Caproate formation in mixed-culture fermentative hydrogen production

Bioresource Technology 101 (2010) 9550–9559

Contents lists available at ScienceDirect

Bioresource Technology journal homepage: www.elsevier.com/locate/biortech

Caproate formation in mixed-culture fermentative hydrogen production Hong-Bo Ding, Giin-Yu Amy Tan, Jing-Yuan Wang * School of Civil and Environmental Engineering, Nanyang Technological University, 50 Nanyang Avenue, Singapore 639798, Singapore

a r t i c l e

i n f o

Article history: Received 10 October 2009 Received in revised form 14 July 2010 Accepted 14 July 2010 Available online 17 July 2010 Keywords: Caproate Hexanoic acid Hydrogen production Clostridium kluyveri Valerate

a b s t r a c t Caproate always appears during fermentative H2 production but its formation was not well explained. It possibly results from the secondary fermentation of ethanol and acetate or butyrate by some special species like Clostridium kluyveri. This study attempts to elucidate caproate formation during the fermentation H2 production by using C. kluyveri as an example and evaluating several possible pathways of caproate formation. A detailed energetic analysis of the empirical data of an H2 -producing reactor demonstrated that caproate can be formed from two substrates, either ethanol and acetate or ethanol and butyrate. The analysis showed that at least 5 mol ethanol per mole reaction was essential to support caproate formation under the experimental condition. The analysis also indicated that the secondary fermentation by C. kluyveri might be another pathway to spontaneously produce H2 , butyrate, and acetate in addition to the butyrate-acetate pathway. Co-production of caproate and H2 from ethanol was thermodynamically feasible and contributed to at least 10–20% of total H2 production in the reactor studied. It is also clarified that caproate formation is hydrogenogenic rather than hydrogenotrophic. Ó 2010 Elsevier Ltd. All rights reserved.

1. Introduction Fermentative H2 production is a special case of acidogenesis in anaerobic digestion (AD). In this process, hydrogenotrophs such as methanogens are severely inhibited while hydrogenogenic acidogens are enriched to convert carbohydrates to H2 , volatile fatty acids (VFAs), and alcohols. Due to the accumulation of VFAs and alcohols and the inhibition of methanogens, some interesting phenomena emerge. One of these phenomena is formation of highcarbon VFAs such as caproate, valerate and heptanoate. The concentrations of caproate and valerate reported are summarized in Table 1. Significant amounts of caproate and valerate were always found in several granule-based upflow anaerobic sludge blanket reactors (Yu and Mu, 2006; Zhao and Yu, 2008; Zhao et al., 2008) and also in other granule-based reactors (Lee et al., 2004; Zhang et al., 2007). Caproate is a unique product of only a few species including Eubacterium alactolyticus, E. biforme, E. limosum and E. pyruvativorans, as well as Clostridium kluyveri, Peptococcus niger and Megasphaera elsdenii (Dworkin et al., 2006). Among all the caproate-producing species mentioned thus far, the spore-forming C. kluyveri appears to be the most possible species resistant to heat treatment or chemical treatment used for the enrichment of H2 producing bacteria. Additionally, ethanol produced under acidic

* Corresponding author. Tel.: +65 6790 4100; fax: +65 6792 7319. E-mail addresses: [email protected] (H.-B. Ding), [email protected] (G.-Y.A. Tan), [email protected] (J.-Y. Wang). 0960-8524/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.biortech.2010.07.056

environment in fermentative H2 production only favors the growth of C. kluyveri, while inhibiting the growth of other caproate-producing species, especially M. elsdenii (Dworkin et al., 2006). Caproate formation by C. kluyveri can be simply described by three coupled reactions in Eqs. (1)–(3) (Thauer et al., 1968; Schoberth and Gottschalk, 1969; Seedorf et al., 2008):

CH3 CH2 OH þ H2 O ¼ CH3 COO þ Hþ þ 2H2

DG00 r

¼ þ9:7 kJ mol

CH3 CH2 OH þ CH3 COO ¼ CH3 ðCH2 Þ2 COO þ H2 O

DG00 r

¼ 38:7 kJ mol ¼ þ38:7 kJ mol

ð2Þ

1

CH3 CH2 OH þ CH3 ðCH2 Þ2 COO ¼ CH3 ðCH2 Þ4 COO þ H2 O

DG00 r

ð1Þ

1

ð3Þ

1

Eq. (1) describes the dehydrogenation of one ethanol leading to ATP synthesis via substrate-level phosphorylation (SLP) and H2 formation. Eqs. (2) and (3) describe the dehydrogenation of another ethanol leading to formation of butyrate and caproate from ethanol and acetate and ethanol and butyrate, respectively. The combination of Eqs. (2) and (3) gives Eq. (4) which describes caproate formation via butyrate formation:

2CH3 CH2 OH þ CH3 COO ¼ CH3 ðCH2 Þ4 COO þ 2H2 O

DG00 r

¼ 77:5 kJ mol

ð4Þ

1

Valerate is formed in fermentation of ethanol and propionate according to the reaction in Eq. (5) which also has to be coupled with the reaction in Eq. (1):

9551

H.-B. Ding et al. / Bioresource Technology 101 (2010) 9550–9559 Table 1 Formation of caproate and valerate in the fermentative hydrogen production. Reactors

Growth mode

Substrate

HRT, h

pH

EtOH

HA

HB

HC

HP

HV

Units

References

UASB

Granule

Sucrose

13

6.1–6.5

11–12

27–33

30–38

15–18

6–10

0.4–2.1

% of mass of VFAs and ethanol

Zhao and Yu (2008)

Sucrose

3–6 10–14 12

8.0–8.5 4.4

15 8–11 7–8 380

48–51 10–11 5–10 1110

7–8 8–9 11–15 1500

4–5 0.6–1.0 0.4–0.7 100

11–12 0.8 0.6–1.3 160

mM

Yu and Mu (2006)

3.9

11–13 8 7–8 770

mg L1

Zhao et al. (2008)

500 3–5

1500 5–10

1060 15–25

250 N.A.

750 5–20

450 2–3

Lee et al. (2004)

28–49 20–25

18–33 6–17

18–32 10–15

11–23 5–6

1–2 2–3

Low N.A.

% of COD sucrose consumed mM mM

Sucrose CIGSB

Granule

Lactose Sucrose

UFBR AFBR

Biofilm Granule

Glucose Glucose

1–4

4.5 6–7

2–3 1–4

5–6 4

This study Zhang et al. (2007)

Note: UASB, upflow anaerobic sludge blanket; CIGSB, carrier-induced granular sludge beds; UFBR, upflow fixed-bed reactor; AFBR, anaerobic fluidized bed reactor; EtOH, ethanol; HC, caproate; HA, acetate; HB, butyrate; HP, propionate; HV, valerate.

CH3 CH2 OH þ CH3 CH2 COO ¼ CH3 ðCH2 Þ3 COO þ H2 O

DG00 r

ð5Þ

1

¼ 38:7 kJ mol

A comparison between the DG00 r of Eqs. (2) and (4) shows that the increment in carbon numbers (from butyrate to caproate) gives additional free energy. It implies that caproate formation can be spontaneous since it yields more energy than butyrate formation. Several other pathways (Reactions 1–5 in Table 2) were suggested for caproate formation in the fermentative H2 production (Yu and Mu, 2006; Zhao and Yu, 2008). The argument is that butyrate is the primary source of caproate formation pathways in Reactions 1–3, and H2 is the electron donor of Reactions 2–4 in Table 2. However, Reaction 1 in Table 2 is unlikely to happen due to its positive Gibbs free energy change of the reaction ðDG00 r Þ under the physiological condition (1 atm, pH 7, 25 °C, concentrations: 1 atm if the reactants are gases or 1 M if the reactants are solutes). The major difference between Reactions 2–4 in Table 2 and caproate formation pathways by C. kluyveri is the source of reducing equivalents. In C. kluyveri, NADH2 ðNADH þ Hþ Þ, the major carrier of reducing equivalents, is derived from ethanol dehydrogenation. As for Reactions 2–4 in Table 2, molecular H2 is directly be utilized. Ethanol would be a much better source of reducing equivalents than H2 to support high VFA synthesis and biomass growth (Kenealy and Waselefsky, 1985). Reaction 5 in Table 2 has ethanol and

acetate as reactants and is superficially the same as the reaction in Eq. (4). However, Reaction 5 in Table 2 gives a simplified picture and does not accurately describe the observed quantitative relations among the substrates and products because H2 formation is also observed along with caproate production (Schoberth and Gottschalk, 1969). Its occurrence for exergonic caproate formation has to be energetically coupled with endergonic ethanol dehydrogenation in Eq. (1) (Seedorf et al., 2008). In this study, the secondary fermentation pathways by C. kluyveri were used to explain caproate formation during the fermentative H2 production. Energetic analysis and thermodynamic efficiency were used to determine caproate formation pathways in an upflow fixed-bed reactor (UFBR). The mutual influence between caproate formation and H2 production was discussed. 2. Methods The UFBR had a working volume of 0.8 L with a diameter of 10 cm in diameter and a height of 25 cm. The cylindrical activated carbon pellets (3 mm in diameter and 7 mm in length) were used as biofilm carriers with a void ratio of 60–70%. The reactor was operated at 30  1  C. Seed sludge (45 g  VSS L1 and pH 7:8) was obtained from the methanogenic reactor of a pilot-scale two-phase AD system established on the campus of the university.

Table 2 Possible pathways for butyrate and caproate formation. No.

Pathways

Reaction 1

2CH3 ðCH2 Þ2 COO ¼ CH3 ðCH2 Þ4 COO þ CH3 COO

Combinations

Reaction 2

CH3 ðCH2 Þ2 COO þ CH3 COO þ 2H2 þ Hþ ¼ CH3 ðCH2 Þ4 COO þ 2H2 O

Reaction 3 Reaction 4 Reaction 5

CH3 ðCH2 Þ2 COO þ 2CO2 þ 6H2 ¼ CH3 ðCH2 Þ4 COO þ 4H2 O 3CH3 CH2 OH þ 4H2 þ 2Hþ ¼ CH3 ðCH2 Þ4 COO þ 4H2 O 2CH3 CH2 OH þ CH3 COO ¼ CH3 ðCH2 Þ4 COO þ 2H2 O

Eq. (6)

ðk þ 1ÞCH3 CH2 OH þ ðk  1ÞCH3 COO ¼ kCH3 ðCH2 Þ2 COO þ Hþ þ 2H2 þ ðk  1ÞH2 O

Eq. (7)

ðk þ 1ÞCH3 CH2 OH þ

Eq. (8)

ðk þ 1ÞCH3 CH2 OH þ kCH3 ðCH2 Þ2 COO ¼ kCH3 ðCH2 Þ4 COO þ CH3 COO þ Hþ þ 2H2 þ ðk  1ÞH2 O

k

2

  1 CH3 COO ¼ 2k CH3 ðCH2 Þ2 COO þ Hþ þ 2H2 þ ðk  1ÞH2 O

References

Remark

Yu and Mu (2006)

Unlike to occur H2 is Used as electron and proton donor

Eq. (1)+kEq. (2)

Schoberth and Gottschalk (1969)

Eq. (1)+k/ 2Eq. (4)

Eq. (1)+k Eq. (3)

This study

Incomplete expression C. kluyveri produces butyrate only from ethanol and acetate C. kluyveri produces caproate from ethanol and acetate. Butyrate is considered as an intermediate product. C. kluyveri produces caproate only from ethanol and butyrate.

9552

H.-B. Ding et al. / Bioresource Technology 101 (2010) 9550–9559

Table 3 Steps in butyrate and caproate formation by C. kluyveri. No.

Steps names

Reaction equations

A

Ethanol dehydrogenation

A1 A2 B

ATP formation via SLP

Overall reaction = A1 + A2 Ethanol þ 2NADþ þ CoA ! Acetyl-CoA þ 2ðNADH þ Hþ Þ Ethanol þ 2NADþ ! Acetyldehye-CoA þ ðNADH þ Hþ Þ Acetyldehye þ NADþ þ CoA ! Acetyl-CoA þ ðNADH þ Hþ Þ Overall reaction Acetyl-CoA þ ADP þ Pi ! Acetate þ ATP þ CoA þ Hþ Intermediate reactions Acetyl-CoA þ Pi ! Acetate þ Pi þ -CoA Acetyl-P þ ADP ! Acetate þ ATP þ Hþ

C1

Ferredoxin reduction

ðNADH þ Hþ Þ þ Fdox ! NADþ þ Fdred

C2

Hydrogen formation

Fdred þ 2Hþ ! Fdox þ H2

C3

NADH2 Regeneration and ATP formation via ETP

Fdred þ NADþ þ Hþ ! Fdox þ ðNADH þ Hþ Þ þ DlHþ

2

2 2

Additional ATP is concurrently synthesized by ETP ðDlHþÞ Overall reaction=D1+D2+D3+D4

D

2

2Acetyl-CoA þ ðNADðPÞH þ Hþ Þ þ 2ðNADH þ Hþ Þ þ Fdox ! Butyrul-CoA þ COA þ NADðPÞþ 2NADþ þ Fdred þ H2 O 2Acetyl-CoA ! Acetoacetyl-CoA þ CoA Acetoacetyl-CoA þ ðNADðPÞH þ Hþ Þ ! 3  Hydroxybutyryl-CoA þ NADðPÞþ 3  Hydroxybutyryl-CoA ! Crotonyl-CoA þ H2 O

D1 D2 D3 D4

Butyryl-CoA formation

E

Caproyl-CoA formation

2ðNADH þ Hþ Þ þ crotonyl-CoA þ Fdox ! 2NADþ þ butyryl-CoA þ Fdred Reaction equation is not shown here due to the lack of detailed information in literature

F

Butyrate formation

Butyryl-CoA þ Acetate ! Butyrate þ Acetyl-CoA

G

Caproate formation

Caproyl-CoA þ Acetate ! Caproate þ Acetyl-CoA

2

The AD system was used to treat food waste collected from the university canteens (Wang et al., 2008). The sludge was first acidified with 1 N HCl to pH 3–4, kept for 24 h, and then adjusted to neutral pH by 1 N NaOH before heat treatment at 100 °C for 2 h. The substrate was prepared daily with substrate (g L1 : 20 glucose), buffer ðg L1 : 4 NaHCO3 Þ, macronutrients (g L1 : 1.5 NH4 Cl and 1 NaH2 PO4 ), and micronutrients ðmg L1 : 22 FeCl2  4H2 O; 20 MgCl2  6H2 O;10 MnCl2  4H2 O;5 CoCl2  4H2 O;6 CaCl2  2H2 O; 3:5 CuCl2  H2 O;4:2 ZnCl2 ; and 4:5 NiCl2  6H2 OÞ (Ding and Wang, 2008). Five hundred milliliter of the fresh substrate was seeded with 200 ml of pretreated sludge circulating through the reactor at HRT of 0.5 h for cell cultivation and immobilization. The cultivation cycles were repeated for 20–40 HRT cycles. After successful biomass cultivation, the UFBR was operated at 4.5 h HRT for 21 d to obtain sufficient biomass before the HRT was decreased stepwise from 4.5 to 2 h with a step size of 0.5 h. Each HRT was maintained at least for 7 d. The UFBR experiment provided empirical data for the calculation of Gibbs free energy changes and thermodynamic efficiencies of caproate formation during fermentative H2 production. Gas compositions were analyzed by a gas chromatograph (GC) (HP, 5890A) equipped with a thermal conductivity detector and a fused-silica capillary column ð30  0:53 mmÞ (Supelco, Carboxen™ 1010 Plot). The operation temperatures were 150, 150, and 200 °C for injector, oven, and detector respectively with argon as the carrier gas. COD and VSS were analyzed according to the standard method (APHA, 1998). Carbohydrate was determined using phenol-sulfuric acid assay (Dubois et al., 1956). VFAs and alcohols were determined by another GC (Agilent, 6890) equipped with a flame ionization detector and a fused-silica capillary column (Agilent, DB-FFAP, 30 m  0:32 mm  0:5 lmÞ. Oven temperature was 1 set at 60 °C and increased at 20  C min to 120 °C, followed by a 1 further increase at 30  C min to 240 °C, and maintained for 3 min. Temperatures of injector and detector were both 250 °C. Helium was the carrier gas at 103 kPa. The Gibbs free energy changes of hypothesized reactions were calculated using the equations (Hanselmann, 1991) listed in Table 1 of Appendix. Table A2 lists the standard Gibb free energy of formation ðDG0f Þ of some important organic compounds.

3. Results and discussion 3.1. Theoretical aspects and reaction equations C. kluyveri ferments ethanol and acetate to form butyrate, caproate, and H2 , and converts ethanol and butyrate to acetate, caproate, and H2 . Similarly, it converts ethanol and propionate to acetate, valerate, and H2 with the production of some butyrate, caproate, and heptanoate. The ethanol dehydrogenation in Eq. (1) produces H2 as an essential end product and forms ATP via the energy-conserving SLP process. However, its standard free Gibbs energy is positive. Thus, ethanol dehydrogenation is not thermodynamically feasible, and must be coupled to some exergonic condensation of ethanol and VFAs in Eqs. (2)–(5) to make the overall fermentation exergonic (Thauer et al., 1968). However, the mechanisms underlying these couplings remained an enigma until recently. New genetics and biochemical evidence have unraveled a series of coupled redox cycles, catalyzed by two key enzymes, a cytoplasmic butyryl-CoA dehydrogenase complex (Bcd/EtfAB) and a membrane- bound energy-converting NADH:ferredoxin oxidoreductase (RnfA-E) (Herrmann et al., 2008; Li et al., 2008; Seedorf et al., 2008). The discovery of the membrane-associated RnfA-E contravened the previous assumption that bacteria lacked the ability to carry out a second type of energy conservation known as electron transport phosphorylation (ETP). In ETP, the electron carriers are reoxidized by a terminal acceptor, establishing an electrochemical ion gradient ðDlHþ Þ at the cytoplasmic membrane, which is used for ATP synthesis (Herrmann et al., 2008). Therefore, the 40-year-old notion that ATP is only formed via SLP should be revised.

3.1.1. Caproate formation from ethanol and acetate The fermentation of ethanol and acetate starts with two dehydrogenation steps (Steps A1 and A2 in Table 3), catalyzed by NAD-dependent ethanol dehydrogenase (Adh) and acetaldehyde dehydrogenase (Ald) respectively to form acetyl-CoA via an acetaldehyde intermediate. Acetyl-CoA is diverted into two routes for ATP synthesis. The first route is the classical SLP process (Step B).

9553

H.-B. Ding et al. / Bioresource Technology 101 (2010) 9550–9559 Table 4 Gibbs free energy changes for the possible pathways of butyrate and caproate formation under equilibrium conditions. 1

DH0 ðkJ mol Temperature pH

DG0 ðkJ mol

Þ

25 °C pH 0

25 °C pH 0

Þ

25 °C pH 7

pH 5:5

30 °C pH 7

pH 5:5

    

0.11 48.41 143.32 96.70 77.41

0.11 56.88 143.32 113.65 77.41

1.28 48.11 140.08 94.93 78.24

1.28 56.72 140.08 112.16 78.24

49.52 38.65 38.76 77.41 38.69

48.89 38.48 39.76 78.24 39.62

9.65 38.65 38.76 77.41 38.69

18.12 38.65 38.76 77.41 38.69

8.35 38.48 39.76 78.24 39.62

16.96 38.48 39.76 78.24 39.62

Eq. (6) Butyrate formation from ethanol and acetate 38.49 10.87 k¼1 10.27 27.78 k¼2 59.03 66.43 k¼3 107.79 105.08 k¼4 156.55 143.72 k¼5 205.31 182.37 k¼6 254.07 221.02 k¼7 302.83 259.67 k¼8

10.41 28.07 66.55 105.03 143.51 181.99 220.47 258.95

29.00 67.65 106.3 144.95 183.59 222.24 260.89 299.54

20.53 59.18 97.83 136.48 175.12 213.77 252.42 291.06

30.13 68.61 107.09 145.57 184.05 222.52 261.00 299.48

21.52 60.00 98.48 136.96 175.43 213.91 252.39 290.87

Eq. (7) Caproate formation from ethanol and acetate 59.43 27.89 k¼2 31.61 105.30 k¼4 3.79 182.71 k¼6 24.03 260.11 k¼8

29.35 107.59 185.83 264.07

67.76 145.17 222.58 299.98

59.29 136.70 214.11 291.50

69.89 148.13 226.37 304.60

61.28 139.52 217.76 295.99

Eq. (8) Caproate formation from ethanol and butyrate 108.19 10.76 k¼1 129.13 28 k¼2 150.07 66.76 k¼3 171.01 105.52 k¼4 191.95 144.27 k¼5 212.89 183.03 k¼6 233.83 221.79 k¼7 254.77 260.55 k¼8

9.13 30.63 70.39 110.15 149.91 189.67 229.43 269.19

29.11 67.87 106.63 145.39 184.14 222.90 261.66 300.42

20.64 59.4 98.16 136.92 175.67 214.43 253.19 291.94

31.41 71.17 110.93 150.69 190.45 230.21 269.97 309.73

22.80 62.56 102.32 142.08 181.83 221.59 261.35 301.11

Equations for possible pathways listed in Table 2 Reaction 1 69.7 0.11 Reaction 2 66.31 88.28 Reaction 3 336.55 143.32 Reaction 4 202.32 176.44 Reaction 5 27.82 77.41 Basic equations for C. kluyveri fermentation Eq. (1) 87.25 Eq. (2) 48.76 Eq. (3) 20.94 Eq. (4) 27.82 Eq. (5) 16.89

30 °C pH 0

1

In this energy conservation process, the acetyl group from acetylCoA is transferred to a phosphate group by phosphotransacetylase (Pta) to form acetylphosphate. SLP then occurs where acetate kinase (Ack) phosphorylates ADP to form ATP, along with acetate and Hþ production. The second ATP synthesis route is the ETP process. This process consists of a series of coupled redox cycles, culminating in the generation of a proton motive force for ATP synthesis. The first redox cycle consists of Steps D and F. Acetyl-CoA enters the redox cycle and reacts with additional acetyl-CoA to form butyryl-CoA (Steps D1–D4). The reaction is catalyzed by acetoacetyl-CoA thiolase (Thl), NAD- and NADP-dependent 3-hydroxybutyryl-CoA dehydrogenase (Hbd), 3-hydroxybutyryl-CoA dehydratase (Crt), and Bcd/ EtfAB, and proceeds via acetoacetyl-CoA, 3-hydroxybutyryl-CoA and crotonyl-CoA as intermediates. NADðPÞþ and H2 O are formed as byproducts. Butyryl-CoA subsequently reacts with acetate to form butyrate, along with the regeneration of acetyl-CoA (Step F). The enzyme involved in Step F is acetate CoA-transferase (Cat3). The formation of crotonyl-CoA is an important juncture of the first redox cycle because the exergonic reduction of crotonyl1 CoA to butyryl-CoA ðE00 ¼ 10 mV;DG00 ¼ 60 kJ mol Þ by NADH ðE00 ¼ 320 mVÞ is coupled to the second redox cycle (Step C1) which is the endergonic reduction of ferredoxin to reduced ferredoxin ðE00 ¼ 410 mVÞ. This coupling results in a free energy of 1 DG00 ¼ 40 kJ mol and is mediated by the cytoplasmic Bcd/EtfAB enzyme complex, as first proposed by Herrmann et al. (2008). The

hypothesis was later confirmed by Li and coworkers. Li et al. (2008) purified Bcd/EtfAB complex from C. kluyveri and demonstrated that the enzyme complex was indeed responsible for catalyzing crotonyl-CoA-dependent reduction of ferredoxin with NADH2 (Step C1). Ferredoxin is regenerated from reduced ferredoxin via two mechanisms. The first mechanism is a dehydrogenase reaction, catalyzed by ferredoxin-dependent hydrogenase (Hyd). Some of 1 the reduced ferredoxin reduces protons to H2 ðDG00 ¼ 0 kJ mol Þ (Step C2). The second mechanism is catalyzed by membrane-associated RnfA-E and is also the mechanism where the ETP energy conservation occurs. In this energy-converting mechanism, the remaining reduced ferredoxin is utilized for the recycling of 1 NADH2 ðDG00 ¼ þ20 kJ mol Þ, along with the production of a proþ ton motive force ðDlH Þ that drives Hþ across a membrane, creating a proton concentration gradient across the membrane (Step C3). Hþ diffuses along this gradient back to the cytoplasm through the Fo transmembrane domain of ATP synthase (AtpA-I) while ATP synthesis happens concurrently at the F1 cytoplasmic domain of AtpA-I. For every 4 mol Hþ that passes through AtpA-I, 1 mol of ATP is synthesized. Similarly, caproate is formed from caproyl-CoA through CoAtransferase activity (Step G). Caproyl-CoA is derived from the reductive condensation of acetyl-CoA and butyryl-CoA (Step E). Hex-2-enoyl-CoA is formed in Step E as intermediate. Its reduction to caproyl-CoA by NADH2 generates reduced ferredoxin as does the reduction of crotonyl-CoA to butyryl-CoA in Step D4. The regener-

9554

H.-B. Ding et al. / Bioresource Technology 101 (2010) 9550–9559

7

7

a

5

4

4

3

3

Actual HRT Designed HRT

2

2 1 1.0

60 40

b

20 0 10

c

0.4 0.2

9 8

Effluent pH Influent pH

7 6

7 6

5

5

4 20

4 300 250 200

OLR

150

10

100 5

-1 -1

d

VHPR

15

OLR, g L d

pH

8

-1 -1

0.6

0.0 10

9

VHPR, L L d

0.8

CH4, %(V/V)

H2% CO2% CH4%

80

50 0 0.4

0 1.5

e

Hydrogen Yield (HY) Biomass Yield (BY)

0.3

-1

1.0 0.2 0.5 0.1 0.0

-1

H2 and CO2, % (V/V)

1 100

HY, mol mol glucose

6

5

BY, g g glucose

HRT, h

6

0

0

10

20

30

40

50

60

70

Time, d Fig. 1. UFBR performance : (a) HRT; (b) biogas compositions; (c) pHs; (d) volumetric hydrogen production rate (VHPR) and glucose organic loading rate (OLR); (e) hydrogen yield (HY) and biomass yield (BY).

ation of ferredoxin from reduced ferredoxin also reduces Hþ to H2 and couples the recycling of NADH with ATP synthesis via ETP. For the purpose of simplification, Table 3 does not include detailed information of caproate formation. However, it is worth to note that the reduction of hex-2-enoyl-CoA to caproyl-CoA generates similar amount of energy as per the reduction of crotonyl-CoA to butyryl-CoA (Seedorf et al., 2008). It implies that the amount of ATP formed via ETP during the reduction of crotonyl-CoA to butyryl-CoA and during the reduction of hex-2-enoyl-CoA to caproylCoA may be the same. 3.1.2. ATP yield The debate on the mechanism of ATP formation has been ongoing since late 1960s. Thauer et al. (1968) firmly believed that ATP

was produced exclusively by SLP and not by ETP. Their opinion was supported by Schoberth and Gottschalk (1969) who concluded that, for every 2 mol of H2 evolved, 1 mol of acetyl-CoA became available to the cells for ATP synthesis (i.e., 1 mol of ATP per 2 mol of H2 ). However, Herrmann et al. (2008) clarified this 40year-old misconception and pointed out that the amount of ATP generated per reaction was underestimated because of the overestimation of biomass yield per ATP ðY ATP Þ by Thauer et al. (1968) and Schoberth and Gottschalk (1969). Since one-third of the cell carbon of C. kluyveri comes from CO2 and two-thirds come from acetate, YATP is certainly less than 9 g (dry cell) per mol of ATP previously used (Stouthamer, 1979). It is now known that the actual ATP yield per reaction is in fact the summation of ATP synthesis via SLP and ETP (Herrmann et al., 2008).

9555

Molar ratios

Molar percentage, %TVFA Molar percentage, %

Molar percentage, %

Molar concentration, mM

H.-B. Ding et al. / Bioresource Technology 101 (2010) 9550–9559

40

a

Ethanol Acetate Butyrate Caproate

30

40 30

20

20

10

10

0 80

0 80

b

70

70

60

60

50

50

40

40 Total VFAs% Total Solvents%

30 20 70 60

30

Ethanol% Acetate% Butyrate% Caproate%

50

c

20 70 60 50

40

40

30

30

20

20

10

10

0

0 Acetate Caproate

80

Butyrate Others

Propoinate

d

80

60

60

40

40

20

20

0 4

0 4

[Butyrate]/[Acetate] [Ethanol]/[Acetate] [Butyrate]/[Ethanol]

3

e

3

2

2

1

1

0

0 0

10

20

30

40 Time, d

50

60

70

Fig. 2. Soluble byproducts of UFBR: (a) molar concentrations of individual soluble byproducts; (b) molar percentages of total VFAs and total solvents in the total soluble byproducts; (c) molar percentages of individual soluble byproducts in the total soluble byproducts; (d) molar percentages of individual VFAs in the total VFAs; (e) molar ratios of butyrate/acetate, ethanol/acetate, and butyrate/ethanol.

In this study, it is suggested that the amount of ATP yield per reaction is dependent on the number of ethanol ðmÞ consumed. The endergonic oxidation of one mol of ethanol produces 2 mol of H2 , 1 mol of acetate, and 1 mol of ATP via SLP. Thus, the number of ATP produced via ETP is directly proportional to ðm  1Þ. According to Seedorf et al. (2008), this number equals to 0:3  ðm  1Þ. Hence a total of 0:7 þ 0:3m ATP would be generated per reaction. The suggestion made here is based on the fact that the reduction of crotonyl-CoA to butyryl-CoA and the reduction of hex-2-enoylCoA to caproyl-CoA can generate the same amount of energy and yield the same number of ATP via ETP.

from ethanol and acetate. In fermentative H2 production from carbohydrates, butyrate is one of the major products (Table 1). Therefore, C. kluyveri has butyrate as an alternative terminal acceptor for the reducing equivalents during caproate synthesis (Eq. (8) in Table 2). Eq. (6) is obtained by summing up Eq. (1) and k numbers of Eq. (2) while Eq. (8) is a summation of Eq. (1) and k numbers of Eq. (3). Similarly, Eq. (7) is the sum of Eq. (1) and k=2 numbers of Eq. (4) (k is even numbers in this case). k has to be known in order to set up the overall equations and reflects the distribution of reducing equivalents between higher acids synthesis and H2 production. In a similar fashion, the overall equation for valerate formation, Eq. (9), can be worked out as:

3.1.3. Overall reaction equations Schoberth and Gottschalk (1969) gave overall equations for the formation of butyrate and caproate (Eqs. (6) and (7) in Table 2)

ðk þ 1ÞCH3 CH2 OH þ kCH3 CH2 COO ¼ kCH3 ðCH2 Þ2 COO þ CH3 COO þ Hþ þ 2H2 þ ðk  1ÞH2 O

ð9Þ

9556

H.-B. Ding et al. / Bioresource Technology 101 (2010) 9550–9559

Eqn.6 k = 3

Eqn.6 k = 4

Eqn.6 k = 5

Eqn.6 k = 6

Eqn.6 k = 7

Eqn.6 k = 8

70

a

η, %

60

50

40

30 Eqn.7 k = 4

Eqn.7 k = 6

Eqn.7 k = 8

70

b

η, %

60

50

40

30 Eqn.8 k = 3

Eqn.8 k = 4

Eqn.8 k = 5

Eqn.8 k = 6

Eqn.8 k = 7

Eqn.8 k = 8

70

c η,%

60 50 40 30 25

30

35

40

45

50

55

60

65

70

75

Time, d Fig. 3. Thermodynamic efficiencies of energy conservation under the experimental condition: (a) Eq. (6) – formation of butyrate from ethanol and acetate; (b) Eq. (7) – formation of caproate from ethanol and acetate; (c) Eq. (8) – formation of caproate from ethanol and butyrate.

3.2. Thermodynamic efficiency Since C. kluyveri has numerous caproate formation pathways, it is difficult to determine the exact pathways involved. Thermodynamic feasibility studies of the possible pathways, by calculating the Gibbs free energy and thermodynamic efficiency, g using Eq. (10) (Thauer et al., 1977), may provide estimations:

g¼

N  DGATP DGr

ð10Þ

where N=number of ATP produced per reaction, DGATP =required energy for synthesis of 1 mol ATP, and DGr =energy released from the reaction. The various values of the Gibbs free energy changes, under different equilibrium conditions and on the assumption that substrates and products are present at 1 atm if the reactants are gases or 1 M concentration if the reactants are solutes, are listed in Table 4 for Eqs. (6)–(8) at several k values. ATP yields of Eqs. (6)–(8) are

N ¼ 1 þ 0:3k per reaction (since m ¼ k þ 1). Under the physiological conditions (pH 7 and 25 °C), the energy required for ATP formation 1 is between 40 and 50 kJ mol ATP with an average value of 1 45 kJ mol ATP (Thauer et al., 1977). The calculation of the thermodynamic efficiencies by Eq. (10) showed that Eqs. (6) and (8) with k ¼ 5  8 and Eq. (7) with k ¼ 6 and 8 have g ¼ 50 to 60%, an optimum range commonly observed in Clostridium species (Thauer et al., 1977; Herrmann et al., 2008). Eqs. (6)–(9) at k ¼ 1  4 are thermodynamically not possible under the physiological condition since g is much higher than the upper limit of this optimum range (60 %). Temperature and pH have significant influences on the energies released from reactions. Table 4 also shows the Gibbs free energy changes of reactions corrected by pH ðpH 5:5Þ and temperature (30 °C) which are commonly used in fermentative H2 production. Generally, higher temperature increases free energies released from reactions while lower pH decreases them if protons ðHþ Þ are the product of reactions.

H.-B. Ding et al. / Bioresource Technology 101 (2010) 9550–9559

9557

Since acetate, butyrate, and ethanol are the major end products of the fermentative H2 production while caproate formation is a secondary fermentation, the k values of Eqs. (6)–(8) may be different. The empirical data of UFBR were used to estimate the k values. The H2 -producing performance of UFBR is shown in Fig. 1. H2 gradually increased from 25% at HRT of 4.5 h to 45% at HRT of 2 h. The buffering capacity was provided to maintain pH between 5 and 7. The volumetric hydrogen production rate (VHPR) increased with the decrease of HRT. The maximum VHPR found 1 was 14:6 L L1 d at 2 h HRT. The VHPR was positively correlated to glucose organic loading rate. At steady states of low HRTs, glucose utilization was 50–80%. The hydrogen yield (HY) was 1 0:5  1:0 mol mol glucose consumed and biomass yield was about 0:1 g  VSS g1 glucose consumed. The concentrations of soluble byproducts are shown in Fig. 2. Ethanol was the major solvent product while acetate and butyrate were the major VFA products. Propionate (1–3% of total VFAs) was always detected. Valerate concentration (not shown) was very low (<0.5 mM), probably due to the low concentration of propionate. Caproate appeared after 30 d and gradually became the third major VFA with a concentration of 11–23 mM. Its appearance indicated the possible presence of C. kluyveri in the UFBR. After startup, the solvent products represented 50–66% of the total soluble byproducts at HRT of 4.5 h. As HRT gradually decreased below 2.5 h, VFAs replaced solvent products and became the major byproducts (>60%). The Gibbs free energy changes of Eqs. (6)–(8) were calculated by using the experimental values of pH (in Fig. 1), temperature (30 °C), and reactant and product concentrations (in Fig. 2). Thermodynamic efficiencies of Eqs. (6)–(8) after 20 days of UBFR operation were calculated and are shown in Fig. 3. When butyrate was formed from ethanol and acetate (Eq. 6), in Fig. 3a, g was over 60% for k ¼ 3, 50–60% for k ¼ 4  5, 40-50% for k ¼ 6  7, and below 40% for k ¼ 8. When caproate was formed from ethanol and acetate (Eq. 7), in Fig. 3b, g was over 80% for k ¼ 2 (not shown), 50–60% for k ¼ 4, and 40-50% for k ¼ 6 and 8. When caproate was formed from ethanol and butyrate (Eq. 8), in Fig. 3c: g was over 70% for k ¼ 1  2, 60-70% for k ¼ 3, 50–60% for k ¼ 4, and 40-50% for k ¼ 5  8. Overall speaking, when k P 4 Eqs. (6)–(8) are thermodynamically feasible under the experimental condition. In another words, a minimum of 5 mol of ethanol per mol of each reaction of Eqs. (6)– (8) can support the formation of butyrate/caproate by C. kluyveri in the UFBR. The result has one mol less of ethanol for Eqs. (6) and (8) or 2 mol less of ethanol for Eq. (7) than that obtained under the physiological condition. The result also has one mol less of ethanol compared to the typical reaction equation used for the fermentation by C. kluyveri in which 6 mol of ethanol and 3 mol of acetate produce 2 mol of H2 , 3 mol of butyrate, and 1 mol of caproate per mol of reaction (Thauer et al., 1968; Seedorf et al., 2008). If 50–60% is considered as an optimum range for g, Eq. (6) with k ¼ 4  5 and Eqs. (7) and (8) with k ¼ 4 would most likely occur in the experiment.

During caproate formation by C. kluyveri, acetate and butyrate can be either consumed or produced depending on the kind of caproate formation pathways utilized. Eq. (6) indicates the direct production of butyrate while Eqs. (7) and (8) represent the consumption of butyrate, either as an intermediate or from an external source. The fate of acetate is similar: consumed in Eqs. (6) and (7) but produced in Eq. (8). These kinds of cyclic mechanisms may interfere with fermentative H2 production in which H2 , acetate, and butyrate are major products. According to Eqs. (6) and (8), fermentative H2 production is not the only process to produce H2 , butyrate, and acetate at the same time. As such, butyrate/acetate ratio may not be used as an indicative predictor of hydrogen yield from fermentative H2 production. This contravenes the suggestion made by Hawkes et al. (2007) in which hydrogen yield can be correlated to butyrate/acetate ratio. H2 is a common product in Eq. (6). Caproate or valerate formation can contribute to H2 production by releasing certain amount of reducing equivalents from ethanol. According to the above analysis, Eqs. (7) and (8) with k ¼ 4 are the most possible caproate formation pathways with optimum thermodynamic efficiencies under the experimental condition. H2 production for Eqs. (7) and 1 (8) were 1 and 0:5 mol mol caproate formed with k ¼ 4, respectively. H2 produced from the caproate pathways was 10 to 20 mmol L1 substrate medium, given the fact that the average caproate concentration in the effluent was 20 mM in the late 1 stages. Since the average total HY was 1 mol mol glucose consumed and 100 mmol glucose was consumed for every liter of substrate medium, a total of 100 mmol H2 L1 substrate medium was produced. So the contribution of caproate pathways for H2 production accounted for 10–20% of total H2 production from the UFBR. This estimation does not include the H2 contribution of butyrate formation using Eq. 6. Future studies are needed to differentiate H2 and butyrate formation between carbohydrate fermentation and ethanol fermentation by C. kluyveri. It should be emphasized that caproate formation by C. kluyveri is a secondary fermentation of ethanol and VFAs in the fermentative H2 production. In this H2 production process, ethanol represents reducing equivalents that have not been liberated as hydrogen gas (Hawkes et al., 2007). As main species responsible for fermentative H2 production, saccharolytic clostridia tend to produce VFAs and H2 at high pH via acidogenesis and facilitate alcohol production at low pH via solventogenesis. Solventogenesis played a significant role in the UFBR as 40–60% of total metabolite products were ethanol (Fig. 2). It resulted in less H2 production from the UFBR as the observed H2 yield was much lower than 1 the theoretical yield of 4 mol H2 mol glucose. At the same time, ethanol produced from the solventogenesis favored the formation of caproate. Although part of the reducing equivalents from ethanol released as H2 gas during caproate formation, the overall H2 yield of the fermentative H2 production was still low. The misconception in Reactions 2–4 (Table 2) is clarified here in this paper: caproate formation is indeed an H2 -producing secondary fermentation instead of an H2 -consuming one. The appearance of caproate in the final products indicated that significant solventogenesis occurred and thus the yield H2 of the fermentative H2 production was poor.

3.4. Significance of caproate formation in the fermentative H2 production

4. Conclusions

Caproate formation is always accompanied by ethanol. The dual roles of ethanol as a byproduct of fermentative H2 production and the essential substrate for caproate formation indicate that C. kluyveri can form a syntrophy relationship with ethanol-producing bacteria. The reducing equivalents present in ethanol can be redistributed by C. kluyveri to butyrate/caproate/valerate and H2 .

A detailed energetic analysis of our experiment demonstrated that caproate can be formed through the secondary fermentation of two substrates, either ethanol and acetate or ethanol and butyrate. The analysis also indicated that caproate formation did not consume H2 but actually produces H2 . Co-production of caproate and H2 was thermodynamically feasible under the experimental

3.3. Gibbs free energy and thermodynamic efficiencies under the experimental condition

9558

H.-B. Ding et al. / Bioresource Technology 101 (2010) 9550–9559

Table A1 Summary of equations describing ecological sate parameters. Main equations P P i v i :Si ¼ j v j :P j P DG0f ;pj  i v i :DG0f ;Si P P DHr o ¼ j v j :DHf ;P0  i v i :DH0f ;Si

Reaction Equilibrium equation (eq)

Input parameter

DG0f ;pj ; DG0f ;Si ; DH0f ;Pj ; v j ; v j Q v fP j;neq g j Q neq ¼ Qj v

j

Non-equilibrium equation (neq)

DGr ¼

Influent of temperature on equilibrium equation (Gibbs-Helmholtz equation)*

DG0r;T act

Influent of pH-change on equilibrium equation**

DG0r ¼ DG0r  2:303:q:R:T:pHact

pHact ; q

Influent of pH-change on non-equilibrium equation**

DGr ¼ dG0r þ 2:303:R:T:ðlogQ neq  q:pHactÞ

Q neq ¼ fHneq þ q g

DG0r ¼

þ R:T:InQ neq

DG0r;T ref : TT act ref

þ

DH0r;T ref

:



1

T act

1

 T ref



i

fSi;neq g

i

DG0r;T ref ; DH0r;T ref ; T ref Q

Note: *Assume that the temperature changes within the physiological temperature span do not influent DH0r remarkably; **under certain circumstances it is more appropriate to define reference conditions for proton producing and consuming reactions for a pH-value different from pH 7.

Table A2 Standard enthalpy of formation and standard Gibbs free energy of formation at 25 °C. Compounds

Chemical formulae

State

DH0f 1

kJ mol Acetate Propionate Butyrate Valerate Caproate Ethanol Carbon dioxide Hydrogen Hþ ðpH ¼ 7Þ Hþ (1 M or pH ¼ 0Þ Water Note: *Aqueous-state values of Sources: Barrow (1974). Hanselmann (1991). Lide (2001). Sawyer et al. (2003). Thauer et al. (1977).

C2 H3 O 2 C3 H5 O 2 C4 H7 O 2 C5 H9 O 2 C6 H11 O 2 C2 H6 O CO2 H2 Hþ Hþ H2 O

DG0f

aqueous Aqueous Aqueous Aqueous Aqueous Aqueous Gas gas Aqueous Aqueous Liquid

at 25 °C

485.6 510.8 535.55 495.1 515.8 287.02 393.51 0

369.41 361.08 352.63 344.34 335.96 181.75 394.38 0 39.87 0 237.18

0 285.80

Subscriptref Subscriptf

Acknowledgements The authors thank the Nanyang Technological University Research Grant for the financial support. Notation for Table A1 Tables. A1, A2

Superscript00

Subscriptact Subscriptneq Subscriptr

DH0f

DG0f

Lide (2001) (Barrow, 1974)

Thauer et al. (1977)

Hanselmann (1991)* Barrow (1974) Sawyer et al. (2003)

Sawyer et al. (2003) Thauer et al. (1977)

Thauer et al. (1977) (Sawyer et al., 2003)

(Sawyer et al., 2003)

for valerate and caproate are not available; liquid-state values were used and were very close to those aqueous-state values.

condition of the present study and considerably contributed to the overall fermentative H2 production.

Superscript0

Sources

DG0f

Denotes standard conditions with reactants in their pure state present at a pressure of 1 atm if the reactants are gases or 1 M if the reactants are solutes at T ¼ 298:15 K, pressure=1 atm, and pH0ð½Hþ  ¼ 1MÞ, e.g., Denotes standard conditions except pH – 0 or ½Hþ  – 1 M; often referred to the standard condition with pH 7 for biochemical reaction, e.g., At actual conditions, e.g., T act ; pHact At non-equilibrium e.g., Pj;neq ; Q 0neq ; Si;neq Denotes thermodynamic quantity associated with a reaction, e.g., DG0f ; DH0f

Subscripti;j SubscriptP SubscriptS V i;j [] {} Q

R DG0f DGr DG0r DG00 r DH0f

At reference state e.g. T ref ¼ 298:15 K Denotes thermochemical quantity associated with the formation of a substance from elements in their reference state, e.g.,

DG0f ; DH0f Chemical species designation Products Substrates Stoichiometric coefficients of species i, j Designates molar concentration Designates activity,  [ ] if concentrations of reactants and products are low, Product of terms Sum of terms Standard free energy of formation Change of Gibbs free energy of reaction at actual conditions Change of Gibbs free energy of reaction at standard conditions Change of Gibbs free energy of reaction at standard condition ðpH – 0Þ and more often pH 7 Standard enthalpy of formation at standard conditions

H.-B. Ding et al. / Bioresource Technology 101 (2010) 9550–9559

R

Change of enthalpy of reaction at standard conditions Number of protons transferred, + if they are produced,  if they are consumed Ratio of actual activity products of reactants (including Hþ if Hþ is participating the reaction) Ratio of actual activity products of reactants excluding Hþ (or/and electrons in half-reactions) Universal gas

T

constant=8:31451  103 kJ mol Temperature in Kelvin degree

DH0r q Q Q0

1

K1

References APHA, 1998. Standard Methods for the Examination of Water and Wastewater, 20th ed. American Public Health Association/American Water Works Association/ Water Environment Federation, Washington DC, USA. Barrow, G.M., 1974. Physical Chemistry for the Life Sciences. McGraw-Hill, New York. Ding, H.-B., Wang, J.-Y., 2008. Responses of the methanogenic reactor to different effluent fractions of fermentative hydrogen production in a phase-separated anaerobic digestion system. Int. J. Hydrogen Energy 33 (23), 6993–7005. Dubois, M., Gilles, K.A., Hamilton, J.K., Rebers, P.A., Smith, F., 1956. Colorimetric method for determination of sugars and related substances. Anal. Chem. 28 (3), 350–356. Dworkin, M., Falkow, S., Rosenberg, E., Schleifer, K.-H., Stackebrandt, E., 2006. Bacteria: Firmicutes, Cyanobacteria. In: Dworkin, M. (Ed.), The Prokaryotes: A Handbook on the Biology of Bacteria. Springer-Verlag, New York. Hanselmann, K.W., 1991. Microbial energetics applied to waste repositories. Experientia 47 (7), 645–687. Hawkes, F.R., Hussy, I., Kyazze, G., Dinsdale, R., Hawkes, D.L., 2007. Continuous dark fermentative hydrogen production by mesophilic microflora: principles and progress. Int. J. Hydrogen Energy 32 (2), 172–184. Herrmann, G., Jayamani, E., Mai, G., Buckel, W., 2008. Energy conservation via electron-transferring flavoprotein in anaerobic bacteria. J. Bacteriol. 190 (3), 784–791.

9559

Kenealy, W.R., Waselefsky, D.M., 1985. Studies on the substrate range of Clostridium kluyveri; the use of propanol and succinate. Arch. Microbiol. 141 (3), 187–194. Lee, K.-S., Wu, J.-F., Lo, Y.-S., Lo, Y.-C., Lin, P.-J., Chang, J.-S., 2004. Anaerobic hydrogen production with an efficient carrier-induced granular sludge bed bioreactor. Biotechnol. Bioeng. 87 (5), 648–657. Lide, D.R., 2001. CRC Handbook of Chemistry and Physics, 82 ed. CRC Press, Boca Raton, FL. Li, F., Hinderberger, J., Seedorf, H., Zhang, J., Buckel, W., Thauer, R.K., 2008. Coupled ferredoxin and crotonyl coenzyme A (CoA) reduction with NADH catalyzed by the butyryl-CoA dehydrogenase/Etf complex from Clostridium kluyveri. J. Bacteriol. 190 (3), 843–850. Sawyer, C.N., McCarty, P.L., Parkin, G.F., 2003. Chemistry for Environmental Engineering and Science, 5th ed. McGraw-Hill, Boston. Schoberth, S., Gottschalk, G., 1969. Considerations on the energy metabolism of Clostridium kluyveri. Arch. Microbiol. 65 (4), 318–328. Seedorf, H., Fricke, W.F., Veith, B., Bruggemann, H., Liesegang, H., Strittmatter, A., Miethke, M., Buckel, W., Hinderberger, J., Li, F., Hagemeier, C., Thauer, R.K., Gottschalk, G., 2008. The genome of Clostridium kluyveri, a strict anaerobe with unique metabolic features. Proc. Natl. Acad. Sci. USA 105 (6), 2128–2133. Stouthamer, A.H., 1979. The search for correlation between theoretical and experimental growth yields. Int. Rev. Biochem. 21 (1), 1–47. Thauer, R.K., Jungermann, K., Henninger, H., Wenning, J., Decker, K., 1968. The energy metabolism of Clostridium kluyveri. Eur. J. Biochem. 4 (2), 173–180. Thauer, R.K., Jungermann, K., Decker, K., 1977. Energy conservation in chemotrophic anaerobic bacteria. Bacteriol. Rev. 41 (1), 100–180. Thauer, R.K., Jungermann, K., Decker, K., 1977. Energy conservation in chemotrophic anaerobic bacteria. Bacteriol. Rev. 41 (1), 100–180. Wang, J.-Y., Liu, X.-Y., Ding, H.-B., Sreeramachandran, S., Jiao, Y., 2008. Demonstration plant using the hybrid anaerobic solid–liquid system for waste bioconversion. The ISWA/WMRAS World Congress 2008, Singapore. Yu, H.Q., Mu, Y., 2006. Biological hydrogen production in a UASB reactor with granules. II: reactor performance in 3-year operation. Biotechnol. Bioeng. 94 (5), 988–995. Zhang, Z.-P., Tay, J.H., Show, K.Y., Yan, R., Liang, T.D., Lee, D.J., Jiang, W.J., 2007. Biohydrogen production in a granular activated carbon anaerobic fluidized bed reactor. Int. J. Hydrogen Energy 32 (2), 185–191. Zhao, Q.B., Yu, H.Q., 2008. Fermentative H5 production in an upflow anaerobic sludge blanket reactor at various pH values. Bioresour. Technol. 99 (5), 1353– 1358. Zhao, Q.-B., Mu, Y., Wang, Y., Liu, X.-W., Dong, F., Yu, H.-Q., 2008. Response of a biohydrogen-producing reactor to the substrate shift from sucrose to lactose. Bioresour. Technol. 99 (1), 8344–8347.