Green Chemistry PAPER

View Online / Journal Homepage / Table of Contents for this issue Green Chemistry Dynamic Article Links

Cite this: Green Chem., 2011, 13, 2489 www.rsc.org/greenchem PAPER

Ionic liquid pretreatment of lignocellulosic biomass with ionic liquid–water mixtures†

Agnieszka Brandt,a,b MichaelJ.Ray,b,c Trang Q. To,a David J. Leak,b,c Richard J. Murphyb,c and Tom Welton*a

Received 6th April 2011, Accepted 21st June 2011 DOI: 10.1039/c1gc15374a

Ground lignocellulosic biomass (Miscanthus giganteus,pine(Pinus sylvestris) and willow (Salix viminalis)) was pretreated with ionic liquid–water mixtures of 1-butyl-3-methylimidazolium methyl sulfate and 1-butyl-3-methylimidazolium hydrogen sulfate. A solid fraction enriched in cellulose was recovered, which was subjected to enzymatic hydrolysis. Up to 90% of the glucose and 25% of the hemicellulose contained in the original biomass were released by the combined ionic liquid pretreatment and the enzymatic hydrolysis. After the pretreatment, the ionic liquid liquor contained the majority of the lignin and the hemicellulose. The lignin portion was partially precipitated from the liquor upon dilution with water. The amount of hemicellulose monomers in the ionic liquid liquor and their conversion into furfurals was also examined. The performance of ionic liquid–water mixtures containing 1,3-dialkylimidazolium ionic liquids with acetate, methanesulfonate, trifluoromethanesulfonate and chloride anions was investigated. The applicability of the ionic liquid 1-butylimidazolium hydrogensulfate for lignocellulose pretreatment was also examined. It was found that ionic liquid liquors containing methyl sulfate, hydrogen sulfate and methanesulfonate anions were most effective in terms of lignin/cellulose fractionation and enhancement of cellulose digestibility. Downloaded by University of Guelph on 18 June 2012 Published on 27 July 2011 http://pubs.rsc.org | doi:10.1039/C1GC15374A Introduction expected, while requiring less energy and material input for its production.2 Various plant species have been suggested as The rising demand for liquid transportation fuels is placing being suitable dedicated biofuel crops, with factors such as increasing demands on finite oil reserves, raising prices and rapid growth, low fertiliser input, and short harvest cycles. Cur- encouraging the search for oil in more remote locations, often in rently favoured crops include grasses (miscanthus, switchgrass), fragile ecosystems. In addition, the planet’s climate is affected by hardwoods (willow, poplar, eucalyptus) and softwoods (pine, carbon dioxide emitted from the use of fossilised carbon as an fir and spruce). Careful implementation of lignocellulose based energy source. The production of many chemicals and materials technology as a substitute for fossil resources could help reduce is also reliant on fossil fuel resources. man-made carbon dioxide emissions in a sustainable way.3 Lignocellulose, essentially the cell wall material of woody Proposed routes for transforming lignocellulosic biomass into plants, is a porous micro-structured composite mainly consisting useful products are the microbial fermentation of the glucose of cellulose, hemicellulose and lignin. It has been projected and other carbohydrates contained in the biomass and the that lignocellulosic biomass has the potential to be a large- thermo-chemical conversion of the lignocellulose via pyrolysis scale, low-cost and sustainable feedstock for renewable fuels or gasification. For the fermentation route, deconstruction of 1 and chemicals. Compared to starch or vegetable oil substrates the lignocellulosic matrix is necessary before the carbohydrates that are currently used as biofuel and biomaterial feedstocks, can be released. A typical deconstruction sequence producing significantly higher biomass yields per unit area of land are fermentable carbohydrates is: size reduction to chips, a pretreatment that solubilises the hemicellulose and alters/removes lignin,4 followed by detoxification and neutralisation. The aDepartment of Chemistry, Imperial College London, London, UK, SW7 pretreated biomass is subsequently processed using hydrolytic 2AZ. E-mail: [email protected] enzymes (saccharification) to produce sugar monomers. b The Porter Alliance, Imperial College London, London, UK, SW7 2AZ The pretreatment step is responsible for a significant por- cDivision of Biology, Imperial College London, London, UK, SW7 2AZ tion of the energy consumption and cost of the biofuel † Electronic supplementary information (ESI) available. See DOI: 5 10.1039/c1gc15374a production process and improvements are required. A large

This journal is © The Royal Society of Chemistry 2011 Green Chem., 2011, 13, 2489–2499 | 2489 View Online

number of pretreatment options are defined in the literature, Various publications concluded that application of methyl such as dilute acid, concentrated acid, ammonia fibre expan- sulfate containing ionic liquids in lignocellulose pretreatment sion (AFEX), lime and organosolv pretreatment. Different did not enhance cellulose digestibility,17,21,24,27 despite their ability plant groups exhibit distinct tissue structures and varying to dissolve large amounts of lignin. cell wall composition, which leads to variable resistance to Water reduces not only cellulose solubility in ionic liquids,12 deconstruction. but also the effectiveness of ionic liquid pretreatment with 21,28 This paper explores the potential of certain ionic liquids as [C2C1im][MeCO2]. Biomass contains significant quantities pretreatment solvents, in particular their mixtures with water. of water, 2–300% relative to the oven-dried weight. In addition, Ionic liquids are a diverse group of salts that are liquid at ambient ionic liquids are hygroscopic and will absorb significant quan- temperatures or melt at slightly elevated temperatures. In the tities of moisture when exposed to air.29 The drying of ionic last two decades, ionic liquids containing organic cations with liquids requires heat and vacuum, particularly when the ionic

quaternised ammonium, phosphonium and sulfonium cores liquids are strongly hydrogen bond basic, like [C2C1im][MeCO2]. have enjoyed increasing popularity in many ﬁelds of research.6 Therefore, an ionic liquid pretreatment that tolerates moisture Many ionic liquids have negligible vapour pressures under would be beneﬁcial for the overall energy and cost balance of a process-relevant conditions and the nature and combination lignocellulose processing system using ionic liquids. of cation and anion can be tuned to suit a particular appli- An advantage of ionic liquid pretreatment could be the cation. Cellulose and lignocellulose processing are only two recovery of a separate lignin fraction which could be con- out of many recently explored applications for these alternative verted to aromatic, value-added chemicals. Lignin recovery solvents.7 Ionic liquids are polar solvents with varying degrees from ionic liquids has been achieved after treatment of sugar of hydrogen-bonding ability.8 It has been found that the ionic cane bagasse with 1-butyl-3-methylimidazolium alkylbenzene-

liquid needs to contain anions with high hydrogen-bond basicity sulfonate, [C2C1im][ABS], an ionic liquid mixture containing such as chloride, phosphates, phosphonates and carboxylates aromatic sulfonate anions, mainly xylenesulfonate.30 Lignin in order to be able solubilise cellulose.9 The hydrogen-bond recovery has also been observed after pretreatment with

acidity also plays a role. If a hydrogen-bond acidic functionality [C2C1im][MeCO2], when the regeneration solvent was a mixture is incorporated into the ionic liquid structure, it will compete of water and acetone.20,26 for the hydrogen-bond basic site on the anion and reduce This study investigates the inﬂuence of water on the effective- cellulose solubilisation.10,11 Water also decreases the solubility ness of ionic liquid pretreatment. We have devised a notation of cellulose,12 probably for a similar reason. The empirical to indicate the amount of the ionic liquid contained in the Kamlet–Taft solvent descriptors can be used to predict cellulose pretreatment solvent/liquor. This involves a subscript being solubility.13 Cellulose can be reconstituted by adding a protic added to the usual ionic liquid notation indicating the ionic antisolvent, such as water or alcohols, and spun into ﬁbres liquid content in volume percent (vol%), with the remainder

or ﬁlms. A variety of homogenous derivatisations of cellulose being water. An example is [C4C1im][MeSO4]80%, which is a

Downloaded by University of Guelph on 18 June 2012 14 dissolved in ionic liquids can be accomplished. mixture of 80 vol% [C4C1im][MeSO4]and20vol%water. Ionic liquids were initially used in cellulose processing15 Conversions of vol% into weight percent (wt%) and mole Published on 27 July 2011 http://pubs.rsc.org | doi:10.1039/C1GC15374A before their application was extended to lignocellulose pro- percent (mol%) were calculated and are listed in Table 1. When

cessing. The solubility of lignocellulose in ionic liquids has allowing [C4C1im][MeSO4] to equilibrate with the moisture in been reported in various hydrogen-bond basic ionic liquids, the laboratory air a water content of 70,400 ppm or 7.0 wt% suggesting that ionic liquids which are cellulose solvents are also suitable for lignocellulose processing.16,17 Reduced crystallinity Table 1 Ionic liquid concentration in aqueous pretreatment liquors of the cellulose contained in lignocellulose was observed upon precipitation with an antisolvent.18 A correlation between the Volume percent Weight percent Molar percent hydrogen-bond basicity of the anion and the ionic liquid’s Mixture (vol%) (wt%) (mol%) ability to swell and partially dissolve wood chips has been [C C im][MeSO ] 98 98 81 observed.19 The solubility of lignin in ionic liquids also seems 4 1 4 98% [C4C1im][HSO4]95% 95 96 64 17,20 to depend on the anion. It has been shown that Kraft [C4C1im][MeSO4]90% 90 92 44 pulp lignin has a very high solubility in the ionic liquids 1,3- [C4C1im][HSO4]90% 90 92 46 [C4C1im][MeSO4]80% 80 83 26 dimethylimidazolium methyl sulfate, [C1C1im][MeSO4], and 1- 20 [C4C1im][HSO4]80% 80 83 27 butyl-3-methylimidazolium methyl sulfate, [C4C1im][MeSO4]. a [C4C1im][MeSO3]80% 80 82 26 Enhanced glucose release from ionic liquid pretreated wood [C2C1im][MeCO2]80% 80 82 32 a has also been observed, mainly with dialkylimidazolium ionic [C4C1im]Cl80% 80 81 30 liquids containing acetate, chloride and dimethyl phosphate [C4C1im][OTf]80% 80 84 24 [C4C1im][MeSO4]60% 60 65 12 21–23 anions. However, the sugar released by hydrolytic enzymes [C4C1im][HSO4]60% 60 65 12 was often less than 80–90% (which is expected for an effective [C4C1im][MeSO4]40% 40 45 6 pretreatment operation). Recently, the impact of ionic liquid [C4C1im][HSO4]40% 40 45 6 [C C im][MeSO ] 20 23 2 pretreatment on biomass composition has received attention. It 4 1 4 20% [C4C1im][HSO4]20% 20 23 2

was noted that lignin and hemicellulose are partially removed [C4C1im][MeSO4]wet n.a. 93 49 during pretreatment with 1-ethyl-3-methylimidazolium acetate, a These ionic liquids are solid at room temperature. Therefore vol% and 17,24–26 ◦ [C2C1im][MeCO2]. A correlation between lignin removal wt% were calculated using the density at 80 C. and cellulose digestibility was suggested.17,21

2490 | Green Chem., 2011, 13, 2489–2499 This journal is © The Royal Society of Chemistry 2011 View Online

was measured (last entry of Table 1). Although the moisture content of air is variable, the measurement demonstrates the highly hygroscopic nature of this ionic liquid. The aim of this work is to investigate the effect of the composition of the ionic liquid liquor on the pretreatment. Solid recovery, pulp composition, its enzymatic digestibility, the precipitation of a lignin-containing fraction and the production of furfurals in the liquor were investigated. The application of an ionic liquid with a monoalkylated imidazolium cation was also examined. Pretreatment of different feedstocks was carried out to assess their recalcitrance towards pretreatment with ionic liquid–water mixtures.

Fig. 1 Sugar yields obtained from Miscanthus pulp after pretreatment ◦ Results and discussion with [C4C1im][MeSO4]or[C4C1im][HSO4]–water mixtures at 120 C. The [C C im][MeSO ] pretreatment was carried out for 22 h, while Tissue softening of Miscanthus chips 4 1 4 [C4C1im][HSO4] pretreatment lasted 13 h, and the sacchariﬁcation In preliminary experiments, we observed substantial disintegra- 96 h. tion of Miscanthus cross sections immersed in the ionic liquid ment. Similar yields were obtained with mixtures containing 1-butyl-3-methylimidazolium methyl sulfate, [C C im][MeSO ], 4 1 4 40–90 vol% [C C im][MeSO ]. when heated above 80 ◦C. This encouraged us to investigate 4 1 4 the application of this ionic liquid for biomass pretreatment. Water sensitivity of [C4C1im][MeSO4] The use of [C4C1im][MeSO4], dried to a water content below

0.3 wt%, resulted in formation of a degraded biomass-ionic When attempting to recycle [C4C1im][MeSO4], we found that liquid composite that was not enzymatically digestible. In the ionic liquid anion was partially hydrolysed. After recording contrast, using a mixture of 80 vol% ionic liquid and 20 vol% a mass spectrum of the recovered ionic liquid, a high abundance water yielded a biomass fraction that was separable from the of a negatively charged species at m/z = 97 was detected, which - (intensely coloured) ionic liquid fraction and highly digestible. was ascribed to the hydrogen sulfate, [HSO4] , anion. This led to It was concluded that a certain amount of water was necessary the conclusion that the ester bonds in methyl sulfate anions are

for successful pretreatment with [C4C1im][MeSO4]. In the “dry” hydrolytically unstable under the conditions of the pretreatment sample, 0.3 wt% water was contained in the ionic liquid as and mixtures of the ester and the hydrolysed form are produced. residual moisture and 0.7 wt% was introduced with the air-

Downloaded by University of Guelph on 18 June 2012 dried biomass containing 8 wt% moisture, supplying 1.1 wt% or 15 mol% water in total. This was apparently not sufficient to Published on 27 July 2011 http://pubs.rsc.org | doi:10.1039/C1GC15374A obtain an enzymatically digestible pulp. The extent of anion hydrolysis depended upon the water content of the liquor (Fig. 2). The more water present in the Influence of the water content on the saccharification yield after mixture, the greater the anion hydrolysis, with the exception ionic liquid pretreatment with [C C im][MeSO ] 4 1 4 of mixtures where the water content was higher than 90 mol%. A range of ionic liquid water mixtures were used for pretreatment These results suggest that without extreme precautions to protect - - of Miscanthus to explore the effect of the water content in more [MeSO4] containing ionic liquids, [HSO4] will be present and detail. The effect of water on the enzymatic release of glucose and other studies using these ionic liquids should be interpreted in hemicellulose is shown in Fig. 1. The yields are calculated based this light.20 on the glucose and hemicellulose content found in the untreated Miscanthus feedstock (on an oven-dry basis), which were 43.6 wt% and 24.3 wt%, respectively.In preliminary experiments, it was shown that the only detectable hemicellulose sugar released during saccharification was xylose. The best saccharification yields were obtained after pretreatment with mixtures containing 60–90 vol% ionic liquid. Pretreat-

ment with [C4C1im][MeSO4]90%,resultedinthereleaseof92%of the glucose originally contained in the biomass. Pretreatment

with [C4C1im][MeSO4]80% and [C4C1im][MeSO4]60%,resulted in the release of 89% and 87% based on the original glucan content. Glucose yields decreased when the ionic liquid content was higher or lower. The hemicellulose yield was signiﬁcantly lower

- than the glucose yield, regardless of the mixture composition; Fig. 2 Ratioof[MeSO4] anions to ionic liquid cations in the recycled 24% of the hemicellulose sugars (based on the initial hemicellu- ionic liquid after pretreatment of Miscanthus (detected by 1H-NMR), - lose content) were released after [C4C1im][MeSO4]60% pretreat- the remaining anions being [HSO4] .

This journal is © The Royal Society of Chemistry 2011 Green Chem., 2011, 13, 2489–2499 | 2491 View Online

Inﬂuence of the water content on the enzymatic sacchariﬁcation

of [C4C1im][HSO4] treated Miscanthus With the knowledge that the binary 1-butyl-3-methylimidazolium methyl sulfate water mixtures turned into quaternary mixtures of two ionic liquids plus two molecular solvents (water and methanol) we set out to identify the active com- ponent(s). Miscanthus was pretreated with aqueous mixtures

of [C4C1im][HSO4], which allowed us to exclude methyl sulfate and methanol. The sacchariﬁcation yields obtained from the

pulps pretreated with various [C4C1im][HSO4] water mixtures are shown in Fig. 1. The glucose yields were almost identical to the glucose yields obtained with the quaternary mixtures. The Fig. 3 Glucose and hemicellulose yields after enzymatic hydrolysis of pattern of hemicellulose release was also similar, however, after Miscanthus pretreated with [C4C1im][MeSO4]80% and [C4C1im][HSO4]80% ◦ [C4C1im][HSO4]40%-80% pretreatment; although less hemicellulose at 120 C. was recovered than after treatment with the equivalent methyl sulfate containing mixtures. pretreatment, the glucose yield slightly increased to above 85%, A glucose recovery of 90% after ionic liquid pretreatment is but the hemicellulose yield decreased to just over 20%. This a substantial improvement compared with the saccharification experiment shows that the presence or absence of methyl sulfate yields reported after pretreatment with other ionic liquids. It in the pretreatment mixture does not significantly influence the has been reported that 74% glucose was enzymatically released speed of the pretreatment. It is anticipated that the pretreatment

from ground maple wood after [C4C1im][MeCO2] treatment time can be shortened by the application of higher temperatures, at 90 ◦C for 24 h.21 70% glucose was released from maple but it must be balanced with the ionic liquid stability and ◦ 30 wood after [C2C1im][MeCO2] treatment at 90 C for 24 h. potential side reactions. Li et al. reported only 15% glucose release from ground eucalyptus, pretreated with 1-allyl-3-methylimidazolum chloride, The effect of [C4C1im][MeSO4]80% and [C4C1im][HSO4]80% ◦ 22 [C C2C1im]Cl, at 120 Cfor5h, while 55% of the glucose pretreatment on biomass composition was released after 1-ethyl-3-methylimidazolium diethyl phos- The composition of untreated Miscanthus and pretreated pulp phate, [C C im][Et PO ], pretreatment of ground wheat straw at 2 1 2 4 is shown in Table 2 and Fig. 4. The untreated biomass 130 ◦C for 30 min.23 It should be noted that saccharification contained 43.6% glucose, 24.3% hemicellulose and 26.5% lignin. yields obtained from ball-milled lignocellulose samples were After pretreatment with [C C im][MeSO ] for 2 h, the main not considered for this listing because fine milling can have 4 1 4 80% effect was a reduction of the lignin content. Treatment with a considerable effect on cellulose digestibility.22 The use of Downloaded by University of Guelph on 18 June 2012 [C C im][HSO ] for 2 h resulted in the removal of lignin and ground material reduces the economic viability,31 but using fine 4 1 4 80% hemicellulose. After an extended pretreatment for 22 h with Published on 27 July 2011 http://pubs.rsc.org | doi:10.1039/C1GC15374A powders obtained by ball-milling is of very little relevance for an [C C im][HSO ] , most of the lignin and hemicellulose was industrial process. Studies using 3,5-dinitrosalicylic acid (DNS) 4 1 4 80% solubilised and the glucan content increased from 44% in the for the determination of glucose yield were also not considered. untreated biomass to 85% in the pretreated biomass. 91% of The test is not specific for glucose and therefore glucose yields the original glucan was still present in the pulp. The biomass from lignocellulose are often overestimated. recovery after 22 h was less than 46%, showing that more than half of the wood had been solubilised in the ionic liquid. Tan Effect of pretreatment time on the enzymatic saccharification et al. reported a mass recovery between 46% and 55% after pretreatment with [C C im][ABS] at 170–190 ◦C, indicating that Next, we were interested in the optimisation of the pretreat- 2 1 this ionic liquid mixture might be capable of similar biomass ment time. Fig. 3 shows the saccharification yields for both fractionation.30 The simultaneous removal of lignin and hemi- [C C im][MeSO ] and [C C im][HSO ] pretreatment after 4 1 4 80% 4 1 4 80% cellulose has also been reported for [C C im][MeCO ],24,25 albeit various lengths of time. It can be seen that the enhancement 2 1 2 less complete than seen in this study with [C C im][HSO ] . of the cellulose and hemicellulose digestibility mainly occurred 4 1 4 80% within the first 4 h. This was also the period when the mass loss Production of solubilised sugars and furfurals increased significantly (data shown in ESI†). The pretreatment was practically complete after 8 h, achieving around 80% As seen above, the hemicellulose was removed from the biomass

glucose and 30% hemicellulose release. When prolonging the during treatment with [C4C1im][HSO4]and[C4C1im][MeSO4]

◦ Table 2 Composition of untreated Miscanthus and Miscanthus pretreated with [C4C1im][MeSO4]80% and [C4C1im][HSO4]80% at 120 C

Glu Xyl Ara Man Gal Lignin Ash Extracts Mass loss

Untreated 43.6 18.3 3.4 1.1 2.4 26.5 1.3 4.7 0

[MeSO4]2h 45.4 18.3 2.1 1.3 2.2 19.3 1.1 — 10 [HSO4]2h 44.5 8.6 0 0 3.4 14.9 0.6 — 28 [HSO4]22h 39.5 3.3 0 0 1.1 1.9 0.6 — 56

2492 | Green Chem., 2011, 13, 2489–2499 This journal is © The Royal Society of Chemistry 2011 View Online

Fig. 4 Composition of Miscanthus before and after pretreatment with ◦ [C4C1im][HSO4]80% and [C4C1im][MeSO4]80% at 120 C for 2 h or 22 h. Fig. 6 Solubilised carbohydrates (monomers only) and the fraction converted to furfurals after pretreatment with [C4C1im][HSO4]80% and

[C4C1im][MeSO4]80% liquors. water mixtures. It is likely that under the conditions of the pretreatment, (partial) hydrolysis of solubilised hemicellulose Lignin recovery occurred. Therefore the concentration of monomeric carbohydrates in the pretreatment liquor was investigated. Fig. 5 We attempted to recover lignin from the liquor (Fig. 7), as this shows the relative amount of hemicellulose sugars and glucose in has been successfully demonstrated for other ionic liquids.26,30

[C4C1im][HSO4]80% and [C4C1im][MeSO4]80% liquors at different It was found that diluting the ionic liquid liquor with water time points. precipitated a ﬁne powder. The powder was characterised by IR spectroscopy and elemental analysis. Comparison with a reference lignin (alkaline lignin from Aldrich) showed that the precipitate is likely to be mostly lignin (see ESI†). When methanol was used for washing the pulp, instead of water, the majority of the precipitate remained in solution and a 15–20% improvement of precipitate recovery was observed. Therefore washing the pulp with methanol was preferred. The ﬁnal protocol consisted of washing the pulp with methanol, drying the combined ionic liquid fractions by evaporating the methanol, and precipitating the lignin by diluting the dried

Downloaded by University of Guelph on 18 June 2012 ionic liquid liquor with water. The precipitate was washed with copious amounts of water and dried before the yield was Published on 27 July 2011 http://pubs.rsc.org | doi:10.1039/C1GC15374A determined. The data (Fig. 7) show that the yield of precipitate was up to 50% of the Klason lignin content of the untreated Fig. 5 Amount of glucose and hemicellulose monomers found in biomass. More precipitate was obtained when the ionic liquid

[C4C1im][HSO4]80% and [C4C1im][MeSO4]80% liquors during pretreatment content in the pretreatment liquor was high. at 120 ◦C.

The amount of hemicellulose monomers in the liquor increased within the ﬁrst 4 h. The increase was more pronounced

in the [C4C1im][HSO4]80% liquor. The maximum amount of hemicellulose monomers was detected around 4–8 h. This coincided with a major increase of cellulose digestibility after 4–8 h of treatment (Fig. 3). Subsequently, the hemicellulose concentration in the pretreatment liquor decreased, suggesting that conversion of carbohydrate monomers into furfurals was occurring. Furfural was detected in the ionic liquid liquors and quanti- Fig. 7 Yield of precipitate (relative to Klason lignin content ﬁed for selected mixtures (Fig. 6). of the untreated biomass) after pretreatment of Miscanthus with The glucose content was signiﬁcantly lower than the hemicel- ◦ [C4C1im][HSO4]–water mixtures at 120 C for 13 h. lulose sugar content and hardly changed over time. The smaller amount of solubilised glucose is ascribed to the slow hydrolysis of cellulose under the conditions of the pretreatment and the We also examined the time dependency of the precipitate yield decomposition of glucose to HMF. The small amount of HMF and observed that the yield of precipitate plateaued within 8 h

might be due to its decomposition to other degradation products (Fig. 8). The yield was slightly higher from [C4C1im][HSO4]80%

in the presence of water. compared to [C4C1im][MeSO4]80%.

This journal is © The Royal Society of Chemistry 2011 Green Chem., 2011, 13, 2489–2499 | 2493 View Online

Fig. 8 Time course of lignin recovery after pretreatment of Miscanthus ◦ with [C4C1im][MeSO4]80% and [C4C1im][HSO4]80% at 120 C. The lignin Fig. 10 Composition of Miscanthus after pretreatment with [C4Him]- was isolated from the liquor by precipitation with water. ◦ [HSO4]–water mixtures at 120 C. The effect of the ionic liquid cation hemicellulose is reflected by the low xylose yields obtained The use of ionic liquids with mono-alkylated imidazolium during saccharification. Treatment with [C4Him][HSO4]95% not + cations (1-alkylimidazolium, [CnHim] ) is advantageous from only resulted in the solubilisation of lignin and hemicellulose, an industrial point of view, as the ionic liquids are easier to but also in a substantial removal of the cellulose fraction 7 synthesise and thus cheaper to produce. Therefore an exemplary (51% of the glucan), explaining the reduced glucose yield pretreatment of Miscanthus with 1-butylimidazolium hydrogen shown in Fig. 9. The results indicate that pretreatment with sulfate, [C4Him][HSO4], was carried out. The sugar yields after [C4Him][HSO4] was harsher than with [C4C1im][HSO4] under treatment with [C4Him][HSO4]80% and a subsequent enzymatic comparable conditions, potentially due to the increased acidity saccharification are shown in Fig. 9. After 4 h pretreatment, 69% of the [C4Him][HSO4] ionic liquid used in this study compared of the original glucose and 10% of the original hemicellulose to [C4C1im][HSO4]. were enzymatically released. The yield was somewhat improved It was also possible to obtain a precipitate upon dilution of the by prolonging the treatment to 20 h, when 75% of the glucose was ionic liquid liquor (Fig. 11). For the [C4Him][HSO4]80% liquor,the recovered. However, the xylose yield was reduced to only 3%. yield was nearly 100% of the lignin content. For the 95% liquor, Pretreatment with [HC4im][HSO4]95% resulted in significantly the amount of precipitate was almost double the amount of the reduced glucose yields (44%). lignin content. We explain the unusually high precipitate yield with the formation of pseudo-lignin. The formation of water- insoluble carbohydrate degradation products has been observed

Downloaded by University of Guelph on 18 June 2012 during biomass pretreatment under severe acidic conditions32,33 as well as high temperature ionic liquid treatment30 and was Published on 27 July 2011 http://pubs.rsc.org | doi:10.1039/C1GC15374A found to obscure the Klason lignin yield. Therefore it has been termed pseudo-lignin. The formation of such degradation

Fig. 9 Enzymatic sacchariﬁcation yields obtained from Miscanthus

after pretreatment with [C4Him][HSO4]95% and [C4Him][HSO4]80%.Sac- chariﬁcation was carried out for 96 h.

The results of the compositional analysis and the mass loss

of [C4Him][HSO4] treated Miscanthus are presented in Table 3 and Fig. 10. 80–93% of the lignin and more than 95% Fig. 11 Lignin removal and precipitate yield after pretreatment of ◦ of the hemicellulose were removed. The thorough removal of Miscanthus with [C4Him][HSO4]–water mixtures at 120 C.

◦ Table 3 Composition of Miscanthus pretreated with [C4Him][HSO4]80% and [C4Him][HSO4]95% at 120 C. Values are given in %; Glu = glucan, Xyl = xylan, Man = mannan, Gal = galactan, Ara = arabinan

IL content, treatment time Glu Xyl Man Gal Ara Lignin Ash Mass loss

80%, 4 h 40.9 2.9 0 0.7 0.2 5.0 0.8 49.5 80%, 20 h 37.7 1.0 0 1.0 0 5.4 0.6 54.2 95%, 20 h 22.4 0.6 0 0.6 0 1.9 0.4 74.2

2494 | Green Chem., 2011, 13, 2489–2499 This journal is © The Royal Society of Chemistry 2011 View Online

products is undesirable and optimisation of the pretreatment conditions is required to minimise this.

The effect of the ionic liquid anion on the composition of ionic liquid treated Miscanthus

The effect of treatment with [C4C1im][HSO4]80% on the composition of Miscanthus was compared with the effect that other 20/80 vol% dialkylimidazolium ionic liquid–water mixtures have on the composition. The anions that we examined were triﬂuoromethanesulfonate, [OTf]-, methanesulfonate, - - - [MeSO3] , chloride, Cl , and acetate, [MeCO2] . It should be

noted that the acetate containing ionic liquid, [C2C1im][MeCO2], was of commercial quality. Fig. 13 Impact of the ionic liquid anion on glucose and hemicellulose yields after enzymatic saccharification of Miscanthus pulp pretreated Fig. 12 and Table 4 show that the nature of the anion with 80/20 vol% ionic liquid–water mixtures at 120 ◦C for 22 h. has a profound effect on mass loss and pulp composition. [C C im][HSO ] removed lignin and hemicellulose most 4 1 4 80% [C C im][MeSO ] and [C C im][HSO ] pretreatment. The thoroughly, followed by [C C im][MeSO ] andthenby 4 1 3 80% 4 1 4 80% 4 1 3 80% hemicellulose yield behaved slightly differently. The xylose yield [C C im][MeCO ] . Hardly any change of the composition was 2 1 2 80% was the highest after pretreatment with [C C im][MeCO ] . observed when the biomass was treated with [C C im]Cl and 2 1 2 80% 4 1 80% The yield was significantly lower after [C C im][MeSO ] and [C C im][OTf] , despite the fact that high solubility of Kraft 4 1 3 80% 4 1 80% [C C im][HSO ] pretreatment. lignin has been reported for both ionic liquids (in anhydrous 4 1 4 80% Comparatively high hemicellulose yields after form).17,20 The contradiction could be resolved if lignin solubil- [C C im][MeCO ] treatment can also be found in the literature.21 isation and lignin extraction (which usually involves chemical 4 1 2 The increased hemicellulose recovery after [C C im][MeCO ] modifications) were regarded as different properties. 2 1 2 80% treatment could be due to a buffering effect exerted by the basic acetate anion. Its ability to combine with protons to form acetic acid may limit the acid-catalysed hydrolysis of hemicellulose polymers. Inhibition of the hydrolysis of cellobiose by

[C4C1im][MeCO2] has been observed in mixtures of the ionic liquid, water and catalytic amounts of strong acid.34 Binder et al., have also observed inhibition of cellulose depolymerisation

in [C4C1im][MeCO2], despite addition of catalytic amounts

Downloaded by University of Guelph on 18 June 2012 of HCl.35 The methanesulfonate anion appears to have a less protective effect and acid-catalysts which are released from the Published on 27 July 2011 http://pubs.rsc.org | doi:10.1039/C1GC15374A biomass (acetic acid and hydroxycinnamic acids) can aid xylan hydrolysis. Hydrogensulfate increases the amount of available Fig. 12 Effect of the ionic liquid anion on the mass loss and the protons, which could explain the particularly low xylan content composition of the recovered pulp after pretreatment of Miscanthus in the pulp. The glucose and xylose yields obtained after ◦ with 80% ionic liquid–water mixtures at 120 C for 22 h. The data are treatment with [C4C1im]Cl80% and [C4C1im][OTf]80% were low, ordered (left to right) according to the hydrogen-bond basicity of the despite their ability to dissolve cellulose and lignin preparations ionic liquid, which is, in the case of 1,3-dialkylimidazolium ionic liquids, (in the case of triﬂate, only lignin solubility). a property of the anion.

The effect of the anion on the sacchariﬁcation yield The effect of the anion on deligniﬁcation and precipitate recovery

Enzymatic sacchariﬁcation of Miscanthus treated with the ionic The yield of precipitate seems to be related to the ability of the liquid liquors was also carried out (Fig. 13). In general, the liquor to extract lignin (Fig. 14). The best deligniﬁcation and the

enzymatic glucose release appeared to reﬂect the extent of highest precipitate yield was obtained with [C4C1im][HSO4]80%,

compositional change/mass loss achieved during ionic liquid followed by [C4C1im][MeSO3]80% and then [C2C1im][MeCO2]80%. pretreatment. The highest glucose yield was observed after This supports the notion that the precipitate comprises lignin,

Table 4 Composition of pretreated Miscanthus after treatment with 80/20% ionic liquid–water mixtures at 120 ◦C for 22 h. Values are given in %; Glu = glucan, Xyl = xylan, Man = mannan, Gal = galactan, Ara = arabinan

Ionic liquid anion Glu Xyl Man Gal Ara Lignin Ash Mass loss

- [MeCO2] 41.9 7.9 0 4.0 3.4 11.6 0.5 30.6 Cl- 44.5 17.8 0 2.3 2.7 22.5 0.7 9.5 - [MeSO3] 37.1 4.3 0 2.3 0 8.5 1.0 46.8 - [HSO4] 39.5 3.3 0 1.1 0 1.9 0.6 53.6 [OTf]- 43.6 13.7 0 5.1 4 24.3 1.0 8.3

This journal is © The Royal Society of Chemistry 2011 Green Chem., 2011, 13, 2489–2499 | 2495 View Online

It is likely that the acidity of the liquor is responsible for the varying concentrations of sugar monomers and furfural found in the liquor. Like the hydrolysis of glycosidic bonds, the rate of furfural formation depends on the acid concentration.36 Since the acidity/basicity of 1,3-dialkylimidazolium ionic liquids is determined by the anion, its nature should have a profound impact on the fate of the solubilised hemicellulose. The amounts of solubilised glucose and HMF were small in all cases. This is ascribed to the enhanced stability of the cellulose fraction towards hydrolysis under pretreatment conditions and the Fig. 14 Effect of the anion on the lignin removal and precipitate yield propensity of HMF to react with formic and levulinic acid in after pretreatment of Miscanthus with 80/20 vol% ionic liquid–water - thepresenceofwater. mixtures. The higher yield from [HSO4] containing liquors (compared to Fig. 7 and Fig. 8) is ascribed to the larger quantity of ionic liquid and biomass used in this experiment. Values are relative to the lignin content The effect of the biomass type: pretreatment of willow and pine of the untreated biomass. Pretreatment with [C4C1im][HSO4]80% was also performed on although it was shown in Fig. 11 that pseudo-lignin also ground willow (a hardwood species) and pine (a softwood precipitates upon dilution of the ionic liquid liquor. species). For comparison, willow and pine were also pretreated with [C2C1im][MeCO2]80%. The effect of the pretreatment on the biomass composition is shown in Table 5 and Fig. 16. The effect of the anion on the formation of soluble degradation products

The quantities of carbohydrate monomers and dehydration products solubilised in the pretreatment liquors are shown in

Fig. 15. The [C4C1im][HSO4]80% and [C4C1im][MeSO3]80% liquors contained approximately 45% of the total hemicellulose as either

sugar monomers or furfural. In [C4C1im][HSO4]80%, the largest fraction was furfural. Conversion of pentoses into furfurals was

also observed in [C4C1im][MeSO3]80%, but to a lesser extent. This is ascribed to the non-acidic nature of this ionic liquid. Only small quantities of monomers were detected in the acetate containing liquor, which is probably due to the fact that the solubilised carbohydrates are mostly in oligomeric form. No Fig. 16 Composition of willow (3 bar graphs on the left) and pine (on Downloaded by University of Guelph on 18 June 2012 the right) before and after pretreatment with [C4C1im][HSO4]80% and furfural was formed in [C2C1im][MeCO2]80% in our experiment. ◦ [C4C1im][MeCO2]80% for22hat120 C. Published on 27 July 2011 http://pubs.rsc.org | doi:10.1039/C1GC15374A

For both substrates, lignin and hemicellulose removal were

more extensive after [C4C1im][HSO4]80% pretreatment than after

treatment with [C2C1im][MeCO2]80%. The degree of cellulose

enrichment after [C4C1im][HSO4]80% pretreatment of willow was almost as good as the enrichment observed for Miscanthus pulp. A precipitate could be recovered from all samples. Signiﬁcantly

higher yields were obtained from the [C4C1im][HSO4]80% liquors (see ESI†). The glucose yields obtained via enzymatic sacchar- iﬁcation are shown in Fig. 17. More than 80% of the original

glucose was released from [C4C1im][HSO4]80% pretreated willow pulp, approaching the sacchariﬁcation yields obtained from Fig. 15 Sugar monomers and furfurals solubilised in liquors containing Miscanthus pretreated with this liquor. However, enzymatic sac- 80 vol% 1,3-dialkylimidazolium ionic liquids with various anions after chariﬁcation of pine pulp only released up to 30% of the glucose; treatment of Miscanthus at 120 ◦C for 22 h. the type of ionic liquid playing a minor role. The generally higher

Table 5 Composition of untreated willow and pine and the pulps after treatment with [C4C1im][HSO4]80% and [C4C1im][MeCO2]80%

Glu Xyl Man Gal Ara Lignin Ash Extractives Mass loss

Willow 46.7 16.8 3.6 1.9 2.5 24.1 0.7 3.7 0

Willow, [MeCO2] 36.3 6.4 2.9 2.7 1.9 19.9 0.7 — 29

Willow, [HSO4] 39.1 3.4 0 0.8 0.9 3.6 0.5 — 52 Pine 45.8 2.5 12.0 2.6 3.4 25.5 1.3 4.3 0

Pine, [MeCO2] 40.4 2.5 16.1 3.4 2.7 21.1 0.6 — 13 Pine, [HSO4] 37.9 3.2 4.6 0 0 8.8 0.2 — 45

mixture [C4C1im][MeSO4]/[HSO4] can be used to pretreat lignocellulosic biomass. These ionic liquids functioned effectively in the presence of signiﬁcant quantities of water, eliminat- ing the need for anhydrous conditions during pretreatment.

Commercial [C2C1im][MeO2] was also effective in the presence of 20 vol% water, but the saccharification yield was lower. Lignin and hemicellulose were solubilised during pretreatment, leaving behind a solid residue that was highly enriched in cellulose. The enzymatic saccharification of Miscanthus pulp pretreated at 120 ◦C with liquors containing 80 vol% ionic liquid resulted in glucose yields of ca. 90%. The hemicellulose was partially recovered with the solid and readily hydrolysable Fig. 17 Enzymatic saccharification of lignocellulosic feedstocks after during enzymatic saccharification. However, a significant por- pretreatment with [C4C1im][HSO4]80% or [C2C1im][MeCO2]80% for 22 h at tion of the hemicellulose remained in the pretreatment liquor 120 ◦C. as sugar monomers and was partially converted to dehydration products. The amount of furfurals generated during ionic liquid yields obtained after [C4C1im][HSO4]80% pretreatment could be due to the improved lignin and hemicellulose removal by the pretreatment arises from the acidity of the ionic liquid liquors. hydrogen sulfate containing liquor, as observed for Miscanthus. In the presence of 20 vol% water, treatment with [C4C1im]Cl and [C4C1im][OTf] had little effect on the biomass, showing Ionic liquid solvent properties and biomass digestibility that the anion of 1,3-dialkylimidazolium ionic liquids plays an important role in determining the effectiveness of ionic liquid

We measured the Kamlet–Taft polarity of [C4C1im][HSO4]and pretreatment and the tolerance towards water. We could not

[C4C1im][MeSO3] (Table 6), as it has not been reported in the ﬁnd a correlation between the pretreatment effectiveness and literature. Three parameters are used to determine the strength the anion basicity, as previously found for cellulose solubility or of solvent solute interactions. The parameter a describes the wood chips swelling. While the enzymatic sugar release from hydrogen-bond acidity of the solvent, b the hydrogen-bond ba- the grass and hardwood pulps was very good, yields from sicity and p* the polarisability. Our measurements showed that softwood pulp were only moderate. Upon dilution with water, a

the b parameter of [C4C1im][HSO4] is the same as the value for precipitate was recovered that is likely to contain lignin as well

[C4C1im][MeSO4]. The hydrogen-bond acidity is very different, as pseudo-lignin. This study also suggests that mono-alkylated

in fact, the a value cannot be determined for [C4C1im][HSO4], imidazolium ionic liquids, such as [C4Him][HSO4], appear to because it protonates one of the dye probes. be promising, industrially relevant alternatives to dialkylimida- We would like to point out that the high glucose yields zolium ionic liquids. The next challenge is to optimise these

Downloaded by University of Guelph on 18 June 2012 were achieved without complete solubilisation of the biomass. processes.

This is due to the relatively low b values of [C4C1im][MeSO4], Published on 27 July 2011 http://pubs.rsc.org | doi:10.1039/C1GC15374A [C4C1im][HSO4]and[C4C1im][MeSO3], which do not enable cellulose solubilisation. The b parameters are lower than the values

of [C4C1im][MeCO2](b = 1.20), 1-butyl-3-methylimidazolium Experimental

dimethyl phosphate, [C4C1im][Me2PO4], (b = 1.12) and 19 Materials [C4C1im]Cl (b = 0.83). Although [C2C1im][MeCO2]candis- solve cellulose when it is anhydrous, the presence of 20 vol% The lignocellulosic feedstocks used in this study were pine water prevents cellulose solubility. sapwood (Pinus sylvestris, variety SCOES) from East Sussex, de- We also attempted to correlate the glucose yields with barked mixed willow (Salix viminalis, variety TORA) stems and the ionic liquids’ hydrogen-bond basicity. While it is clear Miscanthus giganteus whole stems. The biomass was air-dried, that the nature of the anion affects the sacchariﬁcation ground and sieved (0.18–0.85 mm, -20+80ofUSmeshscale) yield, it could not be correlated with the ionic liquid’s b before use. The moisture content of untreated lignocellulose value. was 8.0% (Miscanthus), 8.9% (pine) and 7.6% (willow) based on oven-dry weight. The biomass was stored in plastic bags at Conclusions room temperature. 1-Butyl-3-methylimidazolium methyl sulfate (Basionic AC01) It has been demonstrated for the ﬁrst time that the ionic was purchased from Sigma-Aldrich. 1-Ethyl-3-methylimida-

liquids [C4C1im][HSO4], [C4C1im][MeSO3] and the ionic liquid zolium acetate (Basionic BC01) was a gift from BASF AG, Lud- wigshafen. 1-Butyl-3-methylimidazolium chloride and 1-butyl- Table 6 Kamlet–Taft parameters of selected ionic liquids used in this 3-methylimidazolium triﬂuoromethanesulfonate were synthe- work sised as described previously.19 The syntheses of 1-butyl-3- abp* methylimidazolium hydrogen sulfate, 1-butylimidazolium hydrogen sulfate and 1-butyl-3-methylimidazolium methanesul- [C4C1im][MeSO3] 0.44 0.77 1.02 fonate are described in the ESI.† The ionic liquids were dried to a 19 [C4C1im][MeSO4] 0.55 0.67 1.05 water content of <0.3 wt%, with the exception of [HC im][HSO ] [C C im][HSO ] — 0.67 1.09 4 4 4 1 4 which had a water content of 1 wt%.

Measurement of Kamlet–Taft parameters Determination of moisture content Kamlet–Taft parameters were measured as reported pre- 100–200 mg air-dried biomass was wrapped in aluminium foil viously.19 of known weight and heated at 105 ◦C overnight. The samples were transferred into a desiccator and the weight determined after 5 min. The moisture content was calculated according to Water sensitivity of [C C im][MeSO ] 4 1 4 eqn (2) and used to determine the oven-dried weight of untreated The ionic liquid liquor obtained after lignin precipitation was and pretreated biomass. The moisture content of the air-dried ◦ dried under vacuum at 40 C. A sample of the dried ionic liquid biomass was in the range 5–12%. was submitted to mass spectrometry. Part of the recovered ionic − 1 mmair dried oven dried liquid was dissolved in DMSO-d6 and a H-NMR spectrum moisture (%)= ⋅100 % (2) recorded. The peaks of the methyl group at 3.40 ppm and of the moven dried C-2 ring hydrogen were used to determine the anion to cation ratio. The pretreatment was carried out in capped vessels, so it is Enzymatic saccharification reasonable to assume that the water content did not change Enzymatic saccharification was performed according to labo- substantially during the pretreatment. The water introduced ratory analytical procedure (LAP) “Enzymatic saccharification by the ionic liquid and the air-dried biomass was taken into of lignocellulosic biomass” (NREL/TP-510-42629). 150 mg un- account, but not water consumed in hydrolytic reactions. treated or pretreated air-dried lignocellulosic biomass was used per saccharification experiment. If a pretreatment condition was Lignocellulose pretreatment and isolation of pulp run in duplicate or triplicate, one saccharification experiment was performed per sample. If the pretreatment condition was 0.500 g biomass (on oven-dry weight basis) was placed in wide- not replicated, the saccharification was performed in duplicate. mouthed Pyrex culture tubes with screw cap and Teflon-lining. The enzyme preparations used for the saccharification were 5 ml dried ionic liquid or the equivalent volume of ionic liquid Celluclast (cellulase mix from T. reseei) and Novozyme 188 plus water were added. Mixing effects were neglected. The ◦ b-glucosidase which also contain hemicellulolytic activity and samples were incubated without stirring in an oven at 120 C. can therefore hydrolyse xylan (both from Sigma). Glucose and After the pretreatment was finished, the samples were cooled to hemicellulose yields were calculated based on the glucose and room temperature and mixed with 10 ml methanol. After 2 h, the hemicellulose content of the untreated biomass, respectively. suspension was filtered through hardened cellulose filter papers (Whatman 541 or equivalent). The supernatant was set aside for Compositional analysis precipitation and analysis of the monomer and furfural content. The pulp was washed with methanol from a wash bottle, placed The compositional analysis (lignin, carbohydrates, ash) was in a vial and incubated with 10 ml fresh methanol overnight. performed according to LAP “Determination of structural Downloaded by University of Guelph on 18 June 2012 The suspension was filtered again, rinsed with methanol and carbohydrates and lignin in biomass” (NREL/TP-510-42618),

Published on 27 July 2011 http://pubs.rsc.org | doi:10.1039/C1GC15374A air-dried on the filter paper overnight. The air-dried weight which can be downloaded from http://www.nrel.gov. 300 mg was recorded and the samples transferred into re-sealable of oven-dried sample were used per experiment. Extractives air-tight sample bags. In order to obtain enough material from untreated pine and willow flour were removed by a one- for compositional analysis the pretreatment experiments were step automated solvent extraction with 95% ethanol using scaled up 2–3 times. the ASE 300 accelerated solvent extractor (Dionex) according to the LAP “Determination of extractives” (NREL/TP-510- Lignin recovery 42619). Extractives from untreated Miscanthus were removed by a two-step solvent extraction using deionised water and The supernatant obtained after pretreatment was dried under subsequently 95% ethanol according to the same LAP. HPLC ◦ mild vacuum at 40 C to remove the organic wash solvent. 10 ml analysis of glucose and hemicellulose sugars was performed on water was added per 5 ml of original liquor. The suspension was an Agilent 1200 system equipped with an Aminex HPX-87P centrifuged and the precipitate washed with 3 ¥ 10 ml distilled column, a de-ashing column and a Carbo-P guard column (all water, air-dried for several days and dried under vacuum at room Biorad). The mobile phase was de-ionised water, the column temperature. The precipitate yield was calculated based on the temperature 80 ◦Candtheflowrate0.6mlmin-1. The content Klason lignin content of untreated biomass using eqn (1). Part of carbohydrates, Klason lignin, ash and extractives (where of the precipitate may be pseudo-lignin. applicable) was expressed as a fraction of the sum (normalised to 100%). m Lignin yield (%)=⋅precipitate 100 % (1) Quantification of solubilised sugars and furfurals mKlason lignin 200 ml pretreatment liquor was mixed with 600 ml deionised The precipitate was characterised by IR spectroscopy using water in 1.5 ml plastic cup, vortexed and centrifuged with a a Spectrum 100 IR machine (Perkin-Elmer) equipped with an table-top centrifuge (Biofuge 13, Heraeus) at maximum speed universal ATR sampling accessory with diamond crystal and by for 10 min. The supernatant was transferred into a clean cup elemental analysis, which was performed in duplicate by Medac and centrifuged for 10 min. The supernatant was transferred Ltd, Chobham, Surrey. into HPLC sample vials and analysed on a Jasco HPLC system

equipped with an Aminex HPX-87H column (Biorad) using a 8 L. Crowhurst, P. R. Mawdsley, J. M. Perez-Arlandis, P. A. Salter and 10 mM sulfuric acid mobile phase. The column oven temperature T. Welton, Phys. Chem. Chem. Phys., 2003, 5, 2790–2794. ◦ -1 9 A. Pinkert, K. N. Marsh, S. Pang and M. P. Staiger, Chem. Rev., was 55 C, the ﬂow rate 0.6 ml min and the acquisition 2009, 109, 6712–6728. time 55 min. 2-Furaldehyde (furfural) and 5-(hydroxymethyl)-2- 10 H. Zhao, G. A. Baker, Z. Y. Song, O. Olubajo, T. Crittle and D. furaldehyde (HMF) standards were prepared in deionised water Peters, Green Chem., 2008, 10, 696–705. to concentrations of 0.01, 0.02, 0.1, 0.2 and 0.4 mg ml-1.The 11 A. Pinkert, K. N. Marsh and S. Pang, Ind. Eng. Chem. Res., 2010, -1 49, 11809–11813. standards for carbohydrates were 0.1, 1, 2 and 4 mg ml .The 12 M. Mazza, D. A. Catana, C. Vaca-Garcia and C. Cecutti, Cellulose, factor fHPLC(S) was obtained from the respective calibration 2009, 16, 207–215. curve. The relative yield of solubilised sugar monomers and 13 A. R. Xu, J.J.Wang and H. Y.Wang, Green Chem., 2010, 12, 268–275. furfurals, wt%(S), was calculated using eqn (3). The molecular 14 T. Liebert and T. Heinze, BioResources, 2008, 3, 576–601. 15 R. P. Swatloski, S. K. Spear, J. D. Holbrey and R. D. Rogers, J. Am. mass transformation factor FT was 1.37 for furfural, 1.28 for Chem. Soc., 2002, 124, 4974–4975. HMF, 0.91 for glucose and 0.88 for hemicellulose sugars. The 16 D. A. Fort, R. C. Remsing, R. P. Swatloski, P. Moyna, G. Moyna mass fraction factor F was 0.243 for hemicellulose sugars and andR.D.Rogers,Green Chem., 2007, 9, 63–69. C 17 S. H. Lee, T. V.Doherty, R. J. Linhardt and J. S. Dordick, Biotechnol. furfural and 0.436 for glucose and HMF. Bioeng., 2009, 102, 1368–1376. AFVFS×× ×() 18 I. Kilpelainen,¨ H. Xie, A. King, M. Granstrom, S. Heikkinen and D. wt%( S )= HPLC D PL T ×100% S. Argyropoulos, J. Agric. Food Chem., 2007, 55, 9142–9148. ×× (3) FSmHPLC() biomass F C 19 A. Brandt, J. P. Hallett, D. J. Leak, R. J. Murphy and T. Welton, Green Chem., 2010, 12, 672–679.

AHPLC:areaofHPLCpeak;FHPLC(S): HPLC calibration factor 20 Y. Q. Pu, N. Jiang and A. J. Ragauskas, J. Wood Chem. Technol., for substance S; F : dilution factor; V : volume of pretreatment 2007, 27, 23–33. D PL 21 T. V. Doherty, M. Mora-Pale, S. E. Foley, R. J. Linhardt and J. S. liquor in ml; mbiomass: biomass (oven-dried weight) in mg; FC: Dordick, Green Chem., 2010, 12, 1967–1975. fraction of glucan or hemicellulose sugars in untreated biomass 22 B. Li, J. Asikkala, I. Filpponen and D. S. Argyropoulos, Ind. Eng. Chem. Res., 2010, 49, 2477–2484. as determined by compositional analysis; FT(S): transformation 23 Q. Li, Y.-C. He, M. Xian, G. Jun, X. Xu, J.-M. Yang and L.-Z. Li, factor accounting for molecular mass differences between start- Bioresour. Technol., 2009, 100, 3570–3575. ing material and product. 24 R. Arora, C. Manisseri, C. Li, M. Ong, H. Scheller, K. Vogel, B. Simmons and S. Singh, BioEnergy Res., 2010, 3, 134–145. 25 C. Li, B. Knierim, C. Manisseri, R. Arora, H. V. Scheller, M. Auer, Acknowledgements K. P. Vogel, B. A. Simmons and S. Singh, Bioresour. Technol., 2010, 101, 4900–4906. We gratefully acknowledge the help of Julian Gianuzzi with 26 N. Sun, M. Rahman, Y.Qin, M. L. Maxim, H. Rodriguez and R. D. pretreatment and sacchariﬁcation experiments and funding of Rogers, Green Chem., 2009, 11, 646–655. the studentship for Agnieszka Brandt by the Porter Institute. 27Q.Li,X.Jiang,Y.He,L.Li,M.XianandJ.Yang,Appl. Microbiol. Biotechnol., 2010, 87, 117–126. 28 S. Varanasi, C.A. Schall, A.P. Dadi, J. Anderson, K. Rao, R. P. References Paripati, G. Kumar, WO Pat., 112291 A2, 2008. Downloaded by University of Guelph on 18 June 2012 29 L. Cammarata, S. G. Kazarian, P. A. Salter and T. Welton, Phys. 1 A. J. Ragauskas, C. K. Williams, B. H. Davison, G. Britovsek, J. Chem. Chem. Phys., 2001, 3, 5192–5200. Published on 27 July 2011 http://pubs.rsc.org | doi:10.1039/C1GC15374A Cairney, C. A. Eckert, W. J. Frederick, J. P. Hallett, D. J. Leak, C. L. 30 S. S. Y.Tan, D. R. MacFarlane, J.Upfal, L. A. Edye, W.O. S. Doherty, Liotta, J. R. Mielenz, R. Murphy, R. Templer and T. Tschaplinski, A. F. Patti, J. M. Pringle and J. L. Scott, Green Chem., 2009, 11, 339– Science, 2006, 311, 484–489. 345. 2U.S.DOE,Breaking the Biological Barriers to Cellulosic Ethanol: A 31 L. Cadoche and G. D. Lopez,´ Biol. Wastes, 1989, 30, 153–157. Joint Research Agenda DOE/SC-0095, 2006. 32 Q. Nguyen, M. Tucker, F. Keller, D. Beaty, K. Connors and F. Eddy, 3 J. Pickett, Sustainable biofuels: prospects and challenges,TheRoyal Appl. Biochem. Biotechnol., 1999, 77, 133–142. Society, 2008. 33 P.Sannigrahi, D. H. Kim, S. Jung and A. Ragauskas, Energy Environ. 4 N. Mosier, C. Wyman, B. Dale, R. Elander, Y.Y.Lee, M. Holtzapple Sci., 2011, 4, 1306–1310. and M. Ladisch, Bioresour. Technol., 2005, 96, 673–686. 34 A. A. Dwiatmoko, J. W. Choi, D. J. Suh, Y.-W. Suh and H. H. Kung, 5L.R.Lynd,Annu. Rev. Energy Environ., 1996, 21, 403–465. Appl. Catal., A, 2010, 387, 209–214. 6 J. P. Hallett and T. Welton, Chem. Rev., 2011, 111, 3508–3576. 35 J. B. Binder and R. T. Raines, Proc. Natl. Acad. Sci. U. S. A., 2010, 7 N. V. Plechkova and K. R. Seddon, Chem.Soc.Rev., 2008, 37, 123– 107, 4516–4521. 150. 36 J. F. Saeman, Ind. Eng. Chem., 1945, 37, 43–52.