UBC Faculty Research and Publications

The fate of lignin during hydrothermal pretreatment Trajano, Heather L; Engle, Nancy L; Foston, Marcus; Ragauskas, Arthur J; Tschaplinski, Timothy J; Wyman, Charles E Aug 1, 2013

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
52383-13068_2013_Article_338.pdf [ 1.08MB ]
Metadata
JSON: 52383-1.0221422.json
JSON-LD: 52383-1.0221422-ld.json
RDF/XML (Pretty): 52383-1.0221422-rdf.xml
RDF/JSON: 52383-1.0221422-rdf.json
Turtle: 52383-1.0221422-turtle.txt
N-Triples: 52383-1.0221422-rdf-ntriples.txt
Original Record: 52383-1.0221422-source.json
Full Text
52383-1.0221422-fulltext.txt
Citation
52383-1.0221422.ris

Full Text

RESEARCH Open AccessThe fate of lignin during hydrothermalpretreatmentHeather L Trajano1,2,6†, Nancy L Engle3,6†, Marcus Foston4,5,6†, Arthur J Ragauskas4,6†, Timothy J Tschaplinski3,6†and Charles E Wyman1,6*†AbstractBackground: Effective enzymatic hydrolysis of lignocellulosic biomass benefits from lignin removal, relocation,and/or modification during hydrothermal pretreatment. Phase transition, depolymerization/repolymerization, andsolubility effects may all influence these lignin changes. To better understand how lignin is altered, Populustrichocarpa x P. deltoides wood samples and cellulolytic enzyme lignin (CEL) isolated from P. trichocarpa x P.deltoides were subjected to batch and flowthrough pretreatments. The residual solids and liquid hydrolysate werecharacterized by gel permeation chromatography, heteronuclear single quantum coherence NMR, compositionalanalysis, and gas chromatography–mass spectrometry.Results: Changes in the structure of the solids recovered after the pretreatment of CEL and the production ofaromatic monomers point strongly to depolymerization and condensation being primary mechanisms for ligninextraction and redeposition. The differences in lignin removal and phenolic compound production from nativeP. trichocarpa x P. deltoides and CEL suggested that lignin-carbohydrate interactions increased lignin extraction andthe extractability of syringyl groups relative to guaiacyl groups.Conclusions: These insights into delignification during hydrothermal pretreatment point to desirable pretreatmentstrategies and plant modifications. Because depolymerization followed by repolymerization appears to be thedominant mode of lignin modification, limiting the residence time of depolymerized lignin moieties in the bulkliquid phase should reduce lignin content in pretreated biomass. In addition, the increase in lignin removal in thepresence of polysaccharides suggests that increasing lignin-carbohydrate cross-links in biomass would increasedelignification during pretreatment.Keywords: Condensation, Depolymerization, Flowthrough pretreatment, Hydrothermal pretreatment,Lignin-carbohydrate complex, Phase transitionBackgroundThe development of transportation fuels with low green-house gas emissions is imperative due to growing demandfor transportation fuels, decreasing conventional petroleumsupplies, and increasing evidence of global climate change.Ethanol produced by fermentation of sugars contained incellulose and hemicellulose in cellulosic biomass couldaddress all of these challenges, but its production requiresexpensive pretreatments and enzymes. Furthermore,efficient enzymatic hydrolysis requires hydrothermalor other pretreatments that alter the composition and/orstructure of biomass [1]. One of the primary plant cellwall components is lignin, an amorphous, phenolic poly-mer which strengthens the cell wall and protects the plantfrom microbial damage [2,3]. Because delignificationimproves enzymatic hydrolysis of the remaining biomass[4], lignin removal is an oft-cited goal for pretreatment,but many cost-effective pretreatments do not lower lignincontent appreciably. However, they do modify lignin tomake biomass more accessible to enzymes, and a betterunderstanding of the mechanisms of lignin alterationsduring pretreatment can provide valuable insights intonew pretreatment or plant modification strategies.* Correspondence: cewyman@engr.ucr.edu†Equal contributors1Department of Chemical and Environmental Engineering and Center forEnvironmental Research and Technology, Bourns College of Engineering,University of California Riverside, 1084 Columbia Ave, Riverside, CA 92507, USA6BioEnergy Science Center, Oak Ridge National Laboratory, PO Box 2008MS6341, Oak Ridge, TN 37831, USAFull list of author information is available at the end of the article© 2013 Trajano et al.; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the CreativeCommons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, andreproduction in any medium, provided the original work is properly cited.Trajano et al. Biotechnology for Biofuels 2013, 6:110http://www.biotechnologyforbiofuels.com/content/6/1/110The primary monomeric structural units of lignin arep-coumaryl, coniferyl, and sinapyl alcohol (Figure 1) [3,5].Hardwoods, such as Populus, typically contain syringyland guaiacyl lignin synthesized from sinapyl and coniferylalcohol, respectively [5], and β-O-4 (β aryl ether) linkagesaccount for approximately 80% of the linkages involvingsyringyl units [6]. Other linkages, such as β-5/α-O-4phenyl-coumaran and spirodienone linkages, are alsopresent, as shown in Figure 2. As the cell wall is lignified,ester, ether, and glycosidic bonds form between lignin andpolysaccharides, resulting in lignin-carbohydrate complexes(LCC) [3,6]. Noncovalent interactions may also link ligninand hemicellulose, but there are few interactions betweenlignin and cellulose in native biomass [6].Lignin cycles between the solid and liquid phase duringpretreatment through a complex mechanism that mayinvolve phase transition, reaction, and/or solubilization.Many researchers [7-11] have observed droplets, deter-mined to be lignin [7], on a wide variety of biomasstypes, including corn stover, switchgrass, wheat straw, andTamarix ramosissima, following hydrothermal or diluteacid pretreatment and hypothesized that these dropletsform as a result of the transition of lignin from glassy stateto rubbery state, followed by coalescence, migration, andextrusion from the cell wall [7]. Upon cooling, thesedroplets harden. This view is somewhat incomplete sincethe effects of increasing the temperature of an amorphoussolid are complex. When an amorphous solid is heated,it passes through a glass transition stage over a rangeof temperatures [12,13]. An amorphous solid withoutcross-linking will undergo rubbery flow in the absenceof thermal degradation, while a cross-linked polymer,such as lignin, can only undergo rubbery flow afterbonds break [12]. A review of the literature indicatesthe glass transition of lignin occurs somewhere in therange between 80 and 193°C [12,14-19]; the breadth ofthis range reflects differences in biomass, sample moisturecontent, lignin isolation procedures, and analyticaltechniques [12,16].In addition to these morphological changes, lignin reactsduring pretreatment. Under acidic conditions, carboniumion intermediates are formed with a high affinity fornucleophiles within the lignin structure [3]. Hydrolysisleads to depolymerization, while reactions between thecarbonium ions and nucleophiles leads to repolymerizationor condensation [20,21]. Evidence of depolymerization dur-ing pretreatment includes the loss of β-O-4 bonds [21,22]and a decrease in the molecular weight of lignin at ex-tended pretreatment times [23,24]. Extensive cleavage of β-O-4 bonds without high yields of lignin monomers suggestsdepolymerization is accompanied by repolymerization[25]. Additional evidence of repolymerization includesan increase in molecular weight during short pretreatments[21,23,24], an increase in lignin carbon-carbon bonds,as shown by infrared spectroscopy [24], and alkalinenitrobenzene oxidation [26]. There are few kinetic modelsof lignin depolymerization. However, as lignin is a solidphase reactant, the rate of depolymerization is likely pro-portional to the area of the solid–liquid interface [27]:r ¼ k 00ρpAsurf f Cð Þ ð1Þwhere k'' is the rate constant per unit surface area, ρp isthe particle density, Asurf is the surface area, and f(C) issome function of reactant concentration.Evidence also suggests that the presence of carbohydratesinfluences the solubility of lignin during pretreatment. Theaddition of carbohydrates, such as pectin or arabinoxylanduring in vitro synthesis of artificial lignin or dehydrogen-ation polymer (DHP), increased the molecular weight ofthe resulting DHP [28,29], likely through the formation ofhydrophobic complexes between DHP and carbohydrates,which prevented precipitate growth [28-30]. Similar hydro-phobic aggregates or the covalent bonds between ligninand hemicellulose may improve lignin solubility duringlignin deconstruction as well. When corn stover wassubjected to flowthrough pretreatment, there was alinear relationship between xylan and lignin removal,leading to the hypothesis that lignin is released to solutionas part of an LCC, and once in solution, the bondswithin the LCC break, producing lignin and carbohydratefragments [31-33].Figure 1 Molecular structures of selected hardwood ligninmonomers: (a) p-hydroxybenzoate, (b) guaiacyl, (c) syringyl,and (d) oxidized syringyl units.Trajano et al. Biotechnology for Biofuels 2013, 6:110 Page 2 of 16http://www.biotechnologyforbiofuels.com/content/6/1/110Observing these changes as a function of time is chal-lenging in traditional batch reactors. It is particularly dif-ficult to follow product evolution as a function of timesince side and degradation reactions generate productssuch as those known as humins that interfere with lignincharacterization [34,35]. Additionally, quenching batchreactors to stop a reaction may create artifacts such asprecipitation of oligomers [36]. These problems are avoidedwith a fixed bed flowthrough reactor. In this system,because solubilized products are quickly and continuouslyremoved from the reactor products can be tracked as afunction of time, the potential for side and degradationreactions is limited, and few solubilized products arepresent in the reactor as the reaction is quenched.Cellulolytic enzyme lignin (CEL) isolated fromPopulus trichocarpa x P. deltoides was pretreated inbatch and flowthrough systems in order to study thechanges in molecular weight, chemical bonds, and func-tional groups of lignin during pretreatment as well as theproduction of soluble aromatic compounds. The differ-ences in lignin removal and soluble aromatic productsfrom P. trichocarpa x P. deltoides wood pretreated at thesame conditions were examined to determine the impactof lignin-carbohydrate bonds on lignin deconstruction.These insights into the fundamental phenomena maysuggest new plant modifications and pretreatmentstrategies.Results and discussionChanges in the molecular weight of cellulolytic enzymelignin following hydrothermal pretreatmentThe relative number and weight average molecularweights of CEL were determined before and afterpretreatment using gel permeation chromatographyand calculated from Equations (2) and (3):Mn— ¼XiNiMiXiNið2ÞMw— ¼XiNiM2iXiNiMið3Þin which Ni is the number of polymers detected havingmolecular weight Mi.The results in Figure 3 for batch (0 mL/min) andflowthrough (20 mL/min) pretreatments show that thenumber average molecular weight of the CEL pretreatedat the same temperature and time but under flow andbatch conditions were similar, while the weight averagemolecular weights of the batch pretreated CEL were 16to 18% higher. Since the weight average molecularweight is more sensitive to the presence of large poly-mers, these results indicate more long chain polymers inFigure 2 Selected hardwood lignin structures: (a) methoxy group, (b) β-O-4 ether, (c) β-5/α-O-4 phenyl-coumaran, and (d) spirodienone.Trajano et al. Biotechnology for Biofuels 2013, 6:110 Page 3 of 16http://www.biotechnologyforbiofuels.com/content/6/1/110the batch pretreated CEL than flow pretreated CEL.Solids had equal residence times during batch andflowthrough pretreatment, however soluble com-pounds had a very short residence time, 0.15 min, dur-ing flowthrough pretreatment. Thus, while solid phasereactions, such as depolymerization, could proceedcontinuously during both pretreatment types, liquidphase reactions, such as repolymerization, were limitedduring flowthrough pretreatment. The increase in thesolids’ molecular weight after flowthrough pretreatment at140°C for 12 minutes also suggested repolymerization re-actions. Others [21,23,24] have reported increases in themolecular weight of lignin following batch pretreatment.Changes in the structure and composition of cellulolyticenzyme lignin following hydrothermal pretreatmentHeteronuclear single quantum coherence nuclear magneticresonance (HSQC-NMR) and well-established ligninspectral assignments [37] were applied to examine theinter-unit linkages and aromatic and aliphatic chemicalmoieties in CEL. HSQC-NMR was not applied to thePopulus samples because the carbohydrate signatureswould obscure the lignin-related signals. Although methodsfor semi-quantification have been proposed [38,39], HSQC-NMR is inherently difficult to quantitate for a polymersystem for a variety of reasons, as outlined by Zhang et al[40]. Moreover, because of the difficulties in dissolvingCEL for NMR, the results in Figures 4 and 5, representingDMSO soluble fraction, are discussed qualitatively.The spectra in Figure 4 reveal that residual carbohydratesdetected in the spectra of the raw CEL were not present inthe spectra of the pretreated materials, indicating that theresidual carbohydrates were removed or reacted to producepseudo-lignin [35]. However, because no Klason ligninwas detected in holocellulose, the cellulose-hemicellulosefraction of the P. trichocarpa x P. deltoides samples,pretreated at the same conditions, pseudo-lignin formationduring these pretreatments is unlikely. In both Figures 4and 5, cross-peaks in the spectra (recorded under identicalconditions) of the pretreated materials were broad anddistorted compared to untreated CEL due to a low signalto noise ratio and altered nuclear relaxation (as indicatedby significant line-broadening [40]). This outcome wasattributed to altered solubility and a modified ligninstructure with low local molecular mobility, presumablydue to condensation reactions. Finally, the relative intensityof signals resulting from the pretreated spectra were lowerin comparison to signals in the spectra of the raw CEL. InFigure 4, the reduction in signal intensity indicates the lossof methoxy groups, β-O-4 ether bonds, β-5/α-O-4 phenyl-coumaran bonds, and spirodienone bonds, while thechanges shown in Figure 5 indicate loss of functionalgroups. These results are similar to those seen by Samuelet al. [37,38] and Yelle et al. [41].The relatively more flexible aliphatic functionality foundin significant proportions in the untreated CEL is generallyassociated with more prevalent monolignol inter-unitlinkages, such as β-O-4 ether bonds; therefore its loss istypically correlated with degradation or depolymerization.However, the loss of this type of inter-unit linkage com-bined with the significant mass loss data (Figure 6) and pro-duction of phenolic compounds (Figures 7 and 8) suggeststhat the observed molecular weights (Figure 3) for thepretreated lignin were higher than one might expect.This was particularly evident for CEL pretreated at 140°Cfor 12 minutes for which the loss of mass and aliphaticfunctional groups was accompanied by an increase inmolecular weight. Changes in solubility and alterationsin nuclear relaxation, the latter indicating reduction inlocal molecular mobility, confirm the presence of largeFigure 3 Molecular weights of cellulolytic enzyme lignin (CEL) as determined by gel permeation chromatography: (a) number averageand (b) weight average. Molecular weights of CEL before pretreatment, after flowthrough pretreatment, and after batch pretreatment areshown with horizontal hatching, solid, and diagonal hatching, respectively. The black dashed line represents a molecular weight of 3000 onboth panels.Trajano et al. Biotechnology for Biofuels 2013, 6:110 Page 4 of 16http://www.biotechnologyforbiofuels.com/content/6/1/110Raw20 mL/min140oC12 min20 mL/min140oC192 min0 mL/min140oC192 min20 mL/min180oC12 min20 mL/min180oC192 min0 mL/min180oC12 mina bc de fgFigure 4 (See legend on next page.)Trajano et al. Biotechnology for Biofuels 2013, 6:110 Page 5 of 16http://www.biotechnologyforbiofuels.com/content/6/1/110polymers, which, when combined with loss of ether bondsand aliphatic functionality, suggest that repolymerizationis also occurring. Lignin repolymerization of this natureresults in a more condensed polymer structure, primarilycomposed of carbon-carbon bonds, in agreement withresults from previous studies [24,26].Lignin removal during hydrothermal pretreatment ofcellulolytic enzyme lignin and Populus trichocarpa x P.deltoidesLignin removal from CEL and P. trichocarpa x P. deltoideswood samples by pretreatment was determined from thedifference in the mass of lignin initially loaded intothe reactor and the mass of lignin recovered followingpretreatment (Figure 6, see Additional file 1, Additionalfile 2, Additional file 3 and Additional file 4 for calculationdetails). More lignin was removed from both substratesduring flowthrough pretreatment than during batchpretreatment at the same temperature and time. Thisindicates the flow of water removes lignin fragmentsfrom the reactor before they are redeposited, confirmingthat flowthrough pretreatment facilitates the observationof product formation as a function of time. Lignin removalfrom CEL by flowthrough pretreatment at 140°C for12 minutes was approximately equal to lignin removalfrom CEL by batch pretreatment at 140°C for 192 minutes,which also suggests that interrupting the redeposition oflignin to the solid phase can shorten pretreatment timeswithout compromising lignin removal.Figure 6 shows that batch pretreatment removed morelignin from CEL than from P. trichocarpa x P. deltoides,possibly due to the formation of pseudo-lignin from car-bohydrates [35]. However, no Klason lignin was detectedin holocellulose pretreated at the same conditions thereforepseudo-lignin formation is unlikely. The difference mayinstead be due to the difficulty in recovering pretreatedCEL from the reactor. More lignin was removed from P.trichocarpa x P. deltoides samples than from CEL bythree of four flowthrough pretreaments (Figure 6: 140°C/192 min, 180°C/12 min, 180°C/192 min). P. trichocarpa xP. deltoides samples remained as discrete particles duringpretreatment but pretreated CEL formed a solid whichmolded to the shape of the reactor therefore, differencesin reactive surface area could lead to differences indelignification, in line with Equation (1). Alternatively, thepresence of polysaccharides could cause increased ligninremoval from P. trichocarpa x P. deltoides samples. Toexamine surface area effects, all parameters in Equation(1) were assumed constant except the surface area.Furthermore, based on the appearance of the residualsolids, the pretreatment of P. trichocarpa x P. deltoideswas modeled as a bed of nonporous spherical lignin parti-cles with constant diameter, and no adjustments weremade for the presence of polysaccharides. The CEL bedduring flowthrough pretreatment was modeled as a tubewith an outer radius rrxtr and an inner radius of 0.5rrxtr.The solid interface during batch pretreatment of CEL wasmodeled as a solid cylinder with only the upper surfaceavailable for reaction. The ratios of these surface areas andthe ratios of percent lignin removed by pretreatment at180°C are presented in Table 1. Comparison of the surfacearea ratios and the percent lignin removed ratios revealsthat differences in surface areas were far too great relativeto the differences in mass removal to account for thedifferences in lignin removal from P. trichocarpa x P.deltoides and CEL samples. Therefore, similar to theaddition of carbohydrates during lignin synthesis [28,29],the presence of carbohydrates increased the extent of lig-nin extraction during flowthrough pretreatment. This isalso in agreement with previous work by Foston et al. [31]and Liu and Wyman [32,33]. However, the mildestflowthrough pretreatment at 140°C for 12 min removedmore lignin from CEL than from P. trichocarpa x P.deltoides. The heterogeneity of biomass results in adistibution of lignin-carbohydrate and lignin-lignin bondtypes and strengths. At these mild conditions, the weakestlignin-lignin bonds in the CEL may break while lignin-carbohydrate bonds in P. trichocarpa x. P. deltoidesremain intact, thus limiting lignin release. As pretreat-ment time increases, the proposed solubility effect ofcarbohydrates becomes significant.Production of phenolic compounds from thehydrothermal pretreatment of cellulolytic enzyme ligninand Populus trichocarpa x P. deltoides.Gas chromatography mass-spectrometry (GCMS) was usedto detect phenolic compounds in the hydrolysate. Thecumulative release of total phenolic compounds normalizedto the mass of raw lignin is plotted as a function offlowthrough pretreatment time for each run with CEL(Figure 7) and P. trichocarpa x P. deltoides (Figure 8).The cumulative release of key individual phenolic com-pounds is also presented. The distribution of phenoliccompounds produced by batch pretreatment of CEL(See figure on previous page.)Figure 4 HSQC NMR spectra of raw and pretreated cellulolytic enzyme lignin in the aliphatic region: (a) raw; (b) 20 mL/min, 140°C,12 min; (c) 20 mL/min, 140°C, 192 min; (d) 20 mL/min, 180°C, 12 min; (e) 20 mL/min, 180°C, 192 min; (f) 0 mL/min, 140°C, 192 min;(g) 0 mL/min, 180°C, 12 min. Identified units include methoxy groups, β-O-4 ethers (Aα, Aβ, Aγ), β-5/α-O-4 phenyl-coumararan (Bα, Bβ, Bγ),and spirodienone (Cα) units. Signal assignments are presented in Table 4.Trajano et al. Biotechnology for Biofuels 2013, 6:110 Page 6 of 16http://www.biotechnologyforbiofuels.com/content/6/1/1100 mL/min180oC12 minRaw20 mL/min140oC12 min0 mL/min140oC192 minf20 mL/min140oC192 min20 mL/min180oC12 min20 mL/min180oC192 mina bcedgFigure 5 (See legend on next page.)Trajano et al. Biotechnology for Biofuels 2013, 6:110 Page 7 of 16http://www.biotechnologyforbiofuels.com/content/6/1/110(Figure 7) and P. trichocarpa x P. deltoides (Figure 8)are also shown as pie charts along with the total nor-malized mass of phenolic compounds detected. In total,eighteen phenolic monomers, including hydroquinone,p-hydroxybenzoic acid, coniferyl alcohol, p-coumarylalcohol, 5-hydroxyconiferyl alcohol, vanillic acid, proto-catechuic acid, syringic acid, syringylglycerol (erythro andthreo), coniferyl aldehyde, sinapyl aldehyde, syringaresinol,vanillin, syringaldehyde, sinapyl alcohol, medioresinol,and pinoresinol were identified. Eighteen additional com-pounds detected at low concentrations are likely alsophenolic compounds, but definitive identifications couldnot be made.There were more monomers in the hydrolysates fromflowthrough pretreatment than batch pretreatment, regard-less of substrate (Figures 7, 8). As with the observationsof solids removal, this result confirms that flowing waterinterrupts lignin repolymerization. When this observationis considered in tandem with other results from batch pre-treatment of CEL: low solids removal, high weight averagemolecular weight, and loss of characteristic bonds andside chains, the case for the condensation reactions be-comes very strong. The CEL product profile after batchpretreatment at 140°C was much more diverse than afterbatch pretreatment at 180°C; in fact, p-hydroxybenzoic acidaccounted for 83% of the phenolic compounds pro-duced during batch pretreatment at 180°C. This resultsuggests that the condensation of phenolic compounds,with the exception of p-hydroxybenzoic acid, was muchgreater at 180°C during batch pretreatment. Greaterrepolymerization at 180°C was also supported by the factthat the weight average molecular weight of CEL recoveredfrom batch pretreatment at 180°C was slightly higher thanthe weight average molecular weight of CEL recoveredfrom batch pretreatment at 140°C (Figure 3b).Comparison of the production rate and total mass ofphenolic compounds produced during flowthroughpretreatment of both substrates at 140°C and 180°C for12 minutes revealed, unsurprisingly, that phenolic com-pounds were produced more rapidly and in greater amountsat 180°C. However, it is surprising that the productionrate and total mass of phenolic compounds released after192 minutes of flowthrough pretreatment at 140°C wasgreater than the rate and mass of phenolic compoundsproduced at 180°C. Since the total mass of lignin removedduring flowthrough pretreatment at 180°C was greaterthan the total mass removed by flowthrough pretreatmentat 140°C (Figure 6) and the molecular weights of the CELsolids recovered from these runs were similar (Figure 3), itis unlikely that condensation reactions were acceleratedrelative to depolymerization reactions at 180°C. Instead,the lower mass of phenolic compounds at 180°C could bethe result of the higher temperature allowing for theproduction of larger, soluble phenolic oligomers thatwere undetectable by GCMS (>1000 Da) due to theirlow volatility.p-Hydroxybenzoic acid was the primary phenol observedin the hydrolysate for the majority of runs. Without HSQCanalysis of P. trichocarpa x P. deltoides samples, therelative content of p-hydroxybenzoate units to syringyland guaiacyl units is unknown. However, the HSQCspectra of raw CEL in Figure 5 show that CEL containsfewer p-hydroxybenzoate groups than syringyl andguaiacyl units. Regardless, based on the CEL results,p-hydroxybenzoate groups were released more easilythan syringyl or guaiacyl groups.It has been previously suggested that due to greater pro-pensity for covalent linkages, guaiacyl units are less easily(See figure on previous page.)Figure 5 HSQC NMR spectra of raw and pretreated cellulolytic enzyme lignin in the aromatic region: (a) raw; (b) 20 mL/min, 140°C,12 min; (c) 20 mL/min, 140°C, 192 min; (d) 20 mL/min, 180°C, 12 min; (e) 20 mL/min, 180°C, 192 min; (f) 0 mL/min, 140°C, 192 min;(g) 0 mL/min, 180°C, 12 min. Identified units include p-hydroxyphenyl (PB2, 6), guaiacyl (G2, G5, G6), syringyl (S2, 6), and oxidized syringyl (S’2, 6)units. Signal assignments are presented in Table 4.01020304050607080P. trichocarpa x P. deltoidesCELQ(ml/min)T(oC)t(min)LigninRemoved(wt%)2014012201401920140192201801220180192018012Figure 6 Percent lignin removal from Populus trichocarpa x P.deltoides wood samples and cellulolytic enzyme lignin (CEL)after batch (0 mL/min, diagonal hatching) and flowthrough(20 mL/min, solid) pretreatments at 140 and 180°C for 12 and192 minutes. Percent lignin removal was calculated from theKlason lignin and total solids mass balance. See Additional file 1,Additional file 2, Additional file 3, and Additional file 4 for details ofthe calculations.Trajano et al. Biotechnology for Biofuels 2013, 6:110 Page 8 of 16http://www.biotechnologyforbiofuels.com/content/6/1/110extracted than syringyl units during hydrothermal pretreat-ment [26,37,42]. It has also been shown that due to the lackof a methoxy group in the C5 position, guaiacyl unitsundergo condensation reactions more easily [26]. This hy-pothesis is supported by greater coniferyl alcohol produc-tion during flowthrough pretreatment at 180°C relative topretreatment at 140°C. Also in agreement with the hypoth-esis, the P. trichocarpa x P. deltoides hydrolysates containmore syringaldehyde, syringylglycerol, syringylglycerolglycoside, and sinapyl alcohol than coniferyl alcohol andvanillin, but contradictorily, the CEL hydrolysates con-tain more coniferyl alcohol and vanillin relative tosinapyl alcohol and syringaldehyde. Therefore, the dif-ferences in the relative amounts of guaiacyl and syringyltype products from CEL and P. trichocarpa x P. deltoidessamples are strong evidence that cross-linking betweenlignin and hemicellulose changes the relative reactivity ofguaiacyl and syringyl units.Lora and Wayman [24] found that autohydrolysisof Populus tremuloides proceeded more slowly thanautohydrolysis of milled wood lignin and attributed thisdifference to the reaction of carbohydrate degradationproducts with lignin products. However in this study, withone exception, more phenolic compounds were producedfrom P. trichocarpa x P. deltoides samples than from CELduring identical pretreatments. Differences between theresults of this study and those by Lora and Wayman [24]could be due to a number of factors. The lignin that Loraand Wayman used was produced by ball milling biomassfor 18 days, followed by a dioxane water extraction, whileFigure 7 Phenolic compounds in hydrolysate generated by flowthrough and batch pretreatment of cellulolytic enzyme lignin.Cumulative release of phenolic compounds from flowthrough pretreatment (20 mL/min) are plotted as a function of time: (a) 140°C, 12 minutes;(b) 140°C, 192 minutes; (c) 180°C, 12 minutes; (d) 180°C, 192 minutes. Distribution of primary phenolic compounds after batch pretreatment:(e) at 140°C for 192 min and (f) at 180°C for 12 min.Trajano et al. Biotechnology for Biofuels 2013, 6:110 Page 9 of 16http://www.biotechnologyforbiofuels.com/content/6/1/110the CEL used in this study was produced under less severeconditions (7 days ball milling, enzymatic digestion ofcarbohydrates), thus reducing the possibility of producingan isolated lignin that is more reactive than native lignin.In addition, Lora and Wayman’s work was conducted ina batch reactor for which condensation reactions aremore significant. As previously noted in this paper, thevery short space time of the flowthrough reactor limitsproduction of degradation products and thus condensationreactions. Thus, assuming that CEL is more representativeof native lignin, the greater production of phenolic com-pounds from P. trichocarpa x P. deltoides samples waslikely due the presence of carbohydrates increasing ligninreactivity or solubility, as was shown during the in situsynthesis of lignin [28-30].Figure 8 Phenolic compounds in hydrolysate generated by flowthrough and batch pretreatment of Populus trichocarpa x P. deltoideswood samples. Cumulative release of phenolic compounds from flowthrough pretreatment (20 mL/min) are plotted as a function of time:(a) 140°C, 12 min; (b) 140°C, 192 min; (c) 180°C, 12 min; (d) 180°C, 192 min. Distribution of primary phenolic compounds after batchpretreatment: (e) at 140°C for 192 min and (f) at 180°C for12 min.Table 1 Ratio of surface area or percent lignin removedduring pretreatment of Populus trichocarpa x P. deltoidesor cellulolytic enzyme ligninRatioSurface area Lignin removed(Asurf, i/Asurf, j) (mi/mj)TD, F/CEL, B 1209 2.36TD, F/ CEL, F 384 1.27CEL, F/ CEL, B 3 1.86The surface areas of solids and percent lignin removed during flowthroughpretreatment of Populus trichocarpa x P. deltoides wood samples (i, j = TD, F)and cellulolytic enzyme lignin in flowthrough (i, j = CEL, F) and batchpretreatment (i, j = CEL, B) at 180°C for 12 min were calculated and compared.Trajano et al. Biotechnology for Biofuels 2013, 6:110 Page 10 of 16http://www.biotechnologyforbiofuels.com/content/6/1/110ConclusionsPrevious studies showed that during hydrothermal pre-treatment, lignin cycles between the solid and liquid phaseand suggested that depolymerization/repolymerizationreactions and lignin-carbohydrate interactions play keyroles, although the mechanisms were not well understood.In order to investigate lignin behavior during pretreatment,Populus trichocarpa x P. deltoides wood samples and cellu-lolytic enzyme lignin (CEL) isolated from P. trichocarpa xP. deltoides were subjected to batch and flowthrough hydro-thermal pretreatment. The residual solids and liquid hy-drolysate were characterized by a wide range of analyticaltechniques.Pretreatment of CEL resulted in loss of characteristiclignin bonds and functional groups from the solids, theproduction of phenolic monomers, and modest reductionsin molecular weight. These observations point strongly todepolymerization and condensation being the primarymechanisms for lignin extraction and redeposition. Nodirect evidence of phase transition was observed.There were some similarities in the pretreatment of P.trichocarpa x P. deltoides wood samples and CEL. Shortliquid residence times, due to the flow of water, limitedcondensation reactions for both substrates, as evidencedby greater lignin extraction by flowthrough pretreatment.More lignin was extracted and more phenolic com-pounds were produced by pretreatment of P. trichocarpa xP. deltoides wood samples than by pretreatment of CEL.The difference in lignin extraction between the two sub-strates was too small to be due to differences in solid–li-quid interfacial area. Therefore, it is likely that lignin-carbohydrate interactions significantly influence lignindeconstruction during pretreatment. Due to theirgreater potential for cross-linking, guaiacyl units arethought to be less easily extracted and more easily con-densed, but guaiacyl based phenolic compounds werethe dominant type produced from CEL. However,syringyl based phenolic compounds were the dominantaromatic constituents released from P. trichocarpa x P.deltoides wood samples, suggesting that cross-linksbetween hemicellulose and lignin modify the reactivityof guaiacyl and syringyl groups.The production of ethanol from lignocellulosic biomassrequires efficient and economical pretreatment and enzym-atic hydrolysis. Effective lignin removal or alteration duringpretreatment leads to improved enzymatic hydrolysis. Theknowledge of lignin depolymerization/repolymerizationkinetics and the impact of lignin composition and lignin-carbohydrate interactions on lignin removal gained in thisstudy will help guide plant modification and pretreatmentstrategies. An increase in lignin-carbohydrate cross-links inbiomass and limiting the residence time of depolymerizedlignin moieties during pretreatment would increase ligninremoval during pretreatment.MethodsSubstratesThe stem wood of hybrid cottonwood poplar Populustrichocarpa x P. deltoides clone (TD) ‘53-239’ ♂ used inthis study was provided by Oak Ridge National Laboratory,TN. Logs were debarked, split with an axe, chipped(Yard Machines 10HP, MTD Products Inc., Cleveland,OH), and knife milled (Model 4 Wiley Mill, ThomasScientific, Swedesboro, NJ) through a 1 mm screen size;all of these operations were performed at the NationalRenewable Energy Laboratory (NREL). After one monthof air-drying at NREL, the chips had a moisture contentof approximately 5 wt%. The material was further milledto particles with dimensions of 0.18 mm to 0.85 mm(Thomas-Wiley Laboratory Mill Model 4, Arthur H.Thomas Company, Philadelphia, PA) before being shipped.The cellulolytic enzyme lignin (CEL) was isolated fromP. trichocarpa x P. deltoides by modifying the proceduresdescribed by Chang et al. [43] and Björkman [44]. P.trichocarpa x P. deltoides was ball-milled for 7 days;this material was then subjected to two consecutiverounds of enzymatic hydrolysis with 500 IU cellulase/gP. trichocarpa x P. deltoides (Novozym 188, Sigma-Aldrich, St. Louis, MO) and 200 IU β-glucosidase/gP. trichocarpa x P. deltoides (Celluclast 1.5 L, Sigma-Aldrich, St. Louis, MO) for seventy-two hours at 50°Cshaken at a frequency of 150 rpm. Following hydrolysis,the solids were washed and then subjected to two rounds ofextraction with dioxane for twenty-four hours. The compos-ition of the untreated substrates determined using theprocedures described below are summarized in Table 2.ReactorsA schematic of the flowthrough reactor system employedin this study is shown in Figure 9. A 2 L feed tank wasused to hold the process water. A positive displacementpump (Prep100, LabAlliance, State College, PA) deliveredwater to the reactor, and the pressure of the system wasset using the backpressure regulator (GO, Spartanburg,SC). The system pressure was monitored by pressuregauges P1 (US Gauge, max P 20.6 MPag), P2 (Ashcroft,max P 4.2 MPag), and P3 (Ashcroft, max P 4.2 MPag).The heating coil and reactor were heated using a fluidizedsand bath (SBL-2D, Techne, Princeton, NJ), and theheating coil was a 2.6 m length of stainless steel tubing(D = 3.18 mm) with a coil diameter of 50.9 mm. Thereactor was constructed of a stainless steel tube (D =12.7 mm) with Swagelok fittings (SS-8-VCR-1, SS-8-VCR-3-8TA, SS-8VCR-6-810, SS-200-R-8, Swagelok, San Diego,CA). The total reactor length was 152 mm. The biomasswas held in the reactor by 5 micrometre gaskets (SS-8-VCR-2-5 M, Swagelok, San Diego, CA). The systemtemperature was monitored with K-type thermocoupleT1 at the reactor outlet and recorded as a function ofTrajano et al. Biotechnology for Biofuels 2013, 6:110 Page 11 of 16http://www.biotechnologyforbiofuels.com/content/6/1/110time using a Digi-Sense DualLogR Thermocouple Meter(15-176-96, Fisher Scientific, Pittsburgh, PA). Data wastransferred from the meter to a computer using an infraredadapter (EW-91100-85, Cole Parmer, Vernon Hills, IL).The cooling coil, a 5.3 m length of stainless steel tubing(D = 3.18 mm) with a coil diameter of 44.5 mm, wassubmerged in a 19 L water bath to cool the hydrolysateprior to sampling. The sampling point is open to theatmosphere so that hydrolysate could be collected con-tinuously during a run.Custom-built 10 mL reactors were used for the batchpretreatments. The reactors were constructed fromstainless steel tubing with an outer diameter of 12.7 mm, alength of 150 mm, and sealed using threaded caps(SS-810-C, Swagelok, San Diego, CA). One reactor had athermocouple (.062-K-U-4″-T3-10 ft TF/TF-MP, WilconIndustries, Lake Elsinore, CA) inserted along the center-line to record the reactor temperature as a function oftime using a Digi-Sense DualLogR ThermocoupleMeter (15-176-96, Fisher Scientific, Pittsburgh, PA).Data was transferred from the meter to a computerusing an infrared adapter (EW-91100-85, Cole Parmer,Vernon Hills, IL).PretreatmentThe pretreatment conditions, equipment set points, andsampling intervals for flowthrough and batch pretreatmentare summarized in Table 3. These pretreatment conditionswere selected in order to observe lignin deconstructionover a wide range of pretreatment severities. The flow-through reactor was loaded with 0.71 g of dry CEL or1.00 g of dry P. trichocarpa x P. deltoides. A representativeflowthrough pretreatment run with CEL at 140°C (Trxtr)for 12 minutes is described; the same procedure wasfollowed for all other runs as summarized in Table 3. Thepump was primed, and the sand bath was heated to 149°C(Tsand) prior to pretreatment. After the reactor was loadedwith CEL it was attached to the flowthrough piping system.The pump was set to 20 mL/min (Q) and started, and theback pressure gauge was adjusted to 0.28 MPag (Pset). Thepressurized system was inspected for leaks at roomtemperature. After repairing any leaks, the reactor andheating coil were lowered into the sand bath, and thecooling coil was lowered into the water bath. A Digi-SenseDualLogR Thermocouple Meter was used to monitor thereactor temperature during the run. The time at whichthe temperature reached 138°C (Trxtr-2°C) was recordedas the start of the reaction. Hydrolysate was collected asthe reactor was heated and then over 3 min (Δt) intervalsduring the run using pre-massed flasks. After eachsampling interval, the filled flask was exchanged with aclean, pre-massed flask. The mass of the filled flaskwas recorded before a sample of the hydrolysate wasviiiiiiiiivivxP1T1ixP2 P3iv vii(i)(ii)(iii)(iv)(v)(vi)(vii)(viii)(ix)(x)PTPressure gaugeThermocoupleFeed tankPositive displacement pumpPressure relief valveHeating coilReactorFluidized sand bathCooling coilWater bathBack pressure regulatorSampling pointFigure 9 Schematic of the flowthrough pretreatment system.Table 2 Composition of raw Populus trichocarpa x P.deltoides and cellulolytic enzyme lignin (CEL)ComponentP. trichocarpa x Cellulolytic EnzymeLignin (CEL)P. deltoidesGlucan (wt%) 40.5 (0.2) 0.591 (0.034)Xylan (wt%) 11.5 (0.4) 2.67 (0.08)Klason lignin (wt%) 22.7 (0.1) 84.4 (2.3)Values represent the average of triplicate samples. The standard deviations areshown in brackets.Trajano et al. Biotechnology for Biofuels 2013, 6:110 Page 12 of 16http://www.biotechnologyforbiofuels.com/content/6/1/110taken for subsequent analysis. After 12 min, the reactorand heating coil were transferred to the water bath andcooled to 70°C, at which point the pump was stoppedand the reactor removed from the piping. The residualsolids in the reactor were collected by filtration, washedwith three 100 mL volumes of deionized water, and driedat 45°C for 24 h. As the CEL solids had fused to thereactor, vigorous scraping was required to recover them.Batch pretreatment of CEL was conducted using onebatch reactor loaded with 0.71 g CEL and 7.6 mL of de-ionized water while batch pretreatment of P. trichocarpa xP. deltoides samples was conducted using two tube reac-tors, each loaded with 0.44 g dry biomass and 7.88 mL ofdeionized water. A representative batch pretreatmentof CEL at 140°C (Trxtr) is described; this procedure wasrepeated according to the conditions described in Table 3.The sand bath was heated to 142°C (Tsand, 1). A batchreactor loaded with CEL and deionized water wassealed, shaken, and allowed to soak for 3.5 hours. Aseparate reactor equipped with a thermocouple wasloaded with 8.34 mL deionized water. Both the tube re-actor and thermocouple reactor were placed into a wirebasket before being lowered into the sand bath, and thetemperature was monitored with a Digi-Sense DualLogRThermocouple Meter. The time at which the reactorsreached 138°C (Trxtr-2°C) was taken as the start of the reac-tion, and the sand bath temperature was reset to 140°C(Tsand, 2). After 192 minutes, the wire basket and reactorswere transferred to a water bath and cooled to 70°C, andthe temperature data was transferred to a computer. Fol-lowing pretreatment, the residual solids were recovered byfiltration using a pre-weighed crucible, and the filtrate wasretained for analysis. Once again, vigorous scraping wasrequired to remove CEL solids that had fused to the re-actor walls. The pretreated solids were washed and driedas described above and were stored at room temperatureuntil analyses were performed.The following P. trichocarpa x P. deltoides pretreatmentswere done in triplicate in order to establish reproducibility:flowthrough pretreatment at 140°C for 192 min, flow-through pretreatment at 180°C for 10 min, batch pretreat-ment at 140°C for 192 min, and batch pretreatment at180°C for 12 min. The maximum standard deviationsassociated with the glucan, xylan, and lignin measurementsduring these runs were 0.6%, 3.2%, and 5.0%, respectively.Gel permeation chromatographyThe relative number average and weight average molecularweights of CEL before and after pretreatment were deter-mined by gel permeation chromatography (GPC), usingthe procedure described by Samuel et al. [37] Prior toGPC, the material was acetylated by combining20 mg of lignin with approximately 1.0 mL of 1:1 anhyd-rous pyridine/acetic anhydride under an inert atmospherefor 48 h at room temperature. Adding 5 mg of ethanolquenched the reaction. The solvent was removed by rotaryevaporation followed by drying in a vacuum oven at 45°C.Data were collected using an Agilent GPC SECurity 1200system with a refractive index detector and a UV de-tector (270 nm) using tetrahydrofuran as an eluent at aflow rate of 1.0 mL/min. Four Waters Styragel columns(HR1, HR2, HR4, HR6) were used to separate acetylatedlignin by molecular weight (Waters Co., Milford, MA).The calibration curve was constructed using eight polystyr-ene standards ranging in molecular weight from 1.5 × 103Table 3 Summary of flowthrough (20 mL/min) and batch(0 mL/min) pretreatment conditionsFlow rate Temperature, Setpressure TimeSampleinterval(Q, mL/min) Sand Bath(Tsand,°C)Reactor(Trxtr,°C)(Pset, MPag) (trxn, min) (Δt, min)20 149 140 0.28 12 320 149 140 0.28 192 150Tsand, 1 = 142140 n/a 192 n/aTsand, 2 = 14020 194 180 1.10 12 320 194 180 1.10 192 150Tsand, 1 = 182180 n/a 12 n/aTsand, 2 = 180Table 4 Assignment of 13C-1H correlation signals detectedin HSQC spectra of cellulolytic enzyme lignin isolatedfrom Populus trichocarpa x P. deltoides wood samples [37]δC/δH (ppm) Assignment53.2/3.5 Cβ/Hβ in phenylcoumarin substructure (Bβ)53.6/3.1 Cβ/Hβ in resinol (β-β) substructure55.7/3.8 C/H in methoxyl group (Methoxy)60.2/3.6 Cγ/Hγ in β-O-4 substructure (Aγ)62.8/3.8 Cβ/Hβ in phenylcoumarin substructure (Bβ)71.5/4.8 Cα/Hα in β-O-4 linkage (Aα)84.8/4.3 Cβ/Hβ in β-O-4 linkage (Aβ)81.4/5.1 Cβ/Hβ in spirodienone substructure (Cβ)84.7/4.7 Cα/Hα in spirodienone substructure (Cα)87.1/5.5 Cα/Hα in phenylcoumarin substructure (Bα)104.3/6.7 C2,6/H2,6 in etherified syringyl units (S2,6)105.5/7.3 C2,6/H2,6 in oxidized Cα=O (S'2,6)113.7/6.3 Cβ/Hβ in cinnamate unit (Eβ)111.4/7.0 C2/H2 in guaiacyl units (G2)115.4/6.77 C5/H5 in guaiacyl units (G5)119.3/6.82 C6/H6 in guaiacyl units (G6)130.0/7.5 C2,6/H2,6 in p-hydroxybenzoate units (PB2,6)144.7/7.5 Cα/Hα in cinnamate unit (Eα)Trajano et al. Biotechnology for Biofuels 2013, 6:110 Page 13 of 16http://www.biotechnologyforbiofuels.com/content/6/1/110to 3.6 × 106 g/mol (Polysciences Inc., Warrington, PA).The data was collected and processed using PolymerStandards Service WinGPC Unity software (Build 6807).Heteronuclear single quantum coherence nuclearmagnetic resonanceHeteronuclear single quantum coherence nuclear magneticresonance (HSQC NMR) was used to analyze the func-tional groups and intra-lignin bonds of CEL. Cross peakswere assigned according to Samuel et al. [37], as listed inTable 4. CEL was prepared for HSQC by dissolving 60 mgsolids in 1 mL of dimethyl sulfoxide-d6 (DMSO-d6 99.9atom% D, Cambridge Isotope Laboratories, Andover, MA).However, given that complete dissolution of the pretreatedCEL was not achieved, the results can only be interpretedqualitatively. 2D 13C-1H HSQC correlation NMR spectrawere recorded on a Bruker DRX 500 spectrometer witha 5 mm z-gradient triple resonance probe with inversegeometry at 60°C (Bruker, Billerica, MA). Analysis wasperformed with a Bruker phase-sensitive gradient-editedHSQC pulse sequence using 1024 data points for a0.11 s acquisition time, a 1.5 s recycle delay, a 1JC-Hcoupling constant of 145 Hz, and acquisition of 256 datapoints in the F1 dimension. The data was processedusing zero-filling to 2048 points and a typical squaredsine-bell apodization in both F2 and F1 dimensions.Carbohydrate and Klason lignin analysisThe structural carbohydrate and Klason lignin contentof the raw P. trichocarpa x P. deltoides wood samples andCEL were determined using the two-step acid hydrolysisprocedure outlined by Sluiter et al. [45] The compositionof the pretreated P. trichocarpa x P. deltoides was alsomeasured. However, since the CEL fused during pretreat-ment to form a single solid particle, the composition ofpretreated CEL could not be measured. The carbohydratecomposition of the hydrolysate produced by pretreatmentwas determined according the procedure outlined bySluiter et al. [46] Sugars were detected by high pressureliquid chromatography (HPLC) using an Aminex HPX-87Hcolumn (BioRad, Hercules, CA) heated to 65°C with aseparation module (Alliance 2695, Waters, Milford, MA)equipped with a refractive index detector (2414, Waters,Milford, MA). The eluent was 0.005 M sulfuric acid in theisocratic mode. These results were used to complete massbalances of glucan, xylan, and Klason lignin for eachsubstrate-pretreatment combination (see Additional file 1,Additional file 2, Additional file 3 and Additional file 4).Gas chromatography–mass spectrometryThe phenolic compounds in the hydrolysates from pre-treatment of CEL and P. trichocarpa x P. deltoides woodsamples were determined by gas chromatography–massspectrometry (GCMS) [47,48]. In brief, the samples werefirst filtered through a 0.45 μm nylon membrane and thendried in a helium stream. Sorbitol (15 μl of a 1 mg/mLaqueous solution) was added as an internal standard.The dried extracts were silylated to produce trimethylsilylderivatives prior to injection on an Agilent TechnologiesInc. 5975C inert XL gas chromatograph-mass spectrom-eter. Key mass/charge (m/z) ratios for identified aromaticmetabolites were extracted from the total ion current toquantify metabolites free from co-eluting interference.Predetermined scaling factors were used to scale theextracted peak areas back up to the total ion current. Theconcentrations were normalized to the quantity of the in-ternal standard (sorbitol) recovered, amount of samplederivitized, and injected. Predetermined response factorsfor each identified aromatic metabolite relative to the in-ternal standard were used to determine actual metaboliteconcentration (μg/mL).Additional filesAdditional file 1: Mass balances for the pretreatment of cellulolyticenzyme lignin at 140°C. Additional file 1 summarizes the glucan, xylan,and Klason lignin mass balances for the pretreatment of cellulolyticenzyme lignin (CEL) at 140°C. The CEL fused during pretreatment to forma single solid particle therefore it was not possible to perform acompositional analysis on the residual solids. The concentrations ofsugars in the hydrolysate were too low to accurately measure. However,as no carbohydrate signals were present in the HSQC-NMR spectra of thepretreated CEL (Figure 4) and no lignin was detected in holocellulose,the cellulose-hemicellulose fraction of the P. trichocarpa x P. deltoidessamples, pretreated at the same conditions, it was assumed that theglucan and xylan were completely removed during pretreatment.Therefore the initial mass of glucan and xylan was substracted from thechange in solid mass to determine the mass of lignin removed.Additional file 2: Mass balances for the pretreatment of cellulolyticenzyme lignin at 180°C. Additional file 2 summarizes the glucan, xylan,and Klason lignin mass balances for the pretreatment of cellulolyticenzyme lignin (CEL) at 180°C. The mass balances were calculated usingthe same process described for Additional file 1.Additional file 3: Mass balances for the pretreatment of Populustrichocarpa x P. deltoides at 140°C. Additional file 3 summarizes theglucan, xylan, and Klason lignin mass balances for the pretreatment ofPopulus trichocarpa x P. deltoides at 140°C. The mass of lignin removedwas calculated as the difference between the Klason lignin in theuntreated and pretreated solids.Additional file 4: Mass balances for the pretreatment of Populustrichocarpa x P. deltoides at 180°C. Additional file 4 summarizes theglucan, xylan, and Klason lignin mass balances for the pretreatment ofPopulus trichocarpa x P. deltoides at 180°C. The mass of lignin removedwas calculated as the difference between the Klason lignin in theuntreated and pretreated solids.AbbreviationsAsurf: Surface area; B: Batch pretreatment; CEL: Cellulolytic enzyme lignin;DHP: Dehydrogenation polymer; DMSO: Dimethyl sulfoxide; f(C): Function ofreactant concentration; F: Flowthrough pretreatment; GCMS: Gas chromatographymass-spectrometry; GPC: Gel permeation chromatography; HSQCNMR: Heteronuclear single quantum coherence nuclear magnetic resonance;k'': Rate constant per unit surface area; LCC: Lignin-carbohydrate complexes;Mi: Molecular weight of polymer i; Mn : Number average molecular weight;Mw : Weight average molecular weight; Ni: Number of polymers with molecularweight Mi; Q: Flow rate; R: Rate of reaction; rrxtr: Reactor radius; rxn: Reaction;t: Time; T: Temperature; TD: Populus trichocarpa x P. deltoides; ρp: Particle density.Trajano et al. Biotechnology for Biofuels 2013, 6:110 Page 14 of 16http://www.biotechnologyforbiofuels.com/content/6/1/110Competing interestsCEW is founding Editor in Chief of this Journal. CEW was cofounder ofMascoma Corporation and until recently, Chief Development Officer andChair of their Scientific Advisory Board. CEW is also member of the ScientificAdvisory Board of Mendel Biotechnology, Inc.Authors’ contributionsHLT performed the pretreatments and structural carbohydrate and Klasonlignin analyses, and drafted the manuscript. NLE and TJT performed the gaschromatography–mass spectrometry. MF performed the gel permeationchromatography and heteronuclear single quantum coherence nuclearmagnetic resonance. AJR and CEW coordinated the research and helpedfinalize the manuscript. All authors read and approved the final manuscript.AcknowledgementsWe thank the Office of Biological and Environmental Research in the DOEOffice of Science for supporting this work through the BioEnergy ScienceCenter (BESC). BESC is a U.S. Department of Energy Bioenergy ReseachCenter supported by the Office of Biological and Environmental Research inthe DOE Office of Science. This manuscript has been co-authored by acontractor of the U.S. Government under contract DE-AC05-00OR22725. Wealso wish to thank Dr. Shilin Cao, previously at the School of Chemistry andBiochemistry, Institute of Paper Science and Technology, Georgia Institute ofTechnology and now at the College of Material Engineering at FujianAgriculture and Forestry University, for preparing the cellulolytic enzymelignin for this study. We acknowledge support by the Ford Motor Companyfor the Chair in Environmental Engineering at the University of CaliforniaRiverside (UCR) that augments our ability to perform such research.Author details1Department of Chemical and Environmental Engineering and Center forEnvironmental Research and Technology, Bourns College of Engineering,University of California Riverside, 1084 Columbia Ave, Riverside, CA 92507, USA.2Current address: Department of Chemical and Biological Engineering, TheUniversity of British Columbia, 2360 East Mall, Vancouver, British Columbia V6T1Z3, Canada. 3Biosciences Division, Oak Ridge National Laboratory, PO Box 2008MS6341, Oak Ridge, TN 37831, USA. 4School of Chemistry and Biochemistry,Institute of Paper Science and Technology, Georgia Institute of Technology, 50010th Street N.W., Atlanta, GA 30332, USA. 5Department of Energy,Environmental & Chemical Engineering, Washington University in St. Louis, 1Brookings Drive, Saint Louis, MO 63130, USA. 6BioEnergy Science Center, OakRidge National Laboratory, PO Box 2008 MS6341, Oak Ridge, TN 37831, USA.Received: 20 May 2013 Accepted: 26 July 2013Published: 1 August 2013References1. Mosier N, Wyman C, Dale B, Elander R, Lee YY, Holtzapple M, Ladisch M:Features of promising technologies for pretreatment of lignocellulosicbiomass. Bioresour Technol 2005, 96:673–685.2. Raven PH, Evert RF, Eichhorn SE: Biology of plants. 7th edition. New York: W.H. Freeman and Company Publishers; 2005.3. Fengel D, Wegener G: Wood: chemistry, ultrastructure, reactions. Berlin: Walterde Gruyter & Co; 1983.4. Kumar R, Wyman CE: Key features of pretreated lignocelluloses biomasssolids and their impact on hydrolysis. In Bioalcohol production: biochemicalconversion of lignocellulosic biomass. Edited by Waldon K. Oxford: WoodheadPublishing Ltd; 2000:73–121.5. Sannigrahi P, Ragauskas AJ, Tuskan GA: Poplar as a feedstock for biofuels:a review of compositional characteristics. Biofuels Bioprod Bior 2010,4:209–226.6. Albersheim P, Darvill A, Roberts K, Sederoff R, Staehelin A: Plant cell wallsfrom chemistry to biology. New York: Garland Science; 2011.7. Donohoe BS, Decker SR, Tucker MP, Himmel ME, Vinzant TB: Visualizinglignin coalescence and migration through maize cell walls followingthermochemical pretreatment. Biotechnol Bioeng 2008, 101:913–925.8. Kristensen JB, Thygesen LG, Felby C, Jørgensen H, Elder T: Cell-wallstructural changes in wheat straw pretreated for bioethanol production.Biotechnology for Biofuels 2008, 1. doi:10.1186/1754-6834-1-5.9. Pingali SV, Urban VS, Heller WT, McGaughey J, O’Neill H, Foston M, MylesDA, Ragauskas A, Evans BR: Breakdown of cell wall nanostructure in diluteacid pretreated biomass. Biomacromolecules 2010, 11:2329–2335.10. Selig MJ, Viamajala S, Decker SR, Tucker MP, Himmel ME, Vinzant TB:Deposition of lignin droplets produced during dilute acid pretreatmentof maize stems retards enzymatic hydrolysis of cellulose. Biotechnol Prog2007, 23:1333–1339.11. Xiao LP, Sun ZJ, Shi ZJ, Xu F, Sun RC: Impact of hot compressed waterpretreatment on the structural changes of woody biomass forbioethanol production. Bioresources 2011, 6:1576–1598.12. Irvine GM: The significance of the glass transition of lignin inthermochemical pulping. Wood Sci Technol 1985, 19:139–149.13. Gutzow IS, Schmelzer JWP: Basic properties and the nature of glasses: anoverview. In Glasses and the glass transition. 1st edition. Edited bySchmelzer JWP, Gutzow IS. Hoboken: Wiley-VCH; 2011:1–89.14. Fang Z, Sato T, Smith RL Jr, Inomata H, Arai K, Kozinski JA: Reactionchemistry and phase behavior of lignin in high temperature andsupercritical water. Bioresour Technol 2008, 99:3424–3430.15. Petridis L, Schulz R, Smith JC: Simulation analysis of the temperaturedependence of lignin structure and dynamics. J Am Chem Soc 2011,133:20277–20287.16. Saake B, Lehnen R: Lignin. In Ullmann’s Encyclopedia of industrial chemistry.Edited by Bohnet M, Bellussi G, Bus J, Cornils B, Drauz K, Greim H, et al.New York: Wiley; 2007.17. Salmén L: Viscoelastic properties of in situ lignin under water-saturatedconditions. J Mater Sci 1984, 19:3090–3096.18. Sarkanen KV, Ludwig CH: Lignins- occurrence, formation, structure, andreactions. Canada: John Wiley & Sons Inc; 1971.19. Shao S, Jin Z, Wen G, Iiyama K: Thermocharacteristics of steam-explodedbamboo (Phyllostachys pubescens) lignin. Wood Sci Technol 2009, 43:643–652.20. Chua MGS, Wayman M: Characterization of autohydrolysis aspen (P.tremuloides) lignins. Part 3. Infrared and ultraviolet studies of extractedautohydrolysis lignin. Can J Chem 1979, 57:2603–2611.21. Li J, Henriksson G, Gellerstedt G: Lignin depolymerization/repolymerizationand its critical role for delignification of aspen wood by steam explosion.Bioresour Technol 2007, 98:3061–3068.22. Marchessault RH, Coulombe S, Morikawa H: Characterization of aspenexploded wood lignin. Can J Chem 1982, 60:2372–2382.23. Chua MGS, Wayman M: Characterization of autohydrolysis aspen(P. Tremuloides) lignins. Part 1. Composition and molecular weightdistribution of extracted autohydrolysis lignin. Can J Chem 1979,57:1141–1149.24. Lora JH, Wayman M: Autohydrolysis of aspen milled wood lignin.Can J Chem 1980, 58:669–676.25. Bardet M, Robert DR, Lundquist K: On the reactions and degradation ofthe lignin during steam hydrolysis of aspen wood. Sven Papperstidn 1985,88:R61–R67.26. Wayman M, Chua MGS: Characterization of autohydrolysis aspen(P. Tremuloides) lignins. Part 2. Alkaline nitrobenzene oxidation studies ofextracted autohydrolysis lignin. Can J Chem 1979, 57:2599–2602.27. Fogler HS: Elements of chemical reaction engineering, Upper Saddle River,Prentice Hall Professional Technical Reference. 4th edition; 2006:645–866.28. Barakat A, Chabbert B, Cathala B: Effect of reaction media concentrationon the solubility and the chemical structure of lignin model compounds.Phytochemistry 2007, 68:2118–2125.29. Cathala B, Monties B: Influence of pectins on the solubility and molarmass distribution of dehyrogenative polymers (DHPs, lignin modelcompounds). Int J Biol Macromol 2001, 29:45–51.30. Barakat A, Winter H, Rondeau C, Saake B, Chabbert B, Cathala B: Studies ofxylan interactions and cross-linking to synthetic lignins formed by bulkand end-wise polymerization: a model study of lignin carbohydratecomplex formation. Planta 2007, 226:267–281.31. Foston MB, McKenzie HL, Wyman CE, Ragauskas AJ: Analysis of 13C enrichedcorn stover by water-only flow-through pretreatment, 33rd Symposium onBiotechnolgy for Fuels and Chemicals: 3 May 2011. Seattle.32. Liu C, Wyman CE: The effect of flow rate of compressed hot water onxylan, lignin, and total mass removal from corn stover. Ind Eng Chem Res2003, 42:5409–5416.33. Liu C, Wyman CE: The effect of flow rate of very dilute sulfuric acid onxylan, lignin, and total mass removal from corn stover. Ind Eng Chem Res2004, 43:2781–2788.Trajano et al. Biotechnology for Biofuels 2013, 6:110 Page 15 of 16http://www.biotechnologyforbiofuels.com/content/6/1/11034. Saeman JF: Kinetics of wood saccharification- hydrolysis of cellulose anddecomposition of sugars in dilute acid at high temperature.Ind Eng Chem 1945, 37:43–52.35. Sannigrahi P, Kim DH, Jung S, Ragauskas A: Pseudo-lignin andpretreatment chemistry. Energy Environ Sci 2011, 4:1306–1310.36. Gray MC, Converse AO, Wyman CE: Solubilities of oligomer mixturesproduced by the hydrolysis of xylans and corn stover in water at 180°C.Ind Eng Chem Res 2007, 46:2383–2391.37. Samuel R, Pu YQ, Raman B, Ragauskas AJ: Structural characterization andcomparison of switchgrass ball-milled lignin before and after dilute acidpretreatment. Appl Biochem Biotechnol 2010, 162:62–74.38. Samuel R, Foston M, Jaing N, Cao S, Ragauskas AJ, et al: HSQC(heteronuclear single quantum coherence) 13C–1H correlation spectra ofwhole biomass in perdeuterated pyridinium chloride–DMSO system: aneffective tool for evaluating pretreatment. Fuel 2011, 90:2836–2842.39. Samuel R, Foston M, Jaing N, Allison L, Ragauskas AJ: Structural changes inswitchgrass lignin and hemicelluloses during pretreatments by NMRanalysis. Polym Degrad Stabil 2011, 96:2002–2009.40. Zhang L, Gellerstedt G: Quantitative 2D HSQC NMR determination ofpolymer structures by selecting suitable internal standard references.Mag Res Chem 2007, 47:37–45.41. Yelle DJ, Kaparaju P, Hunt CG, Hirth K, Kim H, Ralph J, Felby C: Two-dimensional NMR evidence for cleavage of lignin and xylan substituentsin wheat straw through hydrothermal pretreatment and enzymatichydrolysis. Bioenergy Research 2013, 6:211–221.42. Chiang VL, Funaoka M: The differences between guaiacyl and guaiacyl-syringyl lignins in their responses to Kraft delignification. Holzforschung1990, 44:309–313.43. Chang HM, Cowling EB, Brown W, Adler E, Miksche G: Comparative studieson cellulolytic enzyme lignin and milled wood lignin of sweetgum andspruce. Holzforschung 1975, 29:153–159.44. Björkman A: Studies on finely divided wood, Part 1. Extract of lignin withneutral solvents. Svensk Papperstidn 1956, 59:477–485.45. Sluiter A, Hames B, Ruiz R, Scarlata C, Sluiter J, Templeton D, Crocker D:Determination of structural carbohydrates and lignin in biomass laboratoryanalytical procedure. Golden: National Renewable Energy Laboratory; 2011.46. Sluiter A, Hames B, Ruiz R, Scarlata C, Sluiter J, Templeton D: Determinationof sugars, byproducts, and degradation products in liquid fraction processsamples laboratory analytical procedure. Golden: National Renewable EnergyLaboratory; 2006.47. Jung HW, Tschaplinski TJ, Wang L, Glazebroo J, Greenberg JT: Priming insystemic plant immunity. Science 2009, 324:89–91.48. Li Y, Tschaplinski TJ, Engle NL, Hamilton CY, Rodriguez M Jr, Liao JC, SchadtCW, Guss AM, Yang Y, Graham DE: Combined inactivation of theClostridium cellulolyticum lactate and malate dehydrogenase genessubstantially increases ethanol yield from cellulose and switchgrassfermentations. Biotechnology for Biofuels 2012, 5. doi:10.1186/1754-6834-5-2.doi:10.1186/1754-6834-6-110Cite this article as: Trajano et al.: The fate of lignin during hydrothermalpretreatment. Biotechnology for Biofuels 2013 6:110.Submit your next manuscript to BioMed Centraland take full advantage of: • Convenient online submission• Thorough peer review• No space constraints or color figure charges• Immediate publication on acceptance• Inclusion in PubMed, CAS, Scopus and Google Scholar• Research which is freely available for redistributionSubmit your manuscript at www.biomedcentral.com/submitTrajano et al. Biotechnology for Biofuels 2013, 6:110 Page 16 of 16http://www.biotechnologyforbiofuels.com/content/6/1/110

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.52383.1-0221422/manifest

Comment

Related Items