Optimizing hydrogen production from organic wastewater treatment in batch reactors through experimental and kinetic analysis

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 4 ( 2 0 0 9 ) 6 1 7 1 – 6 1 8 0

Avai lab le a t www.sc iencedi rec t .com

j ourna l homepage : www.e lsev ier . com/ loca te /he

Optimizing hydrogen production from organic wastewatertreatment in batch reactors through experimentaland kinetic analysis

Yogesh Sharma, Baikun Li*

Department of Civil and Environmental Engineering University of Connecticut, Storrs, CT 06269, USA

a r t i c l e i n f o

Article history:

Received 9 June 2009

Accepted 13 June 2009

Available online 5 July 2009

Keywords:

Hydrogen production

Anaerobic wastewater treatment

Chemical oxygen demand

pH

Specific hydrogen yield

Hydrogen production rate

* Corresponding author. Tel.: 860 486 2339; faE-mail address: [email protected]

0360-3199/$ – see front matter Published bydoi:10.1016/j.ijhydene.2009.06.031

a b s t r a c t

Anaerobic hydrogen production from organic wastewater, an emerging biotechnology to

generate clean energy resources from wastewater treatment, is critical for environmental

and energy sustainability. In this study, hydrogen production, biomass growth and organic

substrate degradation were comprehensively examined at different levels of two critical

parameters (chemical oxygen demand (COD) and pH). Hydrogen yields had a reverse

correlation with COD concentrations. The highest specific hydrogen yield (SHY) of 2.1 mole

H2/mole glucose was achieved at the lowest COD of 1 g/L and decreased to 0.7 mole H2/

mole glucose at the highest COD of 20 g/L. The pH of 5.5–6.0 was optimal for hydrogen

production with the SHY of 1.6 mole H2/mole glucose, whereas the acidic pH (4.5) and

neutral pH (6.0–7.0) lowered the hydrogen yields. Under all operational conditions, acetate

and butyrate were the main components in the liquid fermentation products. Additionally,

a comprehensive kinetic analysis of biomass growth, substrate degradation and hydrogen

production was performed. The maximum rates of microbial growth (mm) and substrate

utilization (Rsu) were 0.03 g biomass/g biomass/day and 0.25 g glucose/g biomass/day,

respectively. The optimum pH for the rate of hydrogen production (RH2 ) and SHY were 5.89

and 5.74 respectively. Based on the kinetic analysis, the highest RH2 and SHY for batch-

mode anaerobic hydrogen production systems were projected to be 13.7 mL/h and

2.32 mole H2/mole glucose.

Published by Elsevier Ltd on behalf of International Association for Hydrogen Energy.

1. Introduction following methanogenic phase (Reaction (2)). The fermenta-

Wastewater contains significant amount of organic

substances (i.e. carbohydrates) that can be converted to

energy (i.e. methane, hydrogen). Hydrogen is expected to be

a substitute for fossil fuel as a clean energy resource, since it

only generates water when burning [1]. In the two-phase

anaerobic treatment of wastewater, hydrogen is produced in

the acidogenesis phase (Reaction (1)), which is much faster

and more resistant to environmental shocks than the

x: 860 486 2298(B. Li).Elsevier Ltd on behalf of

tion liquid products are mainly volatile fatty acids (VFAs) such

as acetic acid, propionic acid and butyric acid that are easily

biodegradable. Several bacteria species (i.e. Clostridium pas-

teurianum) have been found to degrade glucose and produces

hydrogen along with acetate and butyrate (Reactions (1) and

(3)) [2].

C6H12O6þ 2H2O / 2CH3COOHþ 4H2þ 2CO2 (1)

International Association for Hydrogen Energy.

Nomenclature

VFAs volatile fatty acids

SHY specific hydrogen yield, mole H2/mole glucose

k maximum specific substrate utilization rate,

g glucose/g biomass/day

X biomass concentration, g/L

S chemical oxygen demand, g/L

Ks half velocity constant, g/L

m specific growth rate constant, g biomass/

g biomass/day

RH2 rate of hydrogen production, mL/h

Rm maximum rate of hydrogen production, mL/h

Rsu rate of substrate utilization, g glucose/

g biomass/day, respectively

K1, K2 protonation and deprotonation equilibrium

constants respectively, mole/L

Ki inhibition constant, g/L

[Hþ] hydrogen ion concentration, mole/L

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 4 ( 2 0 0 9 ) 6 1 7 1 – 6 1 8 06172

CH3COOH / CO2þCH4 (2)

C6H12O6 / CH3CH2CH2COOHþ 2H2þ 2CO2 (3)

Anaerobic hydrogen production is critical for environ-

mental and energy sustainability. In the past two decades,

tremendous efforts have been put to investigate a variety of

inoculums (i.e. activated sludge, soil, pure cultures) [3–5],

substrates (i.e. wastewater, starch, beet sugar, cellulose) [6–8],

anaerobic fermentation pathways [9,10], and bioreactors (i.e.

anaerobic baffled reactors, upflow reactors, continuous stirred

tank reactors) [11–13].

Currently, there are three problems of hydrogen production

from anaerobic wastewater treatment. First, the experimental

hydrogen production is much lower than the theoretical

values. The theoretical specific hydrogen yield (SHY) is 4 mole

H2/mole glucose when acetate is the sole fermentation product

(Reaction (1)). The yield decreases to 2 mole H2/mole glucose

when butyrate is the sole fermentation product (Reaction (3)).

But the experimental hydrogen yields are only 0.9–2.0 mole H2/

mole glucose [6,14]. Second, the accumulation of acidic

fermentation products in the solution and hydrogen partial

pressure in the headspace is known to inhibit bacterial activi-

ties, suppress fermentation and reduce hydrogen yields

[25–27]. Some engineering parameters (i.e. COD, pH, tempera-

ture, and hydraulic retention time) have been studied indi-

vidually for their effects on hydrogen production [5,9,12]. But

the understanding of the hydrogen production under different

operational conditions is still very limited. Third, there have

been some preliminary kinetic studies of hydrogen production

using pure cultures of Enterobacter sp. and Clostridium sp. in

batch mode [15–24]. A comprehensive kinetic analysis of the

effects of engineering parameters (i.e. COD and pH) on

hydrogen production, substrate utilization and biomass

growth is limited, especially for the mixed cultures (i.e. acti-

vated sludge). In a biochemical process such as anaerobic

hydrogen production, different operational conditions

substantially alter the rates of hydrogen production, substrate

utilization and biomass growth. There is no kinetic study

describing the effects of COD and pH on the rate of hydrogen

production (RH2 , mL/h) and specific hydrogen yield (SHY, mole

H2/mole glucose) using mixed cultures.

There are two-fold objectives of this study to enhance the

understanding of hydrogen production from organic waste-

water treatment. First, two critical operational parameters

(COD and pH) were examined at different levels in batch

systems to determine their effects on hydrogen production

and substrate removal. The quantitative correlations between

hydrogen production and operational parameters were

determined. Second, a comprehensive kinetic analysis,

elucidating the effect of operational parameters on RH2 , SHY,

and substrate removal (Rsu) was performed using Michaelis–

Menton kinetic analysis. The maximum biomass growth rate

(mm) was determined using the Monod kinetic analysis.

2. Material and method

2.1. Batch-mode hydrogen production system setup

Serum bottles (volume: 150 mL) (Wheaton Scientific, Millville,

NJ) were used as batch reactors in this study. The reactors were

filled to 125 mL with a glucose solution as the organic substrate.

An inorganic medium (consisting of per L of water: 2.0 g

NH4HCO3, 1.0 g KH2PO4, 100 mg MgSO4$7H2O, 10 mg NaCl,

10 mg Na2MoO4$2H2O, 10 mg CaCl2$2H2O, 15 mg MnSO4$7H2O,

and 2.78 mg FeCl2) was added to the solution as supplementary

micronutrients for bacteria to grow [6]. The fraction of head-

space (17% of the bottle volume) was designed to reduce the

duration of pressure build-up in the lag period and was consis-

tent with other studies with the similar reactor configuration [6].

The solution was mixed in the serum bottles and then

buffered with 0.05 M 2-(N-morpholino) ethanesulfonic acid

monohydrate (MES; J.T. Baker, Phillipsburg, NJ). The pH of the

solution was adjusted using 1 M KOH. The solution was

sparged with nitrogen for 15 min to remove oxygen

completely. The bottles were then capped with rubber septum

stoppers (Wheaton Scientific, Millville, NJ). The serum bottles

were constantly stirred at 250 rpm. The soil collected from an

organic farm was used as the inocula. Due to the effectiveness

of heat shock treatment in inhibiting methanogens and

selecting for hydrogen-producing bacteria, the soil was heated

shocked at 100 �C for 2 h before being added into the batch

reactors [6,28]. The soil concentration in the batch reactors

was kept at 20 g/L for all tests.

2.2. Operational condition

Two critical operational parameters including COD and pH

were tested individually. COD concentration varies with

wastewater sources (e.g. storm runoff, agricultural waste-

water) and pH varies with acidic fermentation products in

anaerobic systems. All these variations will affect the pop-

ulations and activities of hydrogen-producing bacteria. Five

levels of COD (1–20 g/L) and five levels of pH (4.5–7.0) were

examined with only one parameter being changed under each

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 4 ( 2 0 0 9 ) 6 1 7 1 – 6 1 8 0 6173

condition, leading to the total eleven operational conditions. A

baseline of COD concentration of 5 g/L and pH of 5.5 was

selected for this study. All tests were conducted at 30 �C.

2.3. Control bottle tests

Since soil contains a variety of organic substances that can be

utilized by anaerobic bacteria, a control test without the

addition of glucose was conducted under each operational

condition in order to identify the contribution of these organic

substances to hydrogen production. The actual hydrogen

volume generated from the degradation of glucose equals the

total hydrogen volume generated in the batch reactors (with

glucose being added) minus the hydrogen volume generated

in the control bottles (without glucose being added).

2.4. Biogas volume measurement

The biogas was collected periodically from the headspace of

the batch reactors [29]. The total volume of biogas was

measured by inserting a glass syringe with a capacity of 20 or

100 mL (Perfektum Syringes; Popper & Sons, Inc., NY) into the

headspace of the batch reactor and allowing biogas to flow

into the syringe against atmospheric pressure. The volume

reading was taken when an equilibrium status was achieved

inside the syringe.

2.5. Analytical methods

The biogas in the headspace was sampled with a gas tight

syringe with a capacity of 50 mL (SGE, Illinois) and analyzed

using a gas chromatograph (Agilent 6890N) equipped with

a packed column and a thermal conductivity detector. The

glucose concentration was measured by performing the

phenol-sulfuric acid assay with a spectrophotometer (Cary 50

Bio; Varian, CA) [30]. The pH of the solution in the batch reac-

tors was measured at the end of each test using an Ag/AgCl

reference electrode (Accumet AP72). The pH of organic soil was

measured according to the standard method [31]. The biomass

concentration was measured using the total solids measure-

ment technique according to the standard method [32].

2.6. Kinetic analysis

To describe the effect of COD concentrations on RH2 , Andrew’s

inhibition model was used (Equation (1)) [15].

R ¼ ðRm � SÞ=�Ks þ Sþ S2=Ki

�(1)

where Rm is the maximum rate of hydrogen production

(mL/h), Ks is the half saturation constant (g/L), Ki is the

inhibition constant (g/L), and S is COD concentration (g/L).

To describe the effect of COD concentrations on SHY, the

Michaelis–Menton kinetic model was employed [33].

SHY ¼ ðSHYm � SÞ=ðKs þ SÞ (2)

where SHYm is the maximum SHY (mole H2/mole glucose), S is

COD concentration (g/L), Ks is half saturation constant (g/L).

To describe the effect of pH on RH2 , SHY and substrate

utilization rate (Rsu), a model of active ionization state was

used (Reaction (4)) [33]. E� represents the active enzyme form,

while E and E2� are inactive enzyme forms that are obtained

by the protonation and deprotonation of E�. The Michaelis pH

function (Equation (3)), derived by determining the fraction of

active enzyme concentration (E�), was used to analyze the

dependence of RH2 on pH [33].

E %�Hþ

HþK1

E� %�Hþ

HþK2

E2� (4)

RH2¼�Rm �

�Hþ����

K1 þ�Hþ�þ�Hþ�2=K2

�(3)

where Rm is maximum RH2(mL/h), K1 and K2 are equilibrium

constants (moles/L) respectively (Reaction (4)), and [Hþ] is

hydrogen ion concentration in the solution (moles/L).

To describe the effect of pH on SHY, the model of active

ionization state was used in formulation of Michaelis–Menton

function (Equation (4)).

SHY ¼�SHYm �

�Hþ����

K1 þ�Hþ�þ�Hþ�2=K2�

(4)

where SHYm is the maximum SHY (mole H2/mole glucose)

To illustrate the effect of pH on Rsu, Equation (5) was used.

Rsu ¼�Rsum �

�Hþ����

K1 þ�Hþ�þ�Hþ�2=K2

�(5)

where Rsum is the maximum Rsu (g glucose/day).

The optimum pH for RH2 , SHY, and Rsu was calculated using

Equation (6) [33].

pH ¼ loghðK1 � K2Þ1=2

i(6)

The Monod kinetic analysis of substrate utilization and

biomass growth were performed in this study. For substrate

utilization kinetics, the Equation (7) was used [34,35].

Rsu ¼ �ðk� X� SÞ=ðKsþ SÞ (7)

where k is maximum Rsu (g substrate/g biomass/day), X is

biomass concentration (mg/L), S is COD in solution (mg/L), and

Ks is half velocity constant (mg/L).

For microbial growth analysis, the Equation (8) was

employed.

m ¼ ðmm � SÞ=ðKsþ SÞ (8)

where mm is maximum specific growth rate (1/day).

All the kinetic analyses were performed using GraphPad

software.

In addition, the regression analyses were performed by the

statistical software MINITAB to illustrate theeffects of different

operational conditions (COD and pH) on hydrogen production.

3. Results and discussion

3.1. Effect of substrate concentration (COD) on specifichydrogen yield (SHY)

Substrate concentration is an important parameter in opti-

mizing the hydrogen yields due to its significant effects on

bacterial growth and enzymatic activity [36]. In the control

3

3.5

4

4.5

5

5.5

0 5 10 15 20

COD (g/L)

pH

Initial pH

Final pH

Fig. 2 – The changes of pH at different COD concentrations

before and after batch tests.

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 4 ( 2 0 0 9 ) 6 1 7 1 – 6 1 8 06174

tests without the addition of glucose, hydrogen production

(w5 mL) was still detected in the batch reactors due to the

degradation of organic substances contained in soil. The

actual specific hydrogen yields (SHY, mole H2/mole glucose)

had a reverse correlation with COD concentrations (Equation

(9)) (Fig. 1). The correlation coefficient (R2) of 0.97 indicates the

high adequacy of correlation. The p-value for the COD effects

was 0.028, supporting the difference of yields at different COD

concentrations at 95% confidence interval.

SHY ¼ 2:380� 0:1601 CODþ 0:003956 COD2 (9)

The SHY values decreased from 2.1 to 0.7 mole H2/mole

glucose and the hydrogen content decreased from 55% to 32%

when the glucose concentration increased from 1 to 20 g/L.

There are three possible reasons for the reverse relationship

between hydrogen production and substrate concentrations.

First, the accumulation of liquid fermentation products at

higher substrate concentrations resulted in the over-acidifi-

cation of bacterial cultures and the inhibition of fermentation

[37]. A lower substrate concentration reduced the accumula-

tion of the acidic fermentation production, and led to higher

specific hydrogen yields. The measurement of volatile fatty

acids (VFAs) showed that 7.3 g/L of total VFAs were detected at

the highest substrate concentration of 20 g/L, while only

0.2 g/L VFAs were detected at the substrate concentration of

1 g/L. Second, the pH of the fermentation solution substan-

tially decreased with higher substrate concentrations. The

initial pH of the solution was set at 5.5 for all substrate

concentrations. The pH dropped sharply to 3.5 at the substrate

concentration of 20 g/L, whereas the pH slightly decreased to

5.4 at the substrate concentration of 1 g/L (Fig. 2). Third, the

total hydrogen production (mole H2) increased at high

substrate concentrations, which led to the high partial pres-

sure of hydrogen in the headspace throughout the batch test

period. The accumulation of hydrogen in the batch reactors

inhibited the fermentation reactions, and thus lowered the

specific hydrogen yields [38].

The SHY values obtained in this study (1.6–2.1 mole H2/

mole glucose) were much higher than the values previously

reported (0.92 mole H2/mole glucose at a COD concentration of

0.6

0.85

1.1

1.35

1.6

1.85

2.1

2.35

0 5 10 15 20 25COD (g/L)

Sp

ecific H

yd

ro

gen

Y

ield

( m

ole

Hyd

ro

gen

/m

ole G

lu

co

se)

0

10

20

30

40

50

60

Hyd

ro

gen

C

on

ten

t (%

)

Specific Hydrogen YieldHydrogen Content

R2 = 0.97

Fig. 1 – The specific hydrogen yield (SHY, mole hydrogen/

mole glucose) and the hydrogen content in biogas (%) at

different COD concentrations (pH: 5.5, temperature: 30 8C).

4 g/L) [6]. The higher SHY in this study might be caused by the

higher temperature (30 �C) than their studies (26 �C). It has

been shown that increasing the temperature within meso-

philic range has a positive effect on hydrogen yields.

3.2. Effect of pH on specific hydrogen yield (SHY)

The pH is an important operational parameter for hydrogen

production, since it affects anaerobic metabolic pathways and

the activities of hydrogenase enzymes [39,40]. The SHY values

were the lowest (1.2 mole H2/mole glucose) at the pH of 4.5,

which indicated that the acidic condition inhibited the

fermentation and hydrogen production. The SHY values

substantially increased to 1.55 mole H2/mole glucose at the pH

of 5.5 and 6.0. However, the SHY values decreased to 1.3 and

1.25 at the higher pH of 6.5 and 7.0 (Fig. 3). It had been found

that under the neutral pH condition, a significant amount of

substrates were consumed by bacterial growth other than

hydrogen production [14], which was verified by the higher

biomass concentration at higher pH than at lower pH. This

could explain the lower hydrogen yield at neutral pH. The

correlation between SHY and pH was established by non-

linear regression analysis, yielding a cubic correlation (Equa-

tion (10)). The correlation obtained in this study adequately

described the hydrogen yield as a function of pH with

a correlation coefficient of 0.93 (Fig. 3).

SHY ¼ �29:61þ 15:39 pH� 2:493 pH2 þ 0:1319 pH3 (10)

The low pH has been known to suppress the activity of

hydrogenase [41,42]. A pH of 5.5 has been reported to be ideal

for hydrogen production [14,26], while the hydrogen yield at

6.0 was slightly higher than at the pH of 5.5 in this study.

Because different types of bacterial inoculums such as acti-

vated sludge, soil and pure culture (i.e. Clostridium sp) were

used in these studies, the optimal pH for hydrogen production

changed slightly with bacterial inoculums. Nevertheless, this

study indicated that for the batch reactors inoculated with

soil, a pH of 5.5–6.0 was optimal for hydrogen production.

The initial and final pH values of the solution were

compared under each pH condition (Fig. 4). When the initial

1.1

1.2

1.3

1.4

1.5

1.6

4 4.5 5 5.5 6 6.5 7 7.5pH

Sp

ecific H

yd

ro

gen

Y

ield

(m

ole

Hyd

ro

gen

/m

ole G

lu

co

se)

50

55

60

65

Hyd

ro

gen

C

on

ten

t (%

)

R2 = 0.93

Specific Hydrogen YieldHydrogen Content

Fig. 3 – The specific hydrogen yield (SHY, mole hydrogen/

mole glucose) and the hydrogen content in biogas (%) at

different pH (COD: 5 g/L, temperature: 30 8C).

0 100 200 300Time (hours)

0

5

10

15

20

25

30

Hyd

ro

gen

V

olu

me (m

L)/ g

CO

D

1

5

10

15

20

a

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 4 ( 2 0 0 9 ) 6 1 7 1 – 6 1 8 0 6175

pH was neutral (6.5–7.0), the final pH was 1.5–2.0 lower than

the initial pH. When the initial pH was 4.5, the final pH was

only 0.1 lower than the initial pH. The difference between the

pH values demonstrated that the fermentation reactions

produced a significant amount of acidic liquid products. At the

higher pH, the fermentation proceeded thoroughly and

generated high amounts of VFAs, which led to the significant

drop of pH in the solution. At the lower pH, the fermentation

production was inhibited, which produced less amounts of

acidic liquid products and led to the slight drop of pH values.

The effects of the acidic liquid products on the pH were well

correlated with the SHY values.

The hydrogen content of biogas was 51–57% at the pH

range of 4.5–6.5 and increased to 63% at highest pH of 7.0

(Fig. 3), with the exception of 51% at the pH of 6.0. The soil

tested had the pH of 6.8, which indicated that the hydrogen-

producing bacteria were acclimatized to the neutral pH in

natural environment. The increased metabolic activities of

the hydrogen-producing bacteria at neutral pH may have

4

5

6

7

4 5 6 7Initial pH

Fin

al p

H

Fig. 4 – The initial and final pH values in the batch reactors

during the pH-level tests.

resulted in higher hydrogen content in biogas than those at

the lower pH.

3.3. The correlation between hydrogen production andglucose degradation

The hydrogen production volume (mL) in the headspace was

measured periodically and correlated with the glucose

degradation efficiency (%) in the solution. In the COD-level

tests (pH: 5.5, temperature 30 �C), there was a lag period of 25 h

in all batch reactors before the biogas production was detec-

ted. After the lag period, the hydrogen production increased

exponentially (Fig. 5a). It was clearly shown that the initial

glucose concentrations determined the duration of exponen-

tial phase. For instance, at the low concentrations of 1 and

5 mg/L, the duration of the exponential phase was 30 h. At the

middle concentration of 10 mg/L, the duration increased to

55 h. At the high concentration of 15–20 g/L, the duration

reached 88 h. The increase in the duration period of the

exponential phase was caused by the availability of organic

substrates at high concentrations. After the exponential

phase, the hydrogen production began to decline. At the 262nd

hour, the hydrogen production stopped in all reactors. In

terms of glucose consumption, at the 72nd hour, 95% of

glucose was degraded when the initial glucose concentrations

were 1 and 5 g/L, 90% of glucose was degraded when the initial

0102030405060708090

100

0 100 200 300

Time (hours)

Glu

co

se R

em

oval E

fficien

cy

(%

)

b

Fig. 5 – The changes of hydrogen production (mL H2/g COD)

(a) and the glucose removal efficiency (%) (b) at different

COD concentrations during batch test periods.

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 4 ( 2 0 0 9 ) 6 1 7 1 – 6 1 8 06176

concentration was 10 g/L, and 70–80% of glucose was degraded

when the initial concentration was 15–20 g/L (Fig. 5b).

Although the glucose removal efficiency decreased at higher

concentrations, the amounts of glucose removed were actu-

ally higher. For example, 0.99–4.75 g/L glucose was removed

when the initial concentrations were 1–5 g/L, whereas 12–14 g/

L glucose was removed when the initial concentrations were

15–20 g/L.

In the pH-level tests (COD: 5 g/L, temperature: 30 �C), the

rate of hydrogen production increased as the pH was

increased from 4.5 to 7.0 (Fig. 6a). The pH of organic soil tested

in this study was 6.8, indicating that the bacteria in the soil

could acclimatize quickly at neutral pH and produce hydrogen

at a faster rate. This quick acclimation at neutral pH resulted

in a shorter duration of exponential phase (22 h) at the pH of

6.0–7.0 than those (30 h) at lower pH of 4.5–5.5. In terms of

glucose degradation, over 97% of glucose was degraded at the

pH of 7.0 after 43 h (Fig. 6b), whereas only 25% of glucose was

degraded at the pH of 4.5 during the same test period. This

clearly indicated that the metabolic activities of bacteria in

soil were faster at neutral pH than at lower pH. After 72 h,

95–99% of glucose was degraded in all reactors. No glucose

was detected at the end of experiment at 262 h.

3.4. The kinetic analysis of hydrogen production

In order to improve hydrogen production, it is critical to

determine the effects of operational conditions on hydrogen

production rate (RH2 ) and specific hydrogen yield (SHY). The

0102030405060708090

100

0 100 200 300

Time (hours)

Glu

co

se R

em

oval E

fficien

cy

(%

)

0 100 200 300

Time (hours)

0

5

10

15

20

25

Hyd

ro

gen

V

olu

me (m

L)/ g

CO

D

76.5

65.54.5

a

b

Fig. 6 – The changes of hydrogen production (mL H2/g COD)

(a) and the glucose removal efficiency (%) (b) at different pH

during batch test periods.

kinetic analysis of the effects of COD concentrations on

hydrogen production indicated that RH2 increased at the COD

of 1–5 g/L, but decreased at higher COD of 10–20 g/L (Fig. 7a).

This inhibition of hydrogen production at high COD was

effectively demonstrated by Andrew’s inhibition model

(Equation (1)). The maximum rate of hydrogen production (Rm)

was 13.7 mL/h, which corresponded well to the reported rate

of 15.1 mL/h (with glucose as the substrate and activated

sludge as inoculum, [43]). Increasing the COD within the range

of 5 g/L<COD< 10 g/L would increase the RH2 up to 13.7 mL/h.

Beyond that point, the RH2 would start to decrease due to

substrate inhibition. The Ks and Ki were 12.23 and 3.24 g/L

respectively. A high correlation coefficient of 0.97 suggested

the adequacy of Andrew’s model in describing the inhibitory

effect of COD on hydrogen production.

The experimental results indicated that SHY had a reverse

relationship with COD concentrations, increasing from

0.7 molesH2/mole glucoseat20 g/L to 2.1 molesH2/mole glucose

at 1 g/L (Fig. 1). The SHY was plotted against 1/S according to

Michaelis–Menton kinetics, with the projected maximum SHY

of 2.32 moles H2/mole glucose (Fig. 7b). This demonstrated that

decreasing the COD concentration below the lowest COD tested

(1 g/L) would result in a further increase in SHY reaching

a maximum of 2.32 mole H2/mole glucose. At this point, the SHY

is limited by other operational parameters such as pH,

temperature, stirringspeed aswell asbiologicalpropertiesof the

system. Therefore, the manipulation of COD alone is not

0 5 10 15 20 250

1

2

3

4

COD (g/L)

RH

2 (m

L/h

ou

r)

0.0 0.5 1.0 1.50.0

0.5

1.0

1.5

2.0

2.5

1/COD (L/g)

Sp

ecific H

yd

ro

gen

Y

ield

(m

ole H

2/m

ole g

lu

co

se)

R2 = 0.97

R2 = 0.99

b

a

Fig. 7 – The kinetic analysis of COD effect on hydrogen

production rate (RH2 ) (a) and specific hydrogen yield (SHY)

(b).

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 4 ( 2 0 0 9 ) 6 1 7 1 – 6 1 8 0 6177

sufficient to achieve theoretical SHY of 4 mole H2/mole glucose.

The transformed Michaelis–Menton kinetic model was found to

be adequate with a correlation coefficient of 0.99.

The effects of pH on RH2 were modeled by Michaelis pH

function (Equation (3)) (Fig. 8a). Through a non-linear regres-

sion of RH2 versus [Hþ], the equilibrium constants K1 and K2

were found to be 4.26� 10�8 and 3.96� 10�5 mole/L, respec-

tively. A high correlation coefficient of 0.85 illustrated the

suitability of Michaelis pH function for analysis of RH2 . Based

on Equation (6), the optimum pH of 5.89 was calculated for RH2 .

This pH value was significantly lower than the reported pH of

6.35 for hydrogen production rate from starch hydrolysate

[44]. This difference in optimum pH can be attributed to use of

pure culture of Enterobacter aerogenes in that study compared

to mixed culture in this study.

The effects of pH on SHY were modeled by Michaelis pH

function (Equation (4)). Through a non-linear regression of

SHY versus [Hþ], the equilibrium constants K1 and K2 were

3.07� 10�8 and 3.5� 10�5 mole/L respectively with a correla-

tion coefficient of 0.78 (Fig. 8b). The optimum pH for SHY was

5.74 (Equation (6)), which was not significantly different than

the optimum pH of 5.89 for RH2 .

3.5. The kinetic analysis of substrate utilization

The kinetic analysis of substrate utilization rate (Rsu) was

performed for different COD concentrations and pH. The Rsu

0.00001 0.00002 0.00003

2.0

2.5

3.0

[H+] (mole/L)

RH

2 (m

L/h

ou

r)

0.00000 0.00001 0.00002 0.00003 0.000041.0

1.2

1.4

1.6

1.8

[H+] (mole/L)

Sp

ecific H

yd

ro

gen

Y

ield

(m

ole H

2/m

ole g

lu

co

se)

R2 = 0.85

R2 = 0.78

b

a

Fig. 8 – The kinetic analysis of pH effect on hydrogen

production rate (RH2 ) (a) and specific hydrogen yield

(SHY) (b).

for different COD concentrations was modeled using the

Monod kinetics (Equation (7)). The Rsu was calculated at each

COD concentration (S ) after 43 h (1.79 days, as shown in

Fig. 5a), at which point the substrate utilization was in the

exponential phase for all COD concentrations. In the next

experimental data point (72 h), the reactor with the COD

concentration of 1 g/L had already entered the stationary

phase (Fig. 5a). To calculate Rsu at different COD concentra-

tions, it was critical to employ the duration period that all the

reactors were in the same phase of substrate degradation

(Exponential Phase, 43 h). Through a non-linear regression

between (Rsu/X ) and S, the values of Ks (8 g/L) and k (0.25 g

glucose/g biomass/day) were determined. The correlation

coefficient R2 was 0.99 (Fig. 9a), suggesting the adequacy of

substrate utilization model in this study.

The Rsu for different pH was modeled using the Michaelis

pH function (Equation (5)) (Fig. 9b). The Rsu was calculated at

each pH value in the exponential phase (43 h). The Rsu

increased with increase in pH from 4.5 to 7.0. A non-linear

regression of Rsu versus [Hþ] yielded the equilibrium

constants K1 and K2 of 1.8� 10�12 and 8.2� 10�6 mole/L

respectively (Equation (5)). Based on Equation (6), an

optimum pH of 8.4 was calculated for Rsu. In this model, the

effect of pH range of 4.5–7.0 on Rsu was simulated. The

experimental results showed that although Rsu increased

with pH within the range of 4.5–7.0, the SHY started to

0.00000 0.00001 0.00002 0.00003 0.000040

2

4

6

[H+] (mole/L)

Rs

u (g

g

lu

co

se/d

ay)

0 5 10 15 20 250.00

0.05

0.10

0.15

0.20

COD (g/L)

Rsu /X

(d

ay

-1)

R2 = 0.99

R2 = 0.97

a

b

Fig. 9 – The kinetic analysis of substrate utilization rate

(Rsu) (a) at different COD concentrations, (b) at different pH.

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 4 ( 2 0 0 9 ) 6 1 7 1 – 6 1 8 06178

decrease at pH> 6 (Fig. 3). It was expected that at the pH

higher than 7.0, the unsuitable conditions for hydrogenic

bacterial activities might decrease the Rsu. Therefore if the

effects of pH> 7 on Rsu were to be incorporated in the kinetic

model, the optimum pH would have been lower than the

obtained optimum pH of 8.4. The correlation coefficient, R2 of

0.97 demonstrated the suitability of Michaelis pH function to

describe Rsu as a function of pH.

It should be noted that the optimum pH (8.4) for Rsu is

significantly higher than optimum pH (5.7–5.8) for RH2 and

SHY. This discrepancy could be caused by the shift of

substrate removal and hydrogen production at different pH.

The activities of hydrogen-producing bacteria decreased at

higher pH (pH> 6.0), but the metabolic activities of various

anaerobic bacteria in the mixed culture were viable up to pH of

7.0–8.4, which result in high substrate consumption and high

biomass growth at higher pH.

3.6. The kinetic analysis of biomass growth

The kinetic analysis for the specific growth rate of biomass (m)

was conducted in the batch reactors. The COD-level tests

showed the SHY decreased at higher COD concentrations

(Fig. 1), which could be caused by substrate degradation by

biomass growth rather than hydrogen production. Through

the non-linear regression of m versus COD concentration (S ),

the values of mm (0.03 day�1) and Ks (16.1 g/L) were obtained

(Fig. 10). The biomass growth rate increased with COD

concentrations, which testified the lower SHY at high COD

(Fig. 1). The high value of correlation coefficient (R2: 0.99)

suggested that Monod kinetics can be adequately used to

evaluate the microbial growth, substrate utilization and

hydrogen production.

The mm obtained was much lower than the value (0.3

day�1) for methanogenic degradation of propylene glycol [45].

The inoculum used in this study was soil, which is rich in

organic matters. The control tests without the addition of

glucose indicated that the organic matters in soil was

degraded and resulted in biogas production (Data not

shown). Since m (mg/L/d/mg/L) was obtained by dividing

biomass growth rate (mg/L/d) with the initial biomass

concentration (mg/L), the higher value of initial biomass with

0 5 10 15 20 250.000

0.005

0.010

0.015

COD (g/L)

µ (d

ay

-1)

R2 = 0.99

Fig. 10 – The kinetic analysis of microbial growth rate (m) at

different COD concentrations.

the existence of the organic matters in soil may lead to the

underestimation of mm. In addition, the Ks value of 16.1 g/L

indicated that half of maximum growth rate could be ach-

ieved at a COD concentration of 16.1 g/L. Therefore, the COD

concentrations higher than 16.1 g/L would result in a higher

growth of biomass and a decline of hydrogen yields, which

was well corresponded with the experimental hydrogen

production results (Fig. 1). Mazijat et al. reported the inhibi-

tion of biomass growth at substrate concentration of 5.4 g/L

(as food, F ) [46]. The biomass concentration in their study

was 1 g/L (as microorganisms, M ), which yielded a ratio of

food to microorganism (F/M ) of 5.4. This high F/M ratio may

result in the inhibition of biomass growth. In this study, the

substrate concentrations ranged from 1.0 to 20 g/L with the

inoculum concentration consistent at 20 g/L, which yielded

the F:M ratios at 0.05–1.0. At these low F:M ratios, the

biomass growth were adequately described by the Monod

kinetics.

4. Conclusions

The study comprehensively investigated the enhancement of

hydrogen production from organic substrates in batch reac-

tors under different operational conditions (COD and pH). A

kinetic analysis of biomass growth, substrate utilization and

hydrogen production was conducted. Two major conclusions

were drawn in this study.

First, the hydrogen production rate and (RH2 ) and specific

hydrogen yields (SHY) were clearly affected by different

operational parameters. The COD-level tests showed that the

SHY values were reversely related with COD concentrations,

and decreased from the highest yield of 2.1 moles H2/mole of

glucose at COD of 1.0 mg/L to 0.7 moles at COD of 20 g/L. The

pH-level tests showed that the pH of 5.5–6 was ideal for

hydrogen production with the SHY of 1.6 moles H2/mole of

glucose.

Second, the kinetic analysis reveals the correlation

between hydrogen production, substrate degradation and

biomass growth. The optimal pH for hydrogen production

and substrate degradation was different with the optimal pH

of 5.89 and 5.74 for rate of RH2 and SHY, and of 8.4 for

substrate degradation. The simulated maximum RH2 was

13.7 mL/h, and the simulated maximum SHY was 2.32 mole

H2/mole glucose. The maximum rates of glucose utilization

(Rsu) and specific microbial growth (m) were 0.25 g glucose/g

biomass/day and 0.03 g biomass/g biomass/day,

respectively.

Acknowledgement

This project was funded by the USGS and Connecticut Water

Resource Center. Part of the student scholarship was sup-

ported by the Multidisciplinary Graduate Research Grant of

Center of Environmental Science and Engineering (CESE). The

assistance of Dr. Kenneth Noll in terms of anaerobic micro-

biology was appreciated.

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 4 ( 2 0 0 9 ) 6 1 7 1 – 6 1 8 0 6179

r e f e r e n c e s

[1] Kloeppel J, Rogerson S. The hydrogen economy. ElectronWorld Wireless World 1991;97:668–71.

[2] Thauer RK, Jungermann K, Decker K. Energy conservation inchemotrophic anaerobic bacteria. Bacteriol Rev 1977;41(1):100–80.

[3] Chin HL, Chen ZS, Chou CP. Fedbatch operation usingClostridium acetobutylicum suspension culture as biocatalyst forenhancing H2 production. Biotechnol Prog 2003;19(2):383–8.

[4] Hussy I, Hawkes FR, Dinsdale R, Hawkes DL. Continuousfermentative hydrogen production from a wheat starch co-product by mixed microflora. Biotechnol Bioeng 2003;84(6):619–26.

[5] Oh S, Van Ginkel S, Logan BE. The relative effectiveness of phcontrol and heat treatment for enhancing biohydrogen gasproduction. Environ Sci Technol 2003;37(22):5186–90.

[6] Logan BE, Oh S, Van Ginkel S, Kim IS. Biological hydrogenproduction measured in batch anaerobic respirometers.Environ Sci Technol 2002;36(11):2530–5.

[7] Wang L, Zhou Q, Li FT. Avoiding propionic acid accumulationin the anaerobic process for biohydrogen production.Biomass Bioenergy 2006;30(2):177–82.

[8] Liu Y, Yu P, Song X, Qu Y. Hydrogen production fromcellulose by co-culture of Clostridium thermocellum JN4 andThermoanaerobacterium thermosaccharolyticum GD17. Int JHydrogen Energy 2008;33(12):2927–33.

[9] Dinopoulou G, Rudd T, Lester JN. Anaerobic acidogenesis ofa complex wastewater. I. The influence of operationalparameters on reactor performance. Biotechnol Bioeng 1988;31(9):958–68.

[10] Hwang MH, Jang NJ, Hyun SH, Kim IS. Anaerobic bio-hydrogen production from ethanol fermentation: the role ofpH. J Biotechnol 2004;111(3):297–309.

[11] Yu H, Zhu Z, Hu W, Zhang H. Hydrogen production from ricewinery wastewater in an upflow anaerobic reactor by usingmixed anaerobic cultures. Int J Hydrogen Energy 2002;27(11–12):1359–65.

[12] Van Ginkel S, Logan BE. Increased biological hydrogenproduction with reduced organic loading. Water Res 2005;39(16):3819–26.

[13] Ren N, Li J, Li B, Wang Y, Liu S. Biohydrogen production frommolasses by anaerobic fermentation with a pilot-scalebioreactor system. Int J Hydrogen Energy 2006;31(15):2147–57.

[14] Fang HHP, Liu H. Effect of pH on hydrogen production fromglucose by a mixed culture. Bioresour Technol 2002;82(1):87–93.

[15] Nath K, Muthukumar M, Kumar A, Das D. Kinetics of twostage fermentation process for the production of hydrogen.Int J Hydrogen Energy 2008;33(4):1195–203.

[16] Ntaikou I, Gavala HN, Kornaros M, Lyberatos G. Hydrogenproduction from sugars and sweet sorghum biomass usingRuminococcus albus. Int J Hydrogen Energy 2008;33(4):1153–63.

[17] Lin PY, Whang LM, Wu YR, Ren WJ, Hsiao CJ, Li SL, et al.Biological hydrogen production of the genus Clostridium:metabolic study and mathematical model simulation. Int JHydrogen Energy 2007;32(12):1728–35.

[18] Lo YC, Chen WM, Hung CH, Chen SD, Chang JS. Dark H2

fermentation from sucrose and xylose using H2-producingindigenous bacteria: feasibility and kinetic studies. WaterRes 2008;42(4–5):827–42.

[19] Kumar N, Das D. Enhancement of hydrogen production byEnterobacter cloacae IIT-BT 08. Process Biochem 2000;35(6):589–93.

[20] Lee KS, Hsu YF, Lo YC, Lin PJ, Lin CY, Chang JS. Exploringoptimal environmental factors for fermentative hydrogen

production from starch using mixed anaerobic microflora.Int J Hydrogen Energy 2008;33(5):1565–72.

[21] Chittibabu G, Nath K, Das D. Feasibility studies on thefermentative hydrogen production by recombinantEscherichia coli BL-21. Process Biochem 2006;41(3):682–8.

[22] Chen WH, Chen SY, Khanal SK, Sung S. Kinetic study ofbiological hydrogen production by anaerobic fermentation.Int J Hydrogen Energy 2006;31(15):2170–8.

[23] O-Thong S, Prasertsan P, Karakashev D, Angelidaki I.Thermophilic fermentative hydrogen production by thenewly isolated Thermoanaerobacterium thermosaccharolyticumPSU-2. Int J Hydrogen Energy 2008;33(4):1204–14.

[24] Soni BK, Jain MK. Influence of pH on butyrate uptake andsolvent fermentation by a mutant strain of Clostridiumacetobutylicum. Bioprocess Eng 1997;17(6):329–34.

[25] Van Ginkel SW, Lay JJ, Sung S. Biohydrogen production asa function of pH and substrate concentration. Environ SciTechnol 2001;35(24):4726–30.

[26] Mu Y, Wang G, Yu HQ, Zheng JC. Response surfacemethodological analysis on biohydrogen production byenriched anaerobic cultures. Enzyme Microb Technol 2005;38(7):905–13.

[27] Jo JH, Lee DS, Park D, Choe WS, Park JM. Optimization of keyprocess variables for enhanced hydrogen production byEnterobacter aerogenes using statistical methods. BioresourTechnol 2008;99(6):2061–6.

[28] Wang J, Wan W. Comparison of different pretreatmentmethods for enriching hydrogen-producing bacteria fromdigested sludge. Int J Hydrogen Energy 2008;33(12):2934–41.

[29] Owen WF, Stuckey DC, Healy Jr JB, Young LY, McCarty PL.Bioassay for monitoring biochemical methane potential andanaerobic toxicity. Water Res 1979;13(6):485–92.

[30] Dubois M, Gilles KA, Hamilton JK, Rebers PA, Smith F.Colorimetric method for determination of sugars and relatedsubstances. Anal Chem 1956;28(3):351–6.

[31] ASTM. Standard test method for pH of soils. WestConshohocken, PA: American Society for Testing andMaterials; 2003.

[32] APHA. Standard methods for the examination of water andwastewater. Washington, DC: American Public HealthAssociation; 1992.

[33] Bailey, Ollis. Biochemical engineering fundamentals. NewYork: McGraw-Hill Inc; 1986.

[34] Metcalf, Eddy. Wastewater engineering, treatment, disposal,and reuse. New York: McGraw-Hill Inc; 1991.

[35] Okpokwasili GC, Nweke CO. Microbial growth and substrateutilization kinetics. Afr J Biotechnol 2005;5(4):305–17.

[36] Lin CY, Cheng CH. Fermentative hydrogen production fromxylose using anaerobic mixed microflora. Int J HydrogenEnergy 2006;31(7):832–40.

[37] Li JZ, Ren NQ. The operational controlling strategy about theoptimal fermentation type of acidogenic phase. Chin EnvironSci 1998;18(5):398–402.

[38] Fennell DE, Gossett JM. Modeling the production of andcompetition for hydrogen in a dechlorinating culture.Environ Sci Technol 1998;32(16):2450–60.

[39] Dabrock B, Bahl H, Gottschalk G. Parameters affectingsolvent production by Clostridium acetobutyricum. ApplEnviron Microbiol 1992;58(4):1233–9.

[40] Lay JJ. Modelling and optimization of anaerobic digestedsludge converting starch to hydrogen. Biotechnol Bioeng2000;68(33):269–78.

[41] Holt RA, Cairns AJ, Morris GJ. Production of butanol byClostridium puniceum in batch and culture. Appl MicrobiolBiotechnol 1988;27(4):319–24.

[42] Ueno Y, Sato S, Morimoto M. Operation of a two-stagefermentation process producing hydrogen and methanefrom organic waste. Environ Sci Technol 2007;41(4):1413–9.

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 4 ( 2 0 0 9 ) 6 1 7 1 – 6 1 8 06180

[43] Wang JL, Wan W. The effect of substrate concentration onbiohydrogen production by using kinetic models. Sci ChinaSer B Chem 2008;51(11):1110–7.

[44] Fabiano B, Perego P. Thermodynamic study and optimizationof hydrogen production by Enterobacter aerogenes. Int JHydrogen Energy 2002;27(2):149–56.

[45] Seok J, Komisar SJ. Sequential kinetic parameter estimationof anaerobic fluidized bed bioreactor using an optimizationtechnique. Biotechnol Lett 2002;24(19):1579–86.

[46] Mazijat A, Mitsunori Y, Mitsunori W, Michimasa N, Jun’ichiro M.Hydrogen gas production from glucose anditsmicrobial kineticsin anaerobic systems. Water Sci Technol 1997;36(6):279–86.