Bioresource Technology 102 (2011) 3833–3839

Contents lists available at ScienceDirect

Bioresource Technology

journal homepage: www.elsevier .com/locate /bior tech

Biohydrogen and methane production by co-digestion of cassava stillageand excess sludge under thermophilic condition

Wen Wang a, Li Xie a,b,⇑, Jinrong Chen a, Gang Luo a, Qi Zhou a,b

a State Key Laboratory of Pollution Control and Resources Reuse, Tongji University, 1239 Siping Road, Shanghai 200092, PR Chinab UNEP-Tongji, Tongji University, 1239 Siping Road, Shanghai 200092, PR China

a r t i c l e i n f o

Article history:Received 16 August 2010Received in revised form 29 November 2010Accepted 1 December 2010Available online 7 December 2010

Keywords:Co-digestionCo-productionHydrogenMethaneThermophilic

0960-8524/$ - see front matter � 2010 Elsevier Ltd. Adoi:10.1016/j.biortech.2010.12.012

⇑ Corresponding author at: State Key LaboratorResources Reuse, Tongji University, 1239 Siping RoadTel.: +86 21 65982692; fax: +86 21 65986313.

E-mail address: sally.xieli@gmail.com (L. Xie).

a b s t r a c t

Thermophilic anaerobic hydrogen and methane production by co-digestion of cassava stillage (CS) andexcess sludge (ES) was investigated in this study. The improved hydrogen and subsequent methane pro-duction were observed by co-digestion of CS with certain amount of ES in batch experiments. Comparedwith one phase anaerobic digestion, two phase anaerobic digestion offered an attractive alternative withmore abundant biogas production and energy yield, e.g., the total energy yield in two phase obtained atVSCS/VSES of 3:1 was 25% higher than the value of one phase. Results from continuous experiments fur-ther demonstrated that VSCS/VSES of 3:1 was optimal for hydrogen production with the highest hydrogenyield of 74 mL/g total VS added, the balanced nutrient condition with C/N ratio of 1.5 g carbohydrate–COD/g protein–COD or 11.9 g C/g N might be the main reason for such enhancement. VSCS/VSES of 3:1was also optimal for continuous methane production considering the higher methane yield of 350 mL/g total VS added and the lower propionate concentration in the effluent.

� 2010 Elsevier Ltd. All rights reserved.

1. Introduction

Cassava ethanol as a promising fossil fuel substitute is widelyproduced in China for its renewability and excellent fuel properties(Leng et al., 2008; Yu and Tao, 2009). Cassava stillage (CS) is a li-quid waste generated after distillation process with high contentof organic compounds and suspended solids (Luo et al., 2009; Yuand Tao, 2009). Up to twenty liters of stillage may be generatedfor each liter of ethanol produced (Wilkie et al., 2000). Such kindof wastewater may cause serious environmental pollution if notproperly treated.

Anaerobic digestion for alcohol stillage treatment has beendemonstrated as an effective and economic treatment technology(Espinosa et al., 1995; Yeoh, 1997), which is very important fromrecycling, energy production and pollutants reduction points ofview. By anaerobic digestion, organic pollutants could be degradedand renewable energy could be generated simultaneously. Ourprevious study demonstrated that CS could be used for methaneproduction by anaerobic sequencing reactor with high perfor-mance (Luo et al., 2009). Nowadays, it was further recognized thatcarbohydrate-rich wastes or wastewater are also potential sources

ll rights reserved.

y of Pollution Control and, Shanghai 200092, PR China.

for hydrogen production (Kaparaju et al., 2009; Kim and Lee, 2010).CS has been used for hydrogen production in our previous study(Luo et al., 2010). However, alkali addition was required to main-tain suitable pH for continuous hydrogen generation, due to thelow pH and the tendency of quick acidification.

Currently, anaerobic co-digestion has attracted more attentiondue to its potential in providing better pH conditions, morebalanced C/N ratio, and increased biogas production (Luste andLuostarinen, 2010; Zhu et al., 2008a). Fezzani and Ben Cheikh stud-ied the co-digestion of olive mill wastewater and olive mill solidwaste, and found the wastes were degraded more efficiently(Fezzani and Ben Cheikh, 2010). Zhu et al. (2008b) observed biohy-drogen production could be enhanced by co-digestion of municipalfood waste and sewage sludge due to the increased buffer capacity.Moreover, in the study of Shanmugam and Horan (2009), munici-pal solid waste was used as auxiliary substrate to optimizemethane production from leathering fleshing waste by adjustmentof pH and C/N ratio. Most studies of co-digestion focused on theperformance of single hydrogen production or single methane pro-duction, and limited information is available on the performance ofco-digestion for hydrogen and methane production in two phaseanaerobic system, especially under thermophilic condition. Inaddition, Sewage sludge (ES) is a common waste with high alkalin-ity and abundant nitrogenous compounds, and it could also be gen-erated during the biological treatment of cassava stillage. It wouldbe interesting to co-digest CS with ES to optimize biohydrogen andmethane production.

3834 W. Wang et al. / Bioresource Technology 102 (2011) 3833–3839

The present study was therefore conducted to investigate andoptimize the thermophilic anaerobic co-production of hydrogenand methane by co-digestion of CS and ES. Feasibility of two phasehydrogen and methane production by co-digestion was studiedunder batch mode operation. One-phase methane production byco-digestion was studied for comparison as well. Moreover, perfor-mances of hydrogen and methane production at different mixingratios were further studied under continuous operation mode,and the involved mechanisms were discussed in this study.

2. Methods

2.1. Inoculum and substrates

Anaerobic granular sludge acquired from a full-scale mesophilicupflow anaerobic sludge blanket reactor was used as inoculumwithout any pretreatment. CS and ES used as substrates were ob-tained respectively from cassava ethanol plant (Jiangsu, China),and the dewatering equipments after secondary settling tank ofQuyang wastewater treatment plant (Shanghai, China). After col-lected, CS and ES were both stored at 4 �C before usage. Table 1summarizes the characteristics of CS and ES. The pH of CS (4.04)was much lower than the value of ES (6.77). The soluble carbohy-drate concentration in CS was as high as 167 mg/g VS while theconcentrations of protein and ammonia nitrogen were relativelylower. Compared with CS, ES contained higher content of proteinand NH3–N but lower carbohydrate concentration (11 mg/g VS).Therefore, the biogas production behavior might be benefit fromthe increased buffer capacity and balanced carbohydrate/proteinratio by co-digestion.

2.2. Experimental design and procedure

2.2.1. Batch experimentsBatch experiments were firstly conducted to evaluate the feasi-

bility of hydrogen and methane production from CS co-digestedwith ES without any extra nutrients or buffer solution addition.Series of identical borosilicate glass bottles (Witeg, Boro 3.3) withvolume of 500 mL were used as reactors for both hydrogen andmethane production at thermophilic temperature.

Initially, five reactors were inoculated with 20 mL anaerobicgranular seed sludge each, and then added with the mixture ofCS and ES. The VS in the reactors were controlled at 30 g/L withVSCS/VSES of 4:0, 3:1, 1:1, 1:3 and 0:4, respectively. Working

Table 1Characteristics of substrates, CS and ES.

Index CS ES

TS 45.9 ± 0.6 g/L 182.0 ± 0.3 g/kgVS 39.5 ± 0.5 g/L 121.8 ± 0.3 g/kgpH 4.04 ± 0.01 6.77 ± 0.02Soluble carbohydrate (mg/g VS) 167 ± 14 11 ± 8Soluble protein (mg/g VS) 33 ± 11 55 ± 12NH3–N (mg/g VS) 9 ± 0.5 23 ± 2TCOD (mg/g VS) 1783 ± 20 1576 ± 9SCOD (mg/g VS) 930 ± 12 530 ± 5TN (mg/g VS) 7.7 ± 2 35.4 ± 1.5TP (mg/g VS) 2.1 ± 0.3 2.0 ± 0.8VFA (mg/g VS) Ethanol 10.8 ± 3.1 ud.

Acetate 20.4 ± 4.8 27.5 ± 5.6Propionate ud. 21.8 ± 6.1Iso-butyrate 4.6 ± 1.2 4.1 ± 0.8n-Butyrate ud. 13.3 ± 2.9Iso-valerate ud. 6.9 ± 0.9n-Valerate ud. 4.2 ± 1.1TVFA 35.8 ± 7.6 77.8 ± 9.4

ud. – undetectable.

volumes were adjusted to 400 mL with distilled water. The initialfermentative pH of the mixed solution in each reactor was adjustedto 6.0 by 2 N NaOH or 2 N HCl. After hydrogen production ceased,300 mL effluent from each hydrogen reactor was further fed intonew reactor for subsequent methane production. For comparison,the one phase methane production was conducted for each mixingratio with total 300 mL fresh substrates. One hundred milliliteranaerobic granular sludge was introduced into each methane reac-tor with final 400 mL of working volume. The initial pH for meth-ane production was initially adjusted to 7.5 by NaHCO3.

2.2.2. Continuous experimentsAfter hydrogen and methane production ceased in the last

batch, the operation was switched to semi-continuous mode andoperated as continuously stirred tank reactor in ten 300 mLborosilicate glass bottles with working volume of 200 mL. Fivereactors were conducted for hydrogen production with substratesat VSCS/VSES of 4:0, 3:1, 1:1, 1:3 and 0:4 respectively. Equal amountof digested and fresh mixing substrate were removed and added tothe reactors periodically and the HRT of hydrogen production wasfirstly maintained at 5 days. After 44 days operation, the HRT ofthese five hydrogen reactors was then adjusted to 3 days. The efflu-ent of hydrogen fermentation reactors at VSCS/VSES of 4:0, 3:1, 1:1,1:3 and 0:4 were, respectively, fed to methane fermentation reac-tors for the second phase methane production. The HRT of meth-ane production was maintained at 12.5 days.

In the above batch and continuous experiments, the cappedreactors with rubber stoppers were placed in a reciprocating waterbath shaker and rotated at 120 rpm with the temperature of 60 �C.Each ratio was conducted in triplicate. The evolved biogas was col-lected with gas bag. The amount of biogas was determined period-ically using syringe and the composition of the biogas wasmeasured at the same time. Parameters including pH, TS, VS,COD, short chain volatile fatty acids (VFA) and ethanol, soluble car-bohydrate, soluble protein and ammonia nitrogen were deter-mined simultaneously.

2.3. Analytical methods

After centrifuged at 10,000 rpm for 10 min, the supernate ofsamples were immediately analyzed. Total solid (TS), volatile solid(VS), COD and ammonia nitrogen were analyzed in accordancewith APHA standard methods (APHA, 1998). Soluble carbohydrateconcentration was measured by the phenol sulfuric acid method,with glucose as standard (Dubois et al., 1956). Soluble proteinwas determined by the Lowry–Folin method with bovine serumalbumin as standard (Lowry et al., 1951). C/N ratio (g/g) was calcu-lated based on the chemical composition. Carbon, hydrogen, nitro-gen and sulfur were undertaken using CHNS analyzers model byelemental analysis (Elementar, vario EL III).

The concentrations of ethanol and VFA (Acetic, propionic, iso-butyric, n-butyric, iso-valeric, n-valeric acids) were determinedby gas chromatograph (Agilent, 6890 N) equipped with a flameionization detector and analytical column CPWAX52CB (30 m �0.25 mm � 0.25 lm). The temperature of the injector and FID were200 and 220 �C, respectively. Nitrogen was the carrier gas with aflow rate of 50 mL/min. The GC oven was programmed to beginat 110 �C and remain there for 2 min, then increase at a rate of10 �C/min to 220 �C, and hold at 220 �C for an additional 2 min.The sample injection volume was 1.0 lL.

Biogas (H2, CH4 and CO2) composition were determined by a gaschromatograph (Agilent, 6890 N) equipped with a thermal conduc-tivity detector (TCD) and analytical column Supelco Hayesp Q (80/100 mesh). The operation temperature at the injection port, thecolumn oven and the detector were 120, 35 and 250 �C, respec-tively. Helium was used as a carrier gas at a flow rate of 20 mL/min.

Table 2Parameters of the modified Gompertz equation obtained at different ratios of CS andES in batch experiments.

VSCS/VSES 4:0 3:1 1:1 1:3 0:4

H (mL) 368 376 378 234 21P (mL) 371.8 378 381 244 21.4Rm (mL h�1) 11.9 33.1 19.5 6.0 5.3k (h) 40.7 35.1 34.0 35.3 35.0R2 0.997 0.999 0.993 0.985 0.999

W. Wang et al. / Bioresource Technology 102 (2011) 3833–3839 3835

2.4. Data analysis

Hydrogen production efficiency was evaluated using cumula-tive hydrogen production (mL), hydrogen concentration in the bio-gas (%), hydrogen yield (HY) (the calculated hydrogen productionper gram VS, mL/g VS). And the maximum hydrogen productionrate, hydrogen production potential were estimated using themodified Gompertz equation (Eq. (1)) which could describe theprogress of hydrogen production under batch experiments (Layet al., 1999):

HðtÞ ¼ P � exp � expRm � e

Pðk� tÞ þ 1

� �� �ð1Þ

where H(t) is the cumulative hydrogen production (mL), P is hydro-gen production potential (mL), Rm is maximum hydrogen produc-tion rate (mL/h), e is 2.718, k is lag-phase time (h) and t is time(h). Origin 8.0 was used for non-linear regression obtaining P andRm values at different VS ratios.

Methane production efficiency was evaluated using cumulativemethane production (mL), methane concentration in the biogas(%), methane yield (MY) (the calculated methane production pergram VS, mL/g VS). Since it is difficult to determine the molaramount of complex mixture, the units of hydrogen yields andmethane yields in this paper were unified to mL/g VS at standardtemperature (0 �C) and pressure (1 atm).

3. Results and discussion

3.1. Batch experiments for hydrogen and methane production

3.1.1. Hydrogen productionFig. 1 describes the cumulative hydrogen production at differ-

ent mixing ratios of CS and ES. Similar cumulative hydrogen pro-duction of 368, 376 and 378 mL were obtained at VSCS/VSES of4:0, 3:1 and 1:1 after 4 days fermentation. Further increase of ESfraction led to the decrease of hydrogen production, and the lowesthydrogen production of 21 mL was observed at VSCS/VSES of 0:4.Gompertz equation was used to simulate the process and the re-sults are presented in Table 2. The hydrogen production rate fol-lowed the sequence of VSCS/VSES 3:1 > 1:1 > 4:0 > 1:3 > 0:4, andthe highest hydrogen production rate at 3:1 was about three timeshigher than that at 4:0 and six times higher than that at 0:4. More-over, the lag phase time was decreased at least 5 h with addition ofES into CS.

0 10 20 30 40 50 60 70 80 90 100 110 120 130

0

50

100

150

200

250

300

350

400

450 VSCS :VSES = 4:0 VSCS :VSES = 3:1 VSCS :VSES = 1:1 VSCS :VSES = 1:3 VSCS :VSES = 0:4

Cum

ulat

ive

H2 p

rodu

ctio

n (m

l)

Time (h)

Fig. 1. Cumulative hydrogen production in batch experiments.

The HY calculated based on the total VS added (30 g/L) are sum-marized in Table 3. The VSCS/VSES of 3:1 and 1:1 obtained the max-imum HY of 31 mL/g total VS added. And ES fermentation aloneobtained the minimum HY of 2 mL/g total VS added, indicatingthat ES itself was not a good substrate for hydrogen production.Protein rich substances were found to be difficult for hydrogenproduction (Li et al., 2009; Okamoto et al., 2000). Thus the ob-served production of hydrogen should be mainly resulted fromthe utilization of carbohydrate-rich CS. The HY based on the CScontributed VS was also calculated, and was 31, 42, 63, 78 and0 mL/g CS–VS added for VSCS/VSES of 4:0, 3:1, 1:1, 1:3 and 0:4,respectively. The values were proportional to the mixing ratio ofES and CS, further demonstrating that co-digestion of ES with CScould stimulate the hydrogen production from CS. Although VSCS/VSES of 1:3 obtained the highest HY of 78 mL/g CS–VS added, thecumulative hydrogen amount was relatively low due to the smallpercentage of CS in the substrates.

Addition of certain amount of ES into CS was found to be bene-ficial for the hydrogen production, and such enhancement mightbe attributable to two reasons. Firstly, mixture of CS with proteinrich ES could provide a nutrient balanced condition for hydrogenproducing bacteria. Secondly, the added ES contains certain alka-linity, and the fermentative hydrogen production might be en-hanced through pH control. As the experiments were conductedunder batch operation mode, the involved mechanisms will be fur-ther investigated in the following continuous operation mode.

The production and distribution of VFA at each mixing ratio areshown in Fig. 2. Relatively high butyrate concentration was ob-served at VSCS/VSES of 4:0, 3:1 and 1:1 with n-butyrate accountingfor 55.3–79.1% of total VFA, which was coincided well with thehigh HYs. By increasing ES ratio to VSCS/VSES of 1:3 and 0:4, acetatebecame the dominant species (72.9–74.3% of total), correspondingto the lower HYs. The distribution of VFA was mostly determinedby substrate types. carbohydrate rich substrates usually take thebutyrate fermentation type and protein rich substrates follow theacetate fermentation type ((Kim et al., 2004; Lee et al., 2010; Luoet al., 2010; Xiao and Liu, 2009; Zhao et al., 2010). In this study,the high HY was correlated well with high n-butyrate concentra-tion. It was further suggested that the hydrogen was mainly pro-duced from the degradation of CS, and co-digestion of CS and ESshould be controlled at certain ratios.

3.1.2. Subsequent methane productionEffluent from hydrogen fermentation reactor with high concen-

tration of VFA/ethanol was used for subsequent methane produc-tion. Table 3 summarizes the cumulative CH4 production,methane yield (MY) at different VSCS/VSES values. The maximumcumulative methane production (1824 mL) and the maximumMY (214 mL/g total VS added) was obtained at VSCS/VSES of 4:0.Compared to ES, CS was more biodegradable. The TCOD, SCODand VS removal efficiency after methane production was morethan 61%, 80%, and 48%, respectively at VSCS/VSES of 4:0 and 3:1(Table 4). The coupled methane production after hydrogen produc-tion was helpful for the completely utilization of organics in thewastes. Although co-digestion of ES with CS did not present more

Tabl

e3

Hyd

roge

n/m

etha

neyi

elds

and

ener

gyyi

elds

.

VS C

S/V

S ES

Bat

ch(t

wo

phas

e)C

onti

nu

ous

(HR

T=

3da

ys)

4:0

3:1

1:1

1:3

0:4

4:0

3:1

1:1

1:3

0:4

Hyd

roge

nre

acto

rsH

2pr

odu

ctio

n(m

L)36

8±

1037

6±

1537

8±

823

4±

1021

±5

221.

0±

9.6

222.

10±

11.0

124.

4±

10.4

0.0

0.0

H2

yiel

d(m

L/g

tota

lVS

adde

d)30

.7±

0.8

31.3

±1.

331

.5±

0.7

19.5

±0.

81.

8±

0.4

73.7

±3.

274

.0±

3.7

41.5

±3.

90.

00.

0H

2yi

eld

(mL/

gC

S–V

Sad

ded)

30.7

±0.

841

.7±

1.6

63.0

±1.

378

.0±

3.3

073

.7±

3.2

93.3

±4.

982

.9±

7.7

0.0

0H

2en

ergy

yiel

d(K

J/kg

tota

lVS)

392.

1±

10.2

400.

1±

16.6

402.

7±

8.9

249.

2±

10.2

22.9

±5.

194

1.4

±40

.989

4.3

±47

.352

9.9

±49

.70.

00.

0C

H4

prod

uct

ion

(mL)

0.0

0.0

0.0

0.0

0.0

0.0

0.0

0.0

14.4

±3.

17.

5±

1.1

CH

4yi

eld

(mL/

gto

talV

Sad

ded)

0.0

0.0

0.0

0.0

0.0

0.0

0.0

0.0

4.8

±1.

22.

5±

0.4

Subs

eque

ntm

etha

nere

acto

rsC

H4

prod

uct

ion

(mL)

1824

±30

1686

±26

1300

±42

750

±38

535

±26

389.

4±

13.3

349.

7±

19.3

255.

9±

7.7

163.

4±

3.3

72.8

±2.

6C

alcu

late

dC

H4

prod

uct

ion

(mL)

1824

±30

1502

±23

1180

±15

858

±8

535

±26

389.

4±

13.3

310.

2±

10.0

231.

1±

6.7

151.

9±

3.3

72.8

±2.

6C

H4

yiel

d(m

L/g

tota

lVS

adde

d)21

4.5

±3.

319

5.5

±2.

815

6.8

±4.

789

.6±

4.2

66.3

±2.

937

9.6

±13

.033

8.9

±18

.722

0.5

±6.

511

0.0

±2.

861

.8±

2.2

CH

4en

ergy

yiel

d(K

J/kg

tota

lVS)

8585

.3±

132.

178

25.5

±11

2.1

6276

.1±

188.

235

87.3

±16

8.1

2655

.7±

116.

115

197.

7±

520.

413

567.

4±

748.

688

27.1

±26

0.2

4405

.2±

112.

124

75.6

±88

.1To

tal

ener

gyyi

eld

(KJ/

kgto

talV

S)89

77.4

±13

2.1

8225

.6±

112.

166

78.8

±18

8.2

3836

.5±

168.

126

78.6

±11

6.1

1613

9.1

±52

0.4

1446

1.8

±74

8.6

9357

.0±

260.

245

97.3

±11

2.1

2575

.1±

88.1

0

1000

2000

3000

4000

5000

6000

7000

8000

9000

10000

VSCS:VSES

=0:4VSCS:VSES

=1:3VSCS:VSES

=1:1VSCS

:VSES=3:1VSCS

:VSES=4:0

VFA

/ eth

anol

(mg/

l)

ethanol acetate propionate iso-butyrate n-butyrate iso-valerate n-valerate

a. Hydrogen reactor

0

1000

2000

3000

4000

5000

6000

7000

8000

9000

10000b. Methane reactor

VSCS:VSES

=0:4VSCS:VSES

=1:3VSCS:VSES

=1:1VSCS:VSES

=3:1VSCS:VSES

=4:0

VFA

/ eth

anol

(mg/

l)

Fig. 2. VFA/ethanol concentrations after hydrogen/methane fermentation in batchexperiments.

3836 W. Wang et al. / Bioresource Technology 102 (2011) 3833–3839

advantages on methane production than CS only, the calculatedmethane production (based on the methane production from sin-gle ES and single CS digestion and their mixing ratios) at mixing ra-tio of 3:1 and 1:1 was lower than their experimental data (Table 3).For example, the experimental methane production at VSCS/VSES of3:1 was 12% higher than the calculated value, indicating the syner-gistic effect from co-digestion of CS and ES.

Fig. 2 describes the distribution of VFA/ethanol obtained after24 days digestion. Accumulation of propionate and acetate wasobserved at each mixing ratio with total concentration of 3000–4000 mg/L. The distribution of VFA in methane phase will be fur-ther investigated in the following continuous operation mode.

3.1.3. OPAD performance and energy yieldFor comparison, methane production from OPAD was also con-

ducted. As shown in Table 3, the cumulative methane productionand MYs obtained from two-phase systems were higher than thevalues obtained from one-phase system. Moreover, the propionateaccumulated in OPAD was higher than that in TPAD, almost 1.5times of that in TPAD at the same VSCS/VSES of 3:1. Park et al. alsoobserved that the propionate concentration in one-phase reactorfluctuated largely and was much higher than that in two-phasereactor (Park et al., 2008).

Based on the observed HY and MY, the calculated energyyields are shown in Table 3 and Fig. 3. The densities of hydrogenand methane used for the calculation were 0.09 mg/mL and0.72 mg/mL, and the heating values used were 142 and 55.6 kJ/g

Tabl

e4

Liqu

idm

etab

olit

etr

ansf

orm

atio

n.

VS C

S/V

S ES

Bat

ch(t

wo

phas

e)C

onti

nu

ous

(HR

T=

3da

ys)

4:0

3:1

1:1

1:3

0:4

4:0

3:1

1:1

1:3

0:4

Hyd

roge

nre

acto

rspH

5.19

±0.

015.

6±

0.02

5.73

±0.

015.

99±

0.01

6.34

±0.

015.

44±

0.06

5.65

±0.

065.

76±

0.03

5.90

±0.

046.

15±

0.04

Solu

ble

carb

ohyd

rate

degr

adat

ion

(%)

14.3

±5.

660

.0±

9.7

70.5

±2.

547

.2±

7.1

15.2

±1.

464

±3

69±

252

±2

1±

9-7

1±

7So

lubl

epr

otei

nde

grad

atio

n(%

)�

9.9

±0.

9–1

8.9

±1.

7�

29.4

±2.

1�

51.4

±2.

8�

26.5

±1.

9�

33±

18�

0.3

±8

10±

712

±8

24±

8TC

OD

degr

adat

ion

(%)

8.4

±2

5.2

±2.

17.

1±

0.6

6.5

±0.

95.

6±

1.2

8.3

±0.

99.

1±

1.1

8.3

±1.

41.

2±

0.5

7.8

±0.

7SC

OD

degr

adat

ion

(%)

6.6

±0.

98.

2±

0.2

6.5

±0.

44.

0±

0.9

4.5

±0.

51.

0±

1.0

2.5

±0.

21.

3±

0.5

7.3

±1.

19.

4±

1.2

VS

degr

adat

ion

(%)

5.5

±0.

14.

2±

0.3

7.9

±0.

27.

0±

0.9

10.4

±0.

72.

0±

0.2

2.5

±0.

52.

2±

0.4

14.0

±1.

914

.5±

0.7

Subs

eque

ntm

etha

nere

acto

rspH

8.2

±0.

028.

2±

0.01

8.1

±0.

018.

1±

0.01

8.2

±0.

027.

88±

0.08

8.01

±0.

117.

85±

0.09

7.70

±0.

077.

29±

0.04

Solu

ble

carb

ohyd

rate

Deg

rada

tion

(%)

95.9

±1.

491

.4±

3.9

85.1

±2.

768

.7±

1.0

20.7

±1.

398

±1

95±

287

±1

25±

6�

99±

17So

lubl

epr

otei

nde

grad

atio

n(%

)0.

5±

1.9

11.8

±0.

811

.7±

1.4

14.2

±2.

120

.4±

1.8

73±

1050

±6

41±

530

±6

24±

4TC

OD

degr

adat

ion

(%)

67.4

±2.

961

.0±

5.5

39.3

±4.

231

.9±

3.3

14.8

±1.

768

.0±

1.3

62.2

±2.

138

.9±

1.9

33.1

±3.

329

.2±

1.9

SCO

Dde

grad

atio

n(%

)84

.0±

4.5

80.4

±1.

963

.1±

3.7

37.3

±5.

113

.1±

3.2

91.6

±2.

084

.2±

1.6

68.2

±2.

344

.0±

5.4

7.0

±2.

7V

Sde

grad

atio

n(%

)55

.6±

4.1

48.2

±2.

031

.0±

3.9

26.1

±5.

922

.7±

2.9

61.8

±1.

749

.8±

3.1

31.2

±7.

635

.3±

4.2

25.6

±0.

8

0

1000

2000

3000

4000

5000

6000

7000

8000

9000

10000

VSCS:VSES

=0:4VSCS:VSES

=1:3VSCS:VSES

=1:1VSCS:VSES

=3:1VSCS:VSES

=4:0

H2 energy yield on two phase CH4 energy yield on two phase CH4 energy yield on one phase

biog

as e

nerg

y yi

elds

(KJ

/ g-to

tal-V

S)

Fig. 3. Biogas energy yields (counting on total VS) on two and one phase anaerobicdigestion in batch experiments.

W. Wang et al. / Bioresource Technology 102 (2011) 3833–3839 3837

respectively (Zhu et al., 2008b). Energy generated from hydrogenin TPAD with different VSCS/VSES was only in the range of 0.9–6.5% of the total energy yields. The result was consistent withZhu et al. (2008b) who studied the hydrogen and methane pro-duction from potato waste and found only about 5% of the energywas generated from hydrogen. Although the contribution ofhydrogen to total energy is minor, the existence of hydrogenphase could stimulate the methane energy yield in the nextphase, and the total energy yields obtained from two-phasereactors were 0.02%, 25%, 18%, 17% and 14% higher than thosefrom one-phase reactors at different VSCS/VSES of 4:0, 3:1, 1:1,1:3 and 0:4 (Fig. 3). The above results suggested that hydrogenfermentation should be coupled with methane fermentation,and co-digestion could be further increase the energy yield. 25%increase in total energy yield indicated that the mixing ratio ofCS and ES of 3:1 was most suitable for co-digestion.

3.2. Performance of hydrogen and methane production undercontinuous operation mode

3.2.1. Hydrogen productionPerformance of hydrogen and methane production at different

mixing ratios was further studied under continuous operationmode. Effects of HRT of 5 and 3 days on hydrogen production werealso investigated in this study. The time courses of biogas produc-tion and pH in hydrogen production reactor are described in Fig. 4.The influent pH was firstly unadjusted to check whether the alka-linity of ES was sufficient enough to maintain suitable pH conditionfor hydrogen production. However, as shown in Fig. 4, the pHdropped gradually below 5 during the first 12 days, resulting inthe decrease and stop of hydrogen production for all mixing ratios.And then, the influent pH was adjusted to 6.0 by NaOH, andhydrogen was generated again. The effluent pH was stable anddiffered from 5.4 to 6.2 (Table 4). The pH declined gradually withthe increase of ES fraction, and the largest pH decline was obtainedin the single CS digested reactor. The pH in co-digestion reactorswith VSCS/VSES of 3:1and 1:1 were within the optimal range buthydrogen was produced with different levels. The above resultssuggested that the existence of ES could help to alleviate the rapiddecrease of pH, although its alkalinity was insufficient to controlthe pH by itself only.

In this study, hydrogen production was observed to fluctuate atHRT of 5 days, but was stable when HRT was shortened to 3 days(Fig. 4). Table 3 summarizes the HY calculated based on the total

0 5 10 15 20 25 30 35 40 45 50 55 60 65 70

0

50

100

150

200

250

300

350VSCS :VSES = 4:0

VSCS :VSES = 3:1 VSCS :VSES = 1:1 VSCS :VSES = 1:3 VSCS :VSES = 0:4

influent pH unadjusted

influent pH=6.0, HRT=3dinfluent pH=6.0, HRT=5d

Hyd

roge

n pr

oduc

tion

(ml)

a

0 5 10 15 20 25 30 35 40 45 50 55 60 65

0

5

10

15

20

25

30

Met

hane

pro

duct

ion

(ml)

b

0 5 10 15 20 25 30 35 40 45 50 55 60 65 704.0

4.2

4.4

4.6

4.8

5.0

5.2

5.4

5.6

5.8

6.0

6.2

6.4

influent pH=6.0, HRT=3dinfluent pH=6.0, HRT=5dinfluent pH unadjusted

pH

Time(d)

c

Fig. 4. Time courses of hydrogen/methane production and pH of hydrogen reactorsin continuous experiments.

0

1000

2000

3000

4000

5000

6000

7000

8000

9000

10000

VSCS:VSES

=1:1VSCS:VSES

=3:1 VSCS:VSES

=0:4VSCS:VSES

=1:3VSCS:VSES

=4:0

VFA

/ eth

anol

(mg/

l)

a. Hydrogen reactor ethanol acetate propionate iso-butyrate n-butyrate iso-valerate n-valerate

0

1000

2000

3000

4000

5000

6000

7000

8000

9000

10000b. Methane reactor

VFA

/ eth

anol

(mg/

l)

VSCS:VSES

=0:4VSCS:VSES

=1:3VSCS:VSES

=1:1VSCS:VSES

=3:1VSCS:VSES

=4:0

Fig. 5. VFA/ethanol concentrations after hydrogen/methane fermentation in con-tinuous experiments.

3838 W. Wang et al. / Bioresource Technology 102 (2011) 3833–3839

VS added (30 g/L) and the CS contributed VS added at HRT of3 days. The highest hydrogen production amount and HY basedon total VS were obtained at VSCS/VSES of 3:1 and 4:0, followedby that at 1:1. At VSCS/VSES of 1:3 and 0:4, only methane wasdetected, probably due to the enriched methanogens and highactivity of homoacetogens and propionate producers according tothe increased acetate and propionate concentration (Fig 5). How-ever, when HY was calculated based on the CS contributed VS,

the highest HY of 93.3 mL/g CS–VS added at 3:1 was 27% higherthan that at 4:0 and 13% higher than that at 1:1 (Table 3). Suchimprovement should be attributable to the more balanced nutrientcondition due to the addition of protein-rich ES. The initial C/N ra-tios at VSCS/VSES of 4:0, 3:1, 1:1, 1:3 and 0:4 were 3.2, 1.5, 0.8, 0.2and 0.1 g carbohydrate–COD/g protein–COD or 21.7, 11.9, 8.5, 6.9,and 6.0 g C/g N (calculated from elemental analysis) respectively.In this study with CS and ES as co-digested substrates for thermo-philic hydrogen production, the most balanced nutrient environ-ment was obtained at VSCS/VSES of 3:1 with C/N ratio of1.5 g carbohydrate–COD/g protein–COD or 11.9 g C/g N. The resultwas also consistent with the finding of Kim et al. (2004) that theoptimized C/N ratio of 1.66 g carbohydrate–COD/g protein–CODwas the best condition for hydrogen production from food wasteand sewage sludge co-digestion under mesophilic condition.

It should be noted that the HY obtained at mixing ratio of 3:1 incontinuous operation was almost two times than that in batchexperiment, indicating that hydrogen producing bacteria mightbe enhanced during continuous operation. Moreover, the increasedbutyrate concentration observed in continuous experiments(shown in Fig. 5) further demonstrated the above speculation.

3.2.2. Subsequent methane productionThe continuous methane production was stable after 25 days

operation for each condition. Table 3 presents the methaneproduction amount and MY obtained at each mixing ratios. Similarwith the results from batch experiments, synergistic effect was ob-served in co-digestion of ES and CS. The methane production at

W. Wang et al. / Bioresource Technology 102 (2011) 3833–3839 3839

VSCS/VSES of 3:1 was 13% higher than the simply calculated data.Thus co-digestion of CS and ES with VSCS/VSES of 3:1 not only opti-mally enhanced the hydrogen yield, but also increased the meth-ane yield. After hydrogen production, the C/N ratio (g C/g N) ineach reactor did not change much, and 11.9 g C/g N (in co-digestedreactor with VSCS/VSES of 3:1) was considered as the optimum C/Nratio for both hydrogen and methane production from CS and ESco-digestion in continuous experiment.

Although accumulation of propionate was observed for eachmixing ratio after methane fermentation in batch experiments,propionate was not detected at mixing ratio of 4:0, and less than400 mg/L propionate was detected at 3:1 in continuous experi-ment with HRT of 12.5 days after methane production (Fig. 5). AtVSCS/VSES of 1:1, the cumulated propionate showed some degrada-tion but remained at a low level (Fig. 5). However, at VSCS/VSES of0:4 and 1:3, the propionate was still cumulated and remained ata high level (Fig. 5b). A possible explanation for such accumulationwas that the degradation of propionate was not only determinedby the HRT, but also probably by the micro nutrients essentialfor the growth of propionate-degrading bacteria that may largelycontained in CS but not in ES. From this point of view, the mixingratio of CS and ES was not only essential for hydrogen and methaneproduction, but also important for the degradation of intermedi-ates produced during the anaerobic process. This further demon-strated that the ES addition into CS as co-substrate should becontrolled at certain level.

4. Conclusions

The study demonstrated that co-digestion of CS with certainamount of ES could improve the hydrogen and methane produc-tion in batch experiments. Two phase anaerobic digestion was con-firmed to have more abundant biogas production and bettereffluent quality compare with one phase anaerobic digestion. Thetotal energy yield obtained at VSCS/VSES of 3:1 in two phase was25% more than that in one phase. Continuous experiments furtherdemonstrated that VSCS/VSES of 3:1 was optimal for hydrogen pro-duction which could be attributed to the most balanced nutrientenvironment under C/N ratio of 11.9 g C/g N or 1.5 g carbohy-drate–COD/g protein–COD. VSCS/VSES of 3:1 was also optimal forcontinuous methane production considering the higher methaneyield and lower propionate concentration in the effluent.

Acknowledgements

This research was supported in part by the Water PollutionControl and Management of China (No. 2008ZX07316-002), theFoundation of The State Key Laboratory of Pollution Control andResource Reuse, China (No. PCRRK09004) and the Bayer Sustain-able Development Foundation.

References

APHA, 1998. Standard Methods for the Examination of Water and Wastewater, 20thed. American Public Health Association, Washington, DC, USA.

Dubois, M., Gilles, K.A., Hamilton, J.K., Rebers, P.A., Smith, F., 1956. Colorimetricmethod for determination of sugar and related substance. Analytical Chemistry28, 350–356.

Espinosa, A., Rosas, L., Ilangovan, K., Noyola, A., 1995. Effect of trace metals on theanaerobic degradation of volatile fatty acids in molasses stillage. Water Scienceand technology 32, 121–129.

Fezzani, B., Ben Cheikh, R., 2010. Two-phase anaerobic co-digestion of olive millwastes in semi-continuous digesters at mesophilic temperature. BioresourceTechnology 101, 1628–1634.

Kaparaju, P., Serrano, M., Thomsen, A.B., Kongjan, P., Angelidaki, I., 2009. Bioethanol,biohydrogen and biogas production from wheat straw in a biorefinery concept.Bioresource Technology 100, 2562–2568.

Kim, M.S., Lee, D.Y., 2010. Fermentative hydrogen production from tofu-processingwaste and anaerobic digester sludge using microbial consortium. BioresourceTechnology 101, S48–S52.

Kim, S.H., Han, S.K., Shin, H.S., 2004. Feasibility of biohydrogen production byanaerobic co-digestion of food waste and sewage sludge. International Journalof Hydrogen Energy 29, 1607–1616.

Lay, J.J., Lee, Y.J., Noike, T., 1999. Feasibility of biological hydrogen productionfrom organic fraction of municipal solid waste. Water Research 33, 2579–2586.

Lee, D.Y., Ebie, Y., Xu, K.Q., Li, Y.Y., Inamori, Y., 2010. Continuous H-2 and CH4

production from high-solid food waste in the two-stage thermophilicfermentation process with the recirculation of digester sludge. BioresourceTechnology 101, S42–S47.

Leng, R., Wang, C.T., Zhang, C., Dai, D., Pu, G.Q., 2008. Life cycle inventory and energyanalysis of cassava-based fuel ethanol in China. Journal of Cleaner Production16, 374–384.

Li, D., Yuan, Z., Sun, Y., Kong, X., Zhang, Y., 2009. Hydrogen productioncharacteristics of the organic fraction of municipal solid wastes by anaerobicmixed culture fermentation. International Journal of Hydrogen Energy 34, 812–820.

Lowry, O.H., Rosebrough, N.J., Farr, A.L., Randall, R.J., 1951. Protein measurementwith the Folin phenol reagent. The Journal of Biological Chemistry 193, 265–275.

Luo, G., Xie, L., Zhou, Q., 2009. Enhanced treatment efficiency of an anaerobicsequencing batch reactor (ASBR) for cassava stillage with high solids content.Journal of Bioscience and Bioengineering 107, 641–645.

Luo, G., Xie, L., Zou, Z.H., Wang, W., Zhou, Q., 2010. Evaluation of pretreatmentmethods on mixed inoculum for both batch and continuous thermophilicbiohydrogen production from cassava stillage. Bioresource Technology 101,959–964.

Luste, S., Luostarinen, S., 2010. Anaerobic co-digestion of meat-processing by-products and sewage sludge – effect of hygienization and organic loading rate.Bioresource Technology 101, 2657–2664.

Okamoto, M., Miyahara, T., Mizuno, O., Noike, T., 2000. Biological hydrogenpotential of materials characteristic of the organic fraction of municipal solidwastes. Water Science and technology 41, 25–32.

Park, Y., Hong, F., Cheon, J., Hidaka, T., Tsuno, H., 2008. Comparison of thermophilicanaerobic digestion characteristics between single-phase and two-phasesystems for kitchen garbage treatment. Journal of Bioscience andBioengineering 105, 48–54.

Shanmugam, P., Horan, N.J., 2009. Optimising the biogas production from leatherfleshing waste by co-digestion with MSW. Bioresource Technology 100, 4117–4120.

Wilkie, A.C., Riedesel, K.J., Owens, J.M., 2000. Stillage characterization and anaerobictreatment of ethanol stillage from conventional and cellulosic feedstocks.Biomass and Bioenergy 19, 63–102.

Xiao, B.Y., Liu, J.X., 2009. Biological hydrogen production from sterilized sewagesludge by anaerobic self-fermentation. Journal of Hazardous Materials 168,163–167.

Yeoh, B.G., 1997. Two-phase anaerobic treatment of cane-molasses alcohol stillage.Water Science and technology 36, 441–448.

Yu, S.R., Tao, J., 2009. Energy efficiency assessment by life cycle simulation ofcassava-based fuel ethanol for automotive use in Chinese Guangxi context.Energy 34, 22–31.

Zhao, Y.X., Chen, Y.G., Zhang, D., Zhu, X.Y., 2010. Waste activated sludgefermentation for hydrogen production enhanced by anaerobic processimprovement and acetobacteria inhibition: the role of fermentation pH.Environmental Science and Technology 44, 3317–3323.

Zhu, H.G., Parker, W., Basnar, R., Proracki, A., Falletta, P., Beland, M., Seto, P.,2008a. Biohydrogen production by anaerobic co-digestion of municipal foodwaste and sewage sludges. International Journal of Hydrogen Energy 33,3651–3659.

Zhu, H.G., Stadnyk, A., Beland, M., Seto, P., 2008b. Co-production of hydrogen andmethane from potato waste using a two-stage anaerobic digestion process.Bioresource Technology 99, 5078–5084.