Article pubs.acs.org/est
Bioreactor Performance and Quantitative Analysis of Methanogenic and Bacterial Community Dynamics in Microbial Electrolysis Cells during Large Temperature Fluctuations Lu Lu, Defeng Xing,* and Nanqi Ren* State Key Laboratory of Urban Water Resource and Environment, School of Municipal and Environmental Engineering, Harbin Institute of Technology, Harbin 150090, China S Supporting Information *
ABSTRACT: The use of microbial electrolysis cells (MECs) for H2 production generally finds H2 sink by undesirable methanogenesis at mesophilic temperatures. Previously reported approaches failed to effectively inhibit methanogenesis without the addition of nongreen chemical inhibitors. Here, we demonstrated that the CH4 production and the number of methanogens in single-chamber MECs could be restricted steadily to a negligible level by continuously operating reactors at the relatively low temperature of 15 °C. This resulted in a H2 yield and production rate comparable to those obtained at 30 °C with less CH4 production (CH4% < 1%). However, this operation at 15 °C should be taken from the initial stage of anodic biofilm formation, when the methanogenic community has not yet been established sufficiently. Maintaining MECs operating at 20 °C was not effective for controlling methanogenesis. The varying degrees of methanogenesis observed in MECs at 30 °C could be completely inhibited at 4 and 9 °C, and the total number of methanogens (mainly hydrogenotrophic methanogens) could be reduced by 68−91% during 32−55 days of operation at the low temperatures. However, methanogens cannot be eliminated completely at these temperatures. After the temperature is returned to 30 °C, the CH4 production and the number of total methanogens can rapidly rise to the prior levels. Analysis of bacterial communities using 454 pyrosequencing showed that changes in temperature had no a substantial impact on composition of dominant electricity-producing bacteria (Geobacter). The results of our study provide more information toward understanding the temperature-dependent control of methanogenesis in MECs.
■
concentrations,7 than in two-chamber MECs with separating membranes (e.g., cation and anion exchange membranes). However, the use of such membranes will lead to large pH and ohmic energy losses in MECs, and no membranes in current use can absolutely inhibit H2 diffusion between the two electrodes. Because the production of 1 mol of CH4 consumes 4 mol of H2 in hydrogenotrophic methanogenesis (4H2 + CO2 = CH4 + H2O), CH4 generation tremendously decreases the H2 yield. In previous MEC studies, several methods have been used to reduce CH4 production, including the use of methanogen inhibitors (e.g., BES and lumazine),8 reactor air exposure,9,10 short hydraulic retention times,4,11 low pH,8,10 release of H2 rapidly to the gas phase,3,12 temperature shock,8 carbonatelimited conditions,11 and indirectly increased applied voltage (through altering the anode potential or the reaction time).13 To date, these approaches have failed to inhibit methanogenesis effectively without the addition of nongreen chemical
INTRODUCTION The microbial electrolysis cell has quickly become the most attractive technology for renewable H2 production over the past several years because of its high H2 yield and impressive H2 production rate as compared to photobiological or darkfermentative bio-H2 production methods.1 Microbial electrolysis cells (MECs) and microbial fuel cells (MFCs) share the same anodic biocatalysts, called exoelectrogens, which oxidize organic substrates and transfer electrons extracellularly to the anode. Under a small applied external potential, electrons are circuited to the cathode, where they reduce H+ to generate H2.2 One major problem facing electrohydrogenesis in MECs using mixed cultures is the associated CH4 production. The responsible methanogens are mainly attached to the anode and to a lesser degree to the cathode and walls of the reactor.3 Most MEC studies showed that hydrogenotrophic methanogens are primarily responsible for the methane generation3−6 because acetoclastic methanogens are generally out-competed by acetate-oxidizing exoelectrogens in low concentrations of acetate. This CH4-producing H2 sink is more obvious in single-chamber MECs without any separators, in which the H2 produced at the cathode is more easily scavenged by hydrogenotrophic methanogens on the anode at nanomolar © 2012 American Chemical Society
Received: Revised: Accepted: Published: 6874
March 6, 2012 May 18, 2012 May 21, 2012 May 21, 2012 dx.doi.org/10.1021/es300860a | Environ. Sci. Technol. 2012, 46, 6874−6881
Environmental Science & Technology
Article
MECs were fed 2 g/L (24.4 mM) acetate in a 50 mM nutrient phosphate buffer solution (NPBS)9 in batch mode. Prior to being fed into reactors, the medium was sparged with nitrogen gas (99.999%) for 15 min to remove any oxygen. A fixed voltage of 0.5 V was applied to the MECs with a programmable power source (3645A, Array, Inc.). Two MECs (MEC4-1 and MEC4-2) were first operated at 30 °C until detectable levels of CH4 production were present and then cooled for a period of operation at 4 °C. Finally, the temperature was restored to 30 °C. Another two reactors (MEC9-1 and MEC9-2) were operated in the same manner as outlined above except for having a low temperature of 9 °C. The MEC15 and MEC20 groups were operated throughout at 15 and 20 °C, respectively, using duplicate reactors in each group. To cause the reactors operated at an initial temperature of 30 °C to have different extents of methanogenesis, MEC4-2 was subjected to anode aeration during each batch interval, and MEC9-2 was operated for a comparatively short time. Analyses and Calculations. The gas produced by the MECs was collected in a gas bag (0.1 L Cali-5-Bond, Calibrate, Inc.), and the total volume was measured using a glass syringe. Gas composition was analyzed using a gas chromatograph [4890D, Agilent, Inc.; details of analysis are in the Supporting Information (SI)]. The chemical oxygen demand (COD) of liquid samples was measured according to a standard method.21 The current in the circuits was determined by measuring the voltage over a high-precision resistor (10 Ω) using a multimeter/data acquisition system (model 2700, Keithley, Inc.) at 10 min intervals. Hydrogen production rate, Q (m3 H2/m3 reactor/day); volumetric current density, IV (A/m3); hydrogen yield, YH2 (mol H2/mol acetate); Coulombic efficiency, CE, calculated on the basis of COD; cathodic hydrogen recovery, rcat; and energy recovery relative to the electrical input, ηE, were used to evaluate the performance of MECs as previously described.22 CH4 yield (mol CH4/mol acetate) was converted to equivalent YH2 using the conversion factor of 4 mol H2/mol CH4. Archaeal Clone Library and QPCR. At different stages of each MEC operation, small amounts of graphite fibers were cut from the anodes using sterile scissors for DNA extraction. According to our tests, slight decreases of the anodic surface area do not affect the current generation in MECs. This may be due to the high anodic surface area of our reactor (8589 m2/m3 liquid volume), which is far in excess of that needed to meet the requirements of our system. Total genomic DNA was extracted using a PowerSoil DNA Isolation Kit (Mo Bio Laboratories, Inc., Carlsbad, CA) according to the manufacturer’s instructions. The DNA samples extracted from MECs with detectable methanogenesis at 30 °C were pooled equally. We used this mixed DNA to construct a 16S rRNA gene archaeal clone library according to the method described in our previous study14 with a pair of universal archaeal primers 787F/1059R (SI, Table S1). The representative sequences in each operational taxonomic unit (OTU) were deposited in GenBank with accession numbers JQ319055, JQ319056, JQ319057, JQ319058, JQ319059, JQ319060, JQ319061, JQ319062, JQ319063, JQ319064. On the basis of results from clone library, QPCR based on TaqMan probes was used to quantify total bacteria, archaea, two hydrogenotrophic methanogen orders [Methanobacteriales (MBT) and Methanomicrobiales (MMB)], and two acetoclastic
inhibitors. We previously observed no methane production and no significant drop of H2 yield in MECs operated at low temperatures (4 or 9 °C), which broke through a bottleneck in which fermentative H2 production is depressed under psychrotolerant conditions.14 However, it is unknown whether these low temperatures could eliminate methanogens completely or if the methanogenesis can recover when these psychrotolerant MECs are returned to operation at mesophilic temperatures, and we have not examined the methanogenic community carefully during this process. Chae et al. have shown that a temperature shock from 30 to 20 °C only caused an unremarkable decrease in CH4 production, and the CH4 concentration returned to the original level after the recovery of temperature.8 However, in this study, the inhibition temperature was still relatively high, and the shock time was very short (three batch cycles). To date, little is known about the dynamics of CH4 production and the methanogenic or bacterial community in MECs that experience wide temperature fluctuations during the long-term operation. On the other hand, it is not economical or practical to inhibit methanogenesis using such low temperatures (4 or 9 °C). We need to examine whether the CH4 production could be reduced in MECs at a relatively higher temperature, such as around 15 °C, which is a common critical value to make a distinction between psychrophilic and psychrotolerant microorganisms.15,16 Another limitation of previous MEC studies of CH 4 inhibition is that any inhibitory interventions were only adopted after the methanogens had become dominant or anodophilic biofilms had been fully established. The results from Freguia et al. showed that the suppression of methanogenesis by periodic aeration was only effective in a newly forming biofilms, and not in established ones.17 It is very difficult to reduce methanogenesis after it becomes dominant in MECs.18,19 Therefore, it is necessary to examine if an inhibition of methanogenic dominance in the first place is helpful for controlling the methanogenesis. To accomplish all of goals above, MECs operated at 30 °C with varying degrees of methanogenesis were switched to operate at 4 or 9 °C for certain periods of time. The reactor temperatures were subsequently returned to 30 °C. Other MECs without methanogenesis were operated continuously at 15 or 20 °C. H2 and CH4 production in all of the reactors were examined. The composition of the archaeal community was first analyzed using a clone library based on the 16S rRNA gene. Quantitative real-time polymerase chain reaction (QPCR) was used to examine the dynamic quantities of the bacteria, archaea, and methanogens (hydrogenotrophic and acetoclastic) during the changes of temperature in the MECs. Bacterial community dynamics was also examined by 454 pyrosequencing based on the 16S rRNA gene.
■
MATERIALS AND METHODS MEC Construction and Operation. Single-chamber, membraneless MECs with graphite brush anodes and carbon cloth air cathodes were constructed as previously described20 with a liquid volume of 26 mL. All MEC anodes were initially enriched in similarly constructed MFCs with cathodes exposed to air and inoculated with an effluent from acetate-fed MFCs at 25 °C. Enrichment was conducted at 30 °C with acetate (1 g/ L) as a substrate. After 5−6 days, all MFCs exhibited reproducible maximum voltages with an average of 586 ± 5 mV (over a 1 kΩ resistor). Then, the anodes were transferred to MECs for H2 production. 6875
dx.doi.org/10.1021/es300860a | Environ. Sci. Technol. 2012, 46, 6874−6881
Environmental Science & Technology
Article
methanogen families [Methanosarcginaceae (MSC) and Methanosaetaceae (MST)]. Details of QPCR tests and preparation of standard curves are given in the Supporting Information. Analysis of Bacterial Community Structures Based on Pyrosequencing. High-throughput 454 GS-FLX pyrosequencing of the 16S rRNA gene was conducted according to standard protocols.23 Raw sequencing data were deposited to the NCBI Sequence Read Archive (SRA) with accession no. SRA052612. We clustered qualified sequences into operational taxonomic units (OTUs) as well as produced rarefaction curves, species richness estimators (Chao1 and ACE), and a Shannon diversity index using the MOTHUR program.24 Sequences were assigned to different taxonomic groups using ̈ Bayesian rRNA classifier.25 The phylogenetic the RDP naive comparison of bacterial communities from different samples was conducted using Fast UniFrac analysis26 based on sequences. We offer details of pyrosequencing and statistical analysis in the Supporting Information.
■
RESULTS Previous Exploration of Methanogen Phylogenetic Groups. The archaeal 16S rRNA gene clone library was previously constructed to explore methanogen phylogenetic groups that may be present in MECs (Figure 1). The DNAs
Figure 2. Hydrogen and methane yields in MECs during temperature fluctuations between 4 and 30 °C (MEC4-1 and MEC4-2) as well as 9 and 30 °C (MEC9-1 and MEC9-2). MEC15 and MEC20 are two groups of MEC (duplicate reactors in each group) operated throughout at 15 and 20 °C, respectively. The methane yield (mol CH4/mol acetate) is converted to an equivalent hydrogen yield (mol H2/mol acetate) using a conversion factor of 4 mol H2/mol CH4. The legend “Sum” represents the total of hydrogen and methane yields. The arrow indicates the sampling point for DNA extraction.
Figure 1. Phylogenetic composition of the archaeal clone library based on the 16S rRNA gene. Percentages of nucleotide similarity are between 99% and 100%.
extracted from four reactors (MEC4-1, MEC4-2, MEC9-1, and MEC9-2) at the DNA-sampling point A (Figure 2) were pooled and used for library construction. Before the DNA extraction, the four MECs had been operated at 30 °C for 25− 54 days. Varying degrees of methanogenesis was detected in these reactors. The rarefaction curve for the 16S rRNA gene clone library was approaching a plateau (SI, Figure S1), suggesting adequate sampling for the assessment of the dominant methanogen phylogenetic groups. The most abundant sequence was related to Methanobrevibacter arboriphilicus and accounted for 68.6% of the total sequences, followed by Methanobacterium congolense (15.7%). Both of these as well as Methanobacterium beijingense (1.0%) belong to the MBT order. Methanocorpusculum sinense accounted for 1.5% of the total composition and belongs to the MMB order. The orders MBT and MMB are hydrogenotrophic methanogens capable of metabolizing hydrogen or formate. Other abundant sequences were most closely related to Methanosaeta concilii (4.1%) of the MST family. M. concilii and Methanosarcina siciliae (0.5%) of the MSC family belong to the acetoclastic (acetate-utilizing) methanogens. The results from the library were used as a basis for choosing appropriate QPCR probe sets.
Gas Production and Methanogenic Community Dynamics during Temperature Fluctuations between 4 and 30 °C. Reactors MEC4-1 and MEC4-2 were operated at 30 °C for H2 production for 54 and 50 days, respectively (Figure 2). At the end of this time, CH4 production accounted for 78% (MEC4-1) and 35% (MEC4-2) of the expected H2 yields from consumed substrate assuming no CH4 was produced, thus resulting in relatively low H2 yields of 0.55 mol H2/mol acetate (MEC4-1) and 1.69 mol H2/mol acetate (MEC4-2). The lower CH4 production in MEC4-2 could be attributed to the inhibition of methanogenesis by aerating the anode9,10,17 in this reactor for 15 min during each batch interval in this stage, while the MEC4-1 anode was always maintained under anaerobic conditions. However, aeration of the anode cannot prevent the CH4 production completely, even when performed before an anodic biofilm has been fully established. Analysis of the microbial composition at this time showed that the copy number of total methanogenic 16S rRNA genes in MEC4-1 (2.64 × 103 copies/ng DNA) was larger than that in MEC4-2 (1.62 × 103 copies/ng DNA) (Figure 3), consistent 6876
dx.doi.org/10.1021/es300860a | Environ. Sci. Technol. 2012, 46, 6874−6881
Environmental Science & Technology
Article
found only after one batch cycle (Figure 2). After 18 days, CH4 production in MEC4-1 returned to the initial level at 30 °C with 75% of total H2 being transferred to CH4, resulting in a low H2 yield of 0.64 mol H2/mol acetate. In MEC4-2, the methanogenesis rebounded to even higher levels than before being switched to a low temperature, with 79% of the total H2 being transferred to CH4, resulting in a H2 yield of 0.55 mol H2/mol acetate. The copy number of total methanogenic 16S rRNA genes in both reactors rebounded to 85% of that at the end of 30 °C operation (Figure 3). It is interesting that CH4 production in MEC4-2 was increased after returning to 30 °C, despite there being no rise in the total number of methanogens; this suggests that the activity of methanogens for CH4 catalysis was enhanced when the temperature was restored. This may be a consequence of the methanogens’ reaction to adverse environments. Gas Production and Methanogenic Community Dynamics during Temperature Fluctuations between 9 and 30 °C. The dynamics of gas production and the methanogenic community during temperature fluctuation between 9 and 30 °C in reactors MEC9-1 and MEC9-2 were generally similar to those in MEC4-1 and MEC4-2. At the end of the first operation at 30 °C, MEC9-1 produced more CH4 (65% of total H2 yields) than did MEC9-2 (28% of total H2 yields) (Figure 2). The lower CH4 production in MEC9-2 was due to a shorter time of operation (25 days) as compared to MEC9-1 (49 days) rather than the use of aeration to inhibit methanogenesis. Corresponding to the performances of the reactors, MEC9-1 contained more methanogens than did MEC9-2 (Figure 4). MBT was also the most abundant population detected among the methanogens. Operations at 9 °C also completely inhibited methanogenesis (Figure 2) and significantly decreased the copy number of total methanogenic 16S rRNA genes in MEC9-1 (76% reduction) and MEC9-2 (91% reduction) significantly (Figure 4).
Figure 3. Quantitative changes in bacterial, archaeal, and methanogenic 16S rRNA gene copy concentrations in the MEC4-1 and MEC42 reactors. The codes on the x-axis represent the sampling points for DNA extraction in Figure 2. Error bars represent standard deviations from triplicate QPCR amplifications.
with the observed CH4 production in the two reactors. These copy numbers of methanogenic 16S rRNA genes accounted for 95−96% of total archaeal 16S rRNA genes in each reactor, suggesting that our QPCR probes adequately covered most methanogen populations in the microbial community. The compositions of methanogens in both reactors were similar. MBT were the most abundant members and accounted for 97− 98% of total methanogens, followed by a small number of MST (1.6−2.2%). The quantities of MMB and MSC were negligibly small. These fractions of methanogen groups from the QPCR data were consistent with those of the previous clone library. After MECs were switched to operate at 4 °C, CH4 production was abruptly reduced (Figure 2). After several batches, methanogenesis was inhibited completely. Correspondingly, the H2 yield rapidly recovered to an average of 2.14 ± 0.18 mol H2/mol acetate in the two reactors. This yield was lower than the 3.01 ± 0.35 mol H2/mol acetate observed in several initial batches at 30 °C, in which methane was undetected. The decreased H2 yield at 4 °C was mainly due to a reduction of cathodic hydrogen recovery (4 °C, 63 ± 8%; 30 °C, 89 ± 6%) caused by unknown H2 or electrons sinks during a longer reaction time other than methanogenesis. A higher applied voltage could increase the H2 production rate and therefore recover H2 as quickly as possible and thus reduce the chance of H2 sink. It should be noted that the copy number of total methanogenic 16S rRNA genes decreased by 68% (MEC4-1) and 80% (MEC4-2) after 53−55 days of operation at 4 °C (Figure 3). This indicated that a low temperature inhibits the methanogenesis not only through restraining the activity of methanogens but also by reducing their number. The lower the initial number of methanogens in the reactor (MEC42), the more their number was reduced. Whether longer operation at low temperatures could completely eliminate methanogens is a topic for future investigation. Finally, two reactors were returned to operate at 30 °C with no measures taken to inhibit their methanogenesis. CH4 was
Figure 4. Quantitative changes in bacterial, archaeal, and methanogenic 16S rRNA gene copy concentrations in the MEC9-1 and MEC92 reactors. The codes on the x-axis represent the sampling points for DNA extraction in Figure 2. Error bars represent standard deviations from triplicate QPCR amplifications. 6877
dx.doi.org/10.1021/es300860a | Environ. Sci. Technol. 2012, 46, 6874−6881
Environmental Science & Technology
Article
Table 1. Performances of MECs at Various Temperatures I (A/m3) temp (°C)
MEC4-1a
30 4 30 ↑e
152 ± 5 53 ± 4 146 ± 6
temp (°C)
MEC4-1a
30 4 30 ↑
I (A/m3)
MEC4-2a
155 ± 9 50 ± 3 144 ± 6 CE (%) MEC4-2a
84 ± 8 90 ± 11 88 ± 8 88 ± 3 86 ± 4 83 ± 5 Q (m3/m3/d)
temp (°C)
MEC4-1a
30 30 4 30 ↑
0.24 ± 0.02c 1.17 ± 0.05d 0.38 ± 0.02 0.22 ± 0.01c
MEC4-2a 0.55 ± 0.03c 1.11 ± 0.07d 0.36 ± 0.02 0.25 ± 0.01c ηE (%)
temp (°C)
MEC9-1a
30 9 30 ↑
159 ± 8 71 ± 6 143 ± 5
temp (°C)
MEC9-1a
30 9 30 ↑
temp (°C) 30 30 9 30 ↑
MEC9-2a
158 ± 4 71 ± 3 142 ± 7 CE (%) MEC9-2a
89 ± 7 96 ± 5 84 ± 8 101 ± 8 85 ± 5 94 ± 3 Q (m3/m3/d) MEC9-1a
MEC9-2a
0.39 ± 0.01c 0.86 ± 0.05c d 1.19 ± 0.06 1.13 ± 0.05d 0.52 ± 0.03 0.49 ± 0.01 0.40 ± 0.03c 0.63 ± 0.04c ηE (%)
temp (°C)
I (A/m3): MEC15b
temp (°C)
I (A/m3): MEC20b
15
95 ± 3
20
114 ± 6
temp (°C)
CE (%): MEC15b
temp (°C)
CE (%): MEC20b
15
87 ± 4
20
80 ± 4
temp (°C)
Q (m3/m3/d): MEC15b
temp (°C)
Q (m3/m3/d): MEC20b
15
0.72 ± 0.06
20
0.47 ± 0.02c
20
0.92 ± 0.03d
temp (°C)
MEC4-1a
MEC4-2a
temp (°C)
MEC9-1a
MEC9-2a
temp (°C)
ηE (%): MEC15b
temp (°C)
ηE (%): MEC20b
30 30 4 30 ↑
53 ± 4c 216 ± 10d 209 ± 16 44 ± 7c
129 ± 3c 219 ± 10d 206 ± 5 53 ± 2c
30 30 9 30 ↑
91 ± 2c 216 ± 8d 213 ± 8 82 ± 5c
163 ± 6c 215 ± 5d 206 ± 2 144 ± 5c
15
211 ± 7
20
106 ± 5c
20
216 ± 9d
Reactors with temperature fluctuation between 30 and 4 °C (MEC4-1 and MEC4-2) and between 30 and 9 °C (MEC9-1 and MEC9-2). bMEC15 and MEC20 represent reactors at constant 15 and 20 °C, respectively. cAveraged values obtained from batches with maximum of CH4 concentrations. dAveraged values obtained from batches with CH4 concentrations
