Anda di halaman 1dari 10

International Journal of Hydrogen Energy 32 (2007) 3274 – 3283

www.elsevier.com/locate/ijhydene

Hydrogen production from diluted molasses by anaerobic hydrogen


producing bacteria in an anaerobic baffled reactor (ABR)
Jianzheng Li a,∗ , Baikun Li b , Gefu Zhu a , Nanqi Ren a , Lixin Bo a , Junguo He a
a School of Municipal and Environmental Engineering, Harbin Institute of Technology, Harbin 150090, China
b Department of Civil and Environmental Engineering, University of Connecticut, Storrs, CT 06269, USA

Received 19 December 2006; received in revised form 10 April 2007; accepted 10 April 2007
Available online 8 June 2007

Abstract
Hydrogen production from diluted molasses by anaerobic fermentation bacteria was investigated in a three-compartment anaerobic baffled
reactor (ABR) with an effective volume of 27.48 L. After being inoculated with aerobic activated sludge and operated at chemical oxygen
demand (COD) of 5000 mg/L and temperature of 35 ◦ C for 26 days, the ABR achieved stable ethanol-type fermentation. The liquid fermentation
products, including volatile fatty acids (VFAs) and ethanol, stabilized at 1254, 2053, and 2761 mg/L in the three compartments, respectively.
Effluent pH, oxidation–reduction potential (ORP), and alkalinity ranged at 4.3–4.4, −241 to −249 mV, and 306.334 mg CaCO3 /L, respectively.
The hydrogen yield of the ABR was 32.51 L/d at the stable operation status, specific hydrogen production rate of anaerobic activated sludge
was 0.13 L/g MLVSS d, and the substrate conversion rate was 0.13 L/g COD. Hydrogen yields, fermentation types, and acclimatization durations
varied in each compartment, with the 1st compartment having lowest hydrogen yield but longest acclimatization duration and the 2nd and
3rd compartments having higher hydrogen yields but shorter acclimatization durations. The study found that the individual compartment
configuration in the ABR system provided a favorable environment for different types of anaerobic bacteria. Compared with complete stirring
tank reactor (CSTR), the ABR system had a better operation stability and microbial activity, which led to higher substrate conversion rate and
hydrogen production ability.
䉷 2007 International Association for Hydrogen Energy. Published by Elsevier Ltd. All rights reserved.

Keywords: Hydrogen production; Anaerobic baffled reactor (ABR); pH; Alkalinity; Oxidation–reduction potential (ORP); Ethanol-type fermentation;
Acidogentic bacteria; Acetogenic bacteria

1. Introduction The production of clean energy from waste treatment meets the
requirement of sustainable strategy and has become worldwide
Energy is vital to global prosperity, yet the reserves of the competitive new technology [3–7].
fossil fuel are limited. The excessive use of fossil fuels is Anaerobic hydrogen production is achieved in the acidogenic
one of the primary causes for global warming and acid rain, phase in fermentation processes. Extensive studies have been
which have affected the earth’s climate, weather, vegetation, carried out to enhance hydrogen production by pure bacteria
and aquatic ecosystems [1]. In order to reduce global pollution and mixed cultures with various types of reactors [5,8–17].
and save the existing energy reserves, non-polluting and renew- Most of anaerobic hydrogen bioreactors tested are traditional
able energy need to be developed. Hydrogen is expected to be complete stirring tank reactors (CSTR) [8,10–12,14,15,18].
a substitute for fossil fuel as clean energy in the 21st century, Although stirring/mixing could improve the efficiency of mass
since it only generates water when burning [2]. Hydrogen can transfer and increase hydrogen productivity in CSTRs, it con-
be produced in anaerobic treatment of organic wastes, which sumed electric energy and had potential problem of biogas leak-
employs renewable biomass and high concentrated wastewater. ing. A high volumetric loading rate (VLR) has been found as a
critical factor to improve hydrogen yield, but it caused the accu-
∗ Corresponding author. Tel.: +86 451 86283761. mulation of liquid fermentation products (e.g. acetic acid, pro-
E-mail address: ljz6677@163.com (J. Li). pionic acid, and butyric acid) [18,19], which led to a significant
0360-3199/$ - see front matter 䉷 2007 International Association for Hydrogen Energy. Published by Elsevier Ltd. All rights reserved.
doi:10.1016/j.ijhydene.2007.04.023
J. Li et al. / International Journal of Hydrogen Energy 32 (2007) 3274 – 3283 3275

pH drop in the reactor. The pH could drop to 3.8 and inhibit the Biogas meter
activity of hydrogenic anaerobic bacteria, which was referred Pump Waterlock
to as “over-acidification” [20]. Previous study [11] found that
the pH in an anaerobic hydrogen-producing CSTR system was Biogas outlet
4.3 at a VLR of 6 kg COD/m3 d without pH adjustment in in-
fluent. pH in the CSTR was still as low as 4.3 at a VLR of Baffle Effluent
20.48 kg COD/m3 d, even the influent pH was increased to 9
by adding lime [21]. When the system was over acidified, the
activity of anaerobic hydrogen-producing sludge was severely
Sampling port
impacted. The hydrogen productivity could not recover even Feed tank
though pH was later adjusted to above 4.0 by lowering influ-
ent loading or adding lime, since the microbial community had
changed dramatically during over-acidification period [5]. Solid outlet
On the other hand, anaerobic baffled reactor (ABR), which
has been developed over 20 years for wastewater treatment, Fig. 1. Schematic diagram of the anaerobic baffled reactor system.
possesses several advantages over CSTR [22]. The ABR com-
prises a series of vertical baffles to force wastewater to flow Table 1
under and over them, and wastewater can flow through ac- Composition of the normal molasses (w/w)
tivated sludge beds in each compartments [23,24]. Although Component Percentage (%)
the mass transfer efficiency may be limited without stirring, it
can be compensated by a sufficient contact between activated Dried materials 78–85
sludge and wastewater. The compartmentalized ABR consists Total sugar 48–58
TOC 28–34
of three zones (acidification, methanation, and buffer zones). TKN 0.2–2.8
High activated sludge concentrations (> 20 g biomass/L) in the P2 O 5 0.02–0.07
reactor could enhance the system’s stability [22,25]. It was CaO 0.15–0.8
found [26] that a steady increase in the feed loading from MgO 0.01–0.1
4.8 to 18 kg COD/m3 d did not significantly reduce the sub- K2 O 2.2–4.5
SiO2 0.1–0.5
strate removal efficiency in a three-compartment ABR system. Al2 O3 0.05–0.06
However, no study has been conducted on hydrogen anaerobic Fe2 O3 0.001–0.02
production in an ABR system. Ash content 4–8
There are several parameters affecting the stability and ef-
ficiency of anaerobic hydrogen production systems, such as
VLR, hydraulic retention time (HRT), pH, oxidation–reduction The trapped solids were discharged periodically from the reac-
potential (ORP), and alkalinity [13–15,18,20,27,28]. This tor. Each compartment was equipped with sampling ports that
study investigated the changes of operational parameters in a allowed biological solids, gas, and liquid samples to be with-
three-compartment ABR system with diluted molasses as the drawn. The reactor was wrapped by electrothermal wire and
substrate. The hydrogen production yield (LH2 /d), specific the temperature was maintained at 35 ± 1◦ C. The evolved bio-
hydrogen production rate per biomass (LH2 /g-biomass d), gas was collected separately from the upper part of each com-
substrate conversation efficiency (LH2 /g substrate), hydrogen partment. Biogas volumes for each compartment were daily
percentage in biogas (%), and liquid fermentation products measured using waterlocks and wet gas meters (Model LML-
were compared in each compartment. The hydrogen produc- 1, Changchun Filter Co., Ltd.). Waterlocks and wet gas meters
tion and operation stability of the ABR system was compared had been filled with water of pH 3 in order to prevent the bio-
with CSTR systems and the difference was analyzed at both gas from dissolution.
biochemical and ecological levels.
2.2. Feeding solution and seed sludge
2. Materials and methods

2.1. Experiment apparatus and processes Molasses were collected from a local beet sugar refinery and
its characteristics are given in Table 1. Molasses were diluted
The continuous fermentative bio-producing hydrogen reac- by water to a chemical oxygen demand (COD) of 5000 mg/L,
tor was constructed from 10 mm thick transparent perspex, and the COD:N:P was maintained at 300–500:5:1 by adding
with internal dimensions of 100 cm × 10 cm × 110 cm. The synthetic fertilizer to supply microorganisms with adequate ni-
reactor contained three compartments of same size with the trogen and phosphorus. The pH and alkalinity of the feeding so-
total effective volume of 27.48 L (Fig. 1). Each compartment lution were adjusted to 6.6 and 270 mg CaCO3 /L by NaHCO3
was constituted with an up-flow zone and a down-flow zone. A powder. The ABR system was operated at a HRT of 13.5 h and
sedimentation tank with a volume of 1.5 L was attached to the a VLR of 8.89 kg COD/m3 d.
last compartment for water level control and trapping solids. The reactor was inoculated with excess sludge taken from
The influent was pumped to the reactor by a peristaltic pump. a secondary settling tank in a local brewing wastewater
3276 J. Li et al. / International Journal of Hydrogen Energy 32 (2007) 3274 – 3283

treatment plant. There were three reasons that the sludge kept increasing in the 1st compartment at the lag stage (1–13
was not inoculated from anaerobic digesters. First, anaerobic days) (Fig. 2a), indicating that anaerobic hydrogen produc-
digested sludge contains more methanogens than hydrogen- ing bacteria started adapting to the ABR system. Both biogas
producing bacteria and most of them are embedded in granules. yield and hydrogen percentage began to increase significantly
Since the pH in the ABR system was maintained below 5.0 for from the 14th day and stabilized on the 25th day. During the
hydrogen production, these methanogens would not survive stabilized stage (25–55 days), hydrogen average yield was
at this low pH condition. Second, even if some methanogens 7.17 L/d and hydrogen percentage in biogas was 51% in the 1st
could survive at the ABR system, they would consume hydro- compartment.
gen produced by acidogens and resulted in a lower hydrogen The 2nd and 3rd compartments showed the similar trend
production. Third, most of microbes in anaerobic digested as the 1st compartment, but had shorter lag stages and higher
sludge are obligate anaerobic with low reproduction rates, hydrogen yields (Fig. 2b and c). Biogas and hydrogen yields
which would prolong the inoculation period for hydrogen gradually increased on the 6th day in these two compartments,
production. while there was no obvious increase of biogas and hydrogen
The ratio of mixed liquor volatile suspend solid (MLVSS) yields in the 1st compartment till the 13th day. The possible
to mixed liquor suspend solid (MLSS) was 0.71 in the in- reason for the shorter lag stage in 2nd and 3rd compartments
oculated sludge. After being diluted to a concentration about was the more stable environment achieved by the buffering ef-
6.00 g MLVSS/L, the activated sludge was pumped into the fect of the 1st compartment. During the stabilized stage (26–55
three compartments one by one. After the sludge inoculation, days), the hydrogen yield in the 2nd and 3rd compartments
the ABR system was started up at a HRT of 13.5 h with a VLR were 15.11 and 10.23 L/d, with the percentage of hydrogen in
of 8.89 kg COD/m3 d, and the temperature was raised from 18 biogas of 60% and 51%, respectively.
to 35◦ C within 24 h. The sludge concentrations of each com-
partment on the first day were 6.07, 6.32, and 6.27 g MLVSS/L, 3.2. Liquid fermentation products
respectively.
Along with the biogas and hydrogen production, liquid
2.3. Analytical methods fermentation products also underwent variations during the
acclimatization period. In the 1st compartment, the to-
COD, MLSS, MLVSS, pH, alkalinity, and ORP were mea- tal concentration of liquid fermentation products ranged at
sured according to standard methods [29]. The biogas con- 725–930 mg/L before the 20th day, then rapidly increased, and
stituents (H2 and CO2 ) were analyzed by a gas chromatogram stabilized at 1220–1296 mg/L after the 26th day (Fig. 3a). The
(Model SC-7, Shandong Lunan Instrument Factory) equipped constituents of liquid fermentation products also significantly
with a thermal conductivity detector and a 2 m stainless changed before the 20th day. On the 6th day, the concen-
column packed with Porapak TDS201 (60/80 mesh). The con- trations of ethanol, acetic acid, propionic acid, and butyric
centrations of the volatile fatty acids (VFAs) were measured acid were 191, 137, 145, and 215 mg/L, which indicated a
using another gas chromatograph (Model GC-122, Shanghai mixed-acid fermentation. On the 10th day, ethanol, acetic acid,
Analytical Apparatus Corporation) with a flame ionization de- propionic acid, and butyric acid changed to 159, 311, 177, and
tector and a 2 m stainless column packed with Porapak GDX103 263 mg/L. The total amount of acetic acid and butyric acid was
(60/80 mesh). COD, pH, alkalinity, ORP, VFAs, biogas yield, 574 mg/L, counting for 63% of total liquid products, which
and its constituents were measured or monitored daily. MLSS was a typical butyric acid fermentation [30]. From the 11th
and MLVSS were measured once every 3 days in the first 33 day, the production of butyric acid gradually decreased while
days and once a week in the later experimental period. ethanol and propionic acid increased. On the 26th day, the 1st
compartment reached a stabile stage, with the concentrations
3. Results of ethanol, acetic acid, propionic acid, butyric acid, and valeric
acid of 443, 432, 245, 107, and 27 mg/L. The total amount
3.1. Biogas and hydrogen production of ethanol and acetic acid was 876 mg/L, counting for 70%
of the total liquid products, which indicated an ethanol-type
The three compartments showed variations of biogas and fermentation [20]. According to the study of Li and Ren [20],
hydrogen yields during the acclimatization period (Fig. 2). For an ethanol-type fermentation was obtained when the mass per-
the 1st compartment, the yields of biogas and hydrogen fluctu- centage of ethanol and acetic acid reached above 60% in the
ated greatly in the first 7 days (Fig. 2a). Biogas yield reached total fermentation end-product. Ethanol-type fermentation had
the first peak of 4.56 L/d on the 3rd day, but gradually dropped a higher hydrogen production ability than mixed acid-, butyric
to 1.00 L/d on the 13th day. After 14 days, biogas yield rapidly acid-, and propionic acid-type fermentations [14,15]. These
increased and reached 14.01 L/d on the 25th day. The 1st com- shifts of fermentation types corresponded with hydrogen pro-
partment stabilized afterwards with an average biogas yield duction (Fig. 2), which showed hydrogen yields significantly
of 14.06 L/d. Based on the biogas yield of the 1st compart- increased after 26 days in the 1st compartment. In addition,
ment, there were three stages for acclimatization, with 1–13 the variation of liquid fermentation products during the ac-
days as lag stage, 14–25 days as acceleration stage, and 26–55 climatization and stabilized stage was found to be related with
days as stabilized stage. The percentage of hydrogen in biogas the shifts of microbial community [11,12,18].
J. Li et al. / International Journal of Hydrogen Energy 32 (2007) 3274 – 3283 3277

30 70
Biogas yield Hydrogen yield
Hydrogen concentration 60

Hydrogen concentration (%)


25

Biogas yields (L/d)


50
20
40
15
30
10
20
5 10

0 0
30 70

60

Hydrogen concentration (%)


25
Biogas yields (L/d)

50
20
40
15
30
10
20
5 10

0 0
30 70

60

Hydrogen concentration (%)


25
Biogas yields (L/d)

50
20
40
15
30
10
20
5 10

0 0
0 5 10 15 20 25 30 35 40 45 50 55
Time (d)

Fig. 2. Biogas and hydrogen yields of each compartment in the ABR system (a, 1st compartment; b, 2nd compartment; c, 3rd compartment).

Like the 1st compartment, the 2nd and 3rd compartments production. The changes of these three factors affected not only
also experienced the transition of fermentation types and finally the anaerobic hydrogen production ability, but also the micro-
stabilized at ethanol-type fermentation (Fig. 3b and c). The bial community and fermentation types. It was found that all
2nd compartment showed a mixed-acid-type fermentation in three compartments underwent significant variations of pH, al-
the first 5 days and butyric-acid-type fermentation during 6–16 kalinity and ORP in the first 25 days (Table 2), which was
days, and stabilized at ethanol-type fermentation at 26–55 days concurrent with the fluctuation of hydrogen and liquid fermen-
(Fig. 3b). The 3rd compartment had the same transition as the tation productions (Figs. 2 and 3). The pH and alkalinity of the
2nd compartment and stabilized at ethanol-type fermentation 1st compartment dropped from 4.8 and 310 mg/L on the 1st
(Fig. 3c). day to 3.5 and 0 mg/L (the lowest values) on the 15th day. The
pH and alkalinity of 2nd and 3rd compartments dropped to the
lowest values on the 10th day, with pH of 3.7 and alkalinity
3.3. The variations of pH, alkalinity, and ORP in the ABR of 0 mg/L for the 2nd compartment, and pH of 3.9 and alka-
system linity of 0 mg/L for the 3rd compartment. As for ORP values,
it rose from −510 mV on the 1st day to −190 mV on the 20th
Previous studies [12,20,31,32] have showed that pH, alka- day in the 1st compartment, and then dropped to −230 mV af-
linity, and ORP are important factors for anaerobic hydrogen ter the 25th day. In the mean time, ORP of the 2nd and 3rd
3278 J. Li et al. / International Journal of Hydrogen Energy 32 (2007) 3274 – 3283

1500

1200

VFAs (mg/L) 900 ethanol acetate


butyrate propionate
valerate total
600

300

0
2500

2000
VFAs (mg/L)

1500

1000

500

0
3500

3000

2500
VFAs (mg/L)

2000

1500

1000

500

0
0 5 10 15 20 25 30 35 40 45 50 55
Time (d)

Fig. 3. The variation of VFAs and ethanol concentrations in each compartment of the ABR system (a, 1st compartment; b, 2nd compartment; c, 3rd compartment).

compartments ranged from −490 to −210 and from −440 to ABR system was operated in a continuous mode and pH under-
−250, respectively. After 25 days, the pH, alkalinity and ORP of went consistent variations in each compartment, it was difficult
the three compartments stabilized with pH of 4.3–4.4, alkalin- to derive a clear correlation between ORP and pH. However, it
ity of 306.334 mg CaCO3 /L, and ORP of −240 to −250 mV, could still be seen from Table 2 that ORP was inversely related
which had been demonstrated as the optimal condition for with pH in most cases, with low ORP corresponding to high pH.
ethanol-type fermentation [12,20]. Kang and Yin [33] proposed a kinetic model of ORP values in
Ethanol-type fermentation has been found to occur at pH of acidogenic fermentation systems: Eh = −0.06pH.0.031 gP H2
4.0–4.5 and ORP of −200 to −400 mV [30–32]. ORP values (Eh is the oxidation–reduction potential, and P H2 is the par-
were mainly affected by pH in an anaerobic system. Because the tial pressure of hydrogen in the acidogenic reactor). This model
J. Li et al. / International Journal of Hydrogen Energy 32 (2007) 3274 – 3283 3279

Table 2
The variation of pH, alkalinity, and ORP in the ABR system during 55 days operation

1st day 5th day 10th day 15th day 20th day 25th day 35th day 45th day

Influent pH ALK (mg CaCO3 /L) ORP (mV) 6.8 6.6 6.7 6.7 6.7 6.5 6.7 6.8
270 280 290 270 260 270 270 270
— — — — — — — —

1st compartment pH ALK (mg CaCO3 /L) ORP (mV) 4.8 4 3.8 3.5 3.9 4.4 4.4 4.4
310 120 70 0 60 270 330 320
−510 −230 −220 −260 −190 −230 −240 −240

2nd compartment pH ALK (mg CaCO3 /L) ORP (mV) 4.4 4.1 3.7 3.9 3.9 4.3 4.4 4.4
1400 190 0 120 40 260 330 320
−490 −320 −310 −400 −210 −240 −250 −250

3rd compartment pH ALK (mg CaCO3 /L) ORP (mV) 4.5 4.7 3.9 4.2 4.2 4.3 4.4 4.4
1240 290 0 360 260 380 370 360
−350 −440 −330 −380 −430 −250 −250 −250

could explain the inverse relation between ORP and pH in the in the ABR system gradually decreased along each compart-
ABR system. ment (Table 4). In the initial days of the start-up period, the
The mixed liquor pH in an anaerobic system was determined COD removal efficiency was higher due to the activity of in-
by VFAs concentration and alkalinity. Because alkalinity was oculated aerobic activated sludge and the absorption of sludge
affected by the balance between [CO2 ] and [HCO− 3 ], and the floc. Influent COD was 5389 mg/L on the 2nd day of inoc-
majority of alkalinity was [HCO− 3 ] at pH lower than 5, low pH ulation, the COD of each compartment was 4308, 3807, and
and alkalinity were expected at high VFA concentration due to 3353 mg/L. The average COD removal efficiency was 24.3% in
the consumption of HCO− 3 . In the first 15 days of ABR oper- the first 5 days and then gradually decreased to 11.8% in 6–25
ation, the average total amounts of VFAs were 832, 1521, and days. It stabilized at 9.4% in 26–55 days, during which influ-
1688 mg/L in each compartment, respectively, and did not have ent COD was 5270 mg/L, and the COD concentrations of each
distinct changes (Fig. 3, Table 3). During this period, mixed compartment were 5097, 4946, and 4776 mg/L. After combin-
liquor pH was below 4.7 (Table 3) and determined by alka- ing with hydrogen production yield of 32.51 L/d at 26–55 days
linity, which mainly consisted HCO− −
3 . [HCO3 ] was positively (Fig. 2), it was obtained that substrate conversation efficiency
correlated with CO2 produced by anaerobic fermentation bac- (LH2 /g substrate) was 0.03, 0.06, and 0.04 L/g COD for each
teria. Biogas production was low in the first 15 days (Table 3), compartment.
which meant that [HCO− 3 ] in mixed liquor was low. There-
fore, both alkalinity and pH underwent steady decrease in the 3.5. Biomass
first 15 days (Table 3). After 15 days, both biogas production
(Fig. 2, Table 3) and VFAs (Fig. 3, Table 3) increased, indicat- Based on the measurement of hydrogen yield, liquid fer-
ing anaerobic bacteria had been adapted to the ABR system. mentation products, pH, alkalinity, and ORP values (Figs. 2, 3
With more CO2 being produced, [HCO− 3 ] became higher and and Table 2), the ABR system reached a stable status on the
alkalinity increased correspondingly. A higher alkalinity en- 26th day. However, biomass concentration (MLVSS) exhibited
hanced the system neutralization capability for VFAs and led a different trend (Fig. 4). The biomass concentration dropped
to a stable pH value. Thereby, even though more VFAs were during the starting-up period and became the lowest on the 6th
produced after the 15th day, pH stabilized at 4.4 in all three day with 5.40, 5.47, and 5.36 g MLVSS/L in each compart-
compartments. ment. Afterwards, it gradually increased to 7.31 g MLVSS/L
on the 18th day in the 1st compartment and stabilized in the
3.4. COD removal later period. However, the concentrations of liquid fermenta-
tion products, which indirectly reflected the shift of microbial
In a traditional anaerobic wastewater treatment process, COD community structure, did not reach a stable status until the 26th
was mainly removed through the conversion of intermediate day (Fig. 3a). In contrast, biomass concentration of the 2nd
products (e.g. acetic acid) to methane by the methanogenic and 3rd compartments stabilized around 10.00 g MLVSS/L
microbes [34]. In an ABR system where acidogenic bacteria on the 33rd day (Fig. 4), while the microbial community
were dominant, COD was removed through the cytogenesis and structure had already reached the stable status on the 26th
gas releases (mainly CO2 and H2 ), while a significant amount day according to the liquid fermentation products (Fig. 3b
of COD was converted to liquid intermediate products (e.g. and c). These results indicated that microbial growth (indicated
ethanol, butyrate, and propionate) and stayed in the system. by MLVSS) and the succession of microbial community
Therefore, COD removal efficiency in acidogenic phase was (indicated by liquid fermentation products) occurred at differ-
lower than methanogenic phase. The study showed that COD ent rates.
3280 J. Li et al. / International Journal of Hydrogen Energy 32 (2007) 3274 – 3283

Table 3
The variation of pH, alkalinity, ORP, VFAs, and CO2 in the ABR system during 55 days operationa

5th day 15th day 25th day 35th day 45th day 55th day

pH 4–4.7 3.6–4.3 4.2–4.4 4.4 4.4 4.3–4.4


Alkalinity (mg CaCO3 /L) 120–290 0–360 360–380 320–360 300–330 240–250
ORP (mV) −230 to −440 −260 to −400 −230 to −250 −240 to −250 −240 to −250 −240 to −250
Total VFAs (mg/L) 731–1650 815–1647 1231–2837 1294–2736 1213–2851 1260–2769
Biogas yield (L/d) 0.81–2.00 3.21–10.01 14.01–24.66 14.11–25.49 13.85–24.87 14.37–25.13
CO2 yield (L/d)b 1.09–2.02 2.44–7.61 6.86–10.09 6.91–10.45 6.51–9.70 6.75–10.05
a The
data ranged from the lowest and highest values in three compartments in the ABR system.
b Biogas
yields were highest in the 1st compartment and lowest in the 3rd compartment during the first 5 days operation period. Afterwards, biogas yields
became highest in the 2nd compartment and lowest in the 1st compartment.

Table 4
The variation of COD in each compartment of the ABR system during 55 days operation

Time (d) Influent (mg/L) 1st compartment (mg/L) 2nd compartment (mg/L) 3rd compartment (mg/L) Total COD removal (%)

2–5 5150 ± 166 4600 ± 298 4305 ± 370 3890 ± 365 24.3 ± 9.1
6–25 5249 ± 275 4991 ± 225 4834 ± 245 4625 ± 291 11.8 ± 3.9
26–55 5270 ± 214 5097 ± 242 4946 ± 277 4776 ± 313 9.4 ± 3.7

12 products (e.g. ethanol, propionate, and butyrate) increased due


to the degradation of molasses in the 1st compartment (Fig. 1),
10 which provided a better condition for the growth of acetogenic
biomass (gMLVSS/L)

bacteria. Although acidogenic bacteria could still grow in these


8
two compartments, they were outnumbered by acetogenic bac-
6 teria. However, since the growth rates of acetogenic bacteria
were lower than acidogenic bacteria, it took longer time for
4 1st compartment 2nd compartment them to achieve a stable status. This was the reason that the
biomass kept increasing until the 33rd day in the 2nd and 3rd
2 3rd compartment
compartments, while fermentation types (Fig. 2) and other oper-
ational parameters (Table 2) had become stable on the 26th day.
0
0 5 10 15 20 25 30 35 40 45 50 55 Based on the results at the stable period (33–55 days), the
time (d) biomass concentrations of each compartment were 7.29, 10.42,
and 10.00 g MLVSS/L, and the hydrogen productions were
Fig. 4. The variation of biomass concentration (MLVSS) in each compartment 7.18, 15.10, and 10.21 L/d, respectively. Therefore, the specific
of the ABR system during 55 days operation. hydrogen production rates for each compartment were 0.11,
0.16, and 0.11 L/g MLVSS d, with the average specific rate of
0.13 L/g MLVSS d in the ABR system. This was higher than a
This disparity between microbial growth and the microbial previous study by Ren et al. [15], in which the maximal spe-
community succession could be explained by the different reac- cific hydrogen production rate of a CSTR system was only
tion rates and growth rates of anaerobic bacteria. There are four 0.08 L/g MLVSS d.
major groups of anaerobic bacteria: acidogenic, acetogenic, ho-
moacetogenic bacteria, and methanogenic bacteria. Acidogenic 4. Discussion
bacteria grow fast and degrade complex organic compounds
(e.g. hydrocarbonate, fatty acid, and protein), while acetogenic 4.1. Comparison of ABR and CSTR in terms of hydrogen
and homoacetogenic bacteria grow slow and can only degrade production
simple compounds (e.g. ethanol, propionate, butyrate, CO2 , and
H2 ) [30,35–37]. In the 1st compartment of the ABR system, Previous study [15] showed that a CSTR system reached
sufficient substrates (mainly molasses) in influent provided a the maximal specific hydrogen production rate of 0.08 L/g
favorable condition for the fast growth of acidogenic bacteria. MLVSS d and the maximal substrate conversion efficiency of
They reached a stable growth on the 18th day. Although pH, 0.10 L/g COD when operated at influent COD of 5000 mg/L,
alkalinity, and ORP kept changing until the 26th day, the mi- HRT of 8 h, and temperature of 356 ◦ C. In this study, the ABR
crobial community in the 1st compartment had been dominated system tested under a longer HRT of 13.5 h. Longer HRT was
by acidogenic bacteria since the 18th day. In the 2nd and 3rd supposed to lead to a lower specific hydrogen production rate
compartments, the concentrations of intermediate fermentation due to less substrate flowing into the system per hour. However,
J. Li et al. / International Journal of Hydrogen Energy 32 (2007) 3274 – 3283 3281

the ABR system achieved a higher hydrogen production with in an ABR system [39]. A previous study used single-strand
specific hydrogen production rate of 0.13 L/g MLVSS d and conformation polymorphism (SSCP) technique targeting the
substrate conversion efficiency of 0.13 L/g COD. There were V1.V3 region of 16S rDNA genes to characterize the micro-
two major reasons for the higher hydrogen production in the bial community of different compartments in the ABR system
ABR system. [40]. The results showed that the structure of bacterial commu-
(1) The configuration of the reactor: The variations of influ- nity did not vary with time during the stable period, but signif-
ent quality and quantity affected hydrogen production in anaer- icantly varied along the three compartments. According to the
obic systems [16,18,25,26]. Although the stirring in CSTRs retrieved results of the operational taxonomic units (OTUs) by
diluted the influent and reduced shocks, this dilution could BLAST (the National Center for Biotechnology Information,
not eliminate shocks. The pH, alkalinity, and ORP values of NCBI), acidogenic fermentation bacteria were dominant in the
CSTRs substantially fluctuated and inhibited hydrogen produc- 1st compartment while the population of hydrogen-producing
tion [18,31,32]. It was found that hydrogen yield of a CSTR acetogenic bacteria increased in the 2nd and 3rd compartments.
system varied from 9.75 to 14.99 L/d even at stable ethanol- We thought the substrate selectivity might be the main rea-
type fermentation [15]. son for this shift of microbial community. In the ABR sys-
In the ABR system, the 1st compartment handled influent tem, hydrogen was not only produced from glucose through
variations, which was the main reason for the good stability Reaction (1) in the 1st compartment, but also produced from
of the system. Its hydrogen production ranged at 6.65–7.61 L/d liquid intermediate products through Reactions (2)–(4) in the
during the stable operation period (26–55 days) (Table 5). Be- 2nd and 3rd compartments. This was the reason that ABR had
sides the dilution of influent, high concentration of floated a higher substrate conversation rate (0.13 LH2 /g COD) than
sludge floc in the sludge bed also played an important role in CSTR (0.10 L/g COD) (Table 5).
the 1st compartment. High content of microorganisms in these
sludge flocs interacted with wastewater and formed a stable 4.2. The variation of liquid fermentation products in the
micro-ecological system. When there was a shock in influent ABR system
quality and quantity, the system could still maintain the sta-
bility through internal adjustment [23,25,26]. This buffering The concentrations of liquid fermentation products varied in
function in the 1st compartment protected the 2nd and 3rd com- each compartment (Table 6). The total concentration of VFAs
partments from shocks. Therefore, these two compartments ex- increased along the three compartments with 1255, 2053, and
hibited much stable hydrogen production with 15.72 ∼ 16.00 2761 mg/L, respectively. This might be caused by the accumu-
and 9.60 ∼ 10.86 L/d, respectively (Table 5). lation effect, which means that VFAs was carried over from the
(2) The fermentation microbial community: Complex organic 1st compartment to the 2nd compartment and then to the 3rd
substances (e.g. starch) in wastewater were first hydrolyzed to compartment. The net increase of VFAs was highest in the 1st
single organic (e.g. glucose). Glucose is further broken down in compartment (1254 mg/L), followed by the 2nd compartment
two-step reactions sequentially carried out by acidogenic fer- (798 mg/L) and 3rd compartment (707 mg/L). Most of the liq-
mentation bacteria (Reaction (1)) and hydrogen-producing ace- uid fermentation products increased along compartments, ex-
togenic bacteria (Reactions (2)–(4)) under anaerobic condition cept propionic acid, which was 245, 388, and 324 mg/L in each
[34–37]. Ren et al. [30,38] have found that 1 mol glucose gen- compartment. The reason for the high level of propionic acid
erated 1 mol acetic acid, 1 mol ethanol as well as 2 mol H2 and in the 2nd compartment might be that propionic acid was car-
2 mol CO2 at ethanol-type fermentation (Reaction (1)). ried over from the 1st compartment. But this accumulation ef-
fect did not apply to the 3rd compartment where propionic acid
C6 H12 O6 + H2 O → CH3 COOH + CH3 CH2 OH
concentration dropped. This indicated that the conversion rate
+ 2CO2 + 2H2 , (1) from propionic acid might be greater than the generation rate of
propionic acid in the 3rd compartment. Propionic acid could be
CH3 CH2 COOH + 2H2 O → CH3 COOH + CO2 + 3H2 , (2)
converted to acetic acid and hydrogen by hydrogen-producing
CH3 CH2 CH2 COOH + 2H2 O → 2CH3 COOH + 2H2 , (3) acetogenic bacteria (Reactions (2)–(4)). Although the concen-
trations of ethanol and butyric acid gradually increased from 1st
CH3 CH2 OH + H2 O → CH3 COOH + 2H2 . (4)
to 3rd compartments, their net yields became lower in 2nd and
The stirring in CSTRs provided a favorable condition for the 3rd compartments, indicating that a portion of ethanol and bu-
fast growing bacteria (acidogenic fermentation bacteria), while tyric acid might be converted to acetic acid and hydrogen. The
the slow growing bacteria (hydrogen-producing acetogenic bac- conversion of intermediate products (ethanol, propionic acid,
teria) were at disadvantageous status. Therefore, the hydrogen and butyric acid) to hydrogen by hydrogen-producing aceto-
production in CSTRs was mainly achieved through Reaction genic bacteria led to higher hydrogen productions and lower
(1) by acidogenic fermentation bacteria, and the intermediate net VFSs concentrations in the 2nd and 3rd compartments than
liquid products could hardly produce hydrogen through Reac- that in the 1st compartment (Tables 5 and 6).
tions (2)–(4) due to the low population of hydrogen-producing Both operational parameters and liquid fermentation prod-
acetogenic bacteria. uct results showed that ABR system provided a favorable
The growth of different types of bacteria in individual com- condition for both acidogenic fermentation bacteria and
partments separated microorganisms with different growth rates hydrogen-producing acetogenic bacteria. Organic substances in
3282 J. Li et al. / International Journal of Hydrogen Energy 32 (2007) 3274 – 3283

Table 5
The average values of hydrogen production parameters in the ABR system at stable status (26–55 days)

Influent 1st compartment 2nd compartment 3rd compartment Average Total

Biogas yield (L/d) — 14 25 20 20 59


H2 percentage (%) — 51 60 51 54 —
H2 yield (L/d) — 7.17 15.11 10.23 10.84 32.51
Biomass (g) — 67 95 92 85 254
Special H2 production rate (L/g MLVSS d) — 0.11 0.16 0.11 0.13 —
COD removal (%) — 3.9 3.6 0.9 2.8 8.5
H2 yield from COD (L/g COD) — 0.03 0.06 0.04 0.04 0.13

Table 6
The average concentrations and net yields of VFAs and ethanol in the ABR at stable status (26–55 days)

Ethanol Acetic acid Propionic acid Butyric acid Valeric acid Total

1st compartment Concentration (mg/L) 443 432 245 108 26 1255


Percentage (%) 35 34 20 9 2 —
Net increase (mg/L) 443 432 24 107 26 1255

2nd compartment Concentration (mg/L) 634 767 388 181 84 2053


Percentage (%) 31 37 19 9 4 —
Net increase (mg/L) 190 335 143 73 57 799

3rd compartment Concentration (mg/L) 908 1094 324 269 166 2761
Percentage (%) 33 40 12 10 6 —
Net increase (mg/L) 274 326 −64 88 83 708

wastewater were converted to hydrogen through reactions acclimatization. The total liquid products content stabi-
carried out by different types of bacteria in each compart- lized at 1200, 2000, and 2800 mg/L in the three com-
ment, thus achieved a higher substrate conversion rate of partments, respectively. Effluent pH, ORP, and alkalinity
0.13 LH2 /g COD (Table 5). ranged from 4.2 to 4.4, −230 to −250 mV, and 240 to
It should be noted that although ABR systems had a 350 mg CaCO3 /L.
higher hydrogen production rate than CSTRs, it was only (2) Compared with CSTR systems, the individual com-
0.13 LH2 /g COD. In this study, hydrogen anaerobic produc- partment configuration in the ABR achieved a separa-
tion was achieved by the acidogenesis of reductant sugar by tion of different anaerobic bacteria. Organic substrates
acidogens. The concentration of reductant sugar in the mo- in influent were efficiently converted to hydrogen in
lasses was only 28%, and there are many other components each compartment. ABR system achieved a hydrogen
(e.g. cellulose, xylogen, and pectin) in molasses, which could yield of 32.51 L/d and a substrate conversion rate of
not be effectively degraded by acidogens within a HRT of 0.13 L/g COD.
13.5 h. Therefore, even though the influent COD was as high (3) Acidogenic fermentation bacteria dominated in the 1st
as 5000 mg/L, the specific hydrogen conversation rate by COD compartment, due to the presence of sufficient organic
was still low due to the presence of other non-easily-degraded compounds, while hydrogen-producing acetogenic bacte-
components. There might be two approaches to increase the ria became dominant in the 2nd and 3rd compartments
hydrogen production rate: (1) to find out a process to biode- where simple organic compounds had been produced
grade cellulose and xylogen effectively to monosaccharide; from the degradation of molasses. The ABR system ex-
(2) to enrich hydrogen-producing acetogens in fermentation hibited a better stability and higher microbial activity than
systems to enhance the conversation rate of volatile fatty acids CSTRs.
(except acetic acid) produced by acidogens to acetic acid and
hydrogen.
Acknowledgements
5. Conclusion
The authors would like to thank the National Natural Sci-
Based on the study of hydrogen production in an ABR system ence Foundation of China (Contract No. 50378025), the Na-
with three compartments, several conclusions were drawn: tional High Technology Research and Development Program of
China (863 Program, Grant No. 2006AA05Z109), and Provin-
(1) With diluted molasses as the substrate, the ABR system cial Science Foundation of Heilongjiang (Grant No. E0305) for
reached stable ethanol-type fermentation after 26 days of their supports for this study.
J. Li et al. / International Journal of Hydrogen Energy 32 (2007) 3274 – 3283 3283

References [21] Li JZ, Ren NQ, Lin M, Wang Y. Hydrogen bio-production by anaerobic
fermentation of organic wastewater in pilot-scale. Acta Energiae Solaris
[1] Dincer I, Rosen MA. A worldwide perspective on energy, environment Sin 2002;23(2):252–6.
and sustainable development. Int J Energy Res 1998;22:1350–61. [22] Barber WP, Stuckey DC. The use of the anaerobic baffled reactor
[2] Tanisho S, Shiwata Y. Continuous hydrogen production from molasses (ABR) for wastewater treatment: a review. Water Res 1999;33(7):
by the bacterium Enterobacter aerogeness. Int J Hydrogen Energy 1559–78.
1994;9:807–12. [23] Bachmann A, Beard VL, McCarty PL. Performance characteristics of
[3] Tanisho S, Kuromoto M, Kadokura N. Effect of CO2 removal the anaerobic baffled reactor. Water Res 1985;19(1):99–106.
on hydrogen production by fermentation. Int J Hydrogen Energy [24] Barber WP, Stuckey DC. Start-up strategies for anaerobic baffled reactors
1998;23(7):559–63. treating a synthetic sucrose feed, In: Proceedings of the 8th international
[4] Lay JJ, Lee YJ, Noike T. Feasibility of biological hydrogen conference on anaerobic digestion, 1997. p. 32–9.
production from organic fraction of municipal solid waste. Water Res [25] Grobicki AMW, Stuckey DC. Performance of the anaerobic baffled
1999;33(11):2579–86. reactor under steady state and shock loading conditions. Biotechnol
[5] Ueno Y, Otsuka S, Morimoto M. Hydrogen production from industrial Bioeng 1991;37:344–55.
wastewater by anaerobic microflora in chemostat culture. J Ferment [26] Nachaiyasit S, Stuckey DC. The effect of shock loads on the
Bioeng 1996;82(2):194–7. performance of an anaerobic baffled reactor (ABR). 1. Step changes in
[6] Lin CY, Chang RC. Hydrogen production during the anaerobic feed concentration at constant retention time. Water Res 1997;31(11):
acidogenic conversion of glucose. J Chem Technol Biotechnol 2737–46.
1999;74:498–500. [27] Hwang MH, Jang NJ, Hyun SH, Kim IS. Anaerobic bio-hydrogen
[7] Ren NQ, Wang XJ, Wiang WS, Lin M, Li JZ, Guo WQ. Hydrogen production from ethanol fermentation: the role of pH. J Biotechnol
production with high yield by immobilized cells of hydrogen-producing 2004;111:297–309.
bacteria strain B49 in a column reactor. High Technol Lett 2002;8(4): [28] Stephanson CS, Flanagan GP. Non-toxic hydride energy source for
21–5. biochemical and industrial venues: ORP and NAD+ reduction analyses.
[8] Ren NQ, Wang BZ, Ma F. Hydrogen bio-production of carbohydrate Int J Hydrogen Energy 2004;29:459–64.
fermentation by anaerobic activated sludge process. In: Proceedings of [29] APHA. Standard methods for the examination of water and wastewater.
the 68th water environment federation annual conference exposure, 1995. 19th ed., Washington, DC: American Public Health Association;
p. 145–53. 1995.
[9] Ueno Y, Kawai T, Sato S, Otsuka S, Morimoto M. Biological production [30] Ren NQ, Wang BZ, Huang JC. Ethanol-type fermentation from
of hydrogen from cellulose by natural anaerobic microflora. J Ferment carbohydrate in high rate acidogenic reactor. Biotechnol Bioeng
Bioeng 1995;79(4):395–7. 1997;54:197–200.
[10] Li JZ, Zhu GF, Ouyang H, Guo XY, Jin HZ. Hydrogen production by [31] Ren NQ, Zhao D, Chen XL, Li JZ. Mechanism and controlling
organic wastewater fermentation in anaerobic baffled reactor. J Chem strategy of the production and accumulation of propionic acid for
Ind Eng (China) 2004;55(Suppl.):70–4. anaerobic wastewater treatment. Sci China (Series B) 2002;45(3):
[11] Li JZ, Li N, Zhang N, Ouyang H. Hydrogen production from organic 319–27.
wastewater by fermentative acidogenic activated sludge under condition [32] Ren NQ, Chen XL, Zhao D. Control of fermentation types in continuous
of continuous flow. J Chem Ind Eng (China) 2004;55(Suppl.):75–9. flow acidogenic reactors: effects of pH and redoxpotential. J Harbin Ins
[12] Li JZ, Ren NQ, Qin Z, Li YF, Bao HX, Jiang P. Startup of hydrogen Technol (New Series) 2001;8(2):116–9.
producing fermentation process with anaerobic activated sludge and [33] Kang YH, Yin YX. Study on kinetic model of oxidation–reduction
acclimatization of the dominant ethanol-type fermentation population. potential in acidogenic fermentation system. J Harbin Univ Commer
High Technol Lett (China) 2004;14(9):90–4. (Natural Sciences Edition) 2003;19(4):408–10.
[13] Wang XJ, Ren NQ, Xiang WS, Lin M, Guo WQ. Partial characteristics [34] Hutnan M, Mrafkova L, Drtil M, Derco J. Methanogenic and
of hydrogen production by fermentative hydrogen-producing bacterial nonmethanogenic activity of granulated sludge in anaerobic baffled
strain B49. High Technol Lett 2003;9(3):65–70. reactor. Chem Papers 1999;53(6):374–8.
[14] Wang Y, Ren NQ, Sun YJ, Li JZ, Sun XW. Analysis on the mechanism [35] van Lier JB, Grolle KCF, Frijters CTMJ, Stams AJM, Lettinga G.
and capacity of two types of hydrogen production—ethanol fermentation Effects of acetate, propionate, and butyrate on the thermophilic anaerobic
and butyric acid fermentation. Acta Energiae Solaris Sin (China) degradation of propionate by methanogenic sludge and defined cultures.
2002;23(3):366–73. Appl Environ Microbiol 1993;59(4):1003–11.
[15] Ren NQ, Qin Z, Li JZ. Comparison and analysis of hydrogen production [36] Fox P, Pohland FG. Anaerobic treatment applications and fundamentals:
capacity with different acidogenic fermentative microflora. Chin J substrate specificity during phase separation. Water Environ Res
Environ Sci (China) 2003;24(1):70–4. 1994;66(5):716–24.
[16] Van Ginkel S, Logan BE. Increased biological hydrogen production with [37] Wang QH, Kuninobu M, Ogawa HI, Kato Y. Degradation of volatile
reduced organic loading. Water Res 2005;39:3819–26. fatty acids in highly efficient anaerobic digestion. Biomass Bioenergy
[17] Van Ginkel S, Oh SE, Logan BE. Biohydrogen gas production from 1999;16:407–16.
food processing and domestic wastewater. Int J Hydrogen Energy [38] Ren NQ, Wang BZ, Ma F. A physiological ecology analysis of acidogenic
2005;30:1535–42. fermentation of organic wastewater. China Biogas 1995;13(1):1–6.
[18] Qin Z, Ren NQ, Li JZ, Yan XF. Superacid state of acidogenic phase [39] Uyanik S, Sallis PJ, Anderson GK. The effect of polymer addition
and controlling strategy for recovery. J Harbin Ins Technol 2003;35(9): on granulation in an anaerobic baffled reactor (ABR). Part 2:
1105–8. compartmentalization of bacterial populations. Water Res 2002;36:
[19] Zoetemeyer RJ, Matthijsen AJCM, Cohen A, Boelhouwer C. Product 944–55.
inhibition in the acid forming stage of the anaerobic digestion process. [40] Y. Zhao, J.Z. Li, Y. Liu, Y.G. Zhao, Dynamics of bacterial community
Water Res 1982;16:633–9. during the stable operation of anaerobic baffled reactor monitored
[20] Li JZ, Ren NQ. The operational controlling strategy about the optimal by single-strand conformation polymorphism. In: Proceedings of
fermentation type of acidogenic phase. Chin Environ Sci 1998;18(5): international workshop on environmental health and pollution control,
398–402. 2006. p. 368–78.

Anda mungkin juga menyukai