Hydrogen supersaturation in extreme-thermophilic (70 °C) mixed culture fermentation

Applied Energy 109 (2013) 213–219

Contents lists available at SciVerse ScienceDirect

Applied Energy journal homepage: www.elsevier.com/locate/apenergy

Hydrogen supersaturation in extreme-thermophilic (70 °C) mixed culture fermentation Yan Zhang a,1, Fang Zhang b,1, Man Chen b, Pei-Na Chu b, Jing Ding b, Raymond J. Zeng a,b,⇑ a b

School of Earth and Space Sciences, University of Science and Technology of China, Hefei 230026, China Department of Chemistry, University of Science and Technology of China, Hefei 230026, China

h i g h l i g h t s

g r a p h i c a l a b s t r a c t

 H2 supersaturation occured in

extreme-thermophilic mixed culture fermentation.  H2 supersaturation ratio (RH2) increased when reducing H2 partial pressure (PH2).  Metabolite distribution changed little under low PH2 due to the high RH2 value.  Mass transfer calculation indicated H2 supersaturation was likely inevitable.  Suggested gas sparging rate vs H2 production rate was 2–10 to improve H2 yield.

a r t i c l e

i n f o

Article history: Received 15 October 2012 Received in revised form 12 March 2013 Accepted 5 April 2013 Available online 4 May 2013 Keywords: Hydrogen supersaturation Mixed culture fermentation Extreme-thermophilic H2 partial pressure Dissolved H2 KLa

a b s t r a c t Hydrogen supersaturation in extreme-thermophilic (70 °C) mixed culture fermentation (MCF) was demonstrated for the first time by membrane inlet mass spectrometry. It was found that hydrogen supersaturation ratio (RH2) increased dramatically (from 1.0 to 20.6) when H2 partial pressure (PH2) was reduced by N2 flushing or sparging. The distribution change of metabolites was insignificant under low PH2 (<0.30 atm) due to the high value of RH2, which indicated that it was more relevant to the concentration of dissolved H2 (H2aq) rather than PH2. To explain the cause of hydrogen supersaturation, the overall volumetric mass transfer coefficients (KLa) for H2 were calculated. KLa changed slightly (7.0/h) with N2 flushing, while it increased from 7.4 to 10.2/h when N2 sparging rate increased from 0.3 to 17.9 mL/ min/L. However, the required KLa values were orders of magnitude higher than the experimental ones when maintaining low RH2 by gas sparging, which indicated that hydrogen supersaturation was likely inevitable in MCF. Moreover, to improve the hydrogen yield of MCF, the gas sparging rate was suggested as 2–10 times of the hydrogen production rate. Ó 2013 Elsevier Ltd. All rights reserved.

1. Introduction Conventional fuels such as petroleum and natural gas are amount to 85% of the world energy requirements, which result in

⇑ Corresponding author at: School of Earth and Space Sciences, University of Science and Technology of China, Hefei 230026, China. Tel.: +86 551 63600203; fax: +86 551 63601592. E-mail address: [email protected] (R.J. Zeng). 1 These authors are contributed equally to this work. 0306-2619/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.apenergy.2013.04.019

the irreversible diminishment of natural reserves in the coming years and the terrible environmental pollutions. Therefore, the alternative technologies to simultaneously produce clean energy and degrade the existing wastes are worldwide needed [1]. Mixed culture fermentation (MCF) is a mature and promising technology to convert organic wastes into the valuable chemicals and biofuels, such as hydrogen, ethanol and polyhydroxyalkanoates (PHAs) [2–5]. Among the above products, hydrogen is an ideal energy carrier for the carbon-free nature and high energy density and so on [6]. H2 plays a very important role in MCF to control the redox

214

Y. Zhang et al. / Applied Energy 109 (2013) 213–219

balances and metabolic pathways of fermentative microorganisms [7,8]. Most researches use H2 partial pressure (PH2) as an indicator to understand the role of H2 in MCF [9–11]. However, H2 in liquid phase which the organisms are actually exposed to was usually ignored as the complicated and unstable measurement of the concentration of dissolved H2 in bulk solution (H2aq) [12,13]. Hydrogen supersaturation has been observed in a few MCF processes [13–15]. Kraemer and Bagley have observed this phenomenon in mesophilic MCF regardless of N2 sparging [13]. Our study of thermophilic MCF showed the hydrogen supersaturation ratio (RH2) increased with the increase of organic loading rate and the decrease of Reynold number [14]. These suggest that H2aq shall be a more proper factor than PH2 to study the effect of H2. Recently, more and more studies were focusing on MCF at extreme-thermophilic conditions (>65 °C) due to better pathogen destruction and thermodynamic conditions [16–18]. Hydrogen supersaturation under these conditions was usually ignored due to higher diffusion coefficient of H2 at higher temperature [19]. However, H2 supersaturation was observed in batch experiments of the extreme-thermophilic (70 °C) Caldicellulosiruptor saccharolyticus [12], which suggested that it may also exist in extreme-thermophilic MCF. On the other hand, pH has a significant effect on microbial metabolism [5,20–22], while the alkaline condition was seldom investigated in MCF. A high H2 yield (3.1 mol/molglucose) in fermentation performed under pH 8.0 was reported recently [23]. Under the alkaline condition, more produced CO2 will be dissolved and the PH2 level must be increased, which might result in the elevation of hydrogen supersaturation. Hydrogen supersaturation in MCF was largely related to different operating conditions, such as stirring and gas sparging [13,14,24]. RH2 increased from 3 to 11 when N2 sparging was used to decrease H2aq in fermentative liquor [13]. It decreased from 2.8 to 1.8 as the stirring rate increased from 120 to 450 rpm [14]. The overall volumetric mass transfer coefficient (KLa) is representative of the rate of gas transfer from liquid to gas, and is clearly specific to a given reactor and mode of operating condition [25,26]. It has been used to explain the relationship between H2 mass transfer and N2 sparging rate, and to optimize H2 production of N2-sparged bioreactors [25]. Hence, KLa may be a useful parameter to investigate the relationship between hydrogen supersaturation and different operating conditions. Therefore, the objective of this study was to investigate hydrogen supersaturation in extreme-thermophilic (70 °C) MCF at pH

8.0. To accomplish this objective, N2 sparging and N2 flushing were performed to change PH2 in a continuous stirred tank reactor (CSTR) under two different stirring rates (150 and 400 rpm). The relationship between hydrogen supersaturation and PH2 was investigated. KLa was calculated to describe the possible effects of different operating conditions to hydrogen supersaturation. The variation of metabolites and the effect of different gas sparging rates were also discussed. 2. Materials and methods 2.1. Inoculum, reactor setup and media The inoculum used in this study was an anaerobic sludge taken from a UASB reactor treating citrate-producing wastewater. As shown in Fig. 1, a glass-made CSTR reactor (1.8 L capacity and 1.35 L working volume) was used. The anaerobic sludge was acclimatized in the reactor before experiments. At the initial period of acclimation, the temperature gradually increased from 30 to 70 °C within 30 days by a water bath, while pH was gradually increased from 7.0 to 8.0 with automatic addition of 2 M NaOH. Hydraulic retention time (HRT) was kept at 10 days for 30 days to ensure the biomass activity. Then, it was gradually decreased to 0.8 day within 30 days. The whole acclimation approximately lasted for 90 days. The equipment was interfaced to computer via an Opto PLC used for data logging and set-point modification. Meanwhile, the system was also equipped with a Hiden HPR-40 DSA Membrane Inlet Mass Spectrometer (MIMS) for online monitoring H2aq. The synthetic feed consisted of 50% solution A and 50% solution B. Solution A contained 10 g/L of glucose and 1 g/L of yeast extract, autoclaved at 110 °C for 20 min. The composition of solution B (mg/L) was as follows: KH2PO4 400; KCl 104; Na2SO4 84; NH4Cl 20; MgCl26H2O 144; CaCl22H2O 20; MnCl24H2O 1.6; CoCl26H2O 2.4; H3BO3 0.4; CuCl22H2O 2.2; NaMO42H2O 0.2; ZnSO47H2O 6.4; FeSO47H2O 6.4; NiCl6H2O 1.0; EDTA 0.5. 2.2. Experimental design N2 flushing or sparging was used to introduce changes in the headspace gas composition at two different stirring rates (400 and 150 rpm). When the stirring rate was 400 rpm, the N2 flushing

Fig. 1. Schematic diagram of the experimental setup.

Y. Zhang et al. / Applied Energy 109 (2013) 213–219

215

rates were set as 0, 0.3, 1.2, 3.6, 8.8 and 22.0 mL/min/L, respectively, while the N2 sparging rates were set as 0, 0.3, 1.1, 3.3, 7.2 and 17.9 mL/min/L, respectively. When the stirring rate was 150 rpm, the N2 flushing rates were set as 0, 0.6 and 8.7 mL/min/ L, respectively, while the N2 sparging rates were set as 0, 0.6 and 11.1 mL/min/L, respectively. The units of mL/min/L, which were used by N2 flushing rates, N2 sparging rates and H2 production rates in this work, represent mL gas per min per L of liquid volume. The experiments without N2 flushing or sparging were used as the control. The reactor was operated at each condition for at least 10 HRTs before changing to the next condition. Biogas and effluent compositions were daily monitored. The evolution of H2aq in liquid phase was quantitatively monitored by MIMS. The MIMS-signals were calibrated using a slightly modified method from BastidasOyanedel et al. [27], which the detail could be found in Supporting Information. Fig. 2. Changes of PH2 and H2aq under different N2 flow rates at 400 rpm.

2.3. Analytical methods Gases in the headspace (H2, CH4, CO2) and the concentrations of volatile fatty acids (VFAs), ethanol, lactic and formic acids were measured according to our previous study [14]. Glucose in the effluent was analyzed by the phenol-sulfuric acid method for reducing sugars [28]. Suspended solids (SS) and volatile suspended solids (VSS) were measured according to the standard methods [29]. The chemical oxygen demand (COD) balance was based on the COD concentration of each metabolite and the influent COD of glucose. A standard biomass composition of CH1.8O0.5N0.2 was assumed [20]. 2.4. Hydrogen supersaturation ratio The hydrogen supersaturation ratio, RH2, is a measure of the extent of supersaturation of dissolved H2. In this study, RH2 was calculated as in the following equation:

RH2 ¼

H2aq H2aq

ð1Þ

H2aq ¼ K H PH2

ð2Þ H2aq

where H2aq and are the concentration of actual dissolved H2 (mol/L) and the saturation concentration of H2 (mol/L), respectively; PH2 is the H2 partial pressure (atm) and KH is Henry’s law constant (mol/L/atm). KH for H2 at 70 °C is 0.000724 mol/L/atm according to Green and Perry [30]. 2.5. KLa calculation

3. Results 3.1. Supersaturation of the dissolved H2 As shown in Fig. 2, PH2 and H2aq decreased when applying either N2 flushing or N2 sparging. Error bars represent the standard deviation of triplicates. In N2 flushing experiments, PH2 decreased quickly from 0.87 to 0.23 atm when the N2 flow rate increased from 0 to 1.2 mL/min/L, while H2aq decreased from 637 to 300 lM. Similarly, in N2 sparging experiments, PH2 decreased quickly from 0.87 to 0.29 atm when the N2 flow rate increased from 0 to 1.1 mL/min/L, while H2aq decreased from 637 to 290 lM. However, the changes of H2aq was little when the N2 flow rate continued to increase. Notably, a lower H2aq was achieved in N2 sparging than flushing (Fig. 2), which was mainly due to the increased agitation of fermentative liquor with gas sparging. RH2 under different operating conditions is illustrated in Fig. 3. Error bars were determined as the standard deviation of triplicates. It was interesting to find that the change trend of RH2 was very similar between different experiments, including different stirring rates and N2 stripping ways. RH2 was approximately equal to 1.0 when PH2 was over 0.30 atm, which indicated hydrogen supersaturation was insignificant. However, the degree of hydrogen supersaturation increased dramatically when further reducing PH2. RH2 increased from 2.0 to 20.6 when PH2 decreased from 0.23 to 0.01 atm by N2 flushing, while it increased from 1.3 to 10.0 when PH2 decreased from 0.29 to 0.01 atm by N2 sparging (Fig. 3). Similarly, Kraemer and Bagley [13] have observed that RH2 increased

KLa varies with the rate of gas flow through the reactor, it can be calculated by the following equations based on the superficial gas velocity lg (m/s) [12,31].

lg ¼ Q =A

ð3Þ

K L a ¼ klcg

ð4Þ

where A and Q are the cross sectional area (m2) of reactor and the total gas flow rate (m3/h), respectively; c is an exponential coefficient and k is a constant depending on the reactor setup and the physical properties of the liquid medium. If the reactor setup and liquid medium are not altered in the fermentation experiments, the following relation can be obtained:

K L a ¼ K L a0 ðF=F 0 Þc

ð5Þ

where KLa0 (1/h) and KLa (1/h) are the volumetric mass transfer coefficient for H2 at a total gas flow rate of F0 (mL/min/L) and F (mL/min/L), respectively.

Fig. 3. Relationship between RH2 and PH2 in different experiments.

216

Y. Zhang et al. / Applied Energy 109 (2013) 213–219

from 3 to 11 when N2 sparging was used to improve the H2 yield of MCF. As the H2aq in fermentation liquor became less responsive to the N2 flow rate with the flow rate increase (Fig. 2), the possible cause of the increase of RH2 is that the mass-transfer limitations of H2 movement from the liquid to the gas phase. The high value of RH2 also indicated that PH2 was a poor indicator of the hydrogen concentration in liquid, especially under low PH2 (<0.30 atm). 3.2. The yield change of metabolites The chemical oxygen demand (COD) balance closed between 90% and 110% in all experiments (Table S1), which indicated that most of the products were measured and the loss of volatile products could be ignored. The metabolites distribution was similar between two different stirring rates (400 and 150 rpm). The yield of primary metabolites at 400 rpm are illustrated in Fig. 4. Error bars mean the standard deviation of triplicates. Except the N2 sparging experiments under low PH2 (<0.29 atm), we have observed increase on the hydrogen (YH2) and biomass (Ybiomass) yields when decreasing PH2 by N2 flushing or sparging, which was similar to the former studies [9,10,25]. YH2 and Ybiomass increased first and then kept stable in N2 flushing experiments when decreasing PH2 (Fig. 4A), while increased first and then decreased in N2 sparging experiments (Fig. 4B). It seemed that the microbial activity might be affected at high N2 sparging rate [9,24]. The ethanol yield (Yethanol) was not influenced under different PH2, about 0.8 mol/molglucose in N2 sparging experiments and 0.9 mol/molglucose in N2 flushing experiments, which suggested that the microbial functional enzymes activity of ethanol production pathway, such as ferredoxin-NAD oxidoreductase, might be

insensitive to H2 [32]. The acetate yield (Yacetate) increased first and then kept generally stable when decreasing PH2 in both N2 flushing and N2 sparging experiments, while the lactate yield (Ylactate) decreased first and then kept generally stable (Fig. 4C and D). Since the intracellular NADH/NAD+ is related to H2, high H2 concentration could shift glucose metabolic pathways to produce the reduced metabolites such as ethanol and lactate [10]. In this study, as Yethanol was not influenced by H2, the ratio of lactate/acetate dropped from 1.4 to 0.5 when the N2 flow rate increased from 0 to 1.1 mL/min/L in N2 sparging experiments. However, at high N2 flow rates, H2aq changed little (from 290 to 115 lM) while PH2 changed dramatically (from 0.29 to 0.01 atm). Meanwhile, the metabolite distribution also did not change much, and the ratio of lactate/acetate kept around 0.5. This strongly indicated that the change of metabolites was more relevant to H2aq rather than PH2. 3.3. Calculation of KLa The mass balance of H2 can be established for the fermentative liquor by Eq. (6), which mainly comprises H2 produced from glucose by microbial fermentation, H2 transferred from liquid phase to gas phase and H2 dissolved in the effluent [12,14,26]. Hence, the relationship between RH2 and KLa can be described as Eq. (7), which is obtained from Eqs. (1) and (6).

DY H2 ðG0  GÞ  K L aðH2aq  H2aq Þ ¼ DH2aq RH2 ¼

DY H2 ðG0  GÞ þ K L aH2aq ðK L a þ DÞH2aq

ð6Þ ð7Þ

Fig. 4. Product yield of primary metabolites under different PH2 at 400 rpm. (A) H2 and biomass yield in N2 flushing experiments; (B) H2 and biomass yield in N2 sparging experiments; (C) ethanol, acetate and lactate yield in N2 flushing experiments; and (D) ethanol, acetate and lactate yield in N2 sparging experiments.

Y. Zhang et al. / Applied Energy 109 (2013) 213–219

217

Table 1 Comparison between the calculated KLa values and the experimentally measured RH2 values under different N2 flow rates at 400 rpma. N2 flow rate (mL/min/L)

KLa (1/h)

RH2 (calculated)

RH2 (measured)

Control test 0

6.8

1.2

1.0 ± 0.1

N2 flushing 0.3 1.2 3.6 8.8 22.0

6.9 6.9 7.0 7.2 7.0

1.5 2.1 3.5 6.2 19.7

1.3 ± 0.1 2.0 ± 0.2 3.3 ± 0.2 6.4 ± 0.4 20.6 ± 1.8

N2 sparging 0.3 1.1 3.3 7.2 17.9

7.4 8.0 8.7 9.4 10.2

1.4 1.9 3.6 5.5 11.9

1.1 ± 0.1 1.3 ± 0.1 4.0 ± 0.3 6.4 ± 0.4 10.0 ± 1.0

a The KLa value was calculated from Eq. (8). The RH2 (calculated) was calculated from Eq. (7) which based on the calculated KLa. The RH2 (measured) was determined from the experimentally measured H2aq and PH2.

where YH2 is the theoretical yield of H2 (mol/molglucose) calculated from liquid products, which mainly include acetate, formate, propionate and butyrate in the present study [12,14]; D, G0 and G are the dilution rate (1/h), the influent glucose concentration (mol/L) and the effluent glucose concentration (mol/L), respectively. The calculated KLa values in different experiments at 400 rpm are illustrated in Table 1. For N2 flushing experiments, KLa can be calculated by Eq. (6). The KLa value changed slightly since the effect of different N2 flushing rates to hydrodynamic conditions was little, which approximately was 7.0 ± 0.2. The exponential parameter c was calculated from KLa obtained at different gas flow rates (Eqs. (4) and (5)), given a value of 0.1. Meanwhile, KLa was determined to be 6.8/h at a N2 flow rate of 0.3 mL/min/L. Hence, Eq. (5) is derived as:

K L a ¼ 6:8ðF=0:3Þ0:1

ð8Þ

where F is the rate of total gases (mL/min/L) flowing through the fermentative liquor. In N2 flushing experiments, total gases mainly included H2 and CO2. However, in N2 sparging experiments, total gases included not only H2 and CO2 produced by microorganisms, but also sparging N2. KLa changed significantly in N2 sparging experiments (Table 1). It increased from 7.4 to 10.2/h with N2 flow rate increased from 0.3 to 17.9 mL/min/L. The calculated KLa values were larger than the values previously reported under similar gas flow rates [12,25,26], which was mainly caused by the high temperature (70 °C) in this study. Notably, KLa became less responsive to the gas flow rate as the flow rate increased (Eq. (8)). Table 1 also illustrates the comparison between experimentally calculated RH2 and RH2 calculated from KLa. Errors of experimentally calculated RH2 were determined as the standard deviation of triplicates. The difference between the calculated and the experimentally calculated RH2 is low, which indicated that calculated KLa was suitable to investigate the relationship between hydrogen supersaturation and different gas flow rates. 4. Discussion

Fig. 5. Required KLa values under different PH2 and RH2. KLa were calculated from Eq. (7), where the parameters of D, YH2, G0 and G were set as 0.051/h, 1.0 mol/ molglucose, 0.028 mol/L and 0.0003 mol/L, respectively.

between hydrogen supersaturation and different operating conditions. The simulated KLa values under different PH2 and RH2 are illustrated in Fig. 5. It was found that KLa should be improved significantly in order to obtain a low PH2 or RH2. For example, KLa required is 390/h when RH2 and PH2 are 1.05 and 0.10 atm, respectively. If RH2 is constant, KLa required will increase to 21,400/h as PH2 reduces to 180 Pa. However, as shown in Table 1, even when the N2 sparging rate was 17.9 mL/min/L, the KLa was just 10.2/h. The required KLa values are orders of magnitude higher than the experimental ones, and it is likely impossible to get such high KLa by conventional physical or chemical methods [14,25]. Therefore, RH2 will increase when PH2 is reduced. The simulated result matches well with the experimental one in current work (Fig. 3), which indicates that hydrogen supersaturation is likely inevitable when reducing PH2 in extreme-thermophilic condition. However, hydrogen supersaturation was insignificant when PH2 was over 0.30 atm, indicating that the existence of it in extremethermophilic MCF was conditional. Ljunggren et al. [12] suggested that RH2 shall be dependent on not only H2 mass transfer rate but also H2 production rate (FH2). In this study, FH2 of all experiments was less than 0.5 mL/min/L, which was mainly caused by the special conditions, i.e. 70 °C and alkaline pH 8.0. If FH2 is increased, hydrogen supersaturation will occur in this work regardless of the PH2 variation. For example, RH2 will be 1.70 if FH2 is 1.0 mL/ min/L (the equivalent YH2 as 2.5 mol/molglucose) in the control test according to Eq. (7), while RH2 was only 1.03 at 0.3 mL/min/L in this study. In addition, the RH2 values were almost same between two different stirring rates (400 and 150 rpm) under similar PH2 (Fig. 3). The possible reason is that the contribution of temperature to KLa is higher than stirring at extreme-thermophilic conditions. This means that low stirring rate may be enough for mass transfer of H2 in extreme-thermophilic MCF, such as 150 rpm. Actually, the low stirring rates can also reduce the overall energy consumption of MCF system. However, it should be noted that the suggestion is estimated from only two stirring rates and still needs further exploration.

4.1. Hydrogen supersaturation in extreme-thermophilic (70 °C) MCF

4.2. Evaluation of gas flushing/sparing rate in MCF

In this study, hydrogen supersaturation was significant when PH2 was low (Fig. 3). It is the direct proof of hydrogen supersaturation existed in extreme-thermophilic MCF for the first time. As mentioned above, KLa can be used to study the relationship

Gas flushing or sparging has been a common method used to alter the relative amounts of metabolites in MCF, especially to improve YH2 [9,10,24,25]. In this study, the degree of hydrogen supersaturation increased dramatically under high N2 flushing or

218

Y. Zhang et al. / Applied Energy 109 (2013) 213–219

Table 2 Comparison of N2 sparging during fermentative H2 production in continuous MCF. N2 sparging rate, FN2 (mL/min/L)

H2 production rate, FH2 (mL/min/L)

H2 yield, YH2 (mol H2/mol hexose_added)

Yield increase (%)

FN2/ FH2

Kim et al. [9]a

0 20 40 60 80

2.4 2.8 3.0 3.0 3.0

0.75 0.87 0.91 0.93 0.91

0 16 21 24 21

0 7.1 13.6 19.9 27.1

Kraemer and Bagley [25]b

0 4 12 24 51 86

2.5 3.8 5.2 5.4 5.5 5.5

0.96 1.47 2.00 2.10 2.13 2.12

0 53 108 119 122 120

0 1.0 2.3 4.4 9.2 15.7

0 0.3 1.1 3.3 7.2 17.9

0.3 0.4 0.5 0.3 0.3 0.3

0.66 0.76 1.11 0.66 0.57 0.46

0 15 68 0.3 14c 30c

0 0.9 2.2 10.4 22.4 61.7

Study

This study

Gas sparging can improve the hydrogen yield in MCF and the optimal ratio of gas sparging rate vs H2 production rate was suggested as 2–10. Acknowledgements

a The H2 production rates were calculated from the organic loading rate and YH2 in reference [9]. b The H2 production rates were estimated from Fig. 2 of reference [25]. c The negative values mean that the YH2 was reduced.

sparging rates, which resulted in the insignificant change of metabolites. It seems that the optimum gas flushing or sparging rates applied in fermentative processes need to be readjusted. Table 2 lists literature values of N2 sparging rates used to improve YH2. It was found that YH2 was not always increased with the increasing N2 sparging rates, while it was reduced under high sparging rates. This indicates that the optimal sparging rate to improve YH2 might be lower than those usually used in the literature. Kraemer and Bagley [33] also found that the relationship between the amount of sparging and the increase of YH2 was not obvious. The existence of hydrogen supersaturation under high gas sparging rate might be responsible for this. Besides, high gas sparging rates bring some adverse effects, such as dilution of gases, inhibition of microbial activity and waste of energy [10,33]. Therefore, Table 2 suggested that the optimal ratio of gas sparging rate vs H2 production rate was 2–10. Moreover, the minor ratio could reduce the energy required for MCF system. It should be noted that the given ratio is only estimated from some studies [9,25], and it still requires further research. In addition, gas flushing is a relatively moderate method to deal with microorganism stress compared to gas sparging [10], and it may be more suitable to improve YH2 at high gas flow rate. Recently, methane (as natural gas and biogas) was suggested as the flushing or sparging gas in H2 production MCF [10]. It might be a more cost-effective gas than nitrogen since methane is not required to be removed from the gas mixture for fuel purpose.

5. Conclusions Hydrogen supersaturation of extreme-thermophilic MCF at pH 8.0 was investigated using online measurements of MIMS. It was found that hydrogen supersaturation was insignificant when PH2 was over 0.30 atm. However, RH2 increased dramatically when further reducing PH2 by either N2 sparging or N2 flushing. The distribution change of metabolite was negligible under low PH2 due to hydrogen supersaturation, which illustrated that it was more relevant to H2aq rather than PH2. KLa calculation showed that hydrogen supersaturation was likely inevitable, as the required KLa values were orders of magnitude higher than the experimental ones.

The authors would like to acknowledge the financial support from Natural Science Foundation of China (50978244), HundredTalent Program of CAS and the Fundamental Research Funds for the Central Universities (Grant No. WK2060190007). Appendix A. Supplementary material Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.apenergy.2013. 04.019. References [1] Srirangan K, Akawi L, Moo-Young M, Chou CP. Towards sustainable production of clean energy carriers from biomass resources. Appl Energy 2012;100: 172–86. [2] Kleerebezem R, van Loosdrecht MCM. Mixed culture biotechnology for bioenergy production. Curr Opin Biotechnol 2007;18:207–12. [3] Angenent LT, Karim K, Al-Dahhan MH, Domiguez-Espinosa R. Production of bioenergy and biochemicals from industrial and agricultural wastewater. Trends Biotechnol 2004;22:477–85. [4] Hallenbeck PC, Ghosh D. Advances in fermentative biohydrogen production: the way forward? Trends Biotechnol 2009;27:287–97. [5] Luo G, Xie L, Zou ZH, Zhou Q, Wang JY. Fermentative hydrogen production from cassava stillage by mixed anaerobic microflora: effects of temperature and pH. Appl Energy 2010;87:3710–7. [6] Chu C-Y, Sen B, Lay C-H, Lin Y-C, Lin C-Y. Direct fermentation of sweet potato to produce maximal hydrogen and ethanol. Appl Energy 2012;100:10–8. [7] Rodriguez J, Kleerebezem R, Lema JM, van Loosdrecht MCM. Modeling product formation in anaerobic mixed culture fermentations. Biotechnol Bioeng 2006;93:592–606. [8] Li CL, Fang HHP. Fermentative hydrogen production from wastewater and solid wastes by mixed cultures. Crit Rev Environ Sci Technol 2007;37:1–39. [9] Kim D, Han S, Kim S, Shin H. Effect of gas sparging on continuous fermentative hydrogen production. Int J Hydrogen Energy 2006;31:2158–69. [10] Bastidas-Oyanedel JR, Mohd-Zaki Z, Zeng RJ, Bernet N, Pratt S, Steyer JP, et al. Gas controlled hydrogen fermentation. Bioresour Technol 2012;110:503–9. [11] Bielen AAM, Verhaart MRA, VanFossen AL, Blumer-Schuette SE, Stams AJM, van der Oost J, et al. A thermophile under pressure: transcriptional analysis of the response of Caldicellulosiruptor saccharolyticus to different H2 partial pressures. Int J Hydrogen Energy 2013;38:1837–49. [12] Ljunggren M, Willquist K, Zacchi G, van Niel EW. A kinetic model for quantitative evaluation of the effect of hydrogen and osmolarity on hydrogen production by Caldicellulosiruptor saccharolyticus. Biotechnol Biofuels 2011;4:31. [13] Kraemer JT, Bagley DM. Supersaturation of dissolved H2 and CO2 during fermentative hydrogen production with N2 sparging. Biotechnol Lett 2006;28: 1485–91. [14] Zhang F, Zhang Y, Chen M, Zeng RJ. Hydrogen supersaturation in thermophilic mixed culture fermentation. Int J Hydrogen Energy 2012;37:17809–16. [15] Kraemer JT, Bagley DM. Measurement of H2 consumption and its role in continuous fermentative hydrogen production. Water Sci Technol 2008;57: 681–5. [16] Abreu AA, Karakashev D, Angelidaki I, Sousa DZ, Alves MM. Biohydrogen production from arabinose and glucose using extreme thermophilic anaerobic mixed cultures. Biotechnol Biofuels 2012;5:6. [17] Kongjan P, O-Thong S, Kotay M, Min B, Angelidaki I. Biohydrogen production from wheat straw hydrolysate by dark fermentation using extreme thermophilic mixed culture. Biotechnol Bioeng 2010;105:899–908. [18] Sigurbjornsdottir MA, Orlygsson J. Combined hydrogen and ethanol production from sugars and lignocellulosic biomass by Thermoanaerobacterium AK54, isolated from hot spring. Appl Energy 2012;97: 785–91. [19] Treybal RE. Mass transfer operations. New York: McGraw Hill; 1987. [20] Temudo MF, Kleerebezem R, van Loosdrecht M. Influence of the pH on (open) mixed culture fermentation of glucose: a chemostat study. Biotechnol Bioeng 2007;98:69–79. [21] Fang HHP, Liu H. Effect of pH on hydrogen production from glucose by a mixed culture. Bioresour Technol 2002;82:87–93. [22] Nissilä ME, Li Y-C, Wu S-Y, Lin C-Y, Puhakka JA. Hydrogenic and methanogenic fermentation of birch and conifer pulps. Appl Energy 2012;100:58–65. [23] Junghare M, Subudhi S, Lal B. Improvement of hydrogen production under decreased partial pressure by newly isolated alkaline tolerant anaerobe, Clostridium butyricum TM-9A: optimization of process parameters. Int J Hydrogen Energy 2012;37:3160–8.

Y. Zhang et al. / Applied Energy 109 (2013) 213–219 [24] Mizuno O, Dinsdale R, Hawkes FR, Hawkes DL, Noike T. Enhancement of hydrogen production from glucose by nitrogen gas sparging. Bioresour Technol 2000;73:59–65. [25] Kraemer JT, Bagley DM. Optimisation and design of nitrogen-sparged fermentative hydrogen production bioreactors. Int J Hydrogen Energy 2008; 33:6558–65. [26] Pauss A, Andre G, Perrier M, Guiot SR. Liquid-to-gas mass transfer in anaerobic processes: inevitable transfer limitations of methane and hydrogen in the biomethanation process. Appl Environ Microbiol 1990;56:1636–44. [27] Bastidas-Oyanedel JR, Mohd-Zaki Z, Pratt S, Steyer JP, Batstone DJ. Development of membrane inlet mass spectrometry for examination of fermentation processes. Talanta 2010;83:482–92. [28] Dubois M, Gilles KA, Hamilton JK, Rebers PA, Smith F. Colorimetric method for determination of sugars and related substances. Anal Chem 1956;28:350–6.

219

[29] APHA. Standard methods for the examination of water and wastewater. 19th ed. New York: American Public Health Association; 1995. [30] Green DW, Perry RH. Perry’s chemical engineers’ handbook. 8th ed. McGrawHill; 2008. [31] Vantriet K. Review of measuring methods and results in nonviscous gas–liquid mass transfer in stirred vessels. Ind Eng Chem Process Des Dev 1979;18: 357–64. [32] Lamed R, Zeikus JG. Ethanol production by thermophilic bacteria: relationship between fermentation product yields of and catabolic enzyme activities in Clostridium thermocellum and Thermoanaerobium brockii. J Bacteriol 1980;144: 569–78. [33] Kraemer JT, Bagley DM. Improving the yield from fermentative hydrogen production. Biotechnol Lett 2007;29:685–95.