (CEL) that is considered to be lignin associatedwith carbohydrates (12, 18). In this way, Ikeda et al. (14) havesuggested the combination of these two preparations (MWL and

Isolation of Lignin from Different Wood Species J. Agric. Food Chem., Vol. 54, No. 26, 2006 9697

CEL) to increase the overall yield of pure lignin.Figure 1 showsthat such enzymatic hydrolysis permits isolating more ligninfrom wood. The yields of CEL were found to be similar to thoseof the respective MWL, except for Douglas fir, where the yieldof CEL was 7 times greater than that of MWL. Albeit beingextracted from the residual wood meal using the same solventsystem as MWL, our data are not supportive of the combinationof MWL and CEL to further increase the yields of isolated ligninas described by Ikeda et al. (14). Moreover, a recent solution-state nuclear magnetic resonance study has also revealed thatsome structural differences exist between these two ligninpreparations (18).

We have recently shown that the combination of enzymaticand mild acidolysis permits isolating lignin that may be morerepresentative of the total lignin present in milled wood (11,21). Because a mild acid hydrolysis can liberate lignin fromlignin-carbohydrate complexes, known to preclude ligninisolation in high yields, it offers the possibility of isolating lessmodified lignin in high yields from milled wood (11, 21). Thedata ofFigure 1 show that the concerted effect of enzymaticand mild acid hydrolysis offered significant yield improvementsover the traditional procedures for isolating lignin. It isnoteworthy that the amount of lignin isolated by using theEMAL protocol was higher for all different species of softwoodsanalyzed so far as well as for hardwood and even for isolatinglignin from compression wood. Ongoing work in our laboratoryhas indicated that the yield of EMAL is also higher for wheatstraw (data not shown). It is also significant that the liberationof lignin from lignin-carbohydrate complexes provided by themild acid hydrolysis step offers the possibility of obtaining highyields without applying more severe mechanical action onto thematerial. The data ofFigure 1 show that the yields of EMALwere from 1.9 to 5.3 times greater than those of the correspond-ing MWL and CEL isolated from the same batch of milledwood. A closer inspection of such data points out that for allsoftwood species evaluated so far the yields of EMAL weregreater than the combined yields of MWL and CEL. This findingindicates that EMAL includes not only the lignin fractionnormally isolated as MWL and CEL, but also macromoleculesthat are not accessed by any other available lignin isolationprotocol.

Other virtues of EMAL appear during the comparison of itsisolation procedure to the MWL and CEL protocols. Both

procedures require a two-step purification stage that is extremelytime-consuming. Such purification steps are not needed in theEMAL procedure since most of the non-lignin contaminants(carbohydrates) may readily be removed by the mild acidolysisstage (11). Moreover, the aqueous dioxane extraction step usedto isolate MWL and CEL (two times, each 24 h) is much longerthan the 2 h ofmild acidolysis required to isolate EMAL. Eventhough faster and simpler than the MWL and CEL protocols,the EMAL isolation procedure must be carried out carefully,since slightly higher concentrations of HCl may seriouslycompromise the structure of the isolated lignin (11).

Another aspect of our data shows that different wood speciesoffer different yields when isolated with the same isolationprocedure; i.e., the yields of EMAL, MWL, and CEL fromE.globuluswere found to be greater than those obtained for theexamined softwoods (Figure 1). Moreover, a comparison ofthe lignin yields from southern pine, redwood, Douglas fir, andwhite fir shows that different species of softwood displaydifferent behaviors when submitted to the same isolationprocedure. More specifically, the yields of isolated lignin fromDouglas and white fir were found to be somewhat lower thanthose obtained from redwood and pine. The data ofFigure 1show that even different species of the same genus (Douglasand white fir) offer different yields when submitted to the sameisolation procedure. The yields of MWL and CEL from normaland compression wood of pine, however, were found to be quitesimilar, while the yield of EMAL from compression wood was6.9% higher than the corresponding yield from normal wood.

The abnormal low yields of lignin isolated from Douglas fir,however, demand further attention. Such low yields might berelated to the high amounts of polysaccharides present in thisspecies of softwood. Willfo¨r et al. (27) have analyzed the contentand composition of carbohydrates comprising polysaccharidesin 12 different species of softwoods and found that Douglas fircontained the largest contents of galactoglucomannans, mannansfor short, and cellulose. Besides the enumerated limitations ofMWL, such a significant amount of mannans may require anenzymatic solution with high mannanase activity during theisolation of CEL and EMAL. As such future efforts with suchspecies may take into account this suggestion. Furthermore,limitations imposed to the milling process for this wood speciesdue to the formation of wood aggregates arising from its highresin content cannot be ruled out.

Figure 1. Gravimetric yields of EMAL (open bars), CEL (black filled bars), and MWL (dashed bars) isolated from the same batch of milled Douglas fir,redwood, white fir, E. globulus, and normal and compression wood of southern pine.

9698 J. Agric. Food Chem., Vol. 54, No. 26, 2006 Guerra et al.

Lignin samples isolated from wood still contain associatedcarbohydrates and other non-lignin contaminants, regardless ofthe isolation and purification procedures applied (11, 12, 14).Lignin-carbohydrate linkages exist in wood and are known tobe of benzyl-ester, benzyl-ether, and phenyl-glucoside types(1). Such interactions between lignin and carbohydrates precludethe isolation of lignin in high yields and purities; i.e., the puritiesof CEL isolated from the different wood species evaluated inthis study and purified according to Bjo¨rkman (16) varied from80% to 85% (Figure 2). All of these lignin-carbohydrate bondsare, however, susceptible to acid hydrolysis. In the conditionsapplied to isolate EMAL, the complete cleavage of the phenyl-glucoside bonds has been shown by model compound studiesto be accomplished, whereas the non-phenolic benzyl-etherones were found to be more stable under such conditions (28,29). The benefits of the mild acid hydrolysis of lignin-carbohydrate linkages were, however, apparent as far as thepurities of the EMALs are concerned. From the data shown inFigure 2 it is obvious that all EMALs evaluated so far havepurities from 3.75% to 10.6% higher than those of theircorresponding CELs, depending upon the wood species theyare isolated from. The purities of MWL samples were notdetermined in this study. Nevertheless, previous work haspointed out that the purities of EMALs from both poplar andspruce are higher than those of the corresponding MWLs (11).The purities reported inFigure 2 are based on the sum of Klasonand UV-soluble lignin contents, which represents the totalamount of lignin contained in CEL and EMAL after removalof non-lignin contaminants through a severe hydrolysis with72% (w/w) H2SO4.

The EMALs isolated from the different wood species werefound to have different purities (Figure 2). More specifically,the EMAL isolated from E. globulus appears to be lesscontaminated by non-lignin materials than the EMALs fromsoftwoods, while the EMAL from Douglas fir was found tohave the highest content of such contaminants. Albeit theprevailing consensus that lignin is cross-linked to differentpolysaccharides in the cell wall and that such cross-linking mightbe one of the reasons for the low MWL yields, the exactfrequency of lignin-carbohydrate bonds in different woodspecies is still a matter of discussion (1, 3, 4, 6-9, 21, 28, 29).Lawoko et al. (28) have recently reported, for example, thatlignin without covalent bonds to carbohydrates does not existin spruce wood. The differences in the yields and purities ofEMALs reported inFigures 1and2, respectively, are supportiveof the existence of a different microstructure and the presence

of a variety and variable abundances of lignin-carbohydratebonds among the different wood species.

Molecular Weight Distribution . It was recently reported thatacetobromination represents a facile and rapid alternative to thecomplete solubilization of sparingly soluble lignin samples,while still allowing for an accurate analysis (21). By dissolvinga lignin sample in neat acetyl bromide diluted with glacial aceticacid (8:92, v/v), the primary alcoholic and the phenolic hydroxylgroups are acetylated, while the benzylicR-hydroxyls aredisplaced by bromide (26). Similarly, benzyl aryl ethers arequantitatively cleaved to yield aryl acetates and acetylatedR-bromo products (26). The concerted effect of acetylation whencoupled with the polarity induced by the selectiveR-brominationcaused every lignin sample examined so far to become highlysoluble in THF, allowing rapid SEC analyses. Comparisonbetween acetobromination and acetylation with acetic anhydride/pyridine has shown minor differences in the UV responses andelution profiles, which is supportive of the viability of usingacetobromination as a derivatization technique to sparinglysoluble lignin (21).

The molecular weight distributions of the acetobrominatedMWL, CEL, and EMAL were therefore compared by SEC usingTHF as the mobile phase and UV detection at 280 nm.Figure3A shows a typical set of SEC chromatograms of such ligninpreparations, where a distinctly higher molecular weight dis-tribution was found for EMAL than for CEL and MWL. Thechromatograms reveal that EMAL is richer in high molecularweight fragments (higher than 100× 103 g/mol), whichappeared in lower abundance in CEL and were completelyabsent in MWL. This result is consistent with our recent workin which we observed that the absence of the high molecularweight fractions made the apparent weight-average molecularweights of both MWL and CEL isolated from spruce signifi-cantly lower than that of the corresponding EMAL (21). Thesedata are also supportive of a previous finding stating that theconcerted effect of cellulolytic action and mild acidolysis allowsfor the isolation of lignin fragments that are not accessible byeither of the alternative lignin isolation procedures (21). Aquestion that emerges at present is whether such material causingthe formation of the aforementioned high molecular weightfractions consists of covalently bound lignin or lignin-ligninassociation. Ongoing work in our laboratory is being carriedout in an endeavor to address this very important fundamentalissue.

For the purposes of the present investigation, lignin associa-tion phenomena were not taken into account to calculate theapparent molecular weight averages reported inTable 1.However, to ensure that such association phenomena would notlead to a misinterpretation of the molecular weight distributions,all analyses were carried out on freshly prepared lignin solutionsanalyzed immediately after derivatization, ensuring that thecomparisons made from sample to sample were valid.

The data ofFigure 3B andTable 1 point out that EMALsisolated from different wood species display different elutionprofiles and, consequently, different apparent molecular weightaverages (Mw andMn) when the wood is pulverized under thesame milling conditions and the lignin extracted by using thesame enzymatic and mild acidolysis sequence. As anticipated,a highly polydisperse behavior is apparent in the SEC chro-matograms of the EMAL samples as far as their molecularweight distributions are concerned (Figure 3B,C). The elutionprofiles, however, were found to be different among the EMALsisolated from different wood species. Specifically, while thechromatograms of EMAL isolated from southern pine (Figure

Figure 2. Purities of EMAL (open bars) and CEL (black filled bars) isolatedfrom the same batch of milled Douglas fir, redwood, white fir, E. globulus,and normal and compression wood of southern pine.

Isolation of Lignin from Different Wood Species J. Agric. Food Chem., Vol. 54, No. 26, 2006 9699

3C) and white fir (Figure 3B) displayed a bimodal behavior,the chromatograms of EMAL from Douglas fir, redwood, andE. globulusshowed only a small shoulder extending over 100× 103 g/mol. Moreover, a high molecular weigh fraction (albeitof low abundance), extending over 500× 103 g/mol, wasapparent in the chromatograms of EMAL from pine and whitefir. Such a fraction, however, was absent in the lignins fromthe other evaluated wood species.

The bimodality and presence of the aforementioned highmolecular weight fraction made the apparent weight-averagemolecular weight of EMAL from pine greater than those fromthe other evaluated wood species (Table 1). Such anMw wasfound to be over 57× 103 g/mol, while theMw of EMAL

isolated from white fir, for which the chromatogram alsodisplayed the high molecular weight fractions but in lowabundance, was 52× 103 g/mol. The molecular weights ofEMALs isolated from Douglas fir, redwood, andE. globulus,in which the chromatograms did not display bimodality, were38 × 103, 30× 103, and 32× 103 g/mol, respectively. It mustbe emphasized, however, that the possibility of such differencesin the apparent weight-average molecular weight being due tolignin association cannot be ruled out. In this scenario, neverthe-less, one should note that lignin from different wood speciescan display different propensities to associate.

The size exclusion chromatograms of the acetobrominatedEMAL isolated from normal and compression wood of southernpine are shown inFigure 3C. Comparison of such lignin revealssimilar molecular weight distributions, with EMAL fromcompression wood displaying slightly higher amounts of highmolecular weight fragments, which make theMw of lignin fromcompression wood slightly greater than that of the correspondinglignin from normal wood (Table 1).

Determination of Units Bearing Free Phenolic HydroxylGroups. 31P NMR spectroscopy is a reliable method todetermine the amounts of various hydroxyl groups within thelignin macromolecule (24, 25). Such hydroxyl groups arerevealed and quantified after phosphitylation of lignin with2-chloro-1,3,2-dioxaphospholane or 2-chloro-4,4,5,5-tetram-ethyl-1,3,2-dioxaphospholane (24, 25). Theâ-aryl ether content(Figure 4) was determined after phosphitylation of the CRhydroxyl groups in these moieties with 2-chloro-1,3,2-diox-aphospholane (25, 30). The condensed and uncondensed phe-nolic hydroxyls as well as the carboxylic acids (Figure 5A-C) were determined by phosphitylation of the lignins with2-chloro-4,4,5,5-tetramethyl-1,3,2-dioxaphospholane (24). Quan-tification was then carried out via peak integration usingN-hydroxynaphthalimide as an internal standard. Details of

Figure 3. Typical SEC chromatograms of lignin samples isolated fromthe same batch of milled white fir (A). SEC chromatograms of EMALisolated from milled Douglas fir, redwood, white fir, and E. globulus (B).SEC chromatograms of EMAL isolated from normal and compression woodof southern pine.

Table 1. Weight-Average Molecular Weight (Mw), Number-AverageMolecular Weight (Mn), and Polydispersity (D) of EMAL, MWL, andCEL Isolated from Different Wood Species

ligninaMw

(g/mol)Mn

(g/mol)D

Douglas FirEMAL 38000 7600 5.0MWL 7400 2500 3.0CEL 21800 5500 4.0

White FirEMAL 52000 6300 8.2MWL 8300 2800 3.0CEL 21700 4700 4.6

RedwoodEMAL 30100 4700 6.4MWL 5900 2400 2.5CEL 23000 5400 4.2

E. globulusEMAL 32000 8700 3.7MWL 6700 2600 2.6CEL 17200 5500 3.1

Normal Wood of Southern PineEMAL 57600 9700 5.9MWL 14900 4700 3.2CEL 29600 7500 3.9

Compression Wood of Southern PineEMAL 63500 9300 6.8MWL 16100 5200 3.1CEL 27500 7200 3.8

a Isolated after 28 days of ball milling. EMAL, CEL, and MWL were isolatedfrom the same batch of milled wood.

9700 J. Agric. Food Chem., Vol. 54, No. 26, 2006 Guerra et al.

signal acquisition, assignment, and integration can be foundelsewhere (24, 25).

The data ofFigures 4and5 show that some differences existin the functional group content of EMAL, MWL, and CEL.Such differences, however, were found to be dependent uponthe wood species from which the lignin was isolated. As shownin Figure 4A, MWL isolated from E. globulushas slightlyhigher amounts ofâ-O-aryl ether linkages than its respectiveEMAL and CEL. Previously, we have also found that MWLisolated from spruce has higher contents of such linkages thanCEL and EMAL (21). However, comparison among the threelignin preparations isolated from pine, Douglas fir, redwood,and white fir showed no significant difference in the totalâ-O-aryl ether contents. It is also significant to note that the nearlyidentical amounts of totalâ-O-aryl ether functional groups ofEMAL, MWL, and CEL indicate no evidence ofâ-aryl etherbond cleavage within the lignin during the mild acid hydrolysisstep of the EMAL protocol.

Figure 5 shows the amounts of phenolic hydroxyls as wellas carboxylic acid groups determined by quantitative31P NMR.In general, MWL was found to contain higher amounts ofphenolic hydroxyl groups than EMAL and CEL (Figure 5C).This finding is not surprising since it is well-known thatphenolic-rich lignin structures are more easily and preferentiallyisolated in the extraction procedure utilized for MWL (14). Thisselective fractionation might explain the much lower contentsof such phenolic hydroxyl groups observed for the CELs, sincein this work CEL was isolated from the residue of MWLisolation. As shown inFigure 5C the total contents of phenolichydroxyl within EMAL were found to be lower than withinMWL, indicating that the aforementioned selective fractionationis minimized by the EMAL protocol.

As expected, MWL was found to contain higher amounts ofcondensed phenolic hydroxyl than CEL and EMAL (Figure5A). This result is consistent with the recent work of Holtmanet al. (18), where the degree of condensation of MWL was foundto be slightly higher than that of CEL, both isolated from loblollypine. Such a difference in the degree of condensation might

result, at least in part, from different morphological origins. Asrecently reported by Hu et al. (20), lignin samples isolated frommilled wood in low yields may be more contaminated by ligninfrom the middle lamellae (CML), which is known to be morecondensed than lignin from the secondary wall (SW). The effectof such contamination, however, diminishes when more ligninfrom the SW is isolated.

The uncondensed phenolic hydroxyl contents of EMAL werefound to be somewhat higher than those of CEL (Figure 5B).The possibility that the higher contents of phenolic hydroxylcome from liberation of such groups fromâ-O-aryl etherlinkages can be ruled out, because there is no evidence of lignindegradation during the isolation of EMAL (Figure 4A andTable 1). The higher contents of uncondensed phenolic hydroxylwithin EMAL could be rationalized, at least in part, on the basisof the hydrolysis of phenyl-glucoside bonds within lignin-carbohydrate complexes. Complete cleavage of such bonds canbe accomplished under the conditions used to isolate EMAL(28), liberating new phenolic hydroxyl groups and increasingthe purities of EMAL, as shown inFigure 2. The contents ofcarboxylic acids groups were also very similar among EMAL,MWL, and CEL (Figure 5D). These similarities provide furthersupport for the effectiveness of the EMAL protocol providingnonoxidized lignin.

In summary, our data are not supportive of the protocol inwhich MWL and CEL are combined to further increase theyields of isolated lignin (14). As was observed inFigure 5, thecondensed, uncondensed, and total phenolic hydroxyl contentsare quite different between such lignin preparations. Moreover,the molecular weight distribution data inTable 1 also showthat MWL and CEL protocols afford isolated lignin fragmentswith different molecular weights. More significantly, however,the combination of enzymatic and mild acid hydrolysis offersthe possibility to isolate lignin samples that are more representa-tive of the total lignin in milled wood. As shown inFigure 1,the yields of EMAL were in most cases higher than thecombined yields of CEL and MWL with no evidence ofstructural alteration due to the mild acidolysis involved in theEMAL protocol. This indicates that the cleavage of the lignin-carbohydrate bonds afforded during the mild acidolysis step ofthe EMAL protocol allows the isolation of lignin fractions thatare not accessed by any other isolation procedures. Furthermore,the liberation of lignin from lignin-carbohydrate complexesoffers the possibility of obtaining high yields using low-intensitymilling, which is desirable to avoid lignin degradation duringthe wood pulverization as alluded to earlier.

In an effort to better understand how the wood species affectsthe lignin structure when isolated with the same method, wecompared the31P NMR data obtained for the different EMALs,since such lignin preparation offers the aforementioned benefitsover MWL and CEL. As expected,E. globuluswas found tocontain moreâ-aryl ether structures than all the softwoodsevaluated so far (Figure 4A andTable 2). The value of 2780µmol of â-aryl ether/g of lignin obtained forE. globulusby 31PNMR corresponds to 58.7%â-aryl ether structures within sucha lignin, considering that the average molecular weight of onephenylpropane unit (C9) in such a lignin is 211 g/mol, derivedfrom the elemental composition ofE. globulusdioxane lignin(22). Such a value correlates very well with the 60%â-arylether structures reported for hardwoods by Adler (15). Amongsoftwoods, Douglas fir was found to contain slightly highercontents of such linkages (1600µmol/g), while southern pineand redwood were observed to contain the lowest values (1340µmol/g). Lignins in compression and normal wood were found

Figure 4. â-O-Aryl ether functional group content (A) and erythro/threoratio (B) of EMALs, MWLs, and CELs isolated from the same batch ofmilled white fir (horizontal dashed bars), redwood (gray filled bars), Douglasfir (diagonal dashed bars), E. globulus (black filled bars), and normal(vertical dashed bars) and compression (open bars) wood of southernpine.

Isolation of Lignin from Different Wood Species J. Agric. Food Chem., Vol. 54, No. 26, 2006 9701

to contain 1340 and 1200µmol of â-O-aryl ether functionalgroups/g of lignin, respectively (Table 2). These results indicatea 10.5% decrease in arylglycerol-â-aryl ether linkages incompression wood of southern pine. Such a decrease cor-roborates recent findings where an approximate 15% decreasein arylglycerol-â-aryl ether linkages in compression wood ofloblolly pine has been reported (31).

Both erythro- and threo-stereoisomeric forms ofâ-O-4structures can also be determined using31P NMR after deriva-tization of lignin with 2-chloro-1,3,2-dioxaphospholane, byintegrating the regions from 135 to 134.2 ppm and from 134 to133.4 ppm, which have been attributed to CR-OH in erythro-and threo-forms of â-O-4 structures, respectively (25). Asexpected,Figure 4B shows that theerythro/threo ratios weresimilar in all species of gymnosperms, while the predominanceof the erythro-form was obvious inE. globulus. These datacorroborate previous findings reported by Akiyama et al. (32),where the proportion and amount oferythro- and threo-formswere described as very similar in softwoods, while, in contrast,for hardwood species theerythro-form of â-O-4 structures wasfound to predominate, the extent being dependent upon the woodspecies. Such selective behavior has been rationalized on thebasis of the widely accepted theory for the formation ofâ-O-4

structures in lignin (15). According to this theory, the first stepin the formation ofâ-O-4 structures is the 4-O-coupling of anoligolignol phenoxy radical to a monolignol radical at its side-chainâ-position to form a quinone methide intermediate. Thenext step is water addition to one of the two faces of the quinonemethide, leading to the formation of theerythro- or threo-form.Such water addition, however, depends on the aromatic ringtype, solvent, and pH (33). Model experimental simulations haveshown that the water addition leading to theerythro-form ispreferred when syringyl-type aromatic rings form the quinonemethide (33).

The major differences in the hydroxyl functional groupsamong the EMALs from different wood species are also listedin Table 2. The total amount of phenolic hydroxyl groups wasslightly higher (6.7%) within lignin from compression woodthan normal wood of southern pine. The total H unit content(p-hydroxyphenyl moieties) was found to be significantly higherin compression wood, while more G unit content (guaiacylmoieties) was detected within lignin from normal wood. Asanticipated, a slightly higher degree of condensation was alsoobserved in lignin from compression wood. Such findings areconsistent with the accepted theory that the biosynthesis ofnormal softwood lignins occursVia radical polymerization of

Figure 5. Condensed phenolic hydroxyl (A), uncondensed phenolic hydroxyl (B), total phenolic hydroxyl (C), and carboxylic acid (D) group contents ofEMALs, MWLs, and CELs isolated from the same batch of milled white fir (horizontal dashed bars), redwood (gray filled bars), Douglas fir (diagonaldashed), E. globulus (black filled bars), and normal (vertical dashed) and compression (open bars) wood of southern pine. Due to overlapping, uncondensedand condensed phenolic hydroxyl units have been integrated together to give the total amount of phenolic units reported for E. globulus.

Table 2. Functional Group Contents, Yields, and Weight-Average Molecular Weights (Mw) for EMAL Isolated from Different Wood Speciesa

functionalgroupb

Douglasfir

whitefir redwood

normal wood ofsouthern pine

compression wood ofsouthern pine

E.globulus

total â-aryl ether 1600 1490 1340 1340 1200 2780syringyl OH 0.00 0.00 0.00 0.00 0.00 620guaiacyl OH 840 930 1060 790 570 350p-hydroxyl OH 100 110 160 120 380 20uncondensed PhOH 940 1040 1220 910 950 overlappedcondensed PhOH 410 560 630 430 480 overlappedtotal PhOH 1350 1600 1850 1340 1430 990carboxylic groups 130 190 160 110 100 150yieldc (%) 24.8 42.9 56.7 56.3 65.1 63.7Mw

d (g/mol) 38000 52000 30100 57600 63500 32000

a Isolated after 28 days of ball milling. b Determined by 31P NMR. Values in µmol/g. Error ±20. c Based on Klason lignin contents of extracted ground wood meal.d Determined by size exclusion chromatography.

9702 J. Agric. Food Chem., Vol. 54, No. 26, 2006 Guerra et al.

coniferyl alcohol, while in compression wood a significantamount ofp-coumaryl alcohol units are also incorporated (1).Due to the absence of an aromatic methoxyl, the possibilitiesfor coupling in the radical polymerization are more complex(1). As a result, compression wood lignin has been found to bemore condensed (31, 34) and with a larger number of highmolecular weight fragments than lignin from normal wood(Figure 3C).

The data ofFigure 5A andTable 2also show thatE. globulushas a lower total phenolic hydroxyl content than all the softwoodspecies. Specifically, whileE. globuluswas found to containless than 1000µmol of phenolic hydroxyl/g of lignin, such acontent in lignin from softwood ranged from 1340 to 1850µmol/g. This finding is not surprising, since the total contents ofphenolic hydroxyl groups in softwoods have been reported tobe somewhat higher than in hardwoods (35). Moreover, earlierobservations of MWL have indicated that the syringyl unitspresent in hardwood lignins are primarily of the non-phenolictype (15, 35). Such a low content of free phenolic units inhardwood lignins has long been used to explain their relativelypoor responses to sulfite treatments used in the preparation ofchemimechanical pulps (36). The amount of condensed anduncondensed phenolic hydroxyls within EMAL fromE. globuluswas not calculated since the condensed and uncondensedsyringyl-type phenolic hydroxyl signals overlap in the31P NMR.Due to such overlapping, both units have been integratedtogether to give the total amount of phenolic units reported inFigure 5C.

A considerable variation was found among the softwoodspecies in the phenolic hydroxyl group content, which decreasesin the order redwood (1850µmol/g), white fir (1600µmol/g),Douglas fir (1350µmol/g), and pine (1340µmol/g). Redwoodand white fir seem to be more condensed than pine and Douglasfir as shown by the total amount of condensed phenolic hydroxylgroups in such lignins (Figure 5A andTable 2).

Determination of Units Bearing Etherified Phenolic Hy-droxyl Groups in â-O-Aryl Ether Linkages. Although quan-titative 31P NMR has contributed significantly to our under-standing of the hydroxyl-bearing functional groups, it cannotoffer any information about the etherified or carbon-carbon-linked bonding pattern of lignin (37). To overcome thislimitation, Tohmura and Argyropoulos (37) have recentlyproposed the combination of DFRC with31P NMR. In this way,when the aryl ether linkages are selectively cleaved by DFRC(26), the corresponding phenolic hydroxyls released can bequantified by 31P NMR. Because31P NMR can distinguishcondensed from uncondensed phenolic hydroxyls, the31P NMRspectra “after DFRC” offer detailed information about condensedand uncondensed units connected throughâ-aryl ether linkagesas well as dibenzodioxocins (21, 37). The total amounts ofuncondensedâ-O-aryl structures determined by DFRC/31P NMR

and thioacidolysis have been shown to be quite similar whenboth techniques are applied to the same sample of isolated lignin(21).

The quantification of the hydroxyl groups released fromâ-aryl ether structures by DFRC is given inTable 3. The totalamount of uncondensedâ-aryl ether structures withinE.globuluswas significantly higher than that within lignin fromsoftwoods. As well-known, lignins from hardwoods have moreuncondensedâ-aryl ether structures than lignin from softwood(2). The value of 1730µmol of uncondensedâ-aryl etherstructures/g of lignin obtained forE. globuluscorresponds to36.5% uncondensedâ-aryl ether structures within lignin,considering that the average molecular weight of one phenylpropane unit (C9 unit) in such a lignin is 211 g/mol (22). Thedata ofTable 3, coupled with the data ofTable 2, show that62.2% of the total amount ofâ-O-aryl ether structures inE.globulusare uncondensed (1730µmol/2780µmol), which is ingood agreement with the value obtained by thioacidolysis forthe same wood species (22). The S/G ratio obtained “afterDFRC” (83/15) (Table 3) is somewhat different from thatobserved “before DFRC” (63/35) (Table 2), indicating that thesyringyl units present inE. globulusare primarily of the non-phenolic type (15, 22).

The data ofTable 3 show that while the total contents ofdibenzodioxocins and condensedâ-O-aryl ether units weresimilar among Douglas fir, white fir, and redwood, the totalcontents of uncondensedâ-O-aryl ether units were found to beslightly different from species to species. Condensedâ-aryl etherbonds refer to structures that are characterized by the covalentattachment of two macromolecules or oligomers that themselvesare interlinked via structures other thanâ-aryl ethers. The totalamount of uncondensedâ-aryl ether structures was found todecrease in the order Douglas fir (940µmol/g), white fir (900µmol/g), and redwood (850µmol/g). The total amount ofuncondensedâ-aryl ether linkages in such softwoods was nearlydouble that of condensed moieties, indicating that about two-thirds of the etherified phenolic moieties inâ-aryl etherstructures present in EMAL from these species are uncondensedunits connected to another phenylpropane unit bearingâ-O-4,â-5, andâ-â linkages. One-third of the etherified phenolicmoieties in theâ-aryl ether structures contained a subsistentgrouportho to the phenolic hydroxyl, with the majority beingdibenzodioxocins. These findings are similar to those reportedby Tohmura and Argyropoulos for MWL lignin from blackspruce (37).

The EMAL from normal wood of southern pine, however,was found to be quite different from that of the other softwoodsevaluated so far. The total amounts of condensedâ-aryl etherlinkages and dibenzodioxocins were somewhat lower while theamount of uncondensedâ-aryl ether structures was higher insouthern pine. The total uncondensed H unit content (uncon-

Table 3. Values of Hydroxyl Moiety Contents Determined by DFRC/31P NMR (µmol/g) for EMAL Isolated from Different Wood Speciesa

phenolic OH content

EMALsource

S units involvedonly in uncondensed

â-O-aryl bonds

G units involvedonly in uncondensed

â-O-aryl bonds

H units involvedonly in uncondensed

â-O-aryl bonds

G + H units involvedonly in condensed

â-O-aryl bonds dibenzodioxocins

totaluncondensed

â-O-aryl bonds

Douglas fir not detected 880 60 510 230 940white fir not detected 860 40 490 240 900redwood not detected 810 41 470 220 851normal pineb not detected 990 25 240 170 1015comp pineb not detected 745 74 284 210 819E. globulus 1430 260 40 overlappedc overlappedc 1730

a Isolated after 28 days of ball milling. b EMAL from normal and compression wood of southern pine. c Inadequate resolution for quantification.

Isolation of Lignin from Different Wood Species J. Agric. Food Chem., Vol. 54, No. 26, 2006 9703

densedp-hydroxyphenyl moieties), however, was higher incompression wood than in normal wood, while the amount ofuncondensed G units was found to be higher in normal wood(Table 3). More specifically, 9.0% of the total amount ofuncondensedâ-aryl ether structures present in compressionwood contain at least onep-hydroxyphenyl moiety (H unit),while in normal wood only 2.5% of this unit is involved inuncondensedâ-aryl ether structures. These results are similarto a recent report by Yeh et al. (23), who proposed that themajor difference in H units between normal and compressionwood is from nonconjugatedp-hydroxyphenyl moieties. Asanticipated, lignin from compression wood was found to containless uncondensedâ-aryl ether structures than lignin from normalwood. The data ofTable 3 show a 19.3% decrease in suchuncondensed structures in compression wood lignin. On theother hand, the amount of condensedâ-aryl ether structuresseems to be at least 18.3% higher in compression wood, withthe majority difference being due to a 23.5% increase indibenzodioxocins in lignin from compression wood whencompared with lignin from normal wood.

Conclusions. Overall, MWL, CEL, and EMAL when isolatedfrom four different species of softwood and one hardwood andthoroughly characterized were found to offer different yields,purities, lignin structures, and molecular weights when isolatedwith the same method. Most significantly, the EMAL protocolwas found to offer much higher gravimetric lignin yields andpurities than those of the corresponding MWL and CEL isolatedfrom the same batch of milled wood. A more detailedcomparison of the lignin yields and purities showed that differentspecies of softwood display different behaviors when submittedto the same isolation procedure.

ACKNOWLEDGMENT

We thank Professor Ronald Sederoff for providing the normaland compression wood of Southern pine. The contributions ofAna Xavier are also acknowledged with respect to the Klasonand UV analyses.

LITERATURE CITED

(1) Fengel, D., Wegener, G., Eds.Wood Chemistry, Ultrastructureand Reactions; Walter de Gruyter: Berlin, 1989; 613 pp.

(2) Ralph, J.; Lundquist, K.; Brunow, G.; Lu, F.; Kim, H.; Schatz,P.; Marita, J.; Hatfield, R.; Ralph, S.; Christensen, J.; Boerjan,W. Lignins: Natural polymers from oxidative coupling of4-hydroxyphenylpropanoids.Phytochemistry2004, 3, 29-60.

(3) Yaku, F.; Yamada, Y.; Koshijima, T. Lignin-carbohydratecomplex. Part IV. Lignin as a side chain of the carbohydrate inBjorkman LCC.Holzforschung1981, 35, 177-181.

(4) Lawoko, M.; Henriksson, G.; Gellerstedt, G. New method forthe quantitative preparation of lignin-carbohydrate complex fromunbleached softwood kraft pulp: lignin-polysaccharide networksI. Holzforschung2003, 57, 69-74.

(5) Eriksson, O.; Goring, D.; Lindgren, B. O. Structural studies onthe chemical bonds between lignin and carbohydrates in spruce.Wood Sci. Technol.1980, 14, 267-279.

(6) Watanabe, T.; Koshijima, T. Evidence for an ester linkagebetween lignin and glucuronic acid in lignin-carbohydratecomplexes by DDQ-oxidation.Agric. Biol. Chem.1988, 52,2953-2955.

(7) Lundquist, K.; Simonson, R.; Tingsvik, K. Lignin-carbohydratelinkages in milled-wood lignin preparations from spruce wood.SVen. Papperstidn.1983, 86, 44-47.

(8) Yaku, F.; Yamada, Y.; Koshijima, T. Lignin-carbohydratecomplex Part II. Enzymatic degradation of acid polysaccharidesin Bjorkman LCC.Holzforschung1976,30, 148-156.

(9) Kondo, R.; Sako, T.; Limori, T.; Imamura, H. Formation ofglycosidic lignin-carbohydrate complex in the dehydrogenativepolymerization of coniferyl alcohol.Mokuzai Gakkaishi1990,36, 332-338.

(10) Xie, Y.; Yasuda, S.; Wu, H.; Liu, H. Analysis of the structureof lignin-carbohydrate complexes by specific13C tracer method.J. Wood Sci.2000, 46, 130-136.

(11) Wu, S.; Argyropoulos, D. S. An improved method for isolatinglignin in high yield and purity.J. Pulp Pap. Sci.2003, 29, 235-240.

(12) Chang, H.; Cowling, E.; Brown, W. Comparative studies oncellulolytic enzyme lignin and milled lignin of sweetgum andspruce.Holzforschung1975, 29, 153-159.

(13) Lee, Z.; Meshitsuka, G.; Cho, N.; Nakano, J. Characteristics ofmilled wood lignins isolated with different milling times.Mokuzai Gakkaishi1981, 27, 671-677.

(14) Ikeda, T.; Holtman, K.; Kadla, J.; Chang, H.; Jameel, H. Studieson the effect of ball milling on lignin structure using a modifiedDFRC method.J. Agric. Food Chem.2002, 50, 129-135.

(15) Adler, E. Lignin chemistry- past, present and future.Wood Sci.Technol.1977, 11, 169-218.

(16) Bjorkman, A. Studies on finely divided wood I. Extraction oflignin with neutral solvents.SVen. Papperstidn.1956, 59, 477-485.

(17) Bjorkman, A. Lignin and lignin-carbohydrate complexessextraction from wood meal with neutral solvents.Ind. Eng. Chem.1957, 49, 1395-1398.

(18) Holtman, K.; Chang, H; Kadla, J. Solution-state nuclear magneticresonance study of the similarities between milled wood ligninand cellulolytic enzyme lignin.J. Agric. Food Chem.2004, 52,720-726.

(19) Fugimoto, A.; Matsumoto, Y.; Chang, H.; Meshitsuka, G.Quantitative evaluation of milling effects on lignin structureduring the isolation process of milled wood lignin.J. Wood Sci.2005, 51, 89-91.

(20) Hu, Z.; Yeh, T.; Chang, H.; Matsumoto, Y.; Kadla, J. Elucidationof the structure of cellulolytic enzyme lignin.Holzforschung2006, 60, 389-397.

(21) Guerra, A.; Filpponen, I.; Lucia, L.; Saquing, C.; Baumberger,S.; Argyropoulos, D. S. Toward a better understanding of thelignin isolation process from wood.J. Agric. Food Chem.2006,54, 5939-5947.

(22) Evtuguin, D.; Neto, C.; Silva, A.; Domingues, P.; Amado, F.;Robert, D.; Faix, O. Comprehensive study on the chemicalstructure of dioxane lignin from plantationEucalyptus globuluswood.J. Agric. Food Chem.2001, 49, 4252-4261.

(23) Yeh, T.; Tatsuhiko, Y.; Capanema, E.; Chang, H; Chiang, V.;Kadla, J. Rapid screening of wood chemical component varia-tions using transmittance near-infrared spectroscopy.J. Agric.Food Chem.2005, 53, 3328-3332.

(24) Granata, A.; Argyropoulos, D. S. 2-Chloro-4,4,5,5-tetramethyl-1,3,2-dioxaphospholate, a reagent for the accurate determinationof the uncondensed and condensed phenolic moieties in lignins.J. Agric. Food Chem.1995, 43, 1538-1544.

(25) Argyropoulos, D. S. Quantitative phosphorus-31 NMR analysisof lignin: a new tool for the lignin chemist.J. Wood Chem.Technol.1994, 14, 45-63.

(26) Lu, F.; Ralph, J. The DRFC method for lignin analysis 2.Monomers from isolated lignins.J. Agric. Food Chem.1998,46, 547-552.

(27) Willfor, S.; Sundberg, A.; Hemming, J.; Holmbom, B. Polysac-charides in some industrial important softwood species.WoodSci. Technol.2005, 39, 245-258.

(28) Lawoko, M.; Henriksson, G.; Gellerstedt, G. Characterizationof lignin-carbohydrate complexes from spruce sulfite pulps.Holzforschung2006, 60, 162-165.

(29) Kosikova, B.; Joniak, D.; Kosakova, L. The properties of benzylether bonds in the lignin-saccharidic complex isolated fromspruce.Holzforschung1979, 33, 11-14.

9704 J. Agric. Food Chem., Vol. 54, No. 26, 2006 Guerra et al.

(30) Akim, L.; Argyropoulos, D.; Jouanin, L.; Leple´, J.-L.; Pilate,G.; Pollet, B.; Lapierre, C. Quantitative31P NMR spectroscopyof lignins from transgenic poplars.Holzforshung2001, 55, 386-390.

(31) Yeh, T.-F.; Goldfarb, B.; Chang, H.-M.; Peszlen, I.; Braun, J.;Kadla, J. Comparison of morphological and chemical propertiesbetween juvenile wood and compression wood of loblolly pine.Holzforshung2005, 59, 669-664.

(32) Akiyama, T.; Goto, H.; Nawawi, D.; Syafii, W.; Matsumoto,Y.; Meshitsuka, G.Erythro/threo ratio of â-O-4 structures asan important structural characteristic of lignin. Part 4: variationin the erythro/threoratio in softwood and hardwood lignins andits relation to syringyl/guaiacyl ratio.Holzforshung2005, 59,276-281.

(33) Brunow, G.; Karlsson, O.; Lundquist, K.; Sipila¨, J. On thedistribution of the diastereomers of the structural elements inlignin: the steric course of reactions mimicking lignin biosyn-thesis.Wood Sci. Technol.1993, 27, 281-286.

(34) Onnerud, H.; Gellerstedt, G. Inhomogeneities in the chemicalstructure of spruce lignin.Holzforschung2003, 57, 165-170.

(35) Lai, Y.-Z.; Guo, X.-P. Variation of the phenolic hydroxyl groupcontent in wood lignins.Wood Sci. Technol.1991, 25, 467-472.

(36) Sinkey, J. D. Sulphonation treatments for chemimechanicalpulping of softwood.Appita 1983, 36, 301-307.

(37) Tohmura, S.; Argyropoulos, D. S. Determination of arylglycerol-â-aryl ethers and other linkages in lignin using DFRC/31P-NMR.J. Agric. Food Chem.2001, 49, 536-542.

Received for review August 23, 2006. Revised manuscript receivedOctober 17, 2006. Accepted October 23, 2006. This work was madepossible by United States Department of Energy Grant NumberDE-FC36-04G014308.

JF062433C

Isolation of Lignin from Different Wood Species J. Agric. Food Chem., Vol. 54, No. 26, 2006 9705