- Email: [email protected]
X0 and shock loading in CSTR
Chemosphere 57 (2004) 1059–1068 www.elsevier.com/locate/chemosphere
H2 production through anaerobic mixed culture: effect of batch S0/X0 and shock loading in CSTR Kuo-Shuh Fan *, Ya-Yun Chen Department of Safety and Environmental Engineering, National Kaohsiung First University of Science and Technology, No. 1 University Street, Yanchau 824, Kaohsiung, Taiwan Received 17 February 2004; received in revised form 12 August 2004; accepted 12 August 2004
Abstract Biological production of H2 has received considerable attention lately. The present study was undertaken to observe the effects of substrate/seeding ratios (S0/X0) on batch H2 generation. The H2-producing seeding spores were obtained from the heat treatment (88 °C for 12 h) of the compost from a grass composting facility. A dehydrated brewery mixture was used as feed substrate. The results indicate that the pattern of the cumulative H2 production with time is similar to the growth curve with a typical lag, exponential and stationary phase; the results were successfully modeled with a modified Gompertz equation. It appears that maximum H2 yield potential (27 ml g1 CODadded) occurs at an S0/X0 ratio of about 4, whereas the maximum specific H2 yield (205 ml g1 VSS d1) occurs at approximately S0/X0 = 3. The S0/X0 ratios higher than 4 would inhibit H2 production. An attempt was made to waste a certain amount of reactor content and replaced it with fresh substrate in order to enhance H2 production. After this medium replacement, the H2 production was initially inhibited and the system then exhibited a long lag before it reached an active H2 production stage. For a continuous-stirred tank-reactor (CSTR) system, the results of replacing 25% of the reactor content indicate that there is still a lag time before a sudden increase in H2 production after the addition of the new substrate feed. The major low molecular weight acids identified are HAc and HBu with total volatile acids of about 6000–8000 mg l1. The ratio of HAc/HBu in the present study is relatively constant (about 5) and appears not significantly affected by the medium replacement. The concentration of total alcohols is about 2000 mg l1. All in all, the CSTR system is able to recover to its previous performance after such a dramatic 25% medium replacement. Ó 2004 Elsevier Ltd. All rights reserved. Keywords: H2 production; H2-producing bacteria; S0/X0; Medium replacement; CSTR
1. Introduction
* Corresponding author. Tel.: +886 7 601 1000x2321; fax: +886 7 601 1061. E-mail address: [email protected] (K.-S. Fan).
Hydrogen production has received considerably attention recently, as energy demand grows and because of a simple fact that H2 is a relatively clean fuel with higher energy efficiency. Particularly, interest is emphasized in biological H2 generation as exemplified by many review papers in this area (e.g., Nandi and Sengupta,
0045-6535/$ - see front matter Ó 2004 Elsevier Ltd. All rights reserved. doi:10.1016/j.chemosphere.2004.08.021
1060
K.-S. Fan, Y.-Y. Chen / Chemosphere 57 (2004) 1059–1068
1998; Noike and Mizuno, 2000; Das and Veziroglu, 2001). In addition to photosynthetic organisms, fermentative bacteria are able to generate high concentrations of H2, along with some low molecular weight intermediates. The microbes responsible for H2 generation include: Enterobacter (e.g., Rachman et al., 1998; Kumar and Das, 2000; Kumar et al., 2000), Clostridum (e.g., Taguchi et al., 1992; Kataoka et al., 1997; Noike et al., 2002), Escherichia (Kanayama and Karube, 1987; Nandi et al., 2001), Actinomyces (Oh et al., 2003a), Bacillus (Kalia et al., 1994), and thermophilic Tnermotoga neapolitana (van Ooteghem et al., 2001). A typical pathway of fermentative H2 production is illustrated in Fig. 1. For example, Clostridum butyricum is able to ferment sugar to H2 with butyrate and acetate as byproducts (Madigan et al., 2000). Using glucose, a maximum of 2 mol of H2 could be produced with butyrate as the main product and 4 mol of H2 with acetate as the major byproduct (Nandi and Sengupta, 1998). The free energy for the latter reaction (206 kJ mol1) is sufficient for microbial growth (Claassen et al., 1999). As high as 60% H2 concentration (Mizuno et al., 2000b; Fang and Liu, 2002), yield of 2.5 mol H2 per mol hexose (Noike and Mizuno, 2000) and production rate of about 30 l H2 l1 reactor d1 (Rachman et al., 1998; Chang et al., 2002) could be obtained from a variety of substrate sources. In general, the H2 production is regulated by many factors, including (1) substrate characteristics, e.g., type and concentration of substrate (Ueki et al., 1991; Noike and Mizuno, 2000; Okamoto et al., 2000; Lay, 2001); (2) type of seeding sources (Lay et al., 1999); (3) environmental conditions of pH (van Ginkel et al., 2001; Fang and Liu, 2002; Oh et al., 2003b) and temperature (Oh et al., 2002, 2003b); (4) operating conditions, e.g., hydraulic retention time (HRT) in continuous-flow systems (Fan and Kuo, 2004) and mixing intensity;
Fig. 1. Typical pathway of fermentative hydrogen production (Tanisho et al., 1998).
(5) reactor configuration (e.g., immobilized system, Chang et al., 2002); (6) metabolites produced; (7) partial pressure of CO2/H2 (Tanisho et al., 1998; Mizuno et al., 2000a); (8) presence of H2 consumers in mixed culture (e.g., sulfate reducing bacteria and methanogens) and other bacteria (e.g., lactic acid bacteria, Noike et al., 2002); and (9) presence of other inorganic compounds, e.g., sulfate (Mizuno et al., 1998), thiosulphate (Nandi et al., 2001) and phosphate (Oh et al., 2002, 2003b) as well as other inhibitors. An easy way to determine the suitability of any substrate for its conversion to H2 by mixed culture is to perform simple batch tests. For example, the H2 production rate can be easily calculated through the gas production rate and the corresponding H2 concentration under anaerobic conditions. In fact, anaerobic batch evaluation of the extent and rate of gas production (e.g., CH4, H2 or N2) has been extensively studied; a few have been performed with respect to the effect of substrate to biomass ratio (S0/X0). The low or high S0/X0 ratios favoring a higher H2 production depend on the type of substrate and seeding bacteria (Lay et al., 1999). In general, higher S0/X0 ratios increase the rate of gas production; further increases in S0/X0 may eventually reduce the gas production rate, due partially to substrate and/or byproduct inhibition. Fed-batch operation, by adding substrate in a batch operation, has been used in industry to maximize product formation (Bailey and Ollis, 1986). To our best knowledge, there is no study performed about the effect of replacing medium volume of a well mixed reactor content with the fresh substrate on H2 production during batch growth stage. This approach is similar to fed-batch operation in concept. Further, during the batch exponential growth phase, addition of substrate may facilitate microbial reactions resulting in increased gas production. This ‘‘shocking’’ approach is analogous to adding substrate to highly concentrated active biomass with the expectation of continuing exponential growth to maximize H2 production. The continuous-flow stirred-tank (CSTR) system has been used widely in microbial studies in general and in observing H2 production in particular. Under constant pH and temperature condition, the system provides steady-state results in substrate, cell mass and byproduct concentrations. One of the key parameters in operation is the effect of shock loading on the CSTR performance. Traditionally, shock loading is conducted by either increasing flow rate or feed concentration or both. However, there is no study on the effect of the loss of reactor content on the subsequent CSTR operation. This is analogous to the loss of activated sludge mixed liquor due to sludge bulking in wastewater treatment plants. It is hypothesized that a direct addition of new substrate to replace the lost content would be a novel strategy to recover the CSTR performance.
K.-S. Fan, Y.-Y. Chen / Chemosphere 57 (2004) 1059–1068
Consequently, this study was undertaken to: (1) observe effects of S0/X0 on H2 production; (2) evaluate the effect of replacing the ‘‘lost’’ medium with the fresh substrate on the overall H2 production in batch studies; and (3) observe the transient state of H2 and byproducts in a CSTR system after the above mentioned medium replacement.
2. Materials and methods 2.1. Materials Serum bottles of 100 and 250 ml were used for the batch tests. The seed was obtained from the compost of a local grass composting facility. The compost was first dried at 85 °C for 12 h to select heat-resistant spore forming H2-producing bacteria which are certainly present under previously 50–60 °C composting conditions. Previous investigators have reported such a feasibility of heat treatment to enrich H2-producing biomass (Lay et al., 1999; Lay, 2001; van Ginkel et al., 2001). A given amount of dried compost was then added into 1 l of distilled water, mixed for 2 h, and the supernatant (X0) was used as the seed for the subsequent batch tests.
1061
The dehydrated fermented brewery mixture after dilution at the desirable strength (S0, 40–160 g l1) with supplemental nutrient addition was used as the feed solution. The dehydrated brewery mixture contains a high fraction of volatile solids, VS (0.79 g VS per g chemical oxygen demand, COD), about 60% volatile suspended solids to VS fraction and the soluble fraction of total COD is about 40%. A given amount of the inorganic nutrient solution was added to ensure the adequacy of other nutrients. The composition of the nutrient solution in 1 l of the medium contained: 80 g NH4HCO3, 40 g KH2PO4, 0.4 g MgSO2 Æ 7 H2O, 0.4 g NaCl, 0.4 g Na2MoO4 Æ 2H2O, 0.4 g CaCl2 Æ 2H2O, 0.6 g MnSO4 Æ 7H2O and 0.11 g FeCl2 (Lay et al., 1999). 2.2. Experiments 2.2.1. Batch There were two phases in the present studies. In Phase I, five different S0 quantities (from 0.4 to 4.8 g COD) were used for each of the five X0 concentrations (0.75–1.05 g VSS) to render a total of 25 different S0/X0 ratios (from 0.4 to 6.4 g COD g1 VSS) to observe H2 production (Table 1). The total volume of seed and substrate was 80 ml in 100 ml serum bottles. After the
Table 1 Lag time, H2 production and H2 maximum rate at different S0/X0 ratios Group
Seed vol (ml)a
Substrate, S0 (g COD l1)
Substrate (ml)
S0/X0 (g COD g1 VSS)
k (d)
P (ml)
Rm (ml d1)
PY (ml g1 COD)
Rs (ml g1 VSS d1)
A1 A2 A3 A4 A5 B1 B2 B3 B4 B5 C1 C2 C3 C4 C5 D1 D2 D3 D4 D5 E1 E2 E3 E4 E5
70 65 60 55 50 70 65 60 55 50 70 65 60 55 50 70 65 60 55 50 70 65 60 55 50
40 40 40 40 40 70 70 70 70 70 100 100 100 100 100 130 130 130 130 130 160 160 160 160 160
10 15 20 25 30 10 15 20 25 30 10 15 20 25 30 10 15 20 25 30 10 14 20 25 30
0.38 0.62 0.89 1.21 1.60 0.67 1.08 1.56 2.12 2.80 0.95 1.54 2.22 3.03 4.00 1.24 2.00 2.89 3.94 5.20 1.52 2.46 3.56 4.85 6.40
0.2 0.3 0.3 0.3 0.4 0.3 0.3 0.4 0.4 0.4 0.3 0.4 0.4 0.4 0.4 0.3 0.4 0.4 0.4 0.4 0.3 0.4 0.4 0.4 0.4
2.3 6.1 6.2 6.6 20.5 6.1 13.4 24.1 33.8 48.0 14.8 31.9 44.9 39.7 53.3 24.4 39.1 46.1 80.8 67.9 23.5 50.1 56.6 89.5 95.2
4.6 17.4 38.3 17.1 86.5 17.2 53.6 105 120 153 61.3 122 158 68.4 81 92 151 109 102 67 88 121 86 88 70
5.7 10.2 7.8 6.6 17.0 8.7 12.8 17.2 19.3 22.8 14.8 21.3 22.5 15.9 17.8 18.8 20.0 17.7 24.9 17.4 14.7 20.9 17.7 22.4 19.8
4 19 43 21 115 16 55 117 145 204 58 125 175 83 108 88 154 121 124 89 84 124 95 106 93
a
At seed concentration = 15 g VSS l1.
1062
K.-S. Fan, Y.-Y. Chen / Chemosphere 57 (2004) 1059–1068
Table 2 H2 production after medium replacementa Group
Replacement fraction (%)
Added substrate (g COD)
X0 remaining (g VSS)b
New S0/X0c (g COD g1 VSS)
Cumulative H2 (ml)
Increased cumulative H2 (%)
1 2 3 4 5 6
0 6 25 44 63 81
0 1 4 7 10 13
0.9 0.85 0.68 0.50 0.33 0.17
0 1.9 5.9 14 30 76
20 21 37 42 34 27
0 6 82 105 70 33
a b c
At S0/X0 = 3 g COD g1 VSS for seed 0.9 g (133 ml at 6.75 g l1) and substrate 2.7 g (27 ml at 100 g l1). Based on the original X0 = 0.9 g. Based on the original X0 and newly added S0.
mixture of substrate and seed, the bottles were purged with a mixture of N2/CO2 gas (7:3, v/v) to ensure anaerobic conditions; thereafter, the bottles were placed in a rotary shaker (30 rpm) at 37 °C. In addition to 25 bottles, several additional bottles with the random S0/X0 ratios were used to serve as duplicates for reproducibility checking. The optimum S0/X0 ratio and the corresponding time for the maximum production of H2 (tmax) during batch runs were identified. In Phase II, cultures were grown at an optimum S0/ X0 ratio, identified from the above tests. The total volume of the seed and substrate was 160 ml in 200 ml bottles. At tmax, different fractions of the bottle content volume were withdrawn, and replaced with the fresh substrate to see if any enhanced H2 production might occur. There were five replacement fractions, ranging from 6% to 81% of the bottle content (Table 2). In all batch studies, the gas volume produced was measured using 10 ml syringes, and the H2 composition in the biogas quantified. 2.2.2. Continuous study The system was initially operated in batch mode (with S0/X0 = 3) and the continuous flow was then initiated after observing H2 production. The system was operated at HRT = 8 h (dilution rate = 3 d1) with the same brewery substrate at concentration 115 g l1 and T = 37 ± 1 °C. The pH was automatically controlled at 5.5–6.0 with 2 N NaOH and ORP (oxidation–reduction potential) continuously monitored. After the system reached pseudo-steady-state, 25% of the reactor volume was wasted and fresh medium was then added. Two methods for substrate addition were used: (1) instantaneous addition; and (2) intermittent addition (about 80 ml every 10 min for 2 h). The latter approach is to minimize the shock loading to the remaining H2-producing bacteria. The transient state of several parameters was observed and recovery time was noted (Fig. 2). Several shock loading runs were conducted after 6–9 HRTs to observe the stability of the system. In all cases, the gas output rate was measured via wet-gas meters, and gas
Fig. 2. Biomass replacement in CSTR experiments. t1 and t2 refer to the increased H2 production period and recovery period, respectively. A and B refer to H2 production before and after the shock treatment and C is the final H2 production.
samples were taken from the sampling port located before the meter for H2 quantification. The liquid samples were also taken for the analysis of volatile fatty acids, alcohols and other parameters. 2.3. Analyses The H2 concentration was quantified in a GC (Shimadzu Model 8A) with a thermal conductivity detector. A 2-m glass column (Porapak Q, 50/80 mesh) was installed and the operating temperatures of the injection port, the oven, and the detector were 100, 60, and 100 °C, respectively. Nitrogen gas served as the carrier gas at a flow rate of 30 ml min1. The measurements of COD and VSS essentially followed those prescribed in the Standard Methods (APHA, 1995). Volatile acids (acetic acid, propionic acid and butyric acid) and alcohols (methanol,
K.-S. Fan, Y.-Y. Chen / Chemosphere 57 (2004) 1059–1068
ethanol, propanol and butanol) were measured in another GC-flame ionization detector (Shimadzu, Model 14B). The correlation coefficients for the calibration curves of all these standard volatile acids and alcohols were greater than 0.998. 2.4. Model analysis The modeling of H2 production during the batch studies is based on the following modified Gompertz equation (Zwietering et al., 1990): Rm e H 2 ¼ P exp exp ðk tÞ þ 1 P
ð1Þ
where H2 is cumulated H2 production (ml); P, H2 equilibrium production (ml); t, reaction time (d); k, lag time (d); and Rm, the maximum H2 production rate (ml d1) as shown in Fig. 3a. The H2 yield potential (PY, ml g1 CODadded) is determined by dividing P by the amount of feed substrate, and the specific H2 production rate (Rs, ml g1 VSS d1) is calculated by dividing Rm by the amount of the initial X0. The reason for using the initial substrate COD is that both substrate and seed contain COD; thus, H2 yield using the amount of COD removed may provide misleading results. A nonlinear regression method was used with the least sum of the square errors used to determine the model parameters.
Fig. 3. Cumulative H2 production in batch experiments: experimental data in solid circles and lines are model-predicted values. (a) Symbol designation for the model in Eq. (1); (b) typical examples for group D data in Table 1.
1063
3. Results and discussion 3.1. Batch H2 production Typical plots of H2 production for five S0/X0 ratios as a function of time are shown in Fig. 3b. The pattern of the cumulated H2 with time is similar to the growth curve with a typical lag, exponential and stationary phase. The resultant results of P, k and Rm for all 25 runs are tabulated in Table 1. For five random duplicate bottles with P up to 100 ml H2, the standard errors were less than 10%; thus the results are reproducible. The model fits experimental data well with the R2 values in Table 1 data all greater than 0.92. Consequently, the model equation is useful and relatively accurate to predict the extent (P) as well as the rate of H2 production (Rm) as was also reported by others (Lay et al., 1999; Lay, 2001; van Ginkel et al., 2001). The lag time ranges from 0.2 to 0.4 d with little variation; lower k values are, however, mostly associated with lower S0/X0 ratios. In a few limited studies, k values ranged from a few hours to 5 days, depending on the type of substrate and the seeding sources (Lay et al., 1999; van Ginkel et al., 2001). Clearly, these parameters and other environmental conditions (pH and temperature) determine the extent and rate of spore germination and subsequent H2 production. The maximum H2 yield potential is about 25 ml g1 CODadded (D4 at S0/X0 = 3.9, Table 1) with a maximum production rate about 158 ml d1 (C3, Table 1), and maximum Rs about 205 ml g1 VSS d1 (B5, Table 1). It appears that there is no clear pattern of the effect of S0/X0 on H2 production. Nonetheless, plots in Fig. 4 indicate that a maximum PY occurs at S0/X0 ratio about 4, whereas the maximum Rs occurs at approximately S0/X0 = 3.5. The fact that that PY is based on COD added and not actual substrate removed and that Rs determined from VSS in which both
Fig. 4. The H2 yield potential (PY, ml g1 CODadded) and the specific H2 production rate (Rs, ml g1 VSS d1) as a function of S0/X0.
1064
K.-S. Fan, Y.-Y. Chen / Chemosphere 57 (2004) 1059–1068
substrate and seed contribute may explain the phenomenon that PY and Rs did not occur at the same S0/X0 ratio. Further discussion of aerobic batch culture with respect to S0/X0 ratios may shed some light as to their effect on H2 production. At lower S0/X0, there is essentially no cell multiplication and at higher ratios, there are many polymer types of byproducts generated as reported by Chudoba et al. (1992) in their activated sludge studies. The S0/X0 ratios also determine the extent of catabolism and anabolism (Chudoba et al., 1992), or energy uncoupling (Liu et al., 1998). In monitoring N2 production for denitrification, the microbial activity increases with increased S0/X0 ratios (Buys et al., 2000). Depending on the culture conditions, a significant quantity of low molecular weight acids and alcohols were generated (not measured in batch studies, but monitored in the subsequent CSTR studies). For example, at low S0/X0 ratios with S0 (glucose) being limiting, Clostridium fallax strain produced acetate and butyrate; with excess S0, lactate was produced after HAc and HBu reached maximum concentrations (Ueki et al., 1991). Hence, these products may in fact inhibit H2 production, directly through product inhibition or indirectly via lower pH. In fact, Lay (2001) in his batch studies reported that the pH decreased from 7 to about 4.2–5.2; the higher the S0/X0 ratio was, the more pH drop. Nevertheless, the type of intermediates depends on substrate and seeding source, e.g., HBu and EtOH (Noike and Mizuno, 2000), HBu and PrOH (Lay et al., 1999), HAc and EtOH (Oh et al., 2003a) as well as HAc, HBu and BuOH (Lay, 2001). The fact that lower pH values due to byproduct accumulation inhibit CH4 production in general and H2 production in particular is well known (e.g., Oh et al., 2003b). van Ginkel et al. (2001) reported that as sucrose increased threefold (from 2 to 6 g) at a constant quantity of heat-treated soil-containing H2-producing spores, both the PY and Rs decreased. Other investigators reported that as S0/X0 increased, the H2 production (Oh et al., 2003b) and the H2 concentration (%) increased (Noike and Mizuno, 2000), but with lower yield (mol H2 mol1 substrate). Incidentally, as S0/X0 increases, the cell yield of a Clostrida strain decreases (Ueki et al., 1991) as would also be expected from aerobic cultures (e.g., Liu et al., 1998). In other mixed culture studies, the S0/X0 ratios, among other factors (e.g., pH), significantly affect the distribution between H2-producing and solvent-producing metabolism (Dabrock et al., 1992), resulting in different H2 productions. Also, the same ratio of S0/X0 = 3 but with different amounts of S0 and X0, e.g., 3 g g1 and 30 g 10 g1, may represent a completely different picture of its overall effects either due to mass transfer limitation or varying metabolic state. Compounding the problem for literature comparison is the fact that the absolute X0 concentration may not be meaningful, since the fraction of H2-producing
spores (bacteria) in the mixed culture is unknown. The fact that many factors influence the effects of S0/X0 on H2 production may explain conflicting results as to the optimum S0/X0 ratios reported by various investigators. Consequently, batch S0/X0 studies may only have practical applications for a particular S0 and X0. For comparison, the max H2 yield is about 180 ml g1 TVS, and the maximum specific H2 production rate is 1000 ml g1 VSS d1 as reported by Lay et al. (1999) for the organic fraction of municipal solid waste and 60 mg g1 cellulous and 400 ml g1 VSS d1 for microcrystalline cellulose (Lay, 2001); all at different S0/X0 ratios. Unfortunately, pH in their study was not specified, as pH determines both the rate and extent of H2 production (van Ginkel et al., 2001; Oh et al., 2003b). The contour lines for the H2 yield (ml H2 g1 CODadded) and the Rs as a function of S0 and S0/X0 are shown in Fig. 5. In general, as S0/X0 increases up to 4, regardless the S0 concentration, the H2 yield increases (Fig. 5a). This may be expected due to cell growth. In other words, when this ratio is low, there is little spore germination. Further increases in S0/X0 would also reduce the H2 yield, as discussed before. Nonetheless, it appears for the maximum H2 yield, the optimum S0 occurs at 120 g l1 with the corresponding S0/X0 = 4. As for Rs (Fig. 5b), the pattern is little different than that of H2 yield in that higher Rs occurs as S0/X0 increased with lower S0.
Fig. 5. Contour lines for H2 yield potential and the specific H2 production rate as a function of S0 (g l1) and S0/X0: (a) H2 yield potential and (b) the specific H2 production rate.
K.-S. Fan, Y.-Y. Chen / Chemosphere 57 (2004) 1059–1068
3.2. Batch medium replacement In Phase II experiments, a certain fraction of reactor content was withdrawn and replaced with the fresh medium at the peak of the active batch growth stage previously conducted at about S0/X0 = 3. The added substrate and the corresponding new S0/X0 are shown in Table 2. It was thought that the active biomass would then immediately utilize the fresh feed resulting in perhaps higher H2 production. Fig. 6 illustrates the cumulative H2 production after the medium replacement at t = 0.48 d. There are many points that can be made in these plots. First, the pattern of these plots is similar to the earlier studies, but the H2 production of about 20 ml and yield 7.4 ml g1 COD is far below those of the same S0/X0 = 3 (2.5 g COD per 0.83 g VSS) in the previous runs (40 ml and 16 ml g1 COD; C4 in Table 1). Thus, the experiments are highly variable for the different sets of the same S0/X0 ratio. Secondly, the reproducibility of these tests before the medium replacement is exceptionally well; all 7 runs exhibited the same pattern with the maximum H2 production occurring at the same peak time (t close to 0.5 d). Thirdly, in all cases, there is still a lag time for H2 production after the medium replacement. Thus, the assumption of an immediately rapid growth (hence, H2 production) with the fresh medium addition is not valid; it still needs approximately another day for subsequently active H2 production. In fact, the H2 production for all five ‘‘replacement’’ cases between time 0.5 and 1.3 d is less than that of the control. The exact reason(s) for this inhibition is unclear. Fourth, it appears that after the medium replacement, there were several peaks of H2 production (see 44% in Fig. 6). Fifth, the additional H2 production quantities are much lower in these tests (6–105% more for substrate addition from 1 to 13 g, Table 2) than those of the previous runs with the substrate addition between 0.4 and 1.9 g (Table 1). When the reactor content was withdrawn, the seed was
Fig. 6. Cumulative H2 production for the batch cases of the replacement of the medium. The dash line indicates the replacement occurring at the peak of the H2 production.
1065
also wasted at the same fraction; thus the new S0/X0 would be much higher as tabulated in Table 2. For example, for the 44% fraction replacement, the seed quantity becomes 56% of the original 0.9 g, or 0.5 g remaining, providing there is not a significant biomass generated. Consequently, the new S0/X0 becomes 14 g COD g1 VSS (7 g added as shown in Table 2), which is too high for any appreciable H2 production based on Fig. 5 results. Nonetheless, it appears that the optimum replacement for a higher additional H2 production is about 44%. From Table 1 data, however, 7 g COD would have provided much higher H2 production if the test were done in the original S0/X0 cases in Phase I. Consequently, the ‘‘shock’’ approach to maximizing H2 production is not a sound strategy. 3.3. CSTR medium replacement The H2 production rate after shock treatment by wasting 25% of the reactor content (1 l) and adding the same amount of substrate is illustrated in Fig. 7. There were two methods used in Fig. 7: Fig. 7a data from the instantaneous addition of the substrate and Fig. 7b from the intermittent addition. In either case, about 115 g COD was added to the CSTR system. In general, the data are reproducible with respect to the response of different parameters to the shock treatment. Since pH was automatically controlled, the relatively consistent pH values were expected. It is noted that ORP of 500 mV is indeed under extremely reducing conditions. There is still a lag time before a sudden increase in H2 production after the new substrate feed. The t1 shown in Fig. 2 ranged from 3–4 h for instantaneous feeding (Fig. 7a) and was approximately 14 h for intermittent addition (Fig. 7b). It appears that H2-producing bacteria were not subject to the shock loading of 115 g substrate addition. Although the H2 production rate appears decreasing, the H2 concentration remains relatively constant. It is interesting to note that data are more scattering for the intermittent feeding case as compared to instantaneous feeding. Further, there is no clear pattern in H2 production rate for the intermittent feeding case (Fig. 7b). After the 1st and 2nd replacements, a peak in the H2 production rate is noticed indicating active growth and then the rate keeps decreasing. However, two peaks are observed after the 3rd replacement and another peak is only observed after a severe inhibition for the 4th replacement data. Nonetheless, the peak H2 production rate after the medium replacement is always higher than the previous baseline data. However, the H2 production rate is one order of magnitude lower than that (13 l l1 d1) reported by Fang et al. (2002) using sucrose as substrate. The reasons may be due to diversity of Fangs microbial consortium (Clostridum and Bacillus species with granules size in 1.6 mm diameter) and partly due to the nature
1066
K.-S. Fan, Y.-Y. Chen / Chemosphere 57 (2004) 1059–1068
Fig. 7. Transient state of H2 production, pH and ORP after 25% reactor content replacement in a CSTR system: (a) instantaneous substrate addition and (b) intermittent addition.
of substrate. The brewery substrate contains a large fraction of solid form (115 g l1 COD contains 90 g l1 VSS) and protein, since H2 production is not easily obtained from protein degradation (Mizuno et al., 2000b). The major low molecular weight acids identified are HAc and HBu with the total volatile acids about 8000 mg l1 in the case of instantaneous feed and about 6000 mg l1 for the intermittent case as shown in Fig. 8a. For CSTR study of newly isolated strains including Clostridum, the concentration of ethanol and acetate is only in the hundreds range of mg l1 (Oh et al., 2003a). Again the ratio of HAc/HBu in the present study is rel-
atively constant (about 5 for both feeding cases) and appears not significantly affected by the shock loading. Such high accumulation of HAc may indicate a lower H2 yield, since in other studies the byproducts consist of approximately equal distribution between HAc and HBu (Kataoka et al., 1997; Lay, 2001; Fang et al., 2002). The concentration of total alcohols is about 2000 mg l1 in both feeding cases and appears not affected by the shock treatment as shown in Fig. 8b. All in all, the CSTR system is able to recover to its previous performance after such a dramatic 25% medium replacement (Fig. 8).
Fig. 8. Transient state of volatile acids and alcohols after 25% reactor content replacement in a CSTR system: (a) instantaneous substrate addition and (b) intermittent addition.
K.-S. Fan, Y.-Y. Chen / Chemosphere 57 (2004) 1059–1068
4. Conclusions The heat treatment (85 °C for 12 h) is effective for selecting the H2-producing spores and eliminating H2consuming bacteria. The batch results clearly show typical growth pattern with lag, log and stationary phase of H2 production. The modified Gompertz equation is accurate in predicting the extent and the rate of H2 production. Although there is no clear pattern of the effect of substrate/seed ratios on H2 production, there exists an optimum S0/X0 ratio for the maximum H2 yield (ml g1 COD) and maximum specific H2 production rate (ml g1 VSS d1). The S0/X0 ratio determines the extent of catabolism and anabolism. As the ratio increases, an active growth results in higher H2 production. Further increases in S0/X0 would result in a large quantity of low molecular weight intermediates generated in the system, which in turn may inhibit the overall H2 production. At the peak of batch experiments, reactor content was withdrawn and replaced with the fresh substrate in the hope that H2 production could be enhanced. The results indicate additional quantity of H2 production is less than that of the original batch tests with the same amount of substrate addition. The reason may be due to high S0/X0 ratio since the denominator X0 becomes lower due to biomass wasting. For the CSTR medium replacement, the results from both instantaneous and intermittent feeding indicate that the system appears to be able to recover to the previous CSTR performance. Acknowledgments This work was supported by Energy Commission, Ministry of Economic Affairs, Republic of China under the project 91D0125. The authors would like to thank Li-Wei Cheng for his contribution to the experimental works, and Chen-Pu Chang for his computation assistance. References APHA, 1995. Standard Methods for Examination of Water and Wastewater, 19th ed. American Public Health Association, Washington, DC, USA. Bailey, J.E., Ollis, D.F., 1986. Biochemical Engineering Fundamentals, second ed. McGraw-Hill, New York, NY. Buys, B.R., Mosquera-Corral, A., Sanchez, M., Mendez, R., 2000. Development and application of a denitrification test based on gas production. Wat. Sci. Technol. 41, 113–120. Chang, J.S., Lee, K.S., Lin, P.J., 2002. Biohydrogen production with fixed-bed bioreactors. Int. J. Hydrogen Energy 27, 1167–1174.
1067
Chudoba, P., Capdeville, B., Chudoba, J., 1992. Explanations of biological meaning of the S0/X0 ratio in batch cultivation. Wat. Sci. Technol. 26, 743–751. Claassen, P.A.M., van Lier, J.B., Contreras, A.M.L., van Niel, E.W.J., Sijtsma, L., Stams, A.J.M., de Vries, S.S., Weusthuis, R.A., 1999. Utilisation of biomass for the supply of energy carriers. Appl. Microbiol. Biotechnol. 52, 741–755. Dabrock, B., Bahl, H., Gottschalk, G., 1992. Parameters affecting solvent production by Clostridum pasteuriamum. Appl. Environ. Microbiol. 58, 1233–1239. Das, D., Veziroglu, T.N., 2001. Hydrogen production by biological processes: a survey of literature. Int. J. Hydrogen Energy 26, 13–28. Fan, K.S., Kuo, T.H., 2004. Feasibility study on anaerobic hydrogenesis–methanogenesis process for high strength brewery wastewater, submitted to Biotechnol. Bioeng. Fang, H.H.P., Liu, H., 2002. Effect of pH on hydrogen production from glucose by a mixed culture. Bioresour. Technol. 82, 87–93. Fang, H.H.P., Liu, H., Zhang, T., 2002. Characterization of a hydrogen-producing granular sludge. Biotechnol. Bioeng. 78, 44–52. Kalia, V.C., Jain, S.R., Kumar, A., Joshi, A.P., 1994. Fermentation of biowaste to H2 by Bacillus-licheniformis. World J. Microbiol. Biotechnol. 10, 224–227. Kanayama, H., Karube, I., 1987. Hydrogen production by Escherichia coli containing a cloned Hydrogenase gene from Citrobacter-FreundII. J. Biotechnol. 6, 61–69. Kataoka, N., Miya, A., Kiriyama, K., 1997. Studies on hydrogen production by continuous culture system of hydrogen-producing anaerobic bacteria. Wat. Sci. Technol. 36, 41–47. Kumar, N., Das, D., 2000. Enhancement of hydrogen production by Enterobacter cloacae IIT-BT 08. Proc. Chem. 35, 589–593. Kumar, N., Monga, P.S., Biswas, A.K., Das, D., 2000. Modeling and simulation of clean fuel production by Enterbacter cloacae IIT-BT 08. Int. J. Hydrogen Energy 25, 945–952. Lay, J.J., 2001. Biohydrogen generation by mesophilic anaerobic fermentation of microcrystalline cellulose. Biotechnol. Bioeng. 74, 280–287. Lay, J.J., Lee, Y.J., Noike, T., 1999. Feasibility of biological hydrogen production from organic fraction of municipal solid waste. Wat. Res. 33, 2579–2586. Liu, Y., Chen, G.H., Paul, E., 1998. Effect of the S0/X0 ratio on energy uncoupling in substrate-sufficient batch culture of activated sludge. Wat. Res. 32, 2883–2888. Madigan, M.T., Martinko, J.M., Parker, J., 2000. Brook Biology of Microorganisms, ninth ed. Prentice Hill, Upper Saddle River, NJ. Mizuno, O., Li, Y.Y., Noike, T., 1998. The behavior of sulfatereducing bacteria in acidogenic phase of anaerobic digestion. Wat. Res. 32, 1626–1634. Mizuno, O., Dinsdale, R., Hawkes, F.R., Hawkes, D.L., Noike, T., 2000a. Enhancement of hydrogen production from glucose by nitrogen gas sparging. Bioresour. Technol. 73, 59–65. Mizuno, O., Ohara, T., Shinya, M., Noike, T., 2000b. Characteristics of hydrogen production from bean curd
1068
K.-S. Fan, Y.-Y. Chen / Chemosphere 57 (2004) 1059–1068
manufacturing waste by anaerobic microflora. Wat. Sci. Technol. 42, 345–350. Nandi, R., Sengupta, S., 1998. Microbial production of hydrogen: an overview. Critical Rev. Microbiol. 24, 61–84. Nandi, R., Ray, S., Sengupta, S., 2001. Thiosulphate improves yield of hydrogen production from glucose by the immobilized formate hydrogenlyase system of Escherichia coli. Biotechnol. Bioeng. 75, 492–494. Noike, T., Mizuno, O., 2000. Hydrogen fermentation of organic municipal wastes. Wat. Sci. Technol. 42, 155–162. Noike, T., Takabatake, H., Mizuno, O., Ohba, M., 2002. Inhibition of hydrogen fermentation of organic wastes by lactic acid bacteria. Int. J. Hydrogen Energy 27, 1367– 1371. Okamoto, M., Miyahara, T., Mizuno, O., Noike, T., 2000. Biological hydrogen potential of materials characteristic of the organic fraction of municipal solid wastes. Wat. Sci. Technol. 41, 25–32. Oh, Y.-K., Seol, E.-H., Lee, E.Y., Park, S., 2002. Fermentative hydrogen production by a new chemoheterotrophic bacterium Rhodopseudomonas palustris P4. Int. J. Hydrogen Energy 27, 1373–1379. Oh, Y.-K., Park, M.S., Seol, E.-H., Lee, S.-J., Park, S., 2003a. Isolation of hydrogen-producing bacteria from granular sludge of an upflow anaerobic sludge blanket reactor. Biotechnol. Bioproc. Eng. 8, 54–57. Oh, Y.-K., Seol, E.-H., Kim, J.R., Park, S., 2003b. Fermentative biohydrogen production by a new chemoheterotrophic
bacterium Citrobacter sp. Y19. Int. J. Hydrogen Energy 28, 1353–1359. Rachman, M.A., Nakashimada, Y., Kakizono, T., Nishio, N., 1998. Hydrogen production with high yield and high evolution rate by self-flocculated cells of Enterobacter aerogenes in a packed-bed reactor. Appl. Microbiol. Biotechnol. 49, 450–454. Taguchi, D., Chang, J.D., Takiguchi, S., Morimoto, M., 1992. Efficient hydrogen production from starch by a bacterium isolated from termites. J. Ferment. Technol. 73, 244–245. Tanisho, S., Kuromoto, M., Kadokura, N., 1998. Effect of CO2 removal on hydrogen production by fermentation. Int. J. Hydrogen Energy 23, 559–563. Ueki, A., Ueki, K., Yanaka, K., Takahashi, R., Takano, Y., 1991. End-products and molar growth-yield of clostridiumfallax isolated from an anaerobic digester. J. Ferment. Bioeng. 72, 274–279. van Ginkel, S., Sung, S., Li, L., Kay, J.-J., 2001. Role of initial sucrose and pH levels on natural, hydrogen-producing, anaerobe germination. In: Proc. 2001 DOE Hydrogen Program Review. van Ooteghem, S.A., Beer, S.K., Yue, P.C., 2001. Hydrogen production by the thermophilic bacterium, Thermotoga neapolitana. In: Proc. 2001 DOE Hydrogen Program Review. Zwietering, M.H., Jongenburger, I., Rombouts, F.M., Vantriet, K., 1990. Modeling of the bacterial-growth curve. Appl. Environ. Microbiol. 56, 1875–1881.
H2 production through anaerobic mixed culture: effect of batch S0/X0 and shock loading in CSTR Kuo-Shuh Fan *, Ya-Yun Chen Department of Safety and Environmental Engineering, National Kaohsiung First University of Science and Technology, No. 1 University Street, Yanchau 824, Kaohsiung, Taiwan Received 17 February 2004; received in revised form 12 August 2004; accepted 12 August 2004
Abstract Biological production of H2 has received considerable attention lately. The present study was undertaken to observe the effects of substrate/seeding ratios (S0/X0) on batch H2 generation. The H2-producing seeding spores were obtained from the heat treatment (88 °C for 12 h) of the compost from a grass composting facility. A dehydrated brewery mixture was used as feed substrate. The results indicate that the pattern of the cumulative H2 production with time is similar to the growth curve with a typical lag, exponential and stationary phase; the results were successfully modeled with a modified Gompertz equation. It appears that maximum H2 yield potential (27 ml g1 CODadded) occurs at an S0/X0 ratio of about 4, whereas the maximum specific H2 yield (205 ml g1 VSS d1) occurs at approximately S0/X0 = 3. The S0/X0 ratios higher than 4 would inhibit H2 production. An attempt was made to waste a certain amount of reactor content and replaced it with fresh substrate in order to enhance H2 production. After this medium replacement, the H2 production was initially inhibited and the system then exhibited a long lag before it reached an active H2 production stage. For a continuous-stirred tank-reactor (CSTR) system, the results of replacing 25% of the reactor content indicate that there is still a lag time before a sudden increase in H2 production after the addition of the new substrate feed. The major low molecular weight acids identified are HAc and HBu with total volatile acids of about 6000–8000 mg l1. The ratio of HAc/HBu in the present study is relatively constant (about 5) and appears not significantly affected by the medium replacement. The concentration of total alcohols is about 2000 mg l1. All in all, the CSTR system is able to recover to its previous performance after such a dramatic 25% medium replacement. Ó 2004 Elsevier Ltd. All rights reserved. Keywords: H2 production; H2-producing bacteria; S0/X0; Medium replacement; CSTR
1. Introduction
* Corresponding author. Tel.: +886 7 601 1000x2321; fax: +886 7 601 1061. E-mail address: [email protected] (K.-S. Fan).
Hydrogen production has received considerably attention recently, as energy demand grows and because of a simple fact that H2 is a relatively clean fuel with higher energy efficiency. Particularly, interest is emphasized in biological H2 generation as exemplified by many review papers in this area (e.g., Nandi and Sengupta,
0045-6535/$ - see front matter Ó 2004 Elsevier Ltd. All rights reserved. doi:10.1016/j.chemosphere.2004.08.021
1060
K.-S. Fan, Y.-Y. Chen / Chemosphere 57 (2004) 1059–1068
1998; Noike and Mizuno, 2000; Das and Veziroglu, 2001). In addition to photosynthetic organisms, fermentative bacteria are able to generate high concentrations of H2, along with some low molecular weight intermediates. The microbes responsible for H2 generation include: Enterobacter (e.g., Rachman et al., 1998; Kumar and Das, 2000; Kumar et al., 2000), Clostridum (e.g., Taguchi et al., 1992; Kataoka et al., 1997; Noike et al., 2002), Escherichia (Kanayama and Karube, 1987; Nandi et al., 2001), Actinomyces (Oh et al., 2003a), Bacillus (Kalia et al., 1994), and thermophilic Tnermotoga neapolitana (van Ooteghem et al., 2001). A typical pathway of fermentative H2 production is illustrated in Fig. 1. For example, Clostridum butyricum is able to ferment sugar to H2 with butyrate and acetate as byproducts (Madigan et al., 2000). Using glucose, a maximum of 2 mol of H2 could be produced with butyrate as the main product and 4 mol of H2 with acetate as the major byproduct (Nandi and Sengupta, 1998). The free energy for the latter reaction (206 kJ mol1) is sufficient for microbial growth (Claassen et al., 1999). As high as 60% H2 concentration (Mizuno et al., 2000b; Fang and Liu, 2002), yield of 2.5 mol H2 per mol hexose (Noike and Mizuno, 2000) and production rate of about 30 l H2 l1 reactor d1 (Rachman et al., 1998; Chang et al., 2002) could be obtained from a variety of substrate sources. In general, the H2 production is regulated by many factors, including (1) substrate characteristics, e.g., type and concentration of substrate (Ueki et al., 1991; Noike and Mizuno, 2000; Okamoto et al., 2000; Lay, 2001); (2) type of seeding sources (Lay et al., 1999); (3) environmental conditions of pH (van Ginkel et al., 2001; Fang and Liu, 2002; Oh et al., 2003b) and temperature (Oh et al., 2002, 2003b); (4) operating conditions, e.g., hydraulic retention time (HRT) in continuous-flow systems (Fan and Kuo, 2004) and mixing intensity;
Fig. 1. Typical pathway of fermentative hydrogen production (Tanisho et al., 1998).
(5) reactor configuration (e.g., immobilized system, Chang et al., 2002); (6) metabolites produced; (7) partial pressure of CO2/H2 (Tanisho et al., 1998; Mizuno et al., 2000a); (8) presence of H2 consumers in mixed culture (e.g., sulfate reducing bacteria and methanogens) and other bacteria (e.g., lactic acid bacteria, Noike et al., 2002); and (9) presence of other inorganic compounds, e.g., sulfate (Mizuno et al., 1998), thiosulphate (Nandi et al., 2001) and phosphate (Oh et al., 2002, 2003b) as well as other inhibitors. An easy way to determine the suitability of any substrate for its conversion to H2 by mixed culture is to perform simple batch tests. For example, the H2 production rate can be easily calculated through the gas production rate and the corresponding H2 concentration under anaerobic conditions. In fact, anaerobic batch evaluation of the extent and rate of gas production (e.g., CH4, H2 or N2) has been extensively studied; a few have been performed with respect to the effect of substrate to biomass ratio (S0/X0). The low or high S0/X0 ratios favoring a higher H2 production depend on the type of substrate and seeding bacteria (Lay et al., 1999). In general, higher S0/X0 ratios increase the rate of gas production; further increases in S0/X0 may eventually reduce the gas production rate, due partially to substrate and/or byproduct inhibition. Fed-batch operation, by adding substrate in a batch operation, has been used in industry to maximize product formation (Bailey and Ollis, 1986). To our best knowledge, there is no study performed about the effect of replacing medium volume of a well mixed reactor content with the fresh substrate on H2 production during batch growth stage. This approach is similar to fed-batch operation in concept. Further, during the batch exponential growth phase, addition of substrate may facilitate microbial reactions resulting in increased gas production. This ‘‘shocking’’ approach is analogous to adding substrate to highly concentrated active biomass with the expectation of continuing exponential growth to maximize H2 production. The continuous-flow stirred-tank (CSTR) system has been used widely in microbial studies in general and in observing H2 production in particular. Under constant pH and temperature condition, the system provides steady-state results in substrate, cell mass and byproduct concentrations. One of the key parameters in operation is the effect of shock loading on the CSTR performance. Traditionally, shock loading is conducted by either increasing flow rate or feed concentration or both. However, there is no study on the effect of the loss of reactor content on the subsequent CSTR operation. This is analogous to the loss of activated sludge mixed liquor due to sludge bulking in wastewater treatment plants. It is hypothesized that a direct addition of new substrate to replace the lost content would be a novel strategy to recover the CSTR performance.
K.-S. Fan, Y.-Y. Chen / Chemosphere 57 (2004) 1059–1068
Consequently, this study was undertaken to: (1) observe effects of S0/X0 on H2 production; (2) evaluate the effect of replacing the ‘‘lost’’ medium with the fresh substrate on the overall H2 production in batch studies; and (3) observe the transient state of H2 and byproducts in a CSTR system after the above mentioned medium replacement.
2. Materials and methods 2.1. Materials Serum bottles of 100 and 250 ml were used for the batch tests. The seed was obtained from the compost of a local grass composting facility. The compost was first dried at 85 °C for 12 h to select heat-resistant spore forming H2-producing bacteria which are certainly present under previously 50–60 °C composting conditions. Previous investigators have reported such a feasibility of heat treatment to enrich H2-producing biomass (Lay et al., 1999; Lay, 2001; van Ginkel et al., 2001). A given amount of dried compost was then added into 1 l of distilled water, mixed for 2 h, and the supernatant (X0) was used as the seed for the subsequent batch tests.
1061
The dehydrated fermented brewery mixture after dilution at the desirable strength (S0, 40–160 g l1) with supplemental nutrient addition was used as the feed solution. The dehydrated brewery mixture contains a high fraction of volatile solids, VS (0.79 g VS per g chemical oxygen demand, COD), about 60% volatile suspended solids to VS fraction and the soluble fraction of total COD is about 40%. A given amount of the inorganic nutrient solution was added to ensure the adequacy of other nutrients. The composition of the nutrient solution in 1 l of the medium contained: 80 g NH4HCO3, 40 g KH2PO4, 0.4 g MgSO2 Æ 7 H2O, 0.4 g NaCl, 0.4 g Na2MoO4 Æ 2H2O, 0.4 g CaCl2 Æ 2H2O, 0.6 g MnSO4 Æ 7H2O and 0.11 g FeCl2 (Lay et al., 1999). 2.2. Experiments 2.2.1. Batch There were two phases in the present studies. In Phase I, five different S0 quantities (from 0.4 to 4.8 g COD) were used for each of the five X0 concentrations (0.75–1.05 g VSS) to render a total of 25 different S0/X0 ratios (from 0.4 to 6.4 g COD g1 VSS) to observe H2 production (Table 1). The total volume of seed and substrate was 80 ml in 100 ml serum bottles. After the
Table 1 Lag time, H2 production and H2 maximum rate at different S0/X0 ratios Group
Seed vol (ml)a
Substrate, S0 (g COD l1)
Substrate (ml)
S0/X0 (g COD g1 VSS)
k (d)
P (ml)
Rm (ml d1)
PY (ml g1 COD)
Rs (ml g1 VSS d1)
A1 A2 A3 A4 A5 B1 B2 B3 B4 B5 C1 C2 C3 C4 C5 D1 D2 D3 D4 D5 E1 E2 E3 E4 E5
70 65 60 55 50 70 65 60 55 50 70 65 60 55 50 70 65 60 55 50 70 65 60 55 50
40 40 40 40 40 70 70 70 70 70 100 100 100 100 100 130 130 130 130 130 160 160 160 160 160
10 15 20 25 30 10 15 20 25 30 10 15 20 25 30 10 15 20 25 30 10 14 20 25 30
0.38 0.62 0.89 1.21 1.60 0.67 1.08 1.56 2.12 2.80 0.95 1.54 2.22 3.03 4.00 1.24 2.00 2.89 3.94 5.20 1.52 2.46 3.56 4.85 6.40
0.2 0.3 0.3 0.3 0.4 0.3 0.3 0.4 0.4 0.4 0.3 0.4 0.4 0.4 0.4 0.3 0.4 0.4 0.4 0.4 0.3 0.4 0.4 0.4 0.4
2.3 6.1 6.2 6.6 20.5 6.1 13.4 24.1 33.8 48.0 14.8 31.9 44.9 39.7 53.3 24.4 39.1 46.1 80.8 67.9 23.5 50.1 56.6 89.5 95.2
4.6 17.4 38.3 17.1 86.5 17.2 53.6 105 120 153 61.3 122 158 68.4 81 92 151 109 102 67 88 121 86 88 70
5.7 10.2 7.8 6.6 17.0 8.7 12.8 17.2 19.3 22.8 14.8 21.3 22.5 15.9 17.8 18.8 20.0 17.7 24.9 17.4 14.7 20.9 17.7 22.4 19.8
4 19 43 21 115 16 55 117 145 204 58 125 175 83 108 88 154 121 124 89 84 124 95 106 93
a
At seed concentration = 15 g VSS l1.
1062
K.-S. Fan, Y.-Y. Chen / Chemosphere 57 (2004) 1059–1068
Table 2 H2 production after medium replacementa Group
Replacement fraction (%)
Added substrate (g COD)
X0 remaining (g VSS)b
New S0/X0c (g COD g1 VSS)
Cumulative H2 (ml)
Increased cumulative H2 (%)
1 2 3 4 5 6
0 6 25 44 63 81
0 1 4 7 10 13
0.9 0.85 0.68 0.50 0.33 0.17
0 1.9 5.9 14 30 76
20 21 37 42 34 27
0 6 82 105 70 33
a b c
At S0/X0 = 3 g COD g1 VSS for seed 0.9 g (133 ml at 6.75 g l1) and substrate 2.7 g (27 ml at 100 g l1). Based on the original X0 = 0.9 g. Based on the original X0 and newly added S0.
mixture of substrate and seed, the bottles were purged with a mixture of N2/CO2 gas (7:3, v/v) to ensure anaerobic conditions; thereafter, the bottles were placed in a rotary shaker (30 rpm) at 37 °C. In addition to 25 bottles, several additional bottles with the random S0/X0 ratios were used to serve as duplicates for reproducibility checking. The optimum S0/X0 ratio and the corresponding time for the maximum production of H2 (tmax) during batch runs were identified. In Phase II, cultures were grown at an optimum S0/ X0 ratio, identified from the above tests. The total volume of the seed and substrate was 160 ml in 200 ml bottles. At tmax, different fractions of the bottle content volume were withdrawn, and replaced with the fresh substrate to see if any enhanced H2 production might occur. There were five replacement fractions, ranging from 6% to 81% of the bottle content (Table 2). In all batch studies, the gas volume produced was measured using 10 ml syringes, and the H2 composition in the biogas quantified. 2.2.2. Continuous study The system was initially operated in batch mode (with S0/X0 = 3) and the continuous flow was then initiated after observing H2 production. The system was operated at HRT = 8 h (dilution rate = 3 d1) with the same brewery substrate at concentration 115 g l1 and T = 37 ± 1 °C. The pH was automatically controlled at 5.5–6.0 with 2 N NaOH and ORP (oxidation–reduction potential) continuously monitored. After the system reached pseudo-steady-state, 25% of the reactor volume was wasted and fresh medium was then added. Two methods for substrate addition were used: (1) instantaneous addition; and (2) intermittent addition (about 80 ml every 10 min for 2 h). The latter approach is to minimize the shock loading to the remaining H2-producing bacteria. The transient state of several parameters was observed and recovery time was noted (Fig. 2). Several shock loading runs were conducted after 6–9 HRTs to observe the stability of the system. In all cases, the gas output rate was measured via wet-gas meters, and gas
Fig. 2. Biomass replacement in CSTR experiments. t1 and t2 refer to the increased H2 production period and recovery period, respectively. A and B refer to H2 production before and after the shock treatment and C is the final H2 production.
samples were taken from the sampling port located before the meter for H2 quantification. The liquid samples were also taken for the analysis of volatile fatty acids, alcohols and other parameters. 2.3. Analyses The H2 concentration was quantified in a GC (Shimadzu Model 8A) with a thermal conductivity detector. A 2-m glass column (Porapak Q, 50/80 mesh) was installed and the operating temperatures of the injection port, the oven, and the detector were 100, 60, and 100 °C, respectively. Nitrogen gas served as the carrier gas at a flow rate of 30 ml min1. The measurements of COD and VSS essentially followed those prescribed in the Standard Methods (APHA, 1995). Volatile acids (acetic acid, propionic acid and butyric acid) and alcohols (methanol,
K.-S. Fan, Y.-Y. Chen / Chemosphere 57 (2004) 1059–1068
ethanol, propanol and butanol) were measured in another GC-flame ionization detector (Shimadzu, Model 14B). The correlation coefficients for the calibration curves of all these standard volatile acids and alcohols were greater than 0.998. 2.4. Model analysis The modeling of H2 production during the batch studies is based on the following modified Gompertz equation (Zwietering et al., 1990): Rm e H 2 ¼ P exp exp ðk tÞ þ 1 P
ð1Þ
where H2 is cumulated H2 production (ml); P, H2 equilibrium production (ml); t, reaction time (d); k, lag time (d); and Rm, the maximum H2 production rate (ml d1) as shown in Fig. 3a. The H2 yield potential (PY, ml g1 CODadded) is determined by dividing P by the amount of feed substrate, and the specific H2 production rate (Rs, ml g1 VSS d1) is calculated by dividing Rm by the amount of the initial X0. The reason for using the initial substrate COD is that both substrate and seed contain COD; thus, H2 yield using the amount of COD removed may provide misleading results. A nonlinear regression method was used with the least sum of the square errors used to determine the model parameters.
Fig. 3. Cumulative H2 production in batch experiments: experimental data in solid circles and lines are model-predicted values. (a) Symbol designation for the model in Eq. (1); (b) typical examples for group D data in Table 1.
1063
3. Results and discussion 3.1. Batch H2 production Typical plots of H2 production for five S0/X0 ratios as a function of time are shown in Fig. 3b. The pattern of the cumulated H2 with time is similar to the growth curve with a typical lag, exponential and stationary phase. The resultant results of P, k and Rm for all 25 runs are tabulated in Table 1. For five random duplicate bottles with P up to 100 ml H2, the standard errors were less than 10%; thus the results are reproducible. The model fits experimental data well with the R2 values in Table 1 data all greater than 0.92. Consequently, the model equation is useful and relatively accurate to predict the extent (P) as well as the rate of H2 production (Rm) as was also reported by others (Lay et al., 1999; Lay, 2001; van Ginkel et al., 2001). The lag time ranges from 0.2 to 0.4 d with little variation; lower k values are, however, mostly associated with lower S0/X0 ratios. In a few limited studies, k values ranged from a few hours to 5 days, depending on the type of substrate and the seeding sources (Lay et al., 1999; van Ginkel et al., 2001). Clearly, these parameters and other environmental conditions (pH and temperature) determine the extent and rate of spore germination and subsequent H2 production. The maximum H2 yield potential is about 25 ml g1 CODadded (D4 at S0/X0 = 3.9, Table 1) with a maximum production rate about 158 ml d1 (C3, Table 1), and maximum Rs about 205 ml g1 VSS d1 (B5, Table 1). It appears that there is no clear pattern of the effect of S0/X0 on H2 production. Nonetheless, plots in Fig. 4 indicate that a maximum PY occurs at S0/X0 ratio about 4, whereas the maximum Rs occurs at approximately S0/X0 = 3.5. The fact that that PY is based on COD added and not actual substrate removed and that Rs determined from VSS in which both
Fig. 4. The H2 yield potential (PY, ml g1 CODadded) and the specific H2 production rate (Rs, ml g1 VSS d1) as a function of S0/X0.
1064
K.-S. Fan, Y.-Y. Chen / Chemosphere 57 (2004) 1059–1068
substrate and seed contribute may explain the phenomenon that PY and Rs did not occur at the same S0/X0 ratio. Further discussion of aerobic batch culture with respect to S0/X0 ratios may shed some light as to their effect on H2 production. At lower S0/X0, there is essentially no cell multiplication and at higher ratios, there are many polymer types of byproducts generated as reported by Chudoba et al. (1992) in their activated sludge studies. The S0/X0 ratios also determine the extent of catabolism and anabolism (Chudoba et al., 1992), or energy uncoupling (Liu et al., 1998). In monitoring N2 production for denitrification, the microbial activity increases with increased S0/X0 ratios (Buys et al., 2000). Depending on the culture conditions, a significant quantity of low molecular weight acids and alcohols were generated (not measured in batch studies, but monitored in the subsequent CSTR studies). For example, at low S0/X0 ratios with S0 (glucose) being limiting, Clostridium fallax strain produced acetate and butyrate; with excess S0, lactate was produced after HAc and HBu reached maximum concentrations (Ueki et al., 1991). Hence, these products may in fact inhibit H2 production, directly through product inhibition or indirectly via lower pH. In fact, Lay (2001) in his batch studies reported that the pH decreased from 7 to about 4.2–5.2; the higher the S0/X0 ratio was, the more pH drop. Nevertheless, the type of intermediates depends on substrate and seeding source, e.g., HBu and EtOH (Noike and Mizuno, 2000), HBu and PrOH (Lay et al., 1999), HAc and EtOH (Oh et al., 2003a) as well as HAc, HBu and BuOH (Lay, 2001). The fact that lower pH values due to byproduct accumulation inhibit CH4 production in general and H2 production in particular is well known (e.g., Oh et al., 2003b). van Ginkel et al. (2001) reported that as sucrose increased threefold (from 2 to 6 g) at a constant quantity of heat-treated soil-containing H2-producing spores, both the PY and Rs decreased. Other investigators reported that as S0/X0 increased, the H2 production (Oh et al., 2003b) and the H2 concentration (%) increased (Noike and Mizuno, 2000), but with lower yield (mol H2 mol1 substrate). Incidentally, as S0/X0 increases, the cell yield of a Clostrida strain decreases (Ueki et al., 1991) as would also be expected from aerobic cultures (e.g., Liu et al., 1998). In other mixed culture studies, the S0/X0 ratios, among other factors (e.g., pH), significantly affect the distribution between H2-producing and solvent-producing metabolism (Dabrock et al., 1992), resulting in different H2 productions. Also, the same ratio of S0/X0 = 3 but with different amounts of S0 and X0, e.g., 3 g g1 and 30 g 10 g1, may represent a completely different picture of its overall effects either due to mass transfer limitation or varying metabolic state. Compounding the problem for literature comparison is the fact that the absolute X0 concentration may not be meaningful, since the fraction of H2-producing
spores (bacteria) in the mixed culture is unknown. The fact that many factors influence the effects of S0/X0 on H2 production may explain conflicting results as to the optimum S0/X0 ratios reported by various investigators. Consequently, batch S0/X0 studies may only have practical applications for a particular S0 and X0. For comparison, the max H2 yield is about 180 ml g1 TVS, and the maximum specific H2 production rate is 1000 ml g1 VSS d1 as reported by Lay et al. (1999) for the organic fraction of municipal solid waste and 60 mg g1 cellulous and 400 ml g1 VSS d1 for microcrystalline cellulose (Lay, 2001); all at different S0/X0 ratios. Unfortunately, pH in their study was not specified, as pH determines both the rate and extent of H2 production (van Ginkel et al., 2001; Oh et al., 2003b). The contour lines for the H2 yield (ml H2 g1 CODadded) and the Rs as a function of S0 and S0/X0 are shown in Fig. 5. In general, as S0/X0 increases up to 4, regardless the S0 concentration, the H2 yield increases (Fig. 5a). This may be expected due to cell growth. In other words, when this ratio is low, there is little spore germination. Further increases in S0/X0 would also reduce the H2 yield, as discussed before. Nonetheless, it appears for the maximum H2 yield, the optimum S0 occurs at 120 g l1 with the corresponding S0/X0 = 4. As for Rs (Fig. 5b), the pattern is little different than that of H2 yield in that higher Rs occurs as S0/X0 increased with lower S0.
Fig. 5. Contour lines for H2 yield potential and the specific H2 production rate as a function of S0 (g l1) and S0/X0: (a) H2 yield potential and (b) the specific H2 production rate.
K.-S. Fan, Y.-Y. Chen / Chemosphere 57 (2004) 1059–1068
3.2. Batch medium replacement In Phase II experiments, a certain fraction of reactor content was withdrawn and replaced with the fresh medium at the peak of the active batch growth stage previously conducted at about S0/X0 = 3. The added substrate and the corresponding new S0/X0 are shown in Table 2. It was thought that the active biomass would then immediately utilize the fresh feed resulting in perhaps higher H2 production. Fig. 6 illustrates the cumulative H2 production after the medium replacement at t = 0.48 d. There are many points that can be made in these plots. First, the pattern of these plots is similar to the earlier studies, but the H2 production of about 20 ml and yield 7.4 ml g1 COD is far below those of the same S0/X0 = 3 (2.5 g COD per 0.83 g VSS) in the previous runs (40 ml and 16 ml g1 COD; C4 in Table 1). Thus, the experiments are highly variable for the different sets of the same S0/X0 ratio. Secondly, the reproducibility of these tests before the medium replacement is exceptionally well; all 7 runs exhibited the same pattern with the maximum H2 production occurring at the same peak time (t close to 0.5 d). Thirdly, in all cases, there is still a lag time for H2 production after the medium replacement. Thus, the assumption of an immediately rapid growth (hence, H2 production) with the fresh medium addition is not valid; it still needs approximately another day for subsequently active H2 production. In fact, the H2 production for all five ‘‘replacement’’ cases between time 0.5 and 1.3 d is less than that of the control. The exact reason(s) for this inhibition is unclear. Fourth, it appears that after the medium replacement, there were several peaks of H2 production (see 44% in Fig. 6). Fifth, the additional H2 production quantities are much lower in these tests (6–105% more for substrate addition from 1 to 13 g, Table 2) than those of the previous runs with the substrate addition between 0.4 and 1.9 g (Table 1). When the reactor content was withdrawn, the seed was
Fig. 6. Cumulative H2 production for the batch cases of the replacement of the medium. The dash line indicates the replacement occurring at the peak of the H2 production.
1065
also wasted at the same fraction; thus the new S0/X0 would be much higher as tabulated in Table 2. For example, for the 44% fraction replacement, the seed quantity becomes 56% of the original 0.9 g, or 0.5 g remaining, providing there is not a significant biomass generated. Consequently, the new S0/X0 becomes 14 g COD g1 VSS (7 g added as shown in Table 2), which is too high for any appreciable H2 production based on Fig. 5 results. Nonetheless, it appears that the optimum replacement for a higher additional H2 production is about 44%. From Table 1 data, however, 7 g COD would have provided much higher H2 production if the test were done in the original S0/X0 cases in Phase I. Consequently, the ‘‘shock’’ approach to maximizing H2 production is not a sound strategy. 3.3. CSTR medium replacement The H2 production rate after shock treatment by wasting 25% of the reactor content (1 l) and adding the same amount of substrate is illustrated in Fig. 7. There were two methods used in Fig. 7: Fig. 7a data from the instantaneous addition of the substrate and Fig. 7b from the intermittent addition. In either case, about 115 g COD was added to the CSTR system. In general, the data are reproducible with respect to the response of different parameters to the shock treatment. Since pH was automatically controlled, the relatively consistent pH values were expected. It is noted that ORP of 500 mV is indeed under extremely reducing conditions. There is still a lag time before a sudden increase in H2 production after the new substrate feed. The t1 shown in Fig. 2 ranged from 3–4 h for instantaneous feeding (Fig. 7a) and was approximately 14 h for intermittent addition (Fig. 7b). It appears that H2-producing bacteria were not subject to the shock loading of 115 g substrate addition. Although the H2 production rate appears decreasing, the H2 concentration remains relatively constant. It is interesting to note that data are more scattering for the intermittent feeding case as compared to instantaneous feeding. Further, there is no clear pattern in H2 production rate for the intermittent feeding case (Fig. 7b). After the 1st and 2nd replacements, a peak in the H2 production rate is noticed indicating active growth and then the rate keeps decreasing. However, two peaks are observed after the 3rd replacement and another peak is only observed after a severe inhibition for the 4th replacement data. Nonetheless, the peak H2 production rate after the medium replacement is always higher than the previous baseline data. However, the H2 production rate is one order of magnitude lower than that (13 l l1 d1) reported by Fang et al. (2002) using sucrose as substrate. The reasons may be due to diversity of Fangs microbial consortium (Clostridum and Bacillus species with granules size in 1.6 mm diameter) and partly due to the nature
1066
K.-S. Fan, Y.-Y. Chen / Chemosphere 57 (2004) 1059–1068
Fig. 7. Transient state of H2 production, pH and ORP after 25% reactor content replacement in a CSTR system: (a) instantaneous substrate addition and (b) intermittent addition.
of substrate. The brewery substrate contains a large fraction of solid form (115 g l1 COD contains 90 g l1 VSS) and protein, since H2 production is not easily obtained from protein degradation (Mizuno et al., 2000b). The major low molecular weight acids identified are HAc and HBu with the total volatile acids about 8000 mg l1 in the case of instantaneous feed and about 6000 mg l1 for the intermittent case as shown in Fig. 8a. For CSTR study of newly isolated strains including Clostridum, the concentration of ethanol and acetate is only in the hundreds range of mg l1 (Oh et al., 2003a). Again the ratio of HAc/HBu in the present study is rel-
atively constant (about 5 for both feeding cases) and appears not significantly affected by the shock loading. Such high accumulation of HAc may indicate a lower H2 yield, since in other studies the byproducts consist of approximately equal distribution between HAc and HBu (Kataoka et al., 1997; Lay, 2001; Fang et al., 2002). The concentration of total alcohols is about 2000 mg l1 in both feeding cases and appears not affected by the shock treatment as shown in Fig. 8b. All in all, the CSTR system is able to recover to its previous performance after such a dramatic 25% medium replacement (Fig. 8).
Fig. 8. Transient state of volatile acids and alcohols after 25% reactor content replacement in a CSTR system: (a) instantaneous substrate addition and (b) intermittent addition.
K.-S. Fan, Y.-Y. Chen / Chemosphere 57 (2004) 1059–1068
4. Conclusions The heat treatment (85 °C for 12 h) is effective for selecting the H2-producing spores and eliminating H2consuming bacteria. The batch results clearly show typical growth pattern with lag, log and stationary phase of H2 production. The modified Gompertz equation is accurate in predicting the extent and the rate of H2 production. Although there is no clear pattern of the effect of substrate/seed ratios on H2 production, there exists an optimum S0/X0 ratio for the maximum H2 yield (ml g1 COD) and maximum specific H2 production rate (ml g1 VSS d1). The S0/X0 ratio determines the extent of catabolism and anabolism. As the ratio increases, an active growth results in higher H2 production. Further increases in S0/X0 would result in a large quantity of low molecular weight intermediates generated in the system, which in turn may inhibit the overall H2 production. At the peak of batch experiments, reactor content was withdrawn and replaced with the fresh substrate in the hope that H2 production could be enhanced. The results indicate additional quantity of H2 production is less than that of the original batch tests with the same amount of substrate addition. The reason may be due to high S0/X0 ratio since the denominator X0 becomes lower due to biomass wasting. For the CSTR medium replacement, the results from both instantaneous and intermittent feeding indicate that the system appears to be able to recover to the previous CSTR performance. Acknowledgments This work was supported by Energy Commission, Ministry of Economic Affairs, Republic of China under the project 91D0125. The authors would like to thank Li-Wei Cheng for his contribution to the experimental works, and Chen-Pu Chang for his computation assistance. References APHA, 1995. Standard Methods for Examination of Water and Wastewater, 19th ed. American Public Health Association, Washington, DC, USA. Bailey, J.E., Ollis, D.F., 1986. Biochemical Engineering Fundamentals, second ed. McGraw-Hill, New York, NY. Buys, B.R., Mosquera-Corral, A., Sanchez, M., Mendez, R., 2000. Development and application of a denitrification test based on gas production. Wat. Sci. Technol. 41, 113–120. Chang, J.S., Lee, K.S., Lin, P.J., 2002. Biohydrogen production with fixed-bed bioreactors. Int. J. Hydrogen Energy 27, 1167–1174.
1067
Chudoba, P., Capdeville, B., Chudoba, J., 1992. Explanations of biological meaning of the S0/X0 ratio in batch cultivation. Wat. Sci. Technol. 26, 743–751. Claassen, P.A.M., van Lier, J.B., Contreras, A.M.L., van Niel, E.W.J., Sijtsma, L., Stams, A.J.M., de Vries, S.S., Weusthuis, R.A., 1999. Utilisation of biomass for the supply of energy carriers. Appl. Microbiol. Biotechnol. 52, 741–755. Dabrock, B., Bahl, H., Gottschalk, G., 1992. Parameters affecting solvent production by Clostridum pasteuriamum. Appl. Environ. Microbiol. 58, 1233–1239. Das, D., Veziroglu, T.N., 2001. Hydrogen production by biological processes: a survey of literature. Int. J. Hydrogen Energy 26, 13–28. Fan, K.S., Kuo, T.H., 2004. Feasibility study on anaerobic hydrogenesis–methanogenesis process for high strength brewery wastewater, submitted to Biotechnol. Bioeng. Fang, H.H.P., Liu, H., 2002. Effect of pH on hydrogen production from glucose by a mixed culture. Bioresour. Technol. 82, 87–93. Fang, H.H.P., Liu, H., Zhang, T., 2002. Characterization of a hydrogen-producing granular sludge. Biotechnol. Bioeng. 78, 44–52. Kalia, V.C., Jain, S.R., Kumar, A., Joshi, A.P., 1994. Fermentation of biowaste to H2 by Bacillus-licheniformis. World J. Microbiol. Biotechnol. 10, 224–227. Kanayama, H., Karube, I., 1987. Hydrogen production by Escherichia coli containing a cloned Hydrogenase gene from Citrobacter-FreundII. J. Biotechnol. 6, 61–69. Kataoka, N., Miya, A., Kiriyama, K., 1997. Studies on hydrogen production by continuous culture system of hydrogen-producing anaerobic bacteria. Wat. Sci. Technol. 36, 41–47. Kumar, N., Das, D., 2000. Enhancement of hydrogen production by Enterobacter cloacae IIT-BT 08. Proc. Chem. 35, 589–593. Kumar, N., Monga, P.S., Biswas, A.K., Das, D., 2000. Modeling and simulation of clean fuel production by Enterbacter cloacae IIT-BT 08. Int. J. Hydrogen Energy 25, 945–952. Lay, J.J., 2001. Biohydrogen generation by mesophilic anaerobic fermentation of microcrystalline cellulose. Biotechnol. Bioeng. 74, 280–287. Lay, J.J., Lee, Y.J., Noike, T., 1999. Feasibility of biological hydrogen production from organic fraction of municipal solid waste. Wat. Res. 33, 2579–2586. Liu, Y., Chen, G.H., Paul, E., 1998. Effect of the S0/X0 ratio on energy uncoupling in substrate-sufficient batch culture of activated sludge. Wat. Res. 32, 2883–2888. Madigan, M.T., Martinko, J.M., Parker, J., 2000. Brook Biology of Microorganisms, ninth ed. Prentice Hill, Upper Saddle River, NJ. Mizuno, O., Li, Y.Y., Noike, T., 1998. The behavior of sulfatereducing bacteria in acidogenic phase of anaerobic digestion. Wat. Res. 32, 1626–1634. Mizuno, O., Dinsdale, R., Hawkes, F.R., Hawkes, D.L., Noike, T., 2000a. Enhancement of hydrogen production from glucose by nitrogen gas sparging. Bioresour. Technol. 73, 59–65. Mizuno, O., Ohara, T., Shinya, M., Noike, T., 2000b. Characteristics of hydrogen production from bean curd
1068
K.-S. Fan, Y.-Y. Chen / Chemosphere 57 (2004) 1059–1068
manufacturing waste by anaerobic microflora. Wat. Sci. Technol. 42, 345–350. Nandi, R., Sengupta, S., 1998. Microbial production of hydrogen: an overview. Critical Rev. Microbiol. 24, 61–84. Nandi, R., Ray, S., Sengupta, S., 2001. Thiosulphate improves yield of hydrogen production from glucose by the immobilized formate hydrogenlyase system of Escherichia coli. Biotechnol. Bioeng. 75, 492–494. Noike, T., Mizuno, O., 2000. Hydrogen fermentation of organic municipal wastes. Wat. Sci. Technol. 42, 155–162. Noike, T., Takabatake, H., Mizuno, O., Ohba, M., 2002. Inhibition of hydrogen fermentation of organic wastes by lactic acid bacteria. Int. J. Hydrogen Energy 27, 1367– 1371. Okamoto, M., Miyahara, T., Mizuno, O., Noike, T., 2000. Biological hydrogen potential of materials characteristic of the organic fraction of municipal solid wastes. Wat. Sci. Technol. 41, 25–32. Oh, Y.-K., Seol, E.-H., Lee, E.Y., Park, S., 2002. Fermentative hydrogen production by a new chemoheterotrophic bacterium Rhodopseudomonas palustris P4. Int. J. Hydrogen Energy 27, 1373–1379. Oh, Y.-K., Park, M.S., Seol, E.-H., Lee, S.-J., Park, S., 2003a. Isolation of hydrogen-producing bacteria from granular sludge of an upflow anaerobic sludge blanket reactor. Biotechnol. Bioproc. Eng. 8, 54–57. Oh, Y.-K., Seol, E.-H., Kim, J.R., Park, S., 2003b. Fermentative biohydrogen production by a new chemoheterotrophic
bacterium Citrobacter sp. Y19. Int. J. Hydrogen Energy 28, 1353–1359. Rachman, M.A., Nakashimada, Y., Kakizono, T., Nishio, N., 1998. Hydrogen production with high yield and high evolution rate by self-flocculated cells of Enterobacter aerogenes in a packed-bed reactor. Appl. Microbiol. Biotechnol. 49, 450–454. Taguchi, D., Chang, J.D., Takiguchi, S., Morimoto, M., 1992. Efficient hydrogen production from starch by a bacterium isolated from termites. J. Ferment. Technol. 73, 244–245. Tanisho, S., Kuromoto, M., Kadokura, N., 1998. Effect of CO2 removal on hydrogen production by fermentation. Int. J. Hydrogen Energy 23, 559–563. Ueki, A., Ueki, K., Yanaka, K., Takahashi, R., Takano, Y., 1991. End-products and molar growth-yield of clostridiumfallax isolated from an anaerobic digester. J. Ferment. Bioeng. 72, 274–279. van Ginkel, S., Sung, S., Li, L., Kay, J.-J., 2001. Role of initial sucrose and pH levels on natural, hydrogen-producing, anaerobe germination. In: Proc. 2001 DOE Hydrogen Program Review. van Ooteghem, S.A., Beer, S.K., Yue, P.C., 2001. Hydrogen production by the thermophilic bacterium, Thermotoga neapolitana. In: Proc. 2001 DOE Hydrogen Program Review. Zwietering, M.H., Jongenburger, I., Rombouts, F.M., Vantriet, K., 1990. Modeling of the bacterial-growth curve. Appl. Environ. Microbiol. 56, 1875–1881.