Official Journal of the European Communities TWENTY-FIFTH COMMISSION DIRECTIVE 2000/11/EC of 10 March 2000 adaptingto technical progress Annex II to Council Directive 76/768/EEC on the approximation of the laws of the Member States relatingto cosmetic products (Text with EEA relevance) THE COMMISSION OF THE EUROPEAN COMMUNITIES,Having regard to the Treaty establishing the European
Pii: s0360-3199(02)00124-6International Journal of Hydrogen Energy 27 (2002) 1381–1390 Pretreatment of Miscanthus for hydrogen production by T. de Vrije∗, G.G. de Haas, G.B. Tan, E.R.P. Keijsers, P.A.M. Claassen Department Bioconversion, Agrotechnological Research Institute (ATO B.V.), P.O. Box17, 6700 AA Wageningen, Netherlands Pretreatment methods for the production of fermentable substrates from Miscanthus, a lignocellulosic biomass, were investigated. Results demonstrated an inverse relationship between lignin content and the e ciency of enzymatic hydrolysis of polysaccharides. High deligniÿcation values were obtained by the combination of mechanical, i.e. extrusion or milling, and chemical pretreatment (sodium hydroxide). An optimized process consisted of a one-step extrusion-NaOH pretreatment at moderate temperature (70◦C). A mass balance of this process in combination with enzymatic hydrolysis showed the following: pretreatment resulted in 77% deligniÿcation, a cellulose yield of more than 95% and 44% hydrolysis of hemicellulose. After enzymatic hydrolysis 69% and 38% of the initial cellulose and hemicellulose fraction, respectively, was converted into glucose, xylose and arabinose. Of the initial biomass, 33% was converted into monosaccharides. Normal growth of Thermotoga elÿi on hydrolysate was observed and high amounts of hydrogen were produced.
? 2002 International Association for Hydrogen Energy. Published by Elsevier Science Ltd. All rights reserved.
Keywords: Enzymatic hydrolysis; Extrusion; Lignocellulose; Fermentation; Biomass conversion Various pretreatment methods are described which pro- mote the accessibility of polysaccharides in a lignocellulose The utilization of hydrogen for the purpose of replac- complex for enzymatic hydrolysis. Examples are steam ing carbonized fuels in the energy and chemical industry is explosion and wet oxidation under alkaline conditions, su- gaining worldwide interest. However, when fossil fuels are percritical CO2 pretreatment, mild and concentrated acid used for the production of hydrogen, there is no advantage hydrolysis and solvent extractions These methods with respect to the reduction of CO2 emission. Therefore, often involve conditions, e.g. high temperatures, which may attempts are being made to produce hydrogen in a CO2 neu- lead to the formation of degradation products which act as tral way. The biological production of hydrogen by fermen- inhibitors in fermentations An alternative mechanical tation using biomass as energy source is one of these new pretreatment method is extrusion. In this study, a corotating twin-screw extruder is used. The biomass is transported via A successful biological conversion of biomass to hydro- transport screws to a reversed screw element (RSE). This gen depends strongly on the processing of raw materials results in accumulation and compression of the material to produce feedstock which can be fermented by the mi- in the space between the transport screws and the RSE.
croorganisms. Lignocellulosics are especially interesting as High compression and shear forces cause deÿbration, ÿb- a source of biomass due to their abundance and low costs.
rillation and shortening of the ÿbers in the biomass The e ciency of this pretreatment method is determined by comparing hydrolysis yields of extruded biomass and ∗ Corresponding author. Tel.: +31-317-475315; fax: +31-317- milled material of di erent particle size.
As an example of a lignocellulosic material, Miscanthus has been tested as substrate for the bioprocess. It is a woody 0360-3199/02/$ 22.00 ? 2002 International Association for Hydrogen Energy. Published by Elsevier Science Ltd. All rights reserved.
T. de Vrije et al. / International Journal of Hydrogen Energy 27 (2002) 1381–1390 rhizomatous C4 perennial grass species which grows rapidly ÿ-glucosidase preparations (Novozymes, Bagsvaerd, Den- and gives high yields per hectare. Extrusion in combination mark). The ÿlter paper activity of Celluclast 1:5 LFG with sodium hydroxide pretreatment caused a substantial (34:5 FPU=ml) was determined according to the Interna- deligniÿcation of Miscanthus material and signiÿcantly im- tional Union of Pure and Applied Chemistry (IUPAC) pro- proved C5 and C6 sugar yields by subsequent enzymatichy- cedure ÿ-Glucosidase activity of Celluclast 1:5 LFG drolysis of cellulose and hemicellulose. Many heterotrophic (12:5 U=ml) and Novozym 188 (172 U=ml) was deter- bacteria are known to produce hydrogen from saccharides mined by measuring the amount of p-nitrophenol liberated Compared to mesophilicand moderate thermophilic from p-nitrophenyl-ÿ-glucopyranoside after 10 min. Both bacteria, extreme and hyperthermophiles produce higher hy- assays were done in citrate bu er (50 mM), pH 4.8 at drogen yields (close to the theoretical maximum of 4 mol 45◦C. One unit was deÿned as the amount of enzyme of H2=mol of hexose) A Miscanthus hydrolysate required to liberate 1 mol of glucose or p-nitrophenol promoted hydrogen production by the extreme thermophilic per minute for ÿlter paper and ÿ-glucosidase activity, bacterium Thermotoga elÿi. There were no indications for the presence of inhibitors in the hydrolysate.
Standard conditions of enzymatic hydrolysis of Miscant- hus were: substrate concentration, 5% w/v; enzyme concen- trations, 1:6 FPU (Celluclast) and 2:3 ÿ-glucosidase units (Celluclast plus Novozym) per gram dry matter; bu er, 50 mM citrate, pH 4.8; temperature, 45◦C; incubation time, 0–72 h. Oxy-tetracycline, gentamycine and cycloheximide (200, 100 and 50 g=ml, respectively) were added as preser- Miscanthus used in this study was collected in the spring vatives to hydrolysates-containing substrates which were not of 2000 and 2001 from a location in Groningen, The Nether- treated with NaOH. Analytical and larger batch experiments lands. Stems were harvested and chopped to a length of 0.5 of 20 and 500 ml, respectively, were carried out in dupli- –5 cm. The dry weight of the harvested material varied be- cate. Samples were collected at 0 and approximately 7, 24, tween 86% and 90%. The samples were stored, protected 48 and 72 h. After pelleting of the remaining solids the lib- from the weather under dry conditions.
erated, soluble sugars in the supernatant were determined by enzymaticmethod (glucose) or by HPLC (total sugars).
The glucose yield was taken as a measure of the hydroly- sis e ciency and is expressed as a percentage of the max- Miscanthus was pretreated by a combination of a mechan- imum amount of glucose obtained after complete chemical ical and chemical method. Mechanical treatment existed of either milling or extrusion. A Retsch mill equipped with a 1 mm or 0:25 mm sieve, or a ZPS50 sifter mill (Hosokawa– Alpine, Germany) was used for size reduction. Lengths of the two latter samples were determined by particle size anal- yses which showed a mean length of 0:22 mm and 17 m, T. elÿi DSM 9442 was purchased from the Deutsche respectively. A Clextral BC45 corotating twin-screw ex- Sammlung von Mikroorganismen und Zellkulturen (Braun- truder (Clextral, Firminy, France) with a screw diameter of schweig, Germany). A modiÿed DSM664 medium 55 mm and a total axis length of 1:25 m was used for the ex- consisted of (per liter): NH4Cl 1:0 g, K2HPO4 0:3 g, trusion experiments. The following conditions were applied: KH2PO4 0:3 g, MgCl2 · 6H2O 0:2 g, CaCl2 · 2H2O 0:1 g, screw speed, 100 rpm; biomass throughput, 15–30 kg dry KCl 0:1 g, NaCl 10:0 g, Na-acetate 0:5 g, yeast extract matter/h; temperature of extruder, 100◦C; reversed screw 4 g, cysteine-HCl · H2O 0:5 g, Na2CO3 2:0 g, resazurine element RSE-15H6; speciÿc energy consumption, approxi- 0:5 mg, trace elements 10:0 ml. Glucose was used as carbon and energy source (7 or 10 g=l in medium with Miscanthus was chemically pretreated with 12% NaOH hydrolysate). The pH was adjusted to 8.0 with HCl. The (w/w dry matter) at variable solid:liquid ratios. Standard trace elements were (per liter): nitrilotriacetic acid 1:5 g, condition of the incubation was at 70◦C for 4 h. Only sam- MgSO4 · 7H2O 3:0 g, MnSO4 · 2H2O 0:5 g, NaCl 1:0 g, ple E1 was treated di erently . NaOH-treated samples were FeSO4 · 7H2O 0:1 g, CoSO4 · 7H2O 0:18 g, CaCl2 · washed with excess water to reduce the alkalinity of the ma- 2H2O 0:1 g, ZnSO4 · 7H2O 0:18 g, CuSO4 · 5H2O 0:01 g, terial prior to enzymatichydrolysis. Pretreated samples were KAl(SO4)2 · 12H2O 0:02 g, H3BO3 0:01 g, Na2MoO4 · stored at 4◦C or at room temperature after drying at 45◦C.
2H2O 0:01 g, NiCl2·6H2O 0:025 g, Na2SeO3·5H2O 0:3 mg.
The medium was made anaerobicby ushing with nitro- gen (100%) and sterilized by autoclaving. Separate sterile, anaerobicstock solutions were prepared of Na2CO3, trace Enzymatichydrolyses of pretreated Miscanthus sam- elements and glucose. An anaerobic, non-sterile Miscant- ples were performed using commercial cellulase and hus hydrolysate was used for the experiments.
T. de Vrije et al. / International Journal of Hydrogen Energy 27 (2002) 1381–1390 100◦C, 80◦C and 50◦C, respectively. N2 was used as the T. elÿi was cultivated in anaerobic serum bottles (100 ml) in culture volumes of 30 ml at 65◦C. The asks were inoc- ulated with approximately 1 ml of a culture in the exponen- tial phase in the same medium (4 g=l glucose). Experiments Results on the chemical composition of Miscanthus showed that polysaccharides represent the largest fraction, The chemical composition of Miscanthus was deter- i.e. 62.5% of the total dry matter (Table The most mined using methods described by the Technical Asso- abundant residue of the polysaccharides was glucose, rep- ciation of the Pulp and Paper Industry (Tappi). Milled resentative of cellulose (glucan). The hemicellulose was of samples (Retsch mill with 0:5 mm sieve) were succes- a xylan type because of a relatively high amount of xylose sively extracted with ethanol/toluene (2:1, v/v) in a Soxtec residues in the remaining polysaccharide fraction. The total system and by hot water (100◦C) Lignin and neu- lignin content was 25.0% and the lignin consisted mainly tral sugars of dried extracted material were determined after sulfuric acid hydrolysis. Samples were either directly hydrolyzed in 1 M H2SO4 (3 h, 100◦C, polysaccharides without cellulose) or ÿrst dispersed in 72% H2SO4 (1 h, 30◦C) followed by hydrolysis in 1 M H2SO4 (3 h, 100◦C, In a ÿrst experiment, conditions for the chemical pretreat- total polysaccharides). Neutral sugars were determined as ment of milled Miscanthus (0:22 mm) were studied. Based alditol acetates by GC or directly by HPLC. The gas chro- on earlier experiments with Miscanthus a solid:liquid matograph was equipped with a CP-SIL 88 WCOT fused ratio of 1:6 and a high NaOH load of 12% (w/w) were silica column (Chrompack, The Netherlands) and a ame applied. Incubation was for 4 h at variable temperatures.
ionization detector. Helium was the carrier gas. Sugars were NaOH treatment resulted in e cient deligniÿcation of Mis- separated by HPLC on a Shodex ionpak KC811 column canthus which increased at higher incubation temperatures (Waters, The Netherlands) at 80◦C with di erential refrac- (Table At the same time, a relative increase of the glu- tometricdetection and 3 mM H2SO4 as the mobile phase can content in the dry solid residue was observed. Control ( ow, 1 ml=min). Acid-insoluble lignin was determined (water) treatment at 95◦C did not show signiÿcant e ects gravimetrically as Klason lignin and acid-soluble lignin by on lignin and glucan contents. Enzymatic hydrolysis of the spectrophotometric analysis Uronicacids were mea- glucan fraction of the residual solids was higher after NaOH sured spectrophotometrically using galacturonic acid as the treatment and increased with increasing temperatures during standard The protein content was determined from the NaOH pretreatment. Because no further signiÿcant increase total nitrogen content (Kjeldahl method) using a conversion of hydrolysis was observed at 95◦C a pretreatment temper- factor of 6.25. Ashes were determined after combustion of ature of 70◦C was applied in further experiments.
The e ect of various pretreatment methods (Table Untreated Miscanthus was analyzed in octuple. The ap- on the composition of Miscanthus and the yield of the plied hydrolysis procedure allowed to discriminate between non-cellulose and total sugars. Pretreated Miscanthus sam- ples were analyzed in duplicate. The glucose content in the total polysaccharide fraction was representative for the cel- Chemical composition of Miscanthus expressed as percentage of lulose (glucan) content and the sum of xylose, arabinose and galactose for the hemicellulose (xylan) content.
Glucose in hydrolysates was determined enzymatically (modiÿcation of the Trinder method Sigma, The Netherlands). Soluble sugars in hydrolysates and fermen- tation medium were analyzed by HPLC as described. The same HPLC method was used for determination of the organic acid content in fermentation medium. Reducing sugars in the ÿlter paper assay were determined using the DNS method with glucose as the standard. Dry weight contents were determined after drying at 105◦C for 24 h.
Hydrogen was measured by GC using a RVS MolSieve 5A, 60/80 mesh, 3 m × 1=8 column. The temperature of the thermal conductivity detector, injector and column was Standard deviations are shown within parenthesis.
T. de Vrije et al. / International Journal of Hydrogen Energy 27 (2002) 1381–1390 ÿ-glucosidase preparations. To determine the optimal con- E ect on lignin and glucan content and on enzymatic hydrolysis ditions of hydrolysis, Celluclast 1:5 LFG and Novozym 188 of glucan of milled Miscanthus (0:22 mm) chemically pretreated were tested in various concentrations and ratios. Glucose yields were measured for an extruded and NaOH pretreated sample (E3) (Fig. At increasing concentrations of Cel- luclast (2.5–10 FPU=g glucan) and Novozym in a ÿxed ra- tio, a signiÿcantly higher hydrolysis rate was observed. In addition, the glucose yield at 72 h incubation was higher, and increased from 50% at a low cellulase concentration to 71% at 10 FPU=g glucan. In all cases, a high initial rate was followed by a decreasing hydrolysis rate. Hydrolysis con- aContents are expressed as percentage of dry solids after pre- tinued after 72 h (results not shown). Higher ÿ-glucosidase activities did not e ect the reaction rate or the glucose yield, bHydrolysis to glucose is expressed as percentage of glucan suggesting that decreasing hydrolysis rates were not due to product inhibition by cellobiose (Fig. The presence of Standard deviations are shown within parenthesis.
1% glucose in the reaction mixture did lower the hydrol- ysis rate and the glucose yield (Fig. In this case cel- residual dry matter has been determined (Table For lobiose accumulated in the hydrolysate suggesting inhibition all samples, except E1, the relative glucan content was in- of ÿ-glucosidase activity by glucose.
creased compared to the initial material (Table Relative Subsequently, all NaOH pretreated milled or extruded xylan contents were not signiÿcantly altered or increased.
Miscanthus samples and some controls were tested for glu- Pretreatment resulted in partial solubilization of the solids.
can hydrolysis applying a low Celluclast concentration of The solubilized products consisted of hydrolyzed hemicellu- 1:6 FPU=g dry matter. The glucose yield of milled NaOH lose and decomposed lignin. Deligniÿcation values of more pretreated samples increased as the particle size of the ma- than 70% have been reached for most of the pretreatment terial decreased (Fig. Maximum glucan hydrolysis of methods, only deligniÿcation of sample E1 was less. Xy- 17 m material after 72 h incubation amounted to 56%.
lan solubilization increased with decreasing particle size and Milled material, not chemically pretreated showed ine - was high for extruded samples, except E1. In general, the cient hydrolysis and the maximum value was reached within glucan yield was high, but there was a notable loss of glucan 24 h of incubation. Also in this case glucan hydrolysis was higher for material with a smaller particle size. Extruded samples which were treated with NaOH during or following extrusion (samples E2–E4) showed a comparable hydrolysis e ciency reaching a maximal value of 50% after 72 h incu- bation (Fig. Glucan hydrolysis was signiÿcantly lower hydrolyzed by commercially available cellulase and for material which had undergone chemical (and steam) aMiscanthus with a length between 0.5 and 5 cm was impregnated with 8% NaOH (w/w) at room temperature for 24 h. After allowing the liquid to drain through a perforated screen the impregnated material was preheated with saturated steam at atmospheric pressure for 10 min.
bMiscanthus (0.5–5 cm) was manually fed in a dry form to the extruder. An NaOH solution was supplied through an injection port cMiscanthus (0.5–5 cm) was ÿrst impregnated with excess water for 16 h at room temperature. After drainage of water the material was fed to the extruder and treated as in footnote b.
The liquid:solid ratio of the NaOH treatment is shown within parenthesis.
T. de Vrije et al. / International Journal of Hydrogen Energy 27 (2002) 1381–1390 E ect of pretreatment of Miscanthus on the composition of the solids remaining after treatment, the yield of solids and the removal of various components from the solid fraction aComposition is shown as percentage of the remaining solid fraction after pretreatment.
bSolid yield is shown as percentage of the initial amount of dry matter.
cLoss of components is shown as percentage of the amount in the initial material.
n.d. Not determined. The maximum standard deviation was 3.7, 1.7 and 1.2 for glucan, xylan and lignin composition values, respectively.
Fig. 1. Enzymatichydrolysis of extruded and NaOH-pretreated Fig. 2. Enzymatichydrolysis of milled Miscanthus. NaOH treated: Miscanthus (sample E3) at substrate concentrations of 5% (w/v) (4) 1 mm; (•) 0:22 mm and ( ) 17 m particle size. Non-treated: and various enzyme concentrations. Cellulase: (4; ) 2.5; (•) 5 (◦) 0:22 mm and ( ) 17 m particle size. Enzyme activities: cel- and ( ) 10 FPU=g glucan and ÿ-glucosidase: (4) 2.3; ( ) 4.2; lulase, 1:6 FPU=g dry matter; ÿ-glucosidase, 2:3 U=g dry matter.
(•) 2.9; and ( ) 3:9 U=g dry matter. ( ) enzyme concentrations The maximum standard deviation was 0.4.
as (4) plus 10 g=l glucose. Glucose yield is expressed as percentage of the maximum amount of glucose in glucan. The maximum low ÿ-xylosidase activity in both Celluclast and Novozym.
Starting with an initial concentration of 50 g=l biomass a pretreatment prior to extrusion (E1) and amounted to 21%.
maximum concentration of 32 g=l monomericsugars was No hydrolysis was observed for steamed extruded material.
reached after 72 h incubation. Under standard conditions The presence of other monomeric sugars in the hy- no cellobiose was present in the hydrolysates, but possibly drolysates was determined (Table Besides glucose xylobiose and other oligomerichydrolysis products of xylan hydrolysis products of hemicellulose, xylose and arabinose were present. This could be caused by the low ÿ-xylosidase were present. The concentration of these sugars was de- activity (10 times lower than ÿ-glucosidase activity under pendent on the applied pretreatment method and showed a similar pattern as the glucose concentration. Hardly any E ciencies of hydrolysis and conversion of glucan, xylan hydrolysis of xylan occurred for material without NaOH and total biomass are depicted in Table The conversion treatment. Enzymatic hydrolysis of xylan could occur be- e ciency is based on the initial biomass taken into account cause of the presence of xylanase activity in Celluclast and a the loss of dry matter by pretreatment. Because of the T. de Vrije et al. / International Journal of Hydrogen Energy 27 (2002) 1381–1390 Fig. 3. Enzymatichydrolysis of extruded and NaOH-pretreated Fig. 4. Relation between lignin content and hydrolysis of Miscanthus. ( ) E1; (◦), E2; ( ) E3 and ( ) E4. (•) ex- glucan (•) and xylan (◦) of extruded Miscanthus under standard truded and steam pretreated Miscanthus. Enzyme activities: cel- lulase, 1:6 FPU=g dry matter; ÿ-glucosidase, 2:3 U=g dry matter.
The maximum standard deviation was 0.7.
and 38% of glucan and xylan, respectively, into monomeric sugars were obtained, which corresponds to 34% of the biomass (sample E3, 10 FPU=g glucan).
Enzymatichydrolysis of NaOH-pretreated Miscanthus and control A linear relationship was observed between lignin con- tent of extruded Miscanthus and the level of enzymatic Glucose Xylose + arabinose Monomeric sugars hydrolysis of glucan and xylan (Fig. At equal lignin contents xylan conversion was on average 4% higher than glucan conversion. Under the standard conditions at low cellulase concentrations, a maximal theoretical conversion can be derived at 0% lignin and amounts to 85% and 89% for glucan and xylan, respectively. For milled samples no relationship could be obtained with lignin content. Here, hydrolysis of polysaccharides seemed to be dependent on Growth of the extreme thermophilicbacterium, T. elÿi, Hydrolysis of biomass (50 g=l) to monomericsugars under stan- on hydrolysate was compared with growth on glucose. A dard amounts of Celluclast (1:6 FPU=g dry matter) and Novozym batch culture on glucose medium showed a normal growth and using a four times higher concentration (6:3 FPU=g) in case curve with a short lag phase followed by an exponential of sample E3. Incubation time was 72 h. Control M2 and M3 were phase. After 3 days of growth a maximal optical density milled Miscanthus, control E1 was extruded and steam pretreated Miscanthus, but without NaOH treatment. The maximum standard deviation was 0.5 and 0.3 for glucose and xylose + arabinose con- glucose was converted to hydrogen and acetate. At the end of the fermentation glucose was only partially consumed possibly because of growth inhibition due to a low pH and/or the accumulation of metabolites. T. elÿi was able minimal loss of glucan, the hydrolysis and conversion to grow on a Miscanthus hydrolysate containing glucose e ciencies are almost equal except for sample M3. The as the main monosaccharide and reached a similar optical conversion e ciency of xylan was much lower than the density as on glucose medium (Fig. Table Hydrogen hydrolysis e ciency, mainly because of the signiÿcant production by T. elÿi on hydrolysate was slightly higher loss of xylan caused by NaOH pretreatment. In general, and acetate production was signiÿcantly higher than on biomass conversion of milled or extruded Miscanthus in glucose medium. Glucose consumption, which was compa- combination with a chemical treatment did not signiÿcantly rable to that on glucose medium, occurred simultaneously di er (except sample E3). Maximal conversions of 69% with the consumption of xylose during growth of T. elÿi on T. de Vrije et al. / International Journal of Hydrogen Energy 27 (2002) 1381–1390 Hydrolysis and conversion e ciencies of glucan, xylan and total Miscanthus biomass Hydrolysis is shown as percentage of the content of pretreated material and conversion as percentage of the content of initial material (glucan and xylan conversion) or of the total biomass (biomass conversion). n.d. Not determined.
Growth, consumption of glucose and xylose, and production of hydrogen and acetate by Thermotoga elÿi on hydrolysate of extruded and NaOH-treated Miscanthus and on glucose (control) medium The OD580 and pH values were reached at the end of the fermentation after 93 h of incubation at 65◦C. Consumption and production values were determined for the whole fermentation period.
hydrolysate. No clear preference for one of the two allowed rapid impregnation of a NaOH solution supplied via an injection port downstream the RSE.
Enzymatichydrolysis of milled or extruded material was low. Signiÿcantly higher yields were obtained for NaOH-treated Miscanthus samples with low lignin con- tents. Only at a very small particle size (approximately The present study describes results on pretreatment and 10 m), the lignin content appeared to be less important enzymatichydrolysis of a lignocellulosicbiomass. The (Fig. Table and the substrate surface area seemed to chemical composition of Miscanthus is similar to that of, become an in uencial factor Additionally, the crys- e.g. willow (hard wood), wheat straw, and bagasse tallinity of cellulose is an important feature determining and shows a relatively high amount of lignin. Lignin con- its susceptibility to enzymatic hydrolysis The ef- tent and e ciency of enzymatic hydrolysis appeared to be fects of the applied mechanical and chemical pretreatment inversely related (Fig. which conÿrms the importance methods on the crystallinity of Miscanthus cellulose were, of biomass deligniÿcation Chemical pretreatment however, not studied. At a 5% (w/v) substrate concentra- is required for deligniÿcation and in this study NaOH tion and a cellulase concentration of 10 FPU=g cellulose, was used to decompose and remove lignin. Conditions for which is often used in laboratory experiments and for tech- NaOH pretreatment of Miscanthus were mainly based on noeconomical evaluations approximately 70% of the earlier results Our results showed that NaOH treat- cellulose of NaOH-treated and extruded Miscanthus was ment of chopped material was less e ective, while milled hydrolyzed in 72 h (Fig. Table Enzyme kinetics material (approximately 1 mm and smaller) was deligni- showed a typical course with a high initial rate resulting in ÿed for 70%. More than 75% deligniÿcation was obtained 55% hydrolysis in the ÿrst 24 h. Higher enzyme concen- when the material was treated with NaOH in combination trations and longer incubation times will probably result with extrusion. In this case, the most practical and e cient in complete hydrolysis but these conditions are no options method is the addition of NaOH to the biomass during for commercial processes. End product inhibition appears extrusion. The pressure drop created in passing the RSE to be a major cause of decreasing hydrolysis rates with T. de Vrije et al. / International Journal of Hydrogen Energy 27 (2002) 1381–1390 monosaccharides with 62.5% being the theoretical maxi- mum. The concentration of monosaccharides in the ÿnal product, the hydrolysate, was 32 g=l. Two side streams were generated in the process. The ‘black liquor’, after concentra- tion of the solids, can be used for generation of energy and recovery of the chemical. The solid residue obtained after the enzymatichydrolysis, with a similar composition as the starting material, can be recycled back to the pretreatment stage. With respect to conversion e ciencies, comparison of this process with other pretreatment methods described in literature is di cult. The type of feedstock determines for a considerable part the results, and enzymatic hydrolysis yields greatly depend on assay conditions.
An important advantage of the extrusion technique over other pretreatment methods is the moderate operation tem- perature, preventing the formation of degradation and oxida- tion products of lignin and saccharides, respectively, which are potential inhibitors of fermentation. A ÿrst experiment on small scale demonstrated that an hydrolysate prepared from extruded Miscanthus was able to sustain growth of the extreme thermophilicbacterium, T. elÿi, at a compara- ble level as glucose medium. C5 and C6 sugars, present in this hydrolysate, were simultaneously consumed and high amounts of hydrogen and acetate were produced. In future experiments performances of hydrolysates in large-scale hy- drogen fermentations under controlled conditions will be The cost for hydrogen produced in a small scale biohy- drogen production plant with a capacity of 500 m3 H2=h has Fig. 5. Growth of Thermotoga elÿi on glucose medium (A) and been estimated In this two-stage bioprocess, H Miscanthus hydrolysate (B) at 65◦C. (•) Glucose consumption; ( ) xylose consumption; ( ) hydrogen production; (4) acetate organic acids are produced from biomass in a ÿrst fermen- tation by thermophilic bacteria. In the second stage, photo- heterotrophic bacteria convert the organic acids to hydrogen with the help of light. This bioprocess theoretically yields 12 mol of hydrogen per mole of hexose. In the cost esti- lignocellulosic materials. Suggested approaches to over- mation, a biomass conversion e ciency of 40% and energy come end product inhibition are simultaneous sacchariÿca- consumption of the extruder of 150 kWh=ton dry biomass tion and fermentation of the substrate or removal of was taken into account. The Miscanthus process described the sugars from the hydrolysate by ultraÿltration A in this article yielded a conversion e ciency of 33% and promising strategy for low lignin-containing materials is the an extruder energy demand of approximately 300 kWh=ton recycling of enzymes in combination with short residence dry Miscanthus. This ÿnding will increase the costs by 3 times in the hydrolysis step (reviewed by Gregg and Saddler Euro cents to 0:24=m3 H2, but leave the costs within the range of sustainable hydrogen costs from other small-scale Although biomass conversion e ciencies for milled Miscanthus are not necessarily less than for extruded ma- terial (in combination with a chemical treatment) the latter process seems to be more favorable. Extrusion and chem- ical pretreatment can be combined in one step, gaining potential higher deligniÿcation values. The mass balance This study was ÿnancially supported by the Dutch Min- of such a process is shown in Fig. Pretreatment resulted istries of Economic A airs (EZ), Education, Culture and in 77% deligniÿcation and a loss of hemicellulose of 44%.
Science (OCenW), and Housing, Spacial Planning and the Cellulose yield was more than 95%. Approximately 70% Environment (VROM) via the Economy, Ecology, Tech- of the polysaccharides of the pretreated biomass was en- nology Programme (project number EETK99116) and the zymatically hydrolyzed. Of the initial cellulose fraction of Ministry of Agriculture, Nature Management and Fisheries.
Miscanthus 69% was converted into glucose. Thirty-three B.H. Dijkink and J.C. van der Putten are acknowledged for percent of the total initial biomass was converted into particle size analyses and analyses of chemical composition, T. de Vrije et al. / International Journal of Hydrogen Energy 27 (2002) 1381–1390 Fig. 6. Flowsheet of pretreatment and enzymatichydrolysis of Miscanthus. Conversion and hydrolysis e ciencies of glucan and xylan and yields of soluble sugars starting with 100 g dry matter. NaOH pretreatment was at 70◦C. Enzymatichydrolysis of pretreated biomass (5% w/v) was carried out at cellulase concentrations of 10 FPU=g glucan.
respectively. We thank Novozymes for providing enzyme  Nandi R, Sengupta S. Microbial production of hydrogen: an preparations and A. Drenth (Agromiscanthus B.V.) for pro- overview. Crit Rev Microbiol 1998;24:61–84.
 Schafer T, Schonheit P. Maltose fermentation to acetate, CO2 and H2 in the anaerobichyperthermophilicarchaeon Pyrococcus furiosus: evidence for the operation of a novel sugar fermentation pathway. Arch Microbiol 1992;158:  Claassen PAM, van Lier JB, Lopez Contreras AM, van Niel EWJ, Sijtsma L, Stams AJM, de Vries SS, Weusthuis RA.
 Schroder C, Selig M, Schonheit P. Glucose fermentation Utilisation of biomass for the supply of energy carriers. Appl to acetate, CO2 and H2 in the anaerobichyperthermophilic Microbiol Biotechnol 1999;52:741–55.
eubacterium Thermotoga maritima: involvement of the  Benemann JR. Hydrogen biotechnology: progress and Embden–Meyerhof pathway. Arch Microbiol 1994;161: prospects. Nat Biotechnol 1996;14:1101–3.
 Benemann JR. Biohydrogen: approaches and potential.
 van Niel EWJ, Budde MAW, de Haas GG, van Proceedings of the 11th Canadian Hydrogen Conference, der Wal FJ, Claassen PAM, Stams AJM. Distinctive Victoria, BC, June 2001. p. 17–20.
properties of high hydrogen producing extreme thermophiles,  Fan LT, Lee Y-H, Gharpuray MM. The nature of Caldicellulosiruptor saccharolyticus and Thermotoga elÿi.
lignocellulosics and their pretreatment for enzymatic Int J Hydrogen Energy 2002;27(11–12):1391–8.
hydrolysis. Adv Biochem Eng Biotechnol 1982;23:158–87.
 Ghose TK. Measurement of cellulase activities. Pure Appl  Parisi F. Advances in lignocellulosics hydrolysis and in the utilization of the hydrolyzates. Adv Biochem Eng Biotechnol  Tappi. T 264 om-88. Preparation of wood for chemical analysis. Test methods 1998–1999. Atlanta: Tappi Press.
 Bjerre AB, Olesen AB, Fernqvist T, Ploger A, Schmidt AS.
 Tappi. T UM 250. Acid-soluble lignin in wood and pulp.
Pretreatment of wheat straw using combined wet oxidation Useful methods 1991. Atlanta: Tappi Press.
and alkaline hydrolysis resulting in convertible cellulose and  El Rayad Ahmed A, Labavitch JM. A simpliÿed method for hemicellulose. Biotechnol Bioeng 1996;49:568–77.
accurate determination of cell wall uronide content. J Food  Zheng Y, Lin H-M, Wen J, Cao N, Yu X, Tsao GT.
Supercritical carbon dioxide explosion as a pretreatment for cellulose hydrolysis. Biotechnol Lett 1995;17:845–50.
 Tappi. T 211 om-93. Ash in wood, pulp, paper and paperboard:  Olsson L, Hahn-Hagerdal B. Fermentation of lignocellulosic combustion at 575◦C. Test methods 1998–1999. Atlanta: hydrolysates for ethanol production. Enzyme Microb Technol  Trinder P. Determination of glucose in blood using glucose  Van Roekel GJ, Lips SJJ, Op den Kamp RGM, Baron G.
oxidase with an alternative oxygen acceptor. Ann Clin Extrusion pulping of true hemp bast ÿbre (Cannabis sativa L.). Tappi Conference, Chicago, USA, 1995. p. 477–85.
 Van Roekel GJ. Bulk papermaking applications for bast ÿbre crops using extrusion pulping, Peterborough, England, PIRA  Iglesias G, Bao M, Lamas J, Vega A. Soda pulping of International, plenary paper, 1996. 29p.
Miscanthus sinensis. E ects of operational variables on T. de Vrije et al. / International Journal of Hydrogen Energy 27 (2002) 1381–1390 pulp yield and lignin solubilization. Bioresource Technol  Gregg DJ, Saddler JN. Factors a ecting cellulose hydrolysis and the potential of enzyme recycle to enhance the e ciency  Koullas DP, Christakopoulos PF, Kekos D, Koukios EG, of an integrated wood to ethanol process. Biotechnol Bioeng Macris BJ. E ect of alkali deligniÿcation on wheat straw sacchariÿcation by Fusarium oxysporum cellulases. Biomass  Szczodrak J, Targonski Z. Simultaneous sacchariÿcation and fermentation of cellulose: e ect of ethanol and cellulases on  Lee J. Biological conversion of lignocellulosic biomass to particular stages. Acta Biotechnol 1989;6:555–64.
ethanol. J Biotechnol 1997;56:1–24.
 Philippidis GP, Spindler DD, Wyman CE. Mathematical  Gharpuray MM, Lee Y-H, Fan LT. Structural modiÿcation modeling of cellulose conversion to ethanol by the of lignocellulosics by pretreatments to enhance enzymatic simultaneous sacchariÿcation and fermentation process. Appl hydrolysis. Biotechnol Bioeng 1983;25:157–72.
Biochem Biotechnol 1992;34/35:543–56.
 Sinitsyn AP, Gusakov AV, Vlasenko EY. E ect of structural  Tan LUL, Yu EKC, Mayers P, Saddler JN. Column and physicochemical features of cellulosic substrates on the cellulose hydrolysis reactor: cellulase adsorption proÿle. Appl e ciency of enzymatic hydrolysis. Appl Biochem Biotechnol Microbiol Biotechnol 1986;25:256–61.
 Tan LUL, Yu EKC, Mayers P, Saddler JN. Column cellulose  Thompson DN, Chen HC, Grethlein HC. Comparison of hydrolysis reactor: the e ect of retention time, temperature, pretreatment methods on the basis of available surface area.
cellulase concentration and exogenously added cellobiose Bioresource Technol 1992;39:155–63.
on the overall process. Appl Microbiol Biotechnol 1992;26:  Koullas DP, Christakopoulos PF, Kekos D, Macris BJ, Koukios EG. E ect of cellulose crystallinity on the enzymatic  Claassen PAM, van Groenestijn JW, Janssen AJH, van hydrolysis of lignocellulosics by Fusarium oxysporum Niel EWJ, Wij els RH. Feasibility of biological hydrogen cellulases. Cellul Chem Technol 1990;24:469–74.
production from biomass for utilization in fuel cells.
 Koullas DP, Christakopoulos P, Kekos D, Macris BJ, In: Kyritsis S, Beenackers AACM, Helm P, Grassi A, Koukios EG. Correlating the e ect of pretreatment on the Chiaramonti D, editors. Proceedings of the First World enzymatichydrolysis of straw. Biotechnol Bioeng 1992;39: Conference on Biomass for Energy and Industry. London: James & James, 2001. p. 1665–7.
The Role of Topical Antibiotic Prophylaxis toPrevent Endophthalmitis after IntravitrealInjectionPhilip Storey, MD, Michael Dollin, John Pitcher, MD,Sahitya Reddy, Joseph Vojtko, James Vander, MD,Jason Hsu, Sunir J. Garg, MD,the Post-Injection Endophthalmitis Study TeamTo compare the incidence of endophthalmitis after intravitreal injection with and without topicalpostinjection antibiotic prophy