w a t e r r e s e a r c h 4 3 ( 2 0 0 9 ) 1 4 1 4 – 1 4 2 4

Avai lab le at www.sc iencedi rect .com

journa l homepage : www.e lsev i er . com/ loca te /wat res

Biohydrogen production from xylose at extreme thermophilictemperatures (70 8C) by mixed culture fermentation

Prawit Kongjan, Booki Min, Irini Angelidaki*

Department of Environmental Engineering, Technical University of Denmark, DK-2800 Lyngby, Denmark

a r t i c l e i n f o

Article history:

Received 11 July 2008

Received in revised form

3 December 2008

Accepted 11 December 2008

Published online 24 December 2008

Keywords:

Biohydrogen

Extreme thermophilic

Mixed culture

Xylose

* Corresponding author. Tel.: þ45 4525 1429;E-mail address: ria@env.dtu.dk (I. Angelid

0043-1354/$ – see front matter ª 2008 Elsevidoi:10.1016/j.watres.2008.12.016

a b s t r a c t

Biohydrogen production from xylose at extreme thermophilic temperatures (70 �C) was

investigated in batch and continuous-mode operation. Biohydrogen was successfully

produced from xylose by repeated batch cultivations with mixed culture received from

a biohydrogen reactor treating household solid wastes at 70 �C. The highest hydrogen yield

of 1.62� 0.02 mol-H2/mol-xyloseconsumed was obtained at initial xylose concentration of

0.5 g/L with synthetic medium amended with 1 g/L of yeast extract. Lower hydrogen yield

was achieved at initial xylose concentration higher than 2 g/L. Addition of yeast extract in

the cultivation medium resulted in significant improvement of hydrogen yield. The main

metabolic products during xylose fermentation were acetate, ethanol, and lactate. The

specific growth rates were able to fit the experimental points relatively well with Haldane

equation assuming substrate inhibition, and the following kinetic parameters were

obtained: the maximum specific growth rate (mmax) was 0.17 h�1, the half-saturation

constant (Ks) was 0.75 g/L, and inhibition constant (Ki) was 3.72 g/L of xylose. Intermittent

N2 sparging could enhance hydrogen production when high hydrogen partial pressure

(>0.14 atm) was present in the headspace of the batch reactors. Biohydrogen could be

successfully produced in continuously stirred reactor (CSTR) operated at 72-h hydraulic

retention time (HRT) with 1 g/L of xylose as substrate at 70 �C. The hydrogen production

yield achieved in the CSTR was 1.36� 0.03 mol-H2/mol-xylosesonsumed, and the production

rate was 62� 2 ml/d$Lreactor. The hydrogen content in the methane-free mixed gas was

approximately 31� 1%, and the rest was carbon dioxide. The main intermediate by-prod-

ucts from the effluent were acetate, formate, and ethanol at 4.25� 0.10, 3.01� 0.11, and

2.59� 0.16 mM, respectively.

ª 2008 Elsevier Ltd. All rights reserved.

1. Introduction et al., 2007). Moreover, this process is considered as a prom-

Fermentative biohydrogen production is an emerging tech-

nology and has increasingly attracted interest in recent years

due to high demand of sustainable energy production (Kalia

et al., 2003). The dark fermentative hydrogen process is envi-

ronmentally friendly, cost-effective, and sustainable (Hawkes

fax: þ45 4593 2850.aki).

er Ltd. All rights reserved

ising treatment technology for organic wastes and/or residues

with simultaneous clean energy production with high effi-

ciency (Kapdan and Kargi, 2006). During the dark fermentation

process, hydrogen is produced simultaneously with CO2 in the

gas phase, and organic acids and solvents in the liquid phase

as the end products (Hawkes et al., 2007). Substrates that have

.

Xylose

Pyruvate

Acetyl-CoA

Acetate Butyrate

ADP

ATP

NAD+

NADH

FdoxPFOR

Fdred H2CO2

NFOR

Fdox

Fdred H2

Hydrogenase

Hydrogenase

Fig. 1 – Metabolic pathway for fermentative hydrogen

production from xylose under anaerobic conditions.

w a t e r r e s e a r c h 4 3 ( 2 0 0 9 ) 1 4 1 4 – 1 4 2 4 1415

been used for hydrogen dark fermentation are mainly carbo-

hydrate-containing feedstocks such as glucose (Kotsopoulos

et al., 2006; Zhang et al., 2008), sucrose (van Niel et al., 2003;

Wu et al., 2007), starch (Fang et al., 2006; Zhang et al., 2003) and

lignocellulosic material (Kapdan and Kargi, 2006; Valdez-

Vazquez et al., 2006).

One of the main concerns in biofuels production is the high

costs of the feedstock. Therefore, second generation tech-

nologies where agricultural residues or wastes are used as

feedstock, have emerged. Lignocellulosic material, which is

the main component of agricultural residues such as straw

and corn stover, is a promising feedstock for cost-effective

energy production. Investigations using lignocellulosic mate-

rial for bioethanol or biohydrogen production are being

intensively carried out (Kapdan and Kargi, 2006; Iqnacio et al.,

2006). In general, lignocellulose is composed of cellulose (40–

50%), hemicellulose (25–30%), and lignin (10–20%) (Iqnacio

et al., 2006). Physico-chemical pretreatment of lignocellulose

material such as hydrothermal treatment or wet oxidation,

results in two phase products, a solid mainly consisting of

hexoses and a hydrolysate mainly containing pentose

(Thomsen et al., 2006). Hexoses can be effectively converted to

bioethanol and the process is carried out with high yield

(around 0.40–0.51 g/g) and productivity (up to 1.0 g/(L h)) by

Saccharomyces cerevisiae. However, microorganisms for the

conversion of pentose to bioethanol are not very effective

(Angelidaki et al., 2007). Due to this limitation, an alternative

way of the pentose substrate utilization needs to be found to

increase the net energy yield of the feedstock. Xylose

conversion to biohydrogen can be an effective way to utilize

this substrate which is wasted otherwise. In addition, Datar

et al. (2007) have verified that hemicellulose from corn stover

could serve as the potential substrate for biohydrogen

production by batch-mode fermentation with mixed culture

at 35 �C.

Xylose is a pentose and the main component of hemi-

cellulose (Thomsen et al., 2006). Theoretically xylose can be

converted to hydrogen with a maximum yield of 3.33 mol-H2/

mol-xylose when acetate is produced as the fermentation by-

product (equation (1)). Alternatively, xylose can be converted

into hydrogen by butyrate pathway as shown in equation (2),

with lower yield of 1.67 mol-H2/mol-xylose.

C5H10O5 þ 1:67 H2O�!1:67C2H3O�2 þ 1:67Hþ þ 1:67CO2 þ 3:33H2

DGo ¼ �195:5 KJ=mole ð1Þ

C5H10O5�!0:83C4H7O�2 þ 0:83Hþ þ 1:67CO2 þ 1:67H2

DGo ¼ �233:9 KJ=mole(2)

The metabolic pathway of xylose fermentation for

hydrogen production under anaerobic condition was shown

in Fig. 1 (Angenent et al., 2004; Temudo et al., 2007; Zhu and

Yang, 2004). Hydrogen is produced by mediation of hydroge-

nase, using electrons from reduced ferredoxin (Fdred) to

reduce protons. Oxidized ferredoxin (Fdox) also obtains elec-

trons from the oxidation of NADH, by NADH:Fd oxidoreduc-

tase (NFOR), and then reduced ferredoxin (Fdred) is oxidized

with proton release for hydrogen production. In this pathway,

acetyl-CoA, a branch point intermediate produced from

pyruvate, is further converted to acetate and butyrate via

acetyl-P and butyryl-P by acetate kinase (AK) and butyrate

kinase (BK), respectively.

Research for biological production of hydrogen was mainly

based on glucose as substrate. Investigations of xylose

fermentation to biohydrogen were mainly concentrated on

fermentation at mesophilic temperatures (Lin et al., 2006; Lin

and Cheng, 2006; Lin et al., 2008) and thermophilic tempera-

tures (Calli et al., 2008; Lin et al., 2008; Lo et al., 2008; Wu et al.,

2008). However, only limited studies have been carried out on

the possibility of utilization of xylose for hydrogen production

at extreme thermophilic temperatures (Kadar et al., 2004;

Yokoyama et al., 2007). The biohydrogen fermentation at

extreme thermophilic temperatures (over 70 �C) has many

advantages over at lower temperature conditions due to better

pathogenic destruction, lower risk of contamination by

methanogenic archaea (van Groenestijn et al., 2002), higher

rate of hydrolysis (Lu et al., 2008), and higher H2 yield (Kadar

et al., 2004). Under the extreme thermophilic temperatures of

70 �C, Kotsopoulos et al. (2006) obtained a yield of 2.4 mol-H2/

mol-glucose by mixed culture fermentation in a UASB reactor.

Although working at extreme thermophilic temperatures may

cause higher costs for heating, a smaller reactor volume due to

short HRT as a consequence of higher hydrolysis and

hydrogen production rate at extreme thermophilic tempera-

ture could compensate the energy expenses from temperature

increased. The costs for operation at higher temperatures will

depend largely on the heat exchange efficiency of the plant,

insulation of reactors etc. It has been found that Danish biogas

plants, operating at thermophilic temperature (55 �C), the

energy cost for operating at 55 �C is approx. 10% of the energy

produced at the plant. The extra energy cost for operating at

thermophilic compared to mesophilic temperature is

marginal (1–2%). Additionally, according to the regulations,

biogas plants (and presumably also biohydrogen plants), have

to include a sanitation step where specific categories of

biomasses are required heated up to 70 �C for 1 h, in order to

secure sanitation of the effluents. Consequently, a combined

sanitation and fermentation step, will probably, reduce the

total operation and construction expenses.

Mixed culture fermentation for hydrogen production is

more suitable for industrial application than pure culture

w a t e r r e s e a r c h 4 3 ( 2 0 0 9 ) 1 4 1 4 – 1 4 2 41416

fermentation because mixed culture fermentation has many

advantages: 1) no sterilization of media 2) increased adapta-

tion capacity offered by its high microbial diversity 3) possi-

bility of mixed substrates co-fermentation (Kleerebezem and

van Loosdrecht, 2007) 4) higher capability for continuous

processing (Temudo et al., 2007). In this study, using mixed

culture we examined the conversion of xylose to hydrogen by

dark fermentation at extreme thermophilic temperatures

(70 �C), in both batch and continuous-mode operations. Batch

experiments were carried out to investigate the metabolic

pathways and establish kinetic parameters for biohydrogen

fermentation at different initial xylose concentrations.

Furthermore, continuous biohydrogen production from

xylose was demonstrated for the first time in CSTR reactor at

extreme thermophilic temperatures (70 �C) conditions.

2. Materials and methods

2.1. Batch experiments

2.1.1. Preparation of the batch cultivations55-mL vials were used for the batch experiments. The vials

were first filled with basic anaerobic (BA) medium (Angelidaki

and Sanders, 2004) containing xylose at different concentra-

tions, and then they were sealed with butyl stoppers and

aluminum crimps. The sealed bottles were purged with a gas

mixture (20/80) of CO2/N2 for 10 min and put into the incubator

(70 �C) for 15 min before the inoculum was transferred into the

bottles. The volume of inoculum was 20% of total liquid

volume (20 mL) in the bottles. For each experiment, the bottles

were prepared in triplicate. A control bottle contained only BA

medium without xylose.

2.1.2. Adaptation of the inoculumBatch experiments were conducted for adaptation of inoc-

ulum with xylose as the carbon and energy source at extreme

thermophilic temperatures (70 �C). The inoculum was origi-

nally received from a lab scale CSTR, which was continuously

fed with household solid waste and operated at 70 �C for

hydrogen production at hydraulic retention time of 3 days (Liu

et al., 2008a). The inoculum for the adaptation study was first

enriched by repeated batch cultivations in 55-mL serum vials

(working volume 20 mL) with initial xylose concentration of

0.25 g/L at 70 �C. For repeated batch cultivations, the mixed

culture from the first generation that had the highest

hydrogen production was used as the inoculum for the second

batch cultivation (second generation). Repeated transfers

were stopped when no further increase of the hydrogen yield

was observed compared to the previous cultivation. The

repeated batch cultivations started from 0.25 g/L xylose and

then increased to 0.5 and 1.0 g/L gradually. The batch with

1.0 g/L xylose concentration was subsequently cultivated in

a large volume bottle (total volume 250 mL; working volume

100 mL) and was used as the inoculum for the kinetic study

after being distributed in 20 ml portions in 55-mL serum vials

with initial xylose concentrations of 0.5, 1.0, 2.0, 3.0, and 4.0 g/

L. During these cultivations BA medium was amended with

1 g/L yeast extract in order to improve the medium

composition for enhanced growth of hydrogen producing

microorganisms (van Niel et al., 2002; Xu et al., 2007).

2.1.3. Effect of spargingIn order to investigate the effect of hydrogen partial pressure

on hydrogen production, the headspace of the bottles was

sparged for 3 min with nitrogen gas. Before the nitrogen

sparging, additional xylose was added to the bottles in order to

prevent possible substrate limitation. The hydrogen concen-

tration in liquid phase before and after sparging was calcu-

lated, assuming equilibrium between liquid and gas phase, by

application of Henrys law equation: KH� Pi, where KH is the

Henry’s law constants for hydrogen (0.7903 mM/atm) and Pi is

the headspace hydrogen partial pressure (atm).

2.2. Operation of CSTR for hydrogen production

Continuous hydrogen fermentation with xylose as substrate

was carried out in a 1-L CSTR (700 mL working volume) oper-

ated at 72-h hydraulic retention time (HRT) at 70 �C. The feed

was composed of two bottles: one with basic anaerobic

nutrient solution (BA medium amended with 1 g/L of yeast

extract) and one with 2.0 g/L of xylose solution. Xylose and the

BA medium were mixed 1:1 (v/v) at an entry point before the

reactor. The reactor was first started up in batch mode by

using 140 mL enriched cultures from the batch experiments.

The CSTR was then switched to continuous-mode operation

after 3 days when the maximum hydrogen content in the gas

phase was obtained. The temperature was controlled at 70 �C

by circulating hot water inside a water jacket surrounding the

CSTR.

2.3. Monitoring and analyses

The composition of the produced biogas from both batch and

CSTR reactors (H2, CH4, and CO2) was routinely monitored.

Liquid sample were taken daily from CSTR and at the end of

the stationary phase from batch reactor for further analysis of

VFAs, alcohols, lactate, formate and xylose. COD balance was

also made for the xylose degradation to products. The

biomass concentration (assumed formula C5H7O2N) used in

COD balance was assumed to be 15% of xylose degradation in

term of COD (Kotsopoulos et al., 2006). Yield (mol-H2/mol-

xyloseconsumed) was calculated by using total amount of

hydrogen produced divided by xylose consumed. The yields

obtained during batch experiments were based on average

from triplicate.

All biogas components (H2, CH4, and CO2) were measured

by gas chromatography (GC) (MicroLab, Arhus, Denmark)

equipped with a thermal conductivity detector (TCD).

Hydrogen was analyzed by GC-TCD fitted with a 4.5 m� 3 mm

s-m stainless column packed with Molsieve SA (10/80).

Nitrogen was used as carrier gas at a flow rate of 20 mL/min.

The temperature of both the injection port and detector was

90 �C. Methane and carbon dioxide were analyzed with GC-

TCD fitted with parallel column of 1.1 m� 3/1600 Molsieve 137,

and 0.7 m� 1/400 chromosorb 108. The column temperature

was 55 �C. Helium was used as carrier gas with a flow rate of

40 ml/min. The gas samples (0.5 mL) were injected in

duplicate.

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1 2 3 4 5

Number of repeated batch

Yie

ld (

mol

-H2/

mol

-xyl

ose c

onsu

med

)

0.00

0.05

0.10

0.15

0.20

0.25

0.30

Residual xylose (g/l)

Hydrogen yieldResidual xylose

Fig. 2 – Hydrogen yield and residual xylose at different

transfers with 1 g/L initial xylose concentration in the

batch experiments cultivated with BA medium without

yeast extract addition for the adaptation study.

w a t e r r e s e a r c h 4 3 ( 2 0 0 9 ) 1 4 1 4 – 1 4 2 4 1417

The VFAs and alcohol were measured using a GC (Hewlett

Packard, HP 5890 series II) equipped with a flame ionization

detector (FID) and HP FFAP column (dimensions

30 m� 0.53 mm� 1.0 mm). The temperature program for the

column was increased from 50 �C to 190 �C with a rate of 15 �C/

min. The temperatures of injection port and detector were 200

and 150 �C, respectively. Nitrogen was used as the carrier gas

at a flow rate of 10 ml/min. Samples were centrifuged at

12,000 rpm for 10 min and acidified with 34% of H3PO4 with

a ratio of 30 mL per mL of sample.

Lactate and formate were analyzed by suppressed ion

exclusion chromatography equipped with a HPLC pump L2100

HITATHI, a HPLC auto-sampler L2200 HITATHI, a suppressor

Dionex AMMS-IEC2, a column ICE-AS1 (9� 250 mm),

a conductivity Detector Waters 432, a prefilter Rheodyne

0.5 mm� 3 mm and a column heating (35 �C). The flow rates of

the eluent (4 mM of heptafluorobutyric acid solution) and the

suppressor (25 mM of tetrabutylammonium hydroxide solu-

tion) were 50 mL/h and 30 mL/h respectively. Samples were

filtered through 0.45 mm nylon filter (Pall Corporation) and

acidified with 17% H3PO4 with a ratio of 50 mL per ml of

samples.

Xylose was analyzed by anthrone-sulfuric acid method

(Dubois et al., 1956), using a UV/VIS spectrometer (Perkin

Elmer Lammda2).

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

40 1000 20 60 80 120

Time (Hr)

H2

prod

ucti

on s

peci

fic

to a

mou

nt o

f xy

lose

add

ed(m

ol-H

2/m

ol-x

ylos

e add

ed)

0.5 g/L1 g/L2 g/L3 g/L4 g/L

Fig. 3 – Hydrogen production with respect to xylose added

from batch experiments at different initial xylose

concentrations under extreme thermophilic temperatures

(70 8C).

3. Results and discussion

3.1. Enrichment of H2 producing bacteria with xylose atextreme thermophilic temperatures

The bacteria collected from a CSTR fed with household solid

wastes at 70 �C were first adapted to the new substrate of

xylose at the same temperature. The initial concentration of

xylose was fairly low (0.25 g/L) for ensuring low risk of over-

loading (inhibition) and high microbial activity (Angelidaki and

Sanders, 2004). The hydrogen production yield from the first

generation was around 1.40� 0.02 mol-H2/mol-xyloseconsumed,

which was around 42% of the theoretical value (3.33 mol-H2/

mol-xylose). The yield of hydrogen production was increased

slightly to 1.46� 0.17 mol-H2/mol-xyloseconsumed at higher

xylose concentration of 0.5 g/l. The hydrogen producing

bacteria were further enriched at higher xylose concentration

of 1.0 g/L (Fig. 2). In the first transfer with 1.0 g/L of xylose, the

hydrogen yield was 0.68� 0.04 mol-H2/mol-xyloseconsumed,

which was lower than at 0.25 and 0.5 g/L of xylose. Repeated

batch cultivations were conducted by transferring the cultures

from the previous batch to the new medium containing the

same xylose concentration (1.0 g/L). In the third transfer, the

hydrogen yield was increased up to 1.05� 0.07 mol-H2/mol-

xyloseconsumed. No further increase of the hydrogen yield was

observed by additional successive transfers. This result indi-

cates that bacteria can adapt to new conditions (substrate,

medium) by repeated batch cultivations so that increased

hydrogen yield can be obtained. Similar observation has been

found by Imachi et al. (2000) reporting that successful bacteria

cultivation with propionate was achieved by more than 10

times repeated transfers.

3.2. Hydrogen production at different initial xyloseconcentrations

Hydrogen production and xylose removal were investigated in

the batch cultivations at different initial xylose concentra-

tions of 0.5, 1.0, 2.0, 3.0, and 4.0 g/L with BA medium amended

with 1 g/L of yeast extract (Fig. 3 and Table 1). For all condi-

tions, hydrogen was immediately produced and could reach

stationary phase within 66 h. Addition of 1 g/L yeast extract

resulted in less deviation of the hydrogen production between

the triplicate vials, and at the same time resulted in the

hydrogen yield of 1.46 mol-H2/mol-xyloseconsumed at xylose

concentration of 1 g/L (Table 1), which was significantly higher

than the highest yield of 1.05 mol-H2/mol-xyloseconsumed

(Fig. 2) achieved at the same initial xylose concentration,

without yeast extract addition. The positive effect of yeast

extract on microbial growth has been widely known (van Niel

et al., 2002; Xu et al., 2007) and is due to its content of proteins,

carbohydrates, and salts (Eurasyp, 2008).

Hydrogen production specific to amount of xylose added

(Fig. 3) depends on both the degree of substrate conversion,

and also on the metabolic conversion pathway. Although this

value doesn’t distinguish between metabolic conversion and

substrate utilization degree, it gives important information

Table 1 – Xylose balance and hydrogen formation during batch enrichment at stationary phase of different initial xyloseconcentrations.

Xylose Conc. (g/L) Xylose balance Yield (mol-H2/mol-xyloseconsumed)

pH Hydrogen partialpressure (atm)

Input (mg) Remain (mg) Consumed (mg) Removal (%)

0.5 10 0.46 9.54 95.42 1.62� 0.02 6.78 0.07

1.0 20 1.48 18.85 94.25 1.46� 0.07 6.69 0.12

2.0 40 2.26 37.74 93.51 0.84� 0.05 6.11 0.13

3.0 60 22.68 37.32 62.20 0.62� 0.04 5.98 0.10

4.0 80 44.72 35.28 44.10 0.45� 0.04 5.92 0.07

The cultivations were conducted with BA medium with 1 g/L of yeast extract addition.

w a t e r r e s e a r c h 4 3 ( 2 0 0 9 ) 1 4 1 4 – 1 4 2 41418

about the potential of a substrate to release a specific amount

of hydrogen. Expression of hydrogen production per amount

of substrate added is often used to describe hydrogen

production efficiency (Han and Shin, 2004; Ueno et al., 2007;

Liu et al., 2008b).

A nearly complete degradation of xylose (w95%) was

observed at the initial xylose concentration between 0.5 g/l

and 2.0 g/l, and the highest hydrogen yield was

1.62� 0.02 mol-H2/mol-xyloseconsumed, occurred at 0.5 g/L

xylose concentration. The lower hydrogen yield achieved in

the vials with higher initial xylose concentration, was prob-

ably due to an increase of hydrogen partial pressure (0.07 and

0.13 atm at initial xylose concentrations of 0.5 and 2.0 g/L,

respectively). Dark fermentative hydrogen production is

thermodynamically limited by high hydrogen partial pressure

(Angenent et al., 2004). The NADH, which is an electron carrier

in the cell, will be oxidized with the production of other by-

products such as lactate or ethanol than hydrogen at higher

hydrogen partial pressure. van Niel et al. (2003) reported that

the lactate production started at hydrogen partial pressure

higher than 0.10 atm.

At initial xylose concentration of 3 and 4 g/L, however, the

xylose was not completely removed (62 and 44% removal,

respectively), and H2 production was lower than that obtained

at initial xylose concentrations of 2 g/L. This might be caused

by the by-products during the fermentation process. High

initial substrate concentration has lead to accumulation of

organic products (acetate, ethanol, and lactate) which has

probably resulted in unfavourable thermodynamic state that

prevented further substrate degradation (Rodriguez et al.,

Table 2 – Product concentrations and COD balance of batch ex

Concentration (mM) COD (mg

Xylose con. (g/L) 0.5 1.0 2.0 3.0 4.0 0.5 1.0 2

Consumed Xylose 3.18 6.28 12.47 12.44 11.76 �508.92 �1005.3 �1994

Acetate 2.62 4.93 9.35 9.06 8.96 167.90 315.77 598

Ethanol 1.14 2.39 4.50 4.56 4.39 109.13 229.91 431

Propionate 0.09 0.19 0.32 0.35 0.35 10.19 21.07 35

Lactate 0.48 1.19 3.51 4.18 4.01 46.46 113.94 337

Formate 0.13 0.52 0.92 0.92 0.99 2.07 8.40 14

Hydrogen 5.15 9.17 10.43 7.65 5.30 82.40 146.64 166

Biomass 0.67 1.33 2.65 2.64 2.50 76.34 150.79 299

Balance �14.44 �18.77 �110

a Assumed value: according to the previous study (Kotsopoulos et al., 20

2006). The relatively low H2 partial pressure measured in our

experiments when the initial xylose concentration was higher

than 3 g/L might be due to low hydrogen yielding metabolic

pathways, such as lactate, ethanol resulting pathways,

implying that low hydrogen yielding reactions or even

hydrogen consuming reactions might have occurred. Alter-

natively, bicarbonate might have been neutralized with

organic acids produced during fermentation to form CO2 and

lowered the hydrogen concentration (Zhang et al., 2008). The

release of H2/CO2 to the headspace is more favorable at lower

pH (Kleerebezem et al., 2008).

Acetate was the main metabolic intermediate by-product

followed by ethanol and lactate (Table 2). Yokoyama et al.

(2007) reported that acetate, ethanol, and lactate were the

main metabolites during xylose fermentation by mixed

culture enriched from cow manure at 75 �C. In other studies,

acetate and lactate were found as the main intermediate

products during the fermentation by Caldicellulosiruptor sac-

charolyticus at 70 �C temperature of sucrose (van Niel et al.,

2003) and of xylose (Kadar et al., 2004). Only minor amounts of

propionate and formate were detected in this study, which is

positive for optimal hydrogen production. When sugar is

metabolized through propionate, hydrogen is consumed

(equation (3)) (Antonopoulou et al., 2008), and metabolism of

sugar to formate results in no hydrogen production (equation

(4)). Formate which is an alternative channel for electron

carrier was produced instead of hydrogen and thus had

negative impact on hydrogen yield (Sparling et al., 2006).

However, no butyrate was detected from our batch experi-

ments, indicating that hydrogen fermentation from xylose

periments at different initial xylose concentrations.

-COD/L) COD distribution (%)

.0 3.0 4.0 0.5 1.0 2.0 3.0 4.0

.90 �1990.43 �1881.71 �100 �100 �100 �100 �100

.60 579.72 573.16 32.99 31.41 30.01 29.13 30.46

.96 443.68 421.08 21.44 22.87 21.65 22.29 22.38

.95 39.68 39.06 2.00 2.10 1.80 1.99 2.08

.34 401.36 385.24 9.13 11.33 16.91 20.16 20.47

.68 14.70 15.87 0.41 0.84 0.74 0.74 0.84

.95 122.64 84.83 16.19 14.59 8.37 6.16 4.51

.23 298.56 282.26 15.00a 15.00 15.00 15.00 15.00

.19 �90.08 �80.21 �2.84 �1.87 �5.52 �4.53 �4.26

06).

0

0.02

0.04

0.06

0.08

0.1

0.12

0.14

0.16

0 0.5 1 1.5 2 2.5 3 3.5 4 4.5

Initial xylose concentration (g/l)

Spec

ific

gro

wth

rat

e (h

r-1)

Experimental

Simulated with xylose

Monod eq.

Fig. 4 – Kinetic analysis for specific growth rates at different

initial xylose concentrations of batch cultivations.

w a t e r r e s e a r c h 4 3 ( 2 0 0 9 ) 1 4 1 4 – 1 4 2 4 1419

through butyrate was unfavorable at extreme thermophilic

temperatures.

C5H10O5 þ 1:67H2�!1:67C3H5O�2 þ 1:67Hþ þ 1:67H2ODGo ¼ �315:1 KJ=mole

(3)

C5H10O5 þHCO�3�!2C2H3O�2 þ 2CHO�2 þ 3Hþ

DGo ¼ �223:1 KJ=mole(4)

COD balance can be used as a tool to check the accuracy

of the analyses and suggests existing metabolic pathways

in anaerobic digestion systems. The COD balance for each

xylose concentration showed relatively good recovery with

maximum error of 6% (Table 2), which suggests that the

measurements of degraded metabolic products from xylose

were accurate. The COD distribution for lactate was

significantly increased when xylose concentration was

increased from 0.5 to 3 g/L, while the hydrogen formation

was decreased. At this time, the COD distribution for

acetate and ethanol was almost constant. The lower yield

of hydrogen during our batch experiments was therefore

due to xylose fermentation through lactate. When the

metabolic pathway of acetate and ethanol via acetyl-CoA

and hydrogen production was saturated, the remaining

NADH should be oxidized with the production of lactate or

propionate via non acetyl-CoA pathway in order to regen-

erate NADþ, which is needed to proceed the glycolysis

process (Temudo et al., 2007).

The highest hydrogen yield (1.62� 0.02 mol-H2/mol-

xyloseconsumed) in this study was lower than the value

(2.24 mol-H2/mol-xyloseconsumed) achieved by pure cultures

of Caldicellulosiruptor saccharolyticus during the fermentation

with xylose concentration of 10 g/L (Kadar et al., 2004). This

could be due to utilization of the substrate to no hydrogen

yielding reactions by some of the bacteria contained in the

mixed culture. Although the yield obtained by mixed

cultures might be lower compared to optimal achieved by

some pure cultures, other advantages such as risk of

contamination of pure cultures, operation complexity etc.,

might overweight the lower yield by mixed culture. Mean-

while, the hydrogen yields obtained from our batch experiments

are comparable with the hydrogen yield (0.56 mol-H2/mol-

xyloseconsumed) reported by Yokoyama et al. (2007), when

the mixed culture were used for xylose fermentation (con-

centration¼ 5.8 g/L).

3.3. Kinetic analysis of xylose fermentation

Xylose concentrations higher than 2 g/L might inhibit some

microorganisms involved in xylose degradation, resulting in

lower specific growth rates (Fig. 4). The specific growth rates

were measured at the initial phase (first 30–40 h), where the

accumulation of intermediate product was still low, indicating

that product inhibition was probably not the main reason for

the observed lower rates. Similarly, van Niel et al. (2003) and

Liu et al. (2008b) have previously reported substrate inhibition

for hydrogen dark fermentation. Kinetic analysis was there-

fore performed with Haldane inhibition equation (equation

(5)) which was used, based on the assumption, that substrate

inhibition was occurring.

m ¼ mmax

1þ Ks

Sþ I

Ki

(5)

where m (h�1) is the specific growth rate, mmax (h�1) is the

maximum specific growth rate in the absence of inhibition, S

(g/L) is the limiting substrate concentration, Ks (g/L) is the

saturation constant: numerically equal to the concentration of

substrate at which the specific growth is equal to half of the

maximum specific growth in the absence of inhibition, I (g/L)

is the inhibitor concentration, and Ki (g/L) is the inhibition

constant: numerically equals to the inhibitor concentration at

which the specific growth rate is equal to half the maximum

specific growth in the absence of inhibition.

In case of fixed product stoichiometry, growth rate can be

relatedtoproduct formation.Althoughthis isnot entirely thecase

for hydrogen fermentation, because the process is more compli-

cated, fixed stoichiometry assumed in ADM1 model is fairly

accurate in intensive growth periods, such as the initial phase

batch hydrogen fermentation (Lin et al., 2007; Peiris et al., 2006).

Our experiments were conducted in batch-mode operation,

where the conditions were relatively constant (such as same

inoculum, temperature, etc.) and could be an acceptable approx-

imation of constant hydrogen to biomass yield (Zheng et al., 2008).

Assuming production of H2 was proportional to cell growth

in the initial phase, growth rates were found from the linear

part of a semilogarithmic plot of the cumulative volume of

hydrogen against time (exponential phase, which was occur-

ring the first 30–50 h of fermentation). The correlation coeffi-

cient from each plot was generally 0.94 or higher. The

simulated rates achieved by function solver in Microsoft excel

2003 using xylose as inhibitor (Haldane equation) and no

inhibitor (Monod equation) and the measured rates are shown

in Fig. 4. We found good agreement between modeled rates and

measured rates when initial xylose concentration was used as

the inhibitor. However, Monod model, assuming no inhibition

gave higher modeled rates at the higher substrate concentra-

tions, and failed to predict the tendency of decreasing rates

with increasing xylose concentration higher than 0.5 g/L. This

supports that substrate inhibition was taking place. Of course

this doesn’t exclude product inhibition later in the batch

growth phase. The kinetic parameters in the Haldane equation

(substrate inhibition) were estimated to: mmax¼ 0.17 h�1,

Ks¼ 0.75 g/L, and Ki¼ 3.72 g/L of initial xylose concentration.

The Haldane kinetics has been shown to successfully model

inhibition in general (Majizat et al., 1997; Zheng et al., 2008).

0

2

4

6

8

10

0 50 100 150 200 250 300 350 400

Time (Hr)

Acc

umul

atio

n hy

drog

en v

olum

e (m

l) 0.5 g/L

1 g/L

2 g/L

3 g/L

4 g/L

Xylose re-feeding

Headspace sparging

Initial batch cultivation

Fig. 5 – Cumulative hydrogen production in the batch experiments at different initial xylose concentrations during regular

batch experiments, xylose re-feeding and headspace sparging.

w a t e r r e s e a r c h 4 3 ( 2 0 0 9 ) 1 4 1 4 – 1 4 2 41420

3.4. Effect of headspace sparging with N2 on thefermentation process

Additional amount of xylose at the same concentration as in

the previous cultivation was added to each batch bottle after

the hydrogen production ceased after approx. 119 h (Fig. 5).

In the vials where the initial xylose concentration was 0.5–2 g/

0.0

0.5

1.0

1.5

2.0

2.5

Time (d

Yie

ld (

mol

-H2/

mol

-xyl

ose c

onsu

med

)

0.0

1.0

2.0

3.0

4.0

5.0

0 2 4 6 8 10 12 14 16

0 2 4 6 8 10 12 14 16

Time (d

Met

abol

ic p

rodu

cts

conc

entr

atio

n (m

M)

AA BA

Fig. 6 – Profiles of CSTR performance at 70 8C, 72-h HRT and

iso-butyrate; ETOH: ethanol; BTOH: butanol; LA: lactate; PA:

L, more hydrogen was produced after additional substrate was

added suggesting that the previous cultivation was limited

with xylose. The highest yield of hydrogen production at 0.5 g/

L xylose was 1.61� 0.04 mol-H2/mol-xyloseadded, but it was

decreased at higher xylose concentrations. This is in agree-

ment with the observations during the first cultivation.

However, no more hydrogen production even after re-feeding

ay)

0

10

20

30

40

50

60

70

80

90

100

% X

ylose degradation

Yield % Xylose degradation

18 20 22 24 26 28 30 32

18 20 22 24 26 28 30 32

ay)

ETOH BTOH LA FA PA

feeding xylose concentration 1 g/L. AA: acetate; BA: n-&

propionate; FA: formate.

Table 3 – Main possible metabolic reactions of xylose degradation under extreme thermophilic temperatures in CSTR.

Reaction DGo (kJ/mol) DG0(kJ/mol)

C5H10O5 þ 1:67H2O�!1:67C2H3O�2 þ 1:67Hþ þ 1:67CO2 þ 3:33H2 (1) �195.52 �293.59

C5H10O5 þ 0:83H2O�!0:83C2H3O�2 þ 0:33CHO�2 þ 1:16Hþ þ 0:83C2H6Oþ 1:33H2 þ 1:33CO2 (2) �196.72 �271.77

C5H10O5 þ 0:83H2O�!0:83C2H3O�2 þ 0:83Hþ þ 0:83C2H6Oþ 1:67H2 þ 1:67CO2 (3) �201.83 �258.42

DGo was calculated from the Gibbs energy of formation of the compounds participating in the reaction.

DG0was calculated under the following conditions: T¼ 70 �C, pH¼ 6.7, pH2¼ 0.31 atm, pCO2¼ 0.69 atm, [C5H10O5]¼ 6.7 mM, [C2H3O2

�]¼ 4.25 mM,

[C2H6O]¼ 2.59 mM, [CHO2�]¼ 3.1 mM, [HCO3

�]¼ 31 mM.

w a t e r r e s e a r c h 4 3 ( 2 0 0 9 ) 1 4 1 4 – 1 4 2 4 1421

xylose into the bottles where higher amount of xylose (3 and

4 g/L) was initially added suggests that the previous cessation

of hydrogen production was caused by either substrate or

intermediates inhibition.

The headspace of the vials was sparged with N2 for 3 min, at

187 h after the initiation of the experiment (Fig. 5). The

hydrogen partial pressure in the headspace of the vials before

nitrogen sparing was 0.14, 0.15, and 0.16 atm (corresponding to

hydrogen liquid phase concentrations of 0.11, 0.12, and

0.13 mM) for the vials with 0.5, 1.0, and 2.0 g/L initial xylose

concentrations, respectively. Just after headspace sparging,

the hydrogen partial pressure was reduced to undetectable

levels. After the headspace sparging, the hydrogen partial

pressure was resumed and reached to 0.02, 0.07, and 0.03 atm

from the additional 0.5, 1, and 2 g/L xylose respectively, indi-

cating there was high hydrogen partial pressure limiting

hydrogen production. During stable hydrogen production after

sparing, the calculated hydrogen concentration in liquid phase

was less than 0.05 mM. These results indicate that headspace

flushing decreasedthe dissolvedH2 concentration so that more

hydrogen could be produced. It has been previously reported

that thermophilic hydrogen producing microorganisms could

be inhibitedby very low hydrogen partial pressure (0.016 atm to

0.75 atm) (Valdez-Vazquez et al., 2006). In our study, headspace

partial pressure of hydrogen higher than 0.14 atm could inhibit

the hydrogen producing microorganisms. However, hydrogen

production was not resumed after sparging at xylose concen-

tration of 4.13 and 6.24 g/L in the vials (xylose was re-added

with 3 and 4 g/L respectively), indicating that other limiting

factors such as substrate or liquid by-products inhibition were

involved in hydrogen production from xylose.

3.5. Continuous hydrogen production in CSTR operation

Continuous hydrogen production from xylose was success-

fully demonstrated in CSTR at extreme thermophilic

temperatures (70 �C) using mixed culture which were previ-

ously enriched with xylose in the batch experiment. The CSTR

reactor was operated at 72 h HRT, and it was continuously fed

with 1 g/L xylose. The profile of hydrogen yield, extend of

xylose degradation, and concentration of intermediate prod-

ucts during 31-day operation are shown in Fig. 6. Average

xylose degradation during the operation was 89.1� 2.0%, and

the achieved hydrogen yield ranged from 0.76 to 2.2 mol-H2/

mol-xyloseconsumed depending on the intermediate products

formed. The reactor pH was 5.75, 6.83 and 6.71 on the

operating day 5, 16, and 27, respectively, on which the meta-

bolic products concentration was also different. During the

last week of operation (day 26–31), the average hydrogen yield

was 1.36� 0.03 mol-H2/mol-xyloseconsumed corresponding to

the hydrogen productivity of 62� 2 mL-H2/d$Lreactor. The

mixed gas was methane free, and it was composed of

hydrogen and carbon dioxide at approx. 31� 1 and 69� 1%,

respectively. The main soluble metabolites were acetate,

formate, and ethanol at 4.25� 0.10, 3.01� 0.11, and

2.59� 0.16 mM, respectively. Lin and Cheng (2006) reported

a lower hydrogen yield of 0.7 mol-H2/mol-xyloseconsumed at

mesophilic temperatures (37 �C) using mixed culture.

However, similar yield as the ones from this study have been

found by others such as 1.4 mol-H2/mol-xyloseconsumedd (Lin et al.,

2008) and 1.0 mol-H2/mol-xyloseconsumed (Wu et al., 2008) at ther-

mophilic temperatures (50 �C).

Maximum hydrogen yield from xylose fermentation can

be generally obtained with acetate and butyrate according to

equations (1) and (2). However, the mixed culture fermen-

tation can provide a variety of end products depending on

the environmental conditions such as hydrogen partial

pressure, reactor pH, substrate concentration, temperature,

etc. (Rodriguez et al., 2006; Temudo et al., 2007). In this study,

it is shown that different metabolic pathways were followed

depending on the pH of the system. Rather high hydrogen

yield could be obtained with acetate as the main metabolic

product, while other by-products were produced in lower

amounts when the pH was around 5.75. However, when the

pH was higher around 6.83, significant increase of formate

and ethanol was observed, leading to relatively low hydrogen

yield. Moreover, the concentrations of butyrate, lactate,

butanol, and propionate became less dominant as the pH

was increased from 5.75 to 6.83. Other studies have also

demonstrated that pH was an important parameter deter-

mining the metabolic pathway for degradation of sugars.

Voolapalli and Stuckey (2001) and Sparling et al. (2006) have

shown that changing the system pH from 6.5 to 7 resulted in

significant formate production. The composition and

concentration of by-products from the effluent of CSTR were

different from that in the batch experiments. This might be

explained by a microbial community change during opera-

tion of the CSTR reactors. Additionally, contrary to the

continuous reactor, batch reactor permits accumulation of

products (van Niel et al., 2003), which could lead to different

conditions and subsequently induce different metabolic

pathways. Moreover, the static batch reactor might have

Table 4 – Product concentrations and COD balance ofCSTR at steady state conditions (day 26–31).

Concentration(mM)

COD(mg-COD/L)

%CODdistribution

Xylose removal 6.1� 0.03 �969.00� 5.61 �100

Acetate 4.25� 0.10 272.00� 4.80 28.07

Butyrate 0.22� 0.10 34.45� 11.22 3.55

Ethanol 2.6� 0.16 249.14� 6.62 25.71

Butanol 0.21� 0.0.02 40.01� 3.06 4.13

Propionate 0.14� 0.02 15.63� 2.30 1.61

Lactate 0.66� 0.03 15.04� 2.95 1.55

Formate 3.01� 0.11 48.22� 1.73 4.98

Hydrogen 8.27� 0.22 132.26� 3.61 13.65

Biomass 1.29� 0.007 145.35� 0.84 15a

Balance �16.91 �1.75

a Assumed value: according to the previous study (Kotsopoulos

et al., 2006).

w a t e r r e s e a r c h 4 3 ( 2 0 0 9 ) 1 4 1 4 – 1 4 2 41422

poor mass transfer efficiency due to lack of mechanical

shaking (Lo et al., 2008).

The main possible reactions for xylose fermentation based

on reduced products obtained from our CSTR experiment and

the calculated Gibbs free energy are shown in Table 3. From

a thermodynamic point of view, those reactions under

extreme thermophilic conditions are more favorable

compared to the corresponding reactions at mesophilic or

thermophilic conditions. Especially, equation (1) in Table 3 is

the most exergonic process (Gibbs free energy¼�293.59 kJ/

mol), and also in this study, the highest yield of

2.16� 0.05 mol-H2/mol-xyloseconsumed could be obtained

when acetate was the main metabolic product during the first

4 days of the CSTR operation. The COD balance during steady

state operations (day 26–31) (Table 4) indicate that carbon

from xylose fermentation under extreme thermophilic

temperatures was mainly converted to acetate, ethanol,

hydrogen, and formate (accounting for approx. 68% of total

products). A stoichiometric equation for hydrogen production

from xylose according to equation (2) is proposed (Table 3).

This stoichiometric equation indicates that a theoretical

hydrogen yield is 1.33 mol-H2/mol-xylose, which is similar to

actual values (1.36� 0.03 mol-H2/mol-xylose) in this experi-

ment. According to equation (3) in Table 3, the higher theo-

retical yield of 1.67 mol-H2/mol-xylose can be achieved with

the productions of hydrogen, acetate, and ethanol.

As aforementioned, the extra energy cost for the fermen-

tation operating at higher temperature is not much higher

compared to that operating at mesophilic temperature. This

could be compensated with the high temperature substrates

such as hydrolysate obtained from hydrothermal pretreat-

ment of lignocelluloses, where the temperature is already

high and additional heating would not be needed. Addition-

ally, for complete utilization of the residual organic matter

from xylose fermentation, a subsequent step such as methane

producing step should be attached, as only approx. 15% of the

energy content (carbon source) in xylose was utilized by the

hydrogen fermentation step. A combined hydrogen and

methane step could be advantageous, for production of

hythane, which is a mixture of hydrogen methane gas and has

been trademarked by Hydrogen Consultants Inc. (Gattrell

et al., 2007). Hydrogen is also a powerful combustion

stimulant for accelerating methane combustion. A mixture of

20% hydrogen and 80% methane will significantly reduce the

emission of CO, CO2 and NOx of natural gas powered vehicles

(Alavandi and Agrawal, 2008).

4. Conclusions

Hydrogen producing culture obtained from a lab scale CSTR

reactor fed with household solid wastes at 70 �C could be

adapted by repeated batch fermentations to hydrogen

production from xylose at 70 �C. During the cultivations at

different initial xylose concentrations with BA medium

amended with 1 g/L of yeast extract, different hydrogen yields

achieved were depended on different accumulation of

hydrogen and intermediates mainly contained with acetate,

ethanol, and lactate. Furthermore, yeast extract could

enhance hydrogen yield during the batch xylose fermenta-

tion. Haldane model, which is usually used to describe

substrate inhibition cases, was able to fit the experimental

points relatively well. Intermittent headspace sparging with

N2 could enhance hydrogen production when hydrogen

partial pressure in the gas phase was higher than around

0.14 atm. Xylose fermentation for hydrogen production could

be successfully achieved in CSTR operated at 70 �C, 3-days

HRT, and influent xylose concentration of 1 g/L. Under steady

state conditions, the hydrogen yield and the production rate

were 1.36� 0.03 mol-H2/mol-xyloseconsumed and 62� 2 mL/

d$Lreactor, respectively. The main reduced products from the

reactor were acetate followed by formate and ethanol

respectively.

Acknowledgements

The authors would like to acknowledge the financial support

of the Ministry of Science and Technology of Thailand and the

Research and Innovation Council under the strategic research

program of Bio. REF. Project No. 2104-06-0004 and the STVF

Project No. 2058-03-0020. We also thank Marcos H. Crespo for

his help in the initial stage of the experiments.

r e f e r e n c e s

Alavandi, S., Agrawal, A., 2008. Experimental study of combustionof hydrogen–syngas/methane fuel mixtures in a porousburner. International Journal of Hydrogen Energy 33 (4), 1407–1415.

Angelidaki, I., Kongjan, P., Thomsen, M.H., Thomsen, A., 2007.Biorefinery for Sustainable Biofuel Production from EnergyCrops: Conversion of Lignocellulose to Bioethanol,Biohydrogen and Biomethane. In: Anaerobic Digestion 11thWorld Congress: Bio-energy for Our Future, Brisbane,Australia.

Angelidaki, I., Sanders, W., 2004. Assessment of the anaerobicbiodegradability of macropollutants. Reviews inEnvironmental Science and Biotechnology 3 (2), 117–129.

Angenent, L., Karim, K., Al-Dahhan, M., Wrenn, B., Domıguez-Espinosa, R., 2004. Production of bioenergy and biochemicals

w a t e r r e s e a r c h 4 3 ( 2 0 0 9 ) 1 4 1 4 – 1 4 2 4 1423

from industrial and agricultural wastewater. Trends inBiotechnology 22 (9), 477–485.

Antonopoulou, G., Gavala, H., Skiadas, I., Angelopoulos, K.,Lyberatos, G., 2008. Biofuels generation from sweet sorghum:fermentative hydrogen production and anaerobic digestionof the remaining biomass. Bioresource Technology 99 (1),110–119.

Calli, B., Schoenmaekers, K., Vanbroekhoven, K., Diels, L., 2008.Dark fermentative H2 production from xylose andlactosedeffects of on-line pH control. International Journal ofHydrogen Energy 33 (2), 522–530.

Datar, R., Huang, J., Maness, P., Mohagheghi, A., Czernik, S.,Chornet, E., 2007. Hydrogen production from the fermentation ofcorn stover biomass pretreated with a steam-explosion process.International Journal of Hydrogen Energy 32 (8), 932–939.

Dubois, M., Gilles, K., Hamilton, J., Rebers, P., Smith, F., 1956.Colorimetric method for determination of sugars and relatedsubstances. Analytical Chemistry 28 (3), 350–356.

Eurasyp, 2008. Yeast Extract. Available from: http://www.eurasyp.org/public.levure.extrait.screen.

Fang, H., Li, C., Zhang, T., 2006. Acidophilic biohydrogenproduction from rice slurry. International Journal of HydrogenEnergy 31 (6), 683–692.

Gattrell, M., Gupta, N., Co, A., 2007. Electrochemical reduction ofCO2 to hydrocarbons to store renewable electrical energy andupgrade biogas. Energy Conversion and Management 48 (4),1255–1265.

Han, H.K., Shin, H.S., 2004. Performance of an innovative two-stage process converting food waste to hydrogen andmethane. Journal of Air and Waste Management Association54, 242–249.

Hawkes, F., Hussy, I., Kyazze, G., Dinsdale, R., Hawkes, D., 2007.Continuous dark fermentative hydrogen production bymesophilic microflora: principles and progress. InternationalJournal of Hydrogen Energy 32 (2), 172–184.

Imachi, H., Sekiguchi, Y., Kamagata, Y., Ohashi, A., Harada, H.,2000. Cultivation and in situ detection of a thermophilicbacterium capable of oxidizing propionate in syntrophicassociation with hydrogenotrophic methanogens ina thermophilic methanogenic granular sludge. Applied andEnvironmental Microbiology 66 (8), 3608.

Iqnacio, B., Ma, J.N., Jose, M.O., Araceli, C., Paloma, M.,Mercedes, B., 2006. Ethanol production from steam-explosionpretreated wheat straw. Applied Biochemistry andBiotechnology 130 (1), 496–508.

Kadar, Z., de Vrije, T., van Noorden, G., Budde, M., Szengyel, Z.,Reczey, K., Claassen, P., 2004. Yields from glucose, xylose, andpaper sludge hydrolysate during hydrogen production by theextreme thermophile Caldicellulosiruptor saccharolyticus.Applied Biochemistry and Biotechnology 114 (1), 497–508.

Kalia, V., Lal, S., Ghai, R., Mandal, M., Chauhan, A., 2003. Mininggenomic databases to identify novel hydrogen producers.Trends in Biotechnology 21 (4), 152–156.

Kapdan, I., Kargi, F., 2006. Biohydrogen production from wastematerials. Enzyme and Microbial Technology 38 (5), 569–582.

Kleerebezem, R., Temudo, M., Loosdrecht, M., Rodrıguez, J., 2008.Modeling mixed culture fermentations; the role of differentelectron carriers. Water Science and Technology: A Journal ofthe International Association on Water Pollution Research 57(4), 493.

Kleerebezem, R., van Loosdrecht, M., 2007. Mixed culturebiotechnology for bioenergy production. Current Opinion inBiotechnology 18 (3), 207–212.

Kotsopoulos, T., Zeng, R., Angelidaki, I., 2006. Biohydrogenproduction in granular up-flow anaerobic sludge blanket(UASB) reactors with mixed cultures under hyper-thermophilic temperature (70 �C). Biotechnology andBioengineering 94 (2), 296–302.

Lin, C., Cheng, C., 2006. Fermentative hydrogen production fromxylose using anaerobic mixed microflora. InternationalJournal of Hydrogen Energy 31 (7), 832–840.

Lin, C., Hung, C., Chen, C., Chung, W., Cheng, L., 2006. Effects ofinitial cultivation pH on fermentative hydrogen productionfrom xylose using natural mixed cultures. ProcessBiochemistry 41 (6), 1383–1390.

Lin, C., Wu, C., Hung, C., 2008. Temperature effects on fermentativehydrogen production from xylose using mixed anaerobiccultures. International Journal of Hydrogen Energy 33 (1), 43–50.

Lin, P., Whang, L., Wu, Y., Ren, W., Hsiao, C., Li, S., Chang, J., 2007.Biological hydrogen production of the genus Clostridium:metabolic study and mathematical model simulation.International Journal of Hydrogen Energy 32 (12), 1728–1735.

Liu, D., Zheng, R., Angelidaki, I., 2008a. Effects of pH and hydraulicretention time on hydrogen production versusmethanogenesis during anaerobic fermentation of organichousehold solid waste under extreme-thermophilictemperature (70 �C). Biotechnology and Bioengineering 100 (6),1108–1114.

Liu, D., Zheng, R., Angelidaki, I., 2008b. Enrichment andadaptation of extreme-thermophilic (70 �C) hydrogenproducing bacteria to organic household solid waste byrepeated batch cultivation. International Journal of HydrogenEnergy, doi:10.1016/j.ijhydene.2008.08.014.

Lo, Y., Chen, W., Hung, C., Chen, S., Chang, J., 2008. Dark H2

fermentation from sucrose and xylose using H2-producingindigenous bacteria: feasibility and kinetic studies. WaterResearch 42 (4–5), 827–842.

Lu, J., Gavala, H., Skiadas, I., Mladenovska, Z., Ahring, B., 2008.Improving anaerobic sewage sludge digestion byimplementation of a hyper-thermophilic prehydrolysis step.Journal of Environmental Management 88 (4), 881–889.

Majizat, A., Mitsunori, Y., Mitsunori, W., Michimasa, N.,Jun’ichiro, M., 1997. Hydrogen gas production from glucoseand its microbial kinetics in anaerobic systems. Water Scienceand Technology 36 (6–7), 279–286.

Peiris, B., Rathnasiri, P., Johansen, J., Kuhn, A., Bakke, R., 2006.ADM 1 simulations of hydrogen production. Water Scienceand Technology 53 (8), 129–137.

Rodriguez, J., Kleerebezem, R., Lema, J., van Loosdrecht, M., 2006.Modeling product formation in anaerobic mixed culturefermentations. Biotechnology and Bioengineering 93, 592–606.

Sparling, R., Islam, R., Cicek, N., Carere, C., Chow, H., Levin, D.,2006. Formate synthesis by Clostridium thermocellum duringanaerobic fermentation. Canadian Journal of Microbiology 52(7), 681–688.

Temudo, M., Kleerebezem, R., van Loosdrecht, M., 2007. Influence ofthe pH on (open) mixed culture fermentation of glucose:a chemostat study. Biotechnology and Bioengineering 98 (1), 69.

Thomsen, A.B., Thygesen, A., Bohn, V., Nielsen, K.V., Pallesen, B.,Jorgensen, M.S., 2006. Effects of chemical–physical pre-treatment processes on hemp fibres for reinforcement ofcomposites and for textiles. Industrial Crops and Products 24(2), 113–118.

Ueno, Y., Fukui, H., Goto, M., 2007. Operation of a two-stagefermentation process producing hydrogen and methane fromorganic waste. Environmental Science and Technology 41 (4),1413–1419.

Valdez-Vazquez, I., Rios-Leal, E., Carmona-Martinez, A., Munoz-Paez, K., Poggi-Varald, H., 2006. Improvement of biohydrogenproduction from solid wastes by intermittent venting and gasflushing of batch reactors headspace. Environmental Scienceand Technology 40 (10), 3409–3415.

van Groenestijn, J., Hazewinkel, J., Nienoord, M., Bussmann, P., 2002.Energy aspects of biological hydrogen production in high ratebioreactors operated in the thermophilic temperature range.International Journal of Hydrogen Energy 27 (11), 1141–1147.

w a t e r r e s e a r c h 4 3 ( 2 0 0 9 ) 1 4 1 4 – 1 4 2 41424

van Niel, E., Budde, M., de Haas, G., van der Wal, F., Claassen, P.,Stams, A., 2002. Distinctive properties of high hydrogenproducing extreme thermophiles, Caldicellulosiruptorsaccharolyticus and Thermotoga elfii. International Journal ofHydrogen Energy 27 (11–12), 1391–1398.

van Niel, E., Claassen, P., Stams, A., 2003. Substrate and productinhibition of hydrogen production by the extremethermophile, Caldicellulosiruptor saccharolyticus. Biotechnologyand Bioengineering 81 (3), 255–262.

Voolapalli, R., Stuckey, D., 2001. Hydrogen production inanaerobic reactors during shock loads-influence of formateproduction and H2 kinetics. Water Research 35 (7), 1831–1841.

Wu, K., Chang, C., Chang, J., 2007. Simultaneous production ofbiohydrogen and bioethanol with fluidized-bed and packed-bed bioreactors containing immobilized anaerobic sludge.Process Biochemistry 42 (7), 1165–1171.

Wu, S., Lin, C., Lee, K., Hung, C., Chang, J., Lin, P., Chang, F., 2008.Dark fermentative hydrogen production from xylose indifferent bioreactors using sewage sludge microflora. Energy &Fuels 22 (1), 113–119.

Xu, L., Ren, N., Wang, X., Li, Y., Zhang, Y., Xu, H., Chen, G., 2007.Nitrogen nutrient and metabolic characteristics of hydrogen-

producing novel species-Ethanoligenens harbinese B49.Chinese Journal of Applied and Environmental Biology 13 (2),265.

Yokoyama, H., Moriya, N., Ohmori, H., Waki, M., Ogino, A.,Tanaka, Y., 2007. Community analysis of hydrogen-producingextreme thermophilic anaerobic microflora enriched fromcow manure with five substrates. Applied Microbiology andBiotechnology 77 (1), 213–222.

Zhang, T., Liu, H., Fang, H., 2003. Biohydrogen production fromstarch in wastewater under thermophilic condition. Journal ofEnvironmental Management 69 (2), 149–156.

Zhang, Z., Show, K., Tay, J., Liang, D., Lee, D., 2008. Biohydrogenproduction with anaerobic fluidized bed reactorsdacomparison of biofilm-based and granule-based systems.International Journal of Hydrogen Energy 33 (5), 1559–1564.

Zheng, H., Zeng, R., Angelidaki, I., 2008. Biohydrogen productionfrom glucose in upflow biofilm reactors with plastic carriersunder extreme thermophilic conditions (70 �C). Biotechnologyand Bioengineering 100 (5), 1034–1038.

Zhu, Y., Yang, S., 2004. Effect of pH on metabolic pathway shift infermentation of xylose by Clostridium tyrobutyricum. Journal ofBiotechnology 110 (2), 143–157.