University of Southern California Dissertations and Theses

Bioelectrochemical treatment of anaerobic process effluents: mitigation of dissolved methane and sulfide

(USC Thesis Other)

Bioelectrochemical treatment of anaerobic process effluents: mitigation of dissolved methane and sulfide

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content 1

Bioelectrochemical treatment of 1
anaerobic process effluents: 2
mitigation of dissolved methane and 3
sulfide 4
5
6
by 7
8
Siming Chen 9
10
11
A Dissertation Presented to the 12
FACULTY OF THE USC GRADUATE SCHOOL 13
In Partial Fulfillment of the 14
Requirements for the Degree 15
DOCTOR OF PHILOSOPHY 16
(Environmental Engineering) 17
18
19
20
21
22
23
24
August 2019 25
26
2

Abstract 27
28
Given the increasing emphasis on resource recovery (e.g., water, energy, and nutrients) from 29
domestic wastewater, high-rate anaerobic processes are drawing attention due to their ability 30
to produce methane-rich biogas, significantly reduce sludge production, and produce an effluent 31
of comparable quality to activated sludge processes. However, anaerobic effluents contain 32
dissolved methane and other constituents (e.g., sulfides and nutrients) that require 33
management prior to discharge or reuse. In particular, dissolved methane poses a risk as a 34
potent greenhouse gas if emitted to the atmosphere. Bioelectrochemical systems such as 35
microbial fuel cells (MFCs) are an attractive technology to recover energy from dissolved 36
methane and sulfide. Specifically, air-cathode MFCs require minimal energy demands due to 37
their ability to passively provide oxygen via diffusion across an atmosphere-exposed cathode. In 38
this dissertation, air-cathode MFCs were investigated to treat dissolved methane and sulfide 39
across a range of operational conditions. First, two replicate bench-scale air-cathode MFCs were 40
operated on a synthetic anaerobic effluent containing dissolved methane as the sole organic 41
electron donor. Up to 85% dissolved methane removal was achieved resulting in 0.5 to 0.6 V and 42
a maximum Coulombic efficiency of 17.7%. Longer hydraulic retention time yielded higher 43
voltage generation and dissolved methane removal due to increased oxygen diffusion. A 44
methanotroph-exoelectrogen interaction was proposed based on distinct colonization of 45
methanotrophs and Geobacter on cathode and anode biofilms, respectively. Sequential addition 46
of likely methanotrophic intermediates suggested that formate served as the electron shuttle 47
between the two populations. Next, the replicate MFCs were subjected to operational 48
temperatures of 25, 20, 15, 10 and 5° C to investigate performance across seasonal temperature 49
3

variations common to temperate climates. Though voltage abruptly decreased at and below 50
10° C, stable dissolved methane removal was achieved at all operational temperatures. High- 51
throughput sequencing revealed methantrophs (e.g., Methylobacillus and Methylomonas) and 52
exoelectrogens (e.g., Geobacter spp. and Ferribacterium) at cathode and anode biofilms, 53
respectively. Spearman rank correlation analysis suggested that fermentative bacteria may play 54
a role mediating interactions between methanotrophs and exoelectrogens. Anode microbial 55
diversity was found to strongly correlate with system performance. Finally, one methane-driven 56
MFC was subjected to sulfide addition at concentrations of 1, 5, 10 and 20 mg/L. Even at 20 57
mg/L, sulfide addition did not negatively impact dissolved methane removal or voltage 58
production, and sulfide removal was 46.8 ± 8.9%. Future research is necessary to optimize 59
sulfide removal and evaluate long-term performance at the pilot-scale on actual anaerobic 60
effluents. 61
62
63
64
65
66
67
68
69
70
4

Acknowledgements 71
72
I would like to acknowledge the following financial sources for their support on this research: 73
• China Scholarship Council 74
• Department of Civil and Environmental Engineering, University of Southern California 75
I would like to thank the Sonny Astani Department of Civil and Environmental Engineering staff 76
and research administrators. I want to thank all of the student assistants, Tong Wu, Pooja Sinha 77
and Qin Dong, for their help with reactor operation and maintenance. I wish to thank all the 78
colleagues and friends from the USC Water and Environmental Lab. Your constant support and 79
amazing friendships benefit me more than research. I would like to sincerely thank my 80
qualifying and defense committee, Drs. Amy Childress, Daniel McCurry, James Boedicker and 81
Kelly Sanders, for their contributions to this dissertation. I would like to especially thank my 82
advisor Dr. Adam Smith for his great guidance and advising in the past several years. This 83
research would have not been possible without his incredible mentoring. 84
At last, I would like to thank my family and friends for their great support. I am forever indebted 85
to you. 86
87
88
89
90
91
92
93
5

Table of Contents 94
Abstract ........................................................................................................................................... 2 95
Acknowledgements ......................................................................................................................... 4 96
List of Figures ................................................................................................................................... 8 97
List of Tables .................................................................................................................................. 11 98
Chapter 1. Introduction and Overview .......................................................................................... 12 99
1.1 Background .......................................................................................................................... 12 100
1.2 Overview of Dissertation ..................................................................................................... 14 101
Chapter 2. Revisiting greenhouse gas mitigation from conventional activated sludge and 102
anaerobic-based wastewater treatment systems ......................................................................... 15 103
Abstract ..................................................................................................................................... 15 104
2.1 Introduction ......................................................................................................................... 16 105
2.2 Anthropogenic GHG emissions from CAS WWTPs .............................................................. 18 106
2.2.1 Overview of GHG generation in wastewater treatment processes ............................. 18 107
2.2.2 Common findings of different quantification methods for plant-wide GHG emissions 108
............................................................................................................................................... 20 109
2.2.3 Specific direct GHG emission sources in CAS WWTPs .................................................. 22 110
2.2.3.3.1 Dissolved oxygen control ........................................................................................ 24 111
2.2.3.3.2 pH ........................................................................................................................... 26 112
2.2.3.3.3 Nitrite ..................................................................................................................... 27 113
2.2.3.3.4 Carbon source availability/COD:N ratio ................................................................. 28 114
2.2.3.3.5 Ammonium shock ................................................................................................... 29 115
2.2.4 Implications of N 2O emissions in CAS WWTPs ............................................................. 29 116
2.3 GHG emission management for mainstream anaerobic treatment ................................... 29 117
2.3.1 GHG emission sources in anaerobic bioreactors .......................................................... 30 118
2.3.2 Utilizing membrane contactors for dissolved CH 4 recovery ......................................... 32 119
2.3.3 Biological approaches ................................................................................................... 38 120
2.3.4 Implication of physical vs. biological systems for CH 4 mitigation ................................ 40 121
2.4 Future GHG management perspectives .............................................................................. 41 122
2.4.1 N 2O as an energy recovery oxidant .............................................................................. 41 123
2.4.2 Technologies for targeting CO 2 capture ....................................................................... 41 124
2.4.3 Methanotroph-based recovery of high-value end products ........................................ 42 125
2.4.4 Methane adsorbent-related technology ...................................................................... 42 126
6

2.5 Perspectives on the direct comparison of CAS- and anaerobic-based GHG emissions ...... 43 127
2.6 Conclusions .......................................................................................................................... 45 128
Chapter 3. Methane-driven microbial fuel cells recover energy and mitigate dissolved methane 129
emissions from anaerobic effluents .............................................................................................. 48 130
Abstract ..................................................................................................................................... 48 131
3.1 Introduction ......................................................................................................................... 48 132
3.2 Materials and methods ....................................................................................................... 53 133
3.2.1 MFC configurations ...................................................................................................... 53 134
3.2.2 Chemical assays ............................................................................................................ 55 135
3.2.3 Microbial community structure and activity ................................................................ 56 136
3.3 Results and discussion ......................................................................................................... 59 137
3.3.1 Preliminary findings suggest methane removal and voltage production driven by co- 138
culture of methanotrophs and Geobacter ............................................................................ 59 139
3.3.2 Dissolved methane removal efficiency strongly correlated with HRT ......................... 63 140
3.3.3 Geobacter and methanotroph activity was spatially distinct in MFCs ......................... 67 141
3.3.4 Geobacter scavenge methanotrophic metabolites enabling electron recovery from 142
methane ................................................................................................................................ 69 143
3.3.5 Methane-driven MFCs outcompete existing approaches for dissolved methane 144
management ......................................................................................................................... 72 145
3.4 Conclusions .......................................................................................................................... 75 146
Chapter 4. Performance and microbial ecology of methane-driven microbial fuel cells at 147
temperatures ranging from 25 to 5° C ........................................................................................... 76 148
Abstract ..................................................................................................................................... 76 149
4.1 Introduction ......................................................................................................................... 77 150
4.2 Materials and methods ....................................................................................................... 80 151
4.2.1 MFC configuration, operation, and monitoring ........................................................... 80 152
4.2.2 Sequencing and RT-qPCR .............................................................................................. 82 153
4.2.3 Bioinformatics .............................................................................................................. 83 154
4.3 Results and Discussion......................................................................................................... 84 155
4.3.1 Voltage production abruptly decreased at 10° C .......................................................... 84 156
4.3.2 Dissolved methane removal efficiency was relatively stable across operational 157
temperature .......................................................................................................................... 87 158
4.3.3 Distinct microbial communities were observed on the anode and cathode ............... 88 159
4.3.4 Network analysis revealed anode and cathode microbial community interactions .... 92 160
7

4.3.5 Voltage production and Coulombic efficiency correlated with microbial diversity on 161
the anode .............................................................................................................................. 95 162
4.4. Conclusions ......................................................................................................................... 96 163
Chapter 5. Impact of sulfide on methane-driven microbial fuel cells during treatment of 164
anaerobic effluents........................................................................................................................ 98 165
Abstract ..................................................................................................................................... 98 166
5.1 Introduction ......................................................................................................................... 99 167
5.2 Materials and Methods ..................................................................................................... 100 168
5.3 Results and Discussion....................................................................................................... 102 169
5.3.1 Dissolved methane removal remained high at all influent sulfide concentrations ... 102 170
5.3.2 Voltage production was stable across all influent sulfide concentrations ................. 105 171
5.3.3 Elemental surface characterization of the anode revealed elevated sulfur relative to 172
control ................................................................................................................................. 106 173
5.4 Conclusions ........................................................................................................................ 109 174
Chapter 6. Conclusions and Outlook ........................................................................................... 110 175
6.1 Overview ............................................................................................................................ 110 176
6.2 Dissolved methane accounts for the majority of the carbon footprint in mainstream 177
anaerobic-based processes ..................................................................................................... 110 178
6.3 Methanotrophs-Geobacter interaction drives methane-driven microbial fuel cells ........ 111 179
6.4 MFCs demonstrate stable dissolved methane removal at varying temperatures ............ 112 180
6.5 Sulfide did not negatively impact methane-driven MFC performance ............................. 112 181
6.6 Future research directions ................................................................................................ 113 182
Appendix ...................................................................................................................................... 116 183
A. Supplementary information .................................................................................................... 116 184
SI 1.0 Tables ............................................................................................................................. 116 185
SI 2.0 Figures ............................................................................................................................ 134 186
References ................................................................................................................................... 142 187
188
189
190
191
8

List of Figures 192
Figure 1. Direct greenhouse gases (CH 4 and N 2O) emissions from conventional wastewater 193
treatment plants employing anaerobic/anoxic/oxic activated sludge processes and anaerobic 194
sludge digestion. (ANA/ANX/AER, anaerobic/anoxic/oxic tanks) ................................................. 20 195
196
Figure 2. Five distinct N 2O generation pathways (NH 2OH oxidation with O 2, NH 2OH N-nitrosation 197
hybrid reaction, unstable decomposition of (NOH), nitrifier denitrification, incomplete 198
heterotrophic denitrification) along with nitrification (ammonia oxidizing bacteria (AOB) and 199
nitrite oxidizing bacteria (NOB)) and denitrification pathways. Key enzymes are identified along 200
each pathway. ............................................................................................................................... 26 201
202
Figure 3. (a) Dissolved methane concentration (mg/L) from anaerobic treatment processes 203
across different studies at varying temperature; (b). Dissolved methane supersaturation ratio 204
from anaerobic treatment processes across different studies at varying temperature; (c). 205
Proportion of dissolved methane over total methane production from anaerobic treatment 206
processes across different studies at varying temperature; ○ represents upflow anaerobic 207
blanket sludge bed reactor (UASB), ◊ represents expanded granular sludge bed reactor (EGSB), ∆ 208
represents anaerobic fluidized membrane bioreactor (AFMBR), □ represents anaerobic 209
membrane bioreactor (AnMBR); dash line represents dissolved methane derived from Henry’s 210
law, with 80% gaseous methane in the headspace ...................................................................... 32 211
212
Figure 4. 1st run while single-chamber MFCs were operated in continuous mode on methane 213
containing media (a) Preliminary data on voltage production (b) dissolved methane removal 214
efficiency and Coulombic efficiency (%) from replicate air-cathode single-chamber MFCs ......... 60 215
216
Figure 5. Relative abundance and relative activity based on 16S rRNA sequencing of single- 217
chamber MFCs anodic biomass sample identified to genus level................................................. 62 218
219
Figure 6. (a) Voltage production while single-chamber MFCs were operated in continuous mode 220
on methane containing medium under varying HRTs; (b) Dissolved methane removal efficiency 221
and Coulombic efficiency (%) while single-chamber MFCs were operated in continuous mode on 222
methane containing medium under varying HRTs ........................................................................ 64 223
224
Figure 7. (a) relative activity of Geobacter 16S rRNA gene copy number /16S rRNA gene copy 225
number and (b) relative activity of pmoA transcript copy number /16S rRNA gene copy number 226
at anode and cathode from single-chamber MFCs under 3 different HRTs ................................. 68 227
228
Figure 8. Dissolved methane consumption pathways/ mass balance in single-chamber MFCs 229
under 8 h HRTs .............................................................................................................................. 70 230
9

Figure 9. Voltage production over time with sequential addition of methanol, formaldehyde and 231
formate into single-chamber MFC operated in batch mode ........................................................ 72 232
233
Figure 10. (A) Influent dissolved methane concentration (bars) and theoretical dissolved 234
methane derived from Henry’s law (line) across different operational temperatures (25, 20, 15, 235
10 and 5° C). (B) Voltage production and Coulombic efficiency from continuous operation of 236
replicate air-cathode MFCs. (C) Dissolved methane removal efficiency (open markers) and 237
average absolute dissolved methane removal (filled markers). A negative correlation between 238
average absolute methane removal and temperature was observed with R
2
values equating to 239
0.8864 and 0.7749 for MFC A and B, respectively (solid lines). Dashed lines represent the linear 240
trend line between dissolved methane removal efficiency and temperature (R
2
values were 241
0.3234 and 0.3509 for MFC A and B, respectively). ...................................................................... 85 242
243
Figure 11. Relative activity of Geobacter 16S rRNA copy number normalized to total 16S rRNA 244
copy number on anode (solid bar) and cathode (pattern filled bar). (B) Relative activity of pmoA 245
transcripts copy number normalized to total 16S rRNA copy number on anode (solid bar) and 246
cathode (pattern filled bar). ND = not detected. .......................................................................... 89 247
248
Figure 12. Relative abundance of the top 30 abundant OTUs based on 16S rRNA gene (Bacteria 249
and Archaea) sequencing identified to the genus level at 25, 15, 10, and 5° C. Results are 250
expressed as a percentage normalized using total of 16S rRNA gene sequences. ....................... 91 251
252
Figure 13. (A) Correlation network analysis between genera preferentially colonizing on the 253
anode based on LDA scores >2. Solid lines represent positive correlations and dashed lines 254
represent negative correlations. Color coding denotes distinct metabolisms: potential 255
exoelectrogens (red), fermentative bacteria (brown), and others (e.g., heterotrophs; black); (B) 256
Correlation network analysis between genera preferentially colonizing on the cathode based on 257
LDA scores >2. Solid lines represent positive correlations and dashed lines represent negative 258
correlations. Color coding denotes distinct metabolisms: potential methane utilizers (green) and 259
others (e.g., heterotrophs; black); (C) Correlation network analysis between genera 260
preferentially colonizing on the anode ( ○) and genera preferentially colonizing on the cathode 261
( □) based on LDA scores >2. Solid lines represent positive correlations and dashed lines 262
represent negative correlations. Color coding denotes distinct metabolisms: potential methane 263
utilizers (green), potential exoelectrogens (red), fermentative bacteria (brown), and others (e.g., 264
heterotrophs; black). ..................................................................................................................... 94 265
266
Figure 14. Main figure shows influent and effluent sulfide concentrations. The orange bar 267
represents effluent sulfide overlaid on a blue bar representing influent sulfide. The red 268
horizontal lines represent targeted sulfide addition. The inset shows sulfide removal efficiency 269
at each influent sulfide concentration: 1, 5, 10 and 20 mg/L. The box plot boundaries indicate 270
the standard deviation while the horizontal line within the box indicates the median. The open 271
10

diamond marker in each box represents mean sulfide removal efficiency. The closed diamond 272
markers represent data from individual measurements. ........................................................... 103 273
274
Figure 15. Voltage production and dissolved methane removal efficiency from MFC A (sulfide 275
addition) and MFC B (control). Solid grey markers represent MFC A dissolved methane removal 276
and solid orange markers represent MFC B dissolved methane removal. Open green markers 277
represent MFC A voltage production and open blue marker represent MFC B voltage production. 278
Microbial samples were collected between different sulfide concentrations creating the gaps in 279
data. ............................................................................................................................................. 105 280
281
Figure 16. A1 and B1 shows microscopic image on anode surface from MFC A (sulfide addition), 282
while B1 and B2 shows the corresponding elemental analysis via EDAX. C1 and D1 shows 283
microscopic image on anode surface from MFC B (control), while C2 and D2 shows the 284
corresponding elemental analysis via EDAX. ............................................................................... 108 285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
11

List of Tables 304
Table 1. Off-gas methane purity in the sweep gas membrane contactor (V gas/V liquid represents 305
ratio of gas velocity to liquid velocity; Q gas/Q liquid represents ratio of gas flowrate to liquid 306
flowrate) ........................................................................................................................................ 36 307
308
Table 2. Vacuum driven membrane contactors energy consumption and electricity converted 309
from recovered methane (several constants were utilized: methane molecular weight, 16 g/mol; 310
η, vacuum pump efficiency, 0.5(Jain et al., Method for economic evaluation of membrane-based 311
air separation 1988); atmospheric pressure, 101.325 kPa; methane energy density by 312
combustion, 55.5 MJ/kg; electricity generation efficiency from methane combustion, 35%; P 2, 313
vacuum pump discharging pressure, 101.325 kPa); Bold text represent ratio of electricity 314
consumption to electricity converted from recovered methane, if less than 1 then indicates 315
operational conditions that achieve positive energy net balance. ............................................... 37 316
317
Table 3. Average and standard deviation of influent dissolved methane, average methane 318
loading per cathode area, average and standard deviation of methane removal, average 319
methane removal per cathode area, average and standard deviation of voltage production, 320
average power density, and Coulombic efficiency at HRTs of 16, 8, and 4 h for Reactor A and B. 321
....................................................................................................................................................... 66 322
323
Table 4. Comparison of dissolved methane management approaches including dissolved 324
methane removal, composition of recovered gas, and energy recovery as a percentage based on 325
dissolved methane loading, recovery gas volume, and methane content. An efficiency of 40% 326
was assumed for electricity recovery from collected methane using cogeneration. ................... 74 327
328
Table 5. Statistically significant Spearman’s rank correlations (ρ) between diversity metrics for 329
the anode microbial community (inverse Simpson index and Shannon index) and reactor 330
performance (CE, average daily voltage production, methane removal efficiency, and absolute 331
dissolved methane removal). ** represents P<0.0001 and * represents P<0.05. No significant 332
correlations were observed between the cathode microbial community and diversity metrics. 96 333
334
335
336
337
338
339
340
341
12

Chapter 1. Introduction and Overview 342
1.1 Background 343
Mainstream anaerobic treatment processes have emerged as an alternative to conventional 344
activated sludge due to their ability to recover energy directly from domestic wastewater, 345
significantly reduce sludge production, and eliminate energy-intensive aeration [1]. 346
Although mainstream anaerobic processes such as upflow anaerobic sludge blanket (UASB) 347
reactors have been deployed at the full-scale in tropical regions (e.g., Latin American and 348
India) [2, 3], low ambient temperatures have been a hurdle to more widespread 349
implementation. Anaerobic membrane bioreactors (AnMBRs), which combine anaerobic 350
treatment with membrane separation, are now being developed as a promising approach to 351
effectively manage domestic wastewater anaerobically at lower operational temperatures. 352
Methane, the major constituent of biogas, remains at substantial amounts dissolved in 353
anaerobic effluents, resulting in a lost energy source and significant greenhouse gas 354
emission if released to the atmosphere [4]. Further, low temperatures have been found to 355
exacerbate dissolved methane in the effluent due to (1) increased solubility and (2) 356
additional reliance on a membrane biofilm for treatment [5, 6]. Hydrogen sulfide produced 357
during anaerobic processes from pre-existing sulfate in wastewater is another concern in 358
both biogas and anaerobic effluents due to odor and corrosion concerns. Sulfide is produced 359
biologically via a metabolism that competes with methanogens for electron donor, thus 360
reducing energy recovery potential in the anaerobic process. Sulfides can also be inhibitory 361
to microbial populations in anaerobic systems [7]. Given these concerns, downstream 362
management of dissolved methane and other constituents (e.g., sulfide and nutrients) is 363
necessary before mainstream anaerobic processes can be conscientiously implemented. 364
13

Microbial fuel cells (MFCs), bioelectrochemical systems that microbially oxidize 365
organic/inorganic compounds and deposit electrons on a solid-surface electrode [8-10], 366
provide the opportunity to directly recover electricity from wastewater [8]. MFCs were 367
initially conceived as two-chamber systems (anode and cathode chambers) separated by a 368
proton exchange membrane. Providing an electron acceptor to the cathode chamber 369
required energy-intensive aeration or the use of an alternative electron acceptor (e.g., 370
ferricyanide) which are costly and require chemical regeneration. However, new air-cathode 371
designs require minimal energy demands as oxygen is passively diffused across an 372
atmosphere-exposed cathode [9]. Despite the reduced efficiency of air-cathode systems, the 373
simpler reactor design and lower material costs make them a more scalable system relative 374
to two-chamber designs. 375
Although mainstream anaerobic processes and MFCs are typically seen as competitors given 376
their similar goal of energy recovery from waste streams, the goal of this dissertation 377
research was to evaluate MFCs as a downstream treatment technology to recover energy 378
from dissolved methane and prevent greenhouse gas emissions if allowed to equilibrate 379
with the environment. Prior to this dissertation research, almost no research had evaluated 380
the possibility for energy recovery from methane using MFCs apart from an early 1965 study 381
of a two-chamber system operated with gaseous methane [11]. Given the likelihood of 382
other constituents such as sulfides in anaerobic effluents, this dissertation also sought to 383
determine if methane and sulfide could be successfully co-managed using air-cathode MFCs. 384
A brief overview of the dissertation chapters hereafter is provided below. 385
386
14

1.2 Overview of Dissertation 387
The following chapters investigate methane-driven MFCs as a potential post-treatment 388
technology for anaerobic effluents. Chapter 2 critically reviews greenhouse gas emissions 389
and mitigation strategies from conventional activated sludge processes and anaerobic-based 390
treatment processes. Chapter 3 describes experimental research demonstrating MFCs 391
operated on dissolved methane at a range of hydraulic retention times. Chapter 4 392
investigates methane-driven MFCs at varying operational temperatures and characterizes 393
microbial ecology of the system. Chapter 5 studies the impact of sulfide addition on 394
methane-driven MFCs. Notably, this dissertation is the first study to apply MFCs for post- 395
treatment of anaerobic effluents. 396
397
398
399
400
401
402
403
404
405
406
407
15

Chapter 2. Revisiting greenhouse gas mitigation from conventional 408
activated sludge and anaerobic-based wastewater treatment 409
systems 410
Abstract 411
Recent literature on carbon dioxide (CO 2), methane (CH 4), and nitrous oxide (N 2O) emissions 412
from wastewater treatment plants (WWTPs) has highlighted the poor consensus in total 413
greenhouse gas (GHG) estimation (ranging from 0.243 to 2.4 kg CO 2e/m
3
). In the present study, 414
the major components of GHG emission variability in both conventional activated sludge (CAS) 415
and mainstream anaerobic WWTPs are systematically investigated as a basis for delineating a 416
roadmap to their future control and minimization. Through analysis of N 2O generation 417
pathways, it was determined that additional research via isotope labelling is necessary to 418
elucidate distinct generation mechanisms in CAS WWTPs (e.g., nitrifier denitrification and 419
hydroxylamine denitrification) and better predict N 2O contributions to total GHGs. Conversely, 420
mainstream anaerobic processes, although a potentially more sustainable alternative to 421
conventional aerobic treatment, introduce effluent dissolved CH 4 as a potentially significant GHG 422
contributor. Sweep gas and vacuum driven membrane contactors are promising dissolved 423
methane management strategies. However, further optimization of gas-to-liquid ratios and 424
transmembrane pressures, respectively, are vital to balancing treatment efficiency with energy 425
neutral/positive operation. Overall, a thorough elucidation of N 2O generation pathways in CAS 426
WWTPs and the development of effective dissolved CH 4 management strategies for mainstream 427
anaerobic processes will define their respective future roles in reducing wastewater-associated 428
GHG emissions. 429
16

2.1 Introduction 430
As the risks of climate change become increasingly acute, the necessity for accurate greenhouse 431
gas (GHG) accounting has led to a renewed focus on wastewater management as a GHG 432
emissions source. The most widely employed wastewater treatment methods, namely aerobic 433
(i.e., activated sludge-based processes) and mainstream anaerobic processes, both significantly 434
contribute to GHG generation in their current forms of implementation. Mainstream anaerobic 435
processes, which lessen WWTP energy costs and biosolids generation, are receiving renewed 436
interest as an alternative to aerobic processes [1]. The EPA Inventory of GHG Emissions and 437
Sinks estimated that US wastewater treatment plants (WWTPs) accounted for approximately 438
0.3% of overall emissions in 2016, with CH 4 and N 2O accounting for 3 and 5 MMT CO 2 439
equivalents, respectively [12]. CH 4 and N 2O are of particular concern due to their relatively high 440
100-year global warming potentials (34 and 298 CO 2eq, respectively) [13], with WWTPs currently 441
estimated to be the sixth largest contributor to N 2O emissions worldwide (approximately 3%) 442
[14]. Despite the magnitude of N 2O emissions from WWTPs, our understanding of formation 443
mechanisms and ability to model or predict emissions remains lacking [15, 16]. 444
Today, conventional activated sludge (CAS) processes coupled with anaerobic digestion are 445
widely used in domestic wastewater treatment, despite their high energy requirements (up to 446
3% of overall US electricity consumption) and lack of large-scale energy and nutrient recovery 447
[17]. Although anaerobic digestion significantly offsets WWTP energy demands and reduces 448
sludge handling requirements, it is unclear whether this conventional approach will remain 449
attractive in light of recent advances in mainstream anaerobic treatment [18]. Mainstream 450
anaerobic systems directly recover energy via biogas production, produce drastically less sludge, 451
and have been proven viable at a range of operational temperatures [19]. Taking advantage of 452
their favorably warm climate, several Latin American countries have long incorporated 453
17

mainstream anaerobic processes, specifically upflow anaerobic sludge blanket (UASBs), for full- 454
scale domestic wastewater management [2, 20]. The integration of membrane separation and 455
anaerobic treatment (i.e., anaerobic membrane bioreactors (AnMBRs)) has greatly expanded 456
interest in mainstream anaerobic processes worldwide [21, 22]. This has led to recent advances 457
in the technology's application at ambient temperatures, as well as promising testing at the 458
pilot-scale [23, 24]. However, loss of dissolved CH 4 in effluents is an outstanding concern. Such 459
losses not only reduce energy recovery, but also pose severe environmental impacts due to GHG 460
emissions [1]. A direct comparison of GHG emissions between aerobic and anaerobic processes 461
is vital to help stakeholders navigate a potential transition to anaerobic treatment. 462
Several review papers have been published to address GHG emissions from conventional 463
WWTPs, many of which have focused specifically on N 2O emissions [15, 25-27]. For mainstream 464
anaerobic treatment, a recent review by Crone et al. evaluated dissolved effluent CH 4 while 465
discussing technologies for recovery [28]. Despite the significant contributions of the 466
aforementioned reviews, N 2O and CH 4 emission quantifications for the purpose of directly 467
comparing CAS and mainstream anaerobic treatment systems’ GWP remain unavailable. The 468
primary purpose of the current study is to systematically focus on outstanding knowledge gaps 469
in GHG emissions limiting direct comparability of CAS and mainstream anaerobic treatment. The 470
issues specifically evaluated in this work include CAS WWTP total GHG estimation, pathway- 471
associated N 2O generation mechanisms in aerobic based WWTPs, and dissolved CH 4 recovery 472
efficiency for anaerobic system effluents. 473
18

2.2 Anthropogenic GHG emissions from CAS WWTPs 474
2.2.1 Overview of GHG generation in wastewater treatment processes 475
GHG emissions are attributable to essentially every unit process in conventional aerobic 476
wastewater treatment coupled with anaerobic digestion (Figure 1). Here, we categorize these 477
emissions as either direct or indirect, where direct emissions include GHGs physically produced 478
by either in-plant or downstream environmental processes and where indirect emissions include 479
electrical energy demands and chemical inputs of the system. From the perspective of indirect 480
emissions, aeration tanks comprise more than 40% of total plant energy demand [17, 29, 30] 481
and are often reported as contributing most significantly to overall GHG emissions 482
(approximately 0.298 kg CO 2e/m
3
based on reported U.S. energy carbon footprint of 0.472 kg 483
CO 2/kWh) [31-34]. Sidestream processes for primary and waste activated sludges also 484
contribute indirect GHG emissions via energy demand and chemical addition during dewatering, 485
transportation, land application, and landfilling [35, 36], which can account for between 0.134 to 486
0.167 kgCO 2e/ m
3
of domestic wastewater [37]. It should be noted, however, that biogas 487
production from anaerobic digestion can significantly offset indirect GHG emissions by lessening 488
reliance on a potentially GHG emission heavy primary energy mix [33, 35]. 489
Regarding direct GHG emissions, N 2O generated during denitrification, either in anoxic tanks (in 490
the case of CAS with biological nitrogen removal) or in the receiving aquatic environment when 491
nitrate-rich effluent is released (in the absence of on-site anoxic treatment), is considered the 492
primary source of direct GHG emissions [12]. The EPA’s Inventory of US Greenhouse Gases and 493
Sinks reports this N 2O emission source as part of effluent emissions due to the majority of plants 494
not employing biological nitrogen removal. However, in scenarios where nitrogen removal is 495
achieved, these emissions are largely confined to within the plant footprint [38]. Aeration tanks 496
19

are also responsible for direct N 2O generation as a result of incomplete nitrification, with their 497
contribution to total N 2O footprint being recently identified as potentially much higher than 498
previously considered [12]. In addition to N 2O, aeration tanks are also responsible for significant 499
generation of CO 2 due to microbial degradation of organic carbon, however, these direct CO 2 500
emissions are not traditionally considered in GHG accounting because of their biogenic origin 501
[39]. Despite this, recent research has shown that approximately 14% of total organic carbon in 502
municipal wastewater is actually of non-biogenic origins due to domestic use of soaps and 503
detergents, leading to an underestimation of direct GHG emissions [40, 41]. Quantifying direct 504
emissions of CAS WWTPs has proven to be the most challenging aspect of GHG estimation, as 505
quantification methods and assumptions are wide ranging in existing literature and 506
governmental reports. In the following, GHG emissions are normalized to volume of domestic 507
wastewater (DWW) treated to better compare parallel studies, regardless of differences in 508
treatment process. 509
20

510
Figure 1. Direct greenhouse gases (CH 4 and N 2O) emissions from conventional wastewater 511
treatment plants employing anaerobic/anoxic/oxic activated sludge processes and anaerobic 512
sludge digestion. (ANA/ANX/AER, anaerobic/anoxic/oxic tanks) 513
514
2.2.2 Common findings of different quantification methods for plant-wide GHG emissions 515
Two common approaches have been reported for quantifying direct GHG emissions: (1) 516
emission factor-based methods derived from dynamic modeling and (2) actual emission values 517
determined from on-site measurement. Generally, model-based studies have reported high 518
variability in overall plant emissions (from 0.24 to 2.4 kg CO 2e/m
3
DWW), with contradictory 519
findings regarding the primary source of emissions [29, 42-44]. A clear consensus has thus not 520
yet been reached identifying the major contributors to total GHG emissions [16]. Models are 521
typically constrained to specific plant configurations and feature inconsistent emission factors. 522
21

Nonetheless, one common observation is that N 2O emissions contribute the greatest 523
uncertainty in emissions estimation [27, 42], particularly due to its excessive GWP and the lack 524
of a comprehensive mechanism-based model of formation [16]. 525
On -site quantification methods have included sampling and subsequent lab analysis [31, 32, 526
45], on-line off-gas collection and analysis [46-49], and tracer dispersion monitoring [50]. 527
Normalization of on-site sampling methodology has enabled the comparison of different 528
treatment processes, such as activated sludge, oxidation ditches, anaerobic/anoxic/aerobic 529
processes (A
2
O), and reverse A
2
O [31, 45]. Research taking advantage of covered treatment 530
units and direct off-gas on-line analyzers showed correlation between operational conditions 531
and GHG emissions. Such trends include emission increases with seasonal water temperature 532
variation [46], changes in aeration rates [47], discharge of reject water to influent streams [47], 533
length of anoxic/oxic phases (in sequencing batch reactors (SBRs)) [48], and influent nitrite 534
variations [49]. The use of less conventional approaches, such as tracer addition and dispersion 535
monitoring, have generally been less accurate compared to on-site measurement. A study by 536
Yoshida et al., for example, that utilized tracers and mobile cavity ring-down spectroscopy 537
sampling, found large variations in emissions over multiple campaigns, with CH 4 generation 538
ranging from 4.99 to 92.3 kg/h and N 2O from 0.37 to 10.5 kg/h [50]. 539
Although aeration energy consumption is the primary contributor to indirect GHG emissions 540
[42, 46, 48], direct emission rates remain less clear and are a significant obstacle to achieving a 541
plant-wide understanding [31, 32, 45]. Further, aeration control strategies impact both energy 542
consumption and N 2O generation (affected by DO levels), implying that tradeoffs exist between 543
direct and indirect GHG emissions [29]. Existing literature has reported between 0 and 14.6% of 544
nitrogen entering WWTPs being converted to N 2O [15, 35, 37, 48, 50], contributing 1% to 78.4% 545
of overall plant carbon footprints [32, 33, 37, 46, 48]. Based on this extreme variability, a more 546
22

thorough evaluation of literature addressing N 2O emissions in aerobic-based WWTPs is 547
necessary. 548
2.2.3 Specific direct GHG emission sources in CAS WWTPs 549
2.2.3.1 Considering total CH 4 emissions 550
Unintentional methanogenic conditions in collections systems, influent piping, grit chambers, 551
primary clarifiers, and anoxic/oxic tank dead zones all contribute to methane-based GHGs [33, 552
51, 52]. Existing studies have shown that this upstream-generated CH 4 is predominantly stripped 553
from the liquid phase upon reaching the aeration tanks, serving as the primary source of CH 4 554
emissions in the mainstream portion of WWTPs (6-18 g CO 2e/m
3
DWW) [53-57]. Further, the 555
low organic carbon and high DO remaining in solution after aeration minimize the potential for 556
any additional evolution and release of CH 4 in WWTP effluents (reportedly <0.1% of total CH 4 557
emissions) [58]. 558
Only a few studies on GHG emissions have incorporated sidestream anaerobic digestion. Two 559
studies by Daelman et al. [46, 59] reported total methane-associated emissions in the range of 560
90-95 g CO 2e/m
3
DWW, showing that fugitive gasses associated with sludge handling, digester 561
effluents, and cogeneration engine gas slip accounted for approximately three quarters (72 ± 562
23%) of WWTP CH 4 emissions. An analysis of studies reporting total digester CH 4 emissions 563
(ranging from 17 to 72 g CO 2e/m
3
DWW) suggested that operational parameters such as WWTP 564
SRT and anaerobic digester residence time likely play a significant role in CH 4 emission rates 565
[60]. Digestion associated CH 4 losses, if fully recovered, could potentially increase energy 566
recovery by 10-30% [35]. These results imply that although CH 4 is a relatively minor component 567
of direct CAS emissions, reducing their losses in sludge treatment processes can significantly 568
improve energy-associated GHG footprints. 569
23

2.2.3.2 N 2O emissions: taking generation pathways into account 570
N 2O generated during biological nutrient removal is one of the most variably reported 571
phenomena known to occur in conventional WWTPs. The current US EPA guidance on national 572
GHG inventories estimates that 0.5% of influent nitrogen will be converted to N 2O, primarily due 573
to denitirification of effluent nitrate in receiving waterways [39, 61]. This emission factor was 574
originally developed as part of a study by Czepiel et al. that did not include in-plant 575
denitrification [62]. More recent work on N 2O emissions from full-scale wastewater treatment 576
systems, however, have reported values ranging from 0 to 14.6% of N. To elucidate source 577
variability, a fundamental understanding of the factors affecting N 2O generation is necessary. 578
In CAS-based treatment, ammonium-containing wastewater is intentionally subjected to aerobic 579
and anoxic conditions to convert nitrogen to dinitrogen gas via nitrification and denitrification. 580
However, this process also has potential to contribute N 2O emissions through multiple distinct 581
and complex pathways (Figure 2). When autotrophic ammonium oxidizing bacteria (AOB) are 582
present at low DO, high nitrite, or high ammonium conditions, AOB will perform denitrification, 583
converting nitrite to N 2O (also known as nitrifier denitrification) [63]. Nitrite can also 584
independently react with coexisting organic or inorganic matter during the nitrification process 585
to produce N 2O. Another intermediate during ammonium oxidation, known as a nitrosyl radical, 586
has also been observed to convert to N 2O, either biologically or chemically [64]. During the 587
denitrification process, N 2O serves as a necessary intermediate and will accumulate as a result 588
of oxygen intrusion into the anoxic environment, high nitrite concentrations, or limited carbon 589
source availability. The cause of this accumulation is most commonly the inhibition of N 2O 590
reductase. In other cases, the presence of a hydroxylamine intermediate during ammonium 591
oxidizing conditions can promote N 2O generation through alternative pathways. This reaction 592
24

can proceed with either oxygen as the electron acceptor (hydroxylamine oxidation) or with 593
nitrite as the electron donor (N-nitrosation). 594
Although nitrifier and heterotrophic denitrification are considered the two main sources of N 2O, 595
other less understood pathways likely play a significant role [65-67]. Recent studies have added 596
inhibitors such as allylthiourea and chlorate to accredit N 2O emissions to nitrifier denitirification, 597
heterotrophic denitrification, or NH 2OH oxidation pathways [65, 68, 69]. An underlying problem 598
with this approach, however, is that these inhibitors also inhibit nitrification, in addition to the 599
processes that generate N 2O via nitrifier denitrification. A promising alternative for N 2O 600
emission source differentiation with high resolution is labelled isotope-based nitrogen species 601
introduction and tracking [70]. 602
2.2.3.3 Operational factors affecting N 2O emissions 603
Environmental conditions, operational parameters, wastewater characteristics, and varying 604
WWTP configurations can (individually or collectively) induce and/or increase N 2O generation. 605
Further, numerous N 2O formation pathways have been identified across a range of microbially 606
selective environments. Elucidating relationships between these variables and known N 2O 607
generation mechanisms remains challenging. In the following, a critical analysis of the potential 608
relationships between these two areas of literature is provided. 609
2.2.3.3.1 Dissolved oxygen control 610
Multiple reviews focusing on N 2O emissions from WWTPs have concluded that DO levels are 611
primarily responsible for its generation—low DO during nitrification and high DO during 612
denitrification [15]. However, existing literature that has investigated N 2O formation during 613
partial and/or full nitrification has reached contradictory conclusions regarding the role of DO. 614
For example, multiple studies on pure culture [69, 71], batch experiments [65, 72], lab-scale 615
25

reactors (SBR and CSTR) [73-77], and pilot/full-scale wastewater treatment plants [76, 78] have 616
observed higher emissions at low DO conditions during nitrification. In most of these studies, 617
the accumulation of NO 2
-
was closely related to high N 2O emissions at low aeration rates. 618
Conversely, other pure culture studies [79, 80], lab-scale experiments [68, 81-83], and a full- 619
scale nitritation-anammox reactor investigation [84, 85] have found elevated N 2O emissions 620
under higher DO conditions. It is likely that these varying observations result from differences in 621
microbial community structure and activity profiles leading to distinctly different formation 622
mechanisms. 623
Recent studies have used nitrification inhibitor addition and/or isotope labelling of N-species to 624
pair N 2O emissions with their specific generation pathways at varying DO levels. The use of 625
nitrification inhibitors, specifically, has revealed decreasing relative contributions of AOB 626
denitrification [65, 72], increasing NH 2OH oxidation contributions, and constant heterotrophic 627
denitrification contributions to overall N 2O emissions at increasing DO.[68] Further insight 628
provided in a study by Peng et al. [82], which used isotopic site preference measurements, 629
showed increases in NH 2OH oxidation-sourced N 2O and decreases in AOB denitrification-induced 630
emissions with rising DO (from 0.2 to 3 mg/L). Based on the cumulative findings of these 631
studies, it can be concluded that although AOB denitrification is commonly the dominant N 2O 632
production pathway, the NH 2OH oxidation pathway could outcompete at high DO (e.g., 3.5 mg 633
O 2/L) when combined with low NO 2
-
(e.g., <10 mg O 2/L) [72]. Ultimately, more applied research 634
quantifying both gaseous and aqueous N 2O is needed, while taking into consideration all 635
possible N 2O generation pathways. 636
637
26

638
Figure 2. Five distinct N 2O generation pathways (NH 2OH oxidation with O 2, NH 2OH N-nitrosation 639
hybrid reaction, unstable decomposition of (NOH), nitrifier denitrification, incomplete 640
heterotrophic denitrification) along with nitrification (ammonia oxidizing bacteria (AOB) and 641
nitrite oxidizing bacteria (NOB)) and denitrification pathways. Key enzymes are identified along 642
each pathway. 643
644
2.2.3.3.2 pH 645
Studies investigating N 2O emissions during nitrification at different pH ranges have generally 646
observed highest production at pH 8-8.5, independent of free ammonia and nitrous acid 647
concentrations [69, 86]. N 2O emissions during denitrification have been conversely observed to 648
decrease with increasing pH across 5-8.5, with concurrent decreases in NO 2
-
[87-89]. One study 649
specifically found that no N 2O formation was detected at pH >6.8 and highest production 650
occurred between 5 and 6 [88]. Nonetheless, given that free nitrous acid is believed to exert a 651
stronger inhibitory effect on N 2O reductase than pH and has been strongly correlated with N 2O 652
production, it is possible that the relationship between pH and N 2O production during 653
denitrification is purely incidental [89]. 654
27

2.2.3.3.3 Nitrite 655
In addition to its oxidation by O 2, NH 2OH can also serve as a precursor to N 2O formation via its 656
reaction with nitrite (known as N-nitrosation). Even when NH 2OH is present as an intermediate 657
in the ammonium oxidation process at low concentrations, the N-nitrosation hybrid reaction has 658
still been observed to proceed in full-scale bioreactors (0.03 to 0.11 mg N/L) [90]. Isotope 659
labelled N
15
O 2
-
and N
15
H 2OH have been used to distinguish respective contributions of nitrifier 660
denitrification, the N-nitrosation hybrid reaction, and NH 2OH oxidation in a partial nitrifying 661
bioreactor [70]. The N-nitrosation reaction was the prominent formation pathway in this study, 662
possibly due to the relatively high DO levels. These results imply that high nitrite concentrations 663
can result in significant N 2O formation in the nitrification process, even in the absence of nitrifier 664
denitrification. 665
Increasing nitrite concentrations during denitrification have also been observed to limit the 666
generation of NO reductase, leading to accumulation of nitric oxide (NO) [91]. This can further 667
impact N 2O emissions, as NO causes an inhibitory effect on enzymes involved in the 668
denitrification process (e.g., nitric acid and N 2O reductases). In a mixed microbial community of 669
both nitrifiers and denitrifiers, for example, Tallec et al. [65] observed up to an eight fold 670
increase in N 2O production with nitrite addition at 1 mg O 2/L. Further, specific tests on oxidized 671
nitrogen in an aerobic granule sludge system have shown specific N 2O generation to be 672
approximately 44% higher in the presence of nitrite as compared to nitrate alone [73]. Although 673
the mechanisms of N 2O formation in nitrification and denitrification processes are distinctly 674
different, nitrite presence plays a significant role in both. 675
28

2.2.3.3.4 Carbon source availability/COD:N ratio 676
As has been reviewed [15, 25, 26], limited availability of carbon sources increases N 2O 677
production during denitrification. Although the exact mechanism by which this occurs is not fully 678
understood, competition for electrons between various denitrification enzymes (i.e., NO 3
-
, NO 2
-
, 679
NO and N 2O reductase) is likely the cause [26]. Specifically, NO 3
-
and NO 2
-
reductases have 680
relatively higher electron affinity than NO reductase and N 2O reductases, which induces 681
incomplete denitrification under carbon limited conditions. Increased N 2O production in carbon 682
source-limited environments can also be due to microbial consumption of internal storage 683
compounds (i.e., poly-β-hydroxybutyrate (PHB)) [15, 26]. In simultaneous 684
nitrification/denitrification and phosphorus removal processes employing denitrifying 685
phosphate accumulating organisms (DPAOs), N 2O generation has been observed to start 686
immediately after the pulse addition of nitrite [92], but further research is needed to determine 687
the intrinsic mechanism of this phenomenon. 688
To maintain the minimum COD:N ratio necessary to accomplish full denitrification (typically 689
considered to be > 4), the addition of external substrate as a carbon source is often required 690
[26]. This practice has been shown in certain instances to also significantly reduce N 2O 691
production (by up to 95%) [83]. As such, a range of external carbon source/substrate types (e.g., 692
acetate, methanol, mannitol, glucose, starch, acetic acid, sludge fermentation liquid) have been 693
investigated for their effectiveness at curbing N 2O formation [83, 93-96]. Resultantly, distinctive 694
changes in both microbial diversity and N 2O production rates have been observed with different 695
substrates. These differing microbial communities, which exert preferential consumption of 696
each carbon source type, will ultimately dictate the enzymatic activity responsible for both NO 697
and N 2O reduction. 698
29

2.2.3.3.5 Ammonium shock 699
Returning ammonium-rich reject water to the headworks can significantly contribute to N 2O 700
emissions due to ammonium shock, especially during downstream transitions from anoxic to 701
aerobic conditions/environments [80, 84]. Given that this transition in redox conditions is often 702
unavoidable, the accumulation of ammonium in anoxic environments should be closely 703
monitored. Ammonium shock can also induce decreases in DO levels, potentially triggering 704
nitrifier denitrification and subsequent elevated N 2O emissions [97]. Lab-scale work 705
investigating this phenomenon has identified a critical ammonium loading rate of approximately 706
1.60 mg NH 3-N/g TSS, beyond which nitrite and N 2O increase significantly [98]. 707
2.2.4 Implications of N 2O emissions in CAS WWTPs 708
Optimization of key operational parameters (i.e., sufficient carbon sourcing, pH, DO, and 709
ammonium levels) is key to achieving predictable and minimized N 2O emission rates. A 710
challenge associated with plant-level N 2O source identification is that nearly all N 2O is emitted 711
from aeration tanks, regardless of formation pathway [78, 99]. Therefore, more research 712
employing isotope labelling is likely necessary to elucidate the underlying mechanisms and their 713
contributions to overall N 2O emissions. With a better understanding of each N 2O generation 714
pathway and its role within treatment systems, specific strategies can be devised to mitigate 715
N 2O emissions and ultimately standardize operational guidelines to reduce nationwide GHG 716
emissions. 717
2.3 GHG emission management for mainstream anaerobic treatment 718
Anaerobic processes are considered a sustainable and energetically favorable alternative to 719
conventional aerobic processes. Anaerobic processes directly convert organics to methane-rich 720
biogas and eliminate energy requirements associated with aeration [19]. However, the release 721
30

of dissolved CH 4 along with discharged effluents remains a significant implementation concern, 722
severely increasing GHG emissions while concomitantly reducing potential energy recovery [1, 723
28]. Such losses, which in extreme cases account for up to 90% of total produced CH 4, pose a 724
severe environmental threat if mainstream anaerobic treatment becomes the norm [2]. Still, 725
successful mitigation of these emissions would enable anaerobic treatment with less GHGs than 726
CAS processes, providing impetus for advancing dissolved CH 4 recovery technologies [100]. 727
2.3.1 GHG emission sources in anaerobic bioreactors 728
Given that the majority of GHG emission-related research on anaerobic treatment has been 729
conducted at the bench- and pilot-scale [5, 6, 101-112], full-scale indirect CO 2-based emissions 730
estimates for electricity consumption remain largely unconfirmed. Nonetheless, energy balances 731
of mainstream anaerobic treatment are generally expected to significantly improve upon 732
current CAS [113]. In addition, an objective comparison of GHG emissions between CAS and 733
mainstream anaerobic processes requires inclusion of downstream nutrient removal processes 734
for anaerobic systems (e.g., partial nitritation-anammox). Such nitrogen removal processes have 735
been shown to emit even higher levels of N 2O than CAS, as reviewed by Massara et al. [25], and 736
necessitate further process optimization to be successfully mitigated. 737
Even accounting for these uncertainties, the most significant GHG-associated threat from 738
mainstream anaerobic treatment remains effluent CH 4 losses. CH 4 saturation relative to Henry’s 739
Law in anaerobic effluents has been observed to range between a factor of 1.0 to 5.2, resulting 740
in the loss of 10-90% of total CH 4 produced [28]. Recent work has demonstrated robust 741
operation (i.e., COD removal) at temperatures as low as 6°C. However, such low temperatures 742
exacerbate GHG emission concerns by inherently increasing CH 4 solubility [104]. Overall, CH 4 743
solubility at low temperatures is largely responsible for increasing trends in dissolved CH 4 744
31

concentrations across all anaerobic bioreactor system types (Figure 3a), despite reactor 745
configuration and biogas composition also playing a role. Although the integration of membrane 746
filtration in AnMBRs has improved effluent quality at such low temperatures, similar CH 4 747
oversaturation is still observed [5, 6]. Studies by Smith et al. [6, 114] on low-temperature 748
AnMBR operation, specifically, have documented the likelihood that high methanogenic activity 749
in membrane biofilms are responsible for dissolved CH 4 oversaturation. Experimental findings 750
suggested that as systems increasingly relied on membrane biofilm-based treatment at 751
decreased temperatures, biofilm methanogens directly emitted CH 4 into the effluent [6]. 752
Although limitations of gas-liquid transfer rates have been identified as an obstacle for recovery, 753
in-situ biogas sparging readily achieves gas-liquid equilibrium and maximizes CH 4 evolution to 754
headspace. Several recent studies have demonstrated effluent CH 4 saturation factors of close to 755
1 and/or reduced dissolved CH 4 content by up to 50% by employing in-situ biogas sparging [106, 756
110]. AnMBRs operating at temperatures above 20° C have also shown relatively low saturation 757
factors (in the range of 1.0 to 1.1) [106, 108, 115]. Yet, other work has shown CH 4 saturation 758
exceeding a factor of 1.5 or greater, even when biogas sparging is sufficiently utilized. 759
The strongest deviation above CH 4 saturation level was observed for scenarios at 15° C and lower 760
(Figure 3b). In some scenarios, no biogas CH 4 was produced, with all produced CH 4 being 761
evolved in the effluents (Figure 3c) [6, 104]. Given that this phenomenon is likely caused by 762
disproportionate biofilm-based CH 4 production, it cannot be easily mitigated by reactor biogas 763
sparging/stripping [5, 6]. Using 34 as a standard GWP factor for CH 4, GHG emissions from 764
dissolved CH 4 were calculated to be in the range of 0.281 to 2.82 kg CO 2e/m
3
DWW. This is 765
generally comparable to the wide-ranging rates of CAS (0.24 to 2.4 kg CO 2e/m
3
) and will 766
ultimately necessitate downstream CH 4 recovery technologies. 767
32

768
769
Figure 3. (a) Dissolved methane concentration (mg/L) from anaerobic treatment processes across 770
different studies at varying temperature; (b). Dissolved methane supersaturation ratio from 771
anaerobic treatment processes across different studies at varying temperature; (c). Proportion of 772
dissolved methane over total methane production from anaerobic treatment processes across 773
different studies at varying temperature; ○ represents upflow anaerobic blanket sludge bed 774
reactor (UASB), ◊ represents expanded granular sludge bed reactor (EGSB), ∆ represents 775
anaerobic fluidized membrane bioreactor (AFMBR), □ represents anaerobic membrane 776
bioreactor (AnMBR); dash line represents dissolved methane derived from Henry’s law, with 80% 777
gaseous methane in the headspace 778
779
2.3.2 Utilizing membrane contactors for dissolved CH 4 recovery 780
Of the physiochemical-driven methods examined for dissolved CH 4 removal and/or recovery 781
from anaerobic effluents, the most widely tested involves membrane contactors for effluent CH 4 782
33

desorption. Relevant operational parameters of these systems include membrane properties, 783
contact area, gas/liquid flow rates, flow direction, vacuum pressure, and gas/liquid supply 784
sources (shell or lumen). In the following, we provide a comparative analysis of the energy 785
use/recovery potential of the two primary modes of membrane contactor operation, namely 786
sweep gas- and vacuum-based desorption. 787
2.3.2.1 Sweep gas membrane contactors 788
When operating membrane contactors in sweep gas mode, dissolved CH 4 removal in the liquid 789
phase is driven by a concentration gradient across a gas permeable membrane into crossflowing 790
nitrogen or air. Several studies have successfully demonstrated dissolved effluent CH 4 removal 791
rates from UASB, AnMBR, and synthetic effluents of up to 98.9% and 92.6% using microporous 792
and nonporous hollow fiber membrane contactors (HFMCs), respectively [115, 116]. The 793
aforementioned studies employed polydimethylsiloxane (PDMS) membranes (both microporous 794
and nonporous types). Another recent study utilized fluorinated silica nanoparticle modified 795
membranes to enhance surface hydrophobicity [117]. These modified membranes attained 796
higher CH 4 recovery fluxes as compared to a commercial polypropylene microporous 797
membranes (400-550 mg CH 4/m
2
∙h vs. 200-350 mg CH 4/m
2
∙h) over 300 h of operation, implying 798
that such surface modifications can alleviate long-term pore wetting issues. 799
2.3.2.2 Vacuum suction (degassing) membrane contactors 800
Membrane contactors operated in vacuum mode rely primarily on a pressure differential 801
without significant gas cross-flow to achieve high-concentration CH 4 recovery. This allows for 802
the direct use of captured CH 4 without further purification, but also requires additional energy 803
input in the form of vacuum pressure. Multiple early studies by Bandara et al. [102, 104, 118] on 804
membrane degasification for UASB effluents using a commercial multi-layer composite 805
polyethylene hollow-fiber membrane contactor (HFMC) successfully desorbed 77% to 86% of 806
34

dissolved CH 4 from UASB effluents into the lumen at vacuum pressures of 50 and 80 kPa [102]. 807
Lumen-side liquid flow (as opposed to shell-side), has generally been observed as more effective 808
at CH 4 desorption due to superior liquid to air transfer rates, however it can be limited by hollow 809
fiber flow-path clogging over long-term operation [28, 119]. 810
2.3.2.3 Analysis of energy demands and recovery by membrane contactors 811
Superior removal rates are achievable by vacuum degasification as compared to sweep gas 812
operation [120]. Further, vacuum desorption has specific advantages associated with direct on- 813
site CH 4 use, which are not achievable by sweep gas contactors. Such advantages, however, 814
must be evaluated in comparison with the greater energy requirements of vacuum-driven 815
transmembrane pressure (TMP). Given the knowledge gaps in literature from the perspective of 816
energy use and recovery, a comparative analysis of the practical limitations of each CH 4 recovery 817
method is necessary to assess each technology’s economic feasibility and overall GWP. 818
As summarized in Table 1, CH 4 concentrations for sweep gas driven membrane contactors are 819
generally less than 2.4% of total off-gas volume, with only one case demonstrating relatively 820
high concentrations of 23.2% with a polypropylene HFMC operated at low gas to liquid (G/L) 821
ratios [105, 115-117, 119, 121]. In most cases, sweep gas driven membrane contactors exhibited 822
increasing effluent removal efficiencies at higher gas to liquid (G/L) ratios, which also lead to 823
decreased CH 4 off-gas purity. The majority of studies to date investigating sweep gas membrane 824
contactor use have had a primary objective of reducing effluent concentrations to eliminate 825
combustion risks in downstream discharge piping. Therefore, the systems were not operated to 826
achieve optimal off gas concentrations. Without post-removal purification of sweep gas, limited 827
approaches are available for utilization, as CH 4 concentrations are generally too low for even 828
basic off-gas flaring (i.e., >5% CH 4). However, recent developments have shown that direct 829
combustion via thermal/catalytic flow reversal reactors, regenerative/catalytic oxidation, or lean 830
35

burn-gas turbine combustion can be achieved at CH 4 concentrations as low as 1% [122]. 831
Implementing such air-based off-gas in on-site cogeneration plant engines has been proposed 832
previously for anaerobic digester dewatering process gasses [59]. These applications, in 833
combination with optimization of G/L ratios, could lead to sweep gas membrane contactors 834
being a viable option for effluent CH 4 GHG mitigation. 835
A recent review by Crone et al. [28] calculated an energy input to recovery ratio of 1.0 for 836
effluent CH 4 recovery using vacuum driven membrane contactors. However, given the high 837
variability in existing literature associated with operational parameters of vacuum 838
degasification, a more comparable evaluation of these variables is necessary. Table 2 provides a 839
normalized summary of studies on vacuum-driven membrane contactors, their energy 840
requirements, and the potential energy content of recovered CH 4 using a unified methodology. 841
Results of multiple studies suggest that systems operated at TMPs between 14 and 50 kPa are 842
generally energy positive while maintaining CH 4 recovery rates between 60 and 90% [102, 118- 843
120]. Further, CH 4 recovery in different scenarios did not improve substantially with increasing 844
vacuum pressure, proving that low vacuum scenarios are generally effective. Based on these 845
observations, operating vacuum driven membrane contactors at relatively low vacuum 846
pressures (< 50 kPa) can enable the entire HFMC system to be energy neutral/positive while 847
achieving sufficient effluent GHG reduction. 848
849
850
851
852
36

853
Table 1. Off-gas methane purity in the sweep gas membrane contactor (V gas/V liquid represents 854
ratio of gas velocity to liquid velocity; Q gas/Q liquid represents ratio of gas flowrate to liquid 855
flowrate) 856
857

Dissolved
methane
concentration
(mg/L)
Dissolved
methane
removal
efficiency
(%)
Vgas/Vliquid
or
Qgas/Qliquid
Sweep gas
flowrate (N2,
m
3
/s)
Recovered
methane
flowrate
(m
3
/s)
Ratio of
methane/nitroge
n in off-gas
Cookney et al. 2016, PDMS,
nonporous
21.0 92.6 825 3.10E-02 1.24E-09 4.01E-08
Cookney et al. 2012, PDMS,
hollow fiber
12.9 72.0 70.0 1.41E-05 6.21E-09 4.42E-04
Wongchitphimon et al.
2017, polymer-fluorinated
silica composite, hollow
fiber (Mo-MT-A)
tap water
saturated with
60:40 CH4/CO2
NA 0.497 3.33E-07 1.15E-10 3.46E-04
Rongwong et al. 2017, in-
house fabricated hollow
fiber membrane
hollow fiber
anaerobic
bioreactor pilot
plant effluent
bubbled with
60:40 CH4/CO2
<65.0 0.106 3.33E-07 4.58E-09 1.37E-02
Henares et al. 2016 (a),
PDMS, nonporous
30.0 74.0 7.51E-06 7.51E-07 3.39E-09 4.52E-03
Henares et al. 2016 (a), PP,
microporous
30.0 98.4 1.94E-04 2.20E-07 5.11E-08 2.32E-01
Henares et al. 2016 (b),
PDMS, nonporous
30.0 75.0 6.94 6.94E-07 3.44E-09 4.95E-03
McLeod et al. 2016,
polypropylene, hollow
fiber
18.0 90.0 1.00 1.70E-06 4.21E-08 2.48E-02
858
859
860
861
862
863
864
865
37

866
Table 2. Vacuum driven membrane contactors energy consumption and electricity converted 867
from recovered methane (several constants were utilized: methane molecular weight, 16 g/mol; 868
η, vacuum pump efficiency, 0.5(Jain et al., Method for economic evaluation of membrane-based 869
air separation 1988); atmospheric pressure, 101.325 kPa; methane energy density by 870
combustion, 55.5 MJ/kg; electricity generation efficiency from methane combustion, 35%; P 2, 871
vacuum pump discharging pressure, 101.325 kPa); Bold text represent ratio of electricity 872
consumption to electricity converted from recovered methane, if less than 1 then indicates 873
operational conditions that achieve positive energy net balance. 874
875
876

Dissolve
d
methane
concentr
ation,
mg/L
Dissolved
methane
removal
efficiency, %
Transmembra
ne pressure,
kPa
Suction
flowrate, Q
(assume pure
methane,
m3/s)
Vacuum
pump/compress
or energy;
𝜆𝑅 𝑇 𝑊 𝜆 − 1
[ (
𝑃 2
𝑃 1
)
𝜆 − 1
𝜆 − 1 ]
[12]
; J/s
Electricity
generated
from
desorbed
gas, J/s
Electricity
consumptio
n/electricity
converted
from
recovered
methane
Cookney et al. 2012,
PDMS, nonporous
12.9 77.0% 30.8 1.15E-05 18.9 1.50E+02 0.126
Bandara et al. 2011,
multi-layered
composite hollow
fiber membrane
12.5-
25.0
<89.0%
50.0 2.81E-10 0.151 0.239 0.632
80.0 5.90E-10 14.7 0.502 29.4
80.0 9.73E-10 24.3 0.854 28.4
80.0 6.27E-10 15.7 0.570 27.5
Henares et al. 2016
(a), PP, microporous
30.0
70.0% 14.0 3.64E-08 0.0143 0.462 0.0309
82.0% 50.0 4.26E-08 0.331 0.541 0.611
94.0% 80.0 4.88E-08 17.6 0.620 28.4
Luo et al. 2014,
Hollow fiber
membrane 3504
~15.0 86.0%
94.0 6.77E-10 1.70E+03 0.572 2.97E+03
94.0 1.58E-09 3.96E+03 1.33 2.97E+03
94.0 2.31E-09 5.80E+03 1.95 2.97E+03
Henares et al. 2016
(b), PDMS,
nonporous
30.0
60.0% 14.0 2.75E-09 0.00108 0.0350 0.0309
70.0% 50.0 3.21E-09 0.0249 0.0408 0.611
75.0% 80.0 3.44E-09 1.24 0.0437 28.4
Bandara et al 2013,
3-layer composite
hollow fiber
membrane
17.3
90.0% 50.0 1.99E-09 0.0155 0.0245 0.632
95.0% 70.0 2.80E-09 0.191 0.0344 5.57
38

2.3.3 Biological approaches 877
Biological oxidation is a promising alternative strategy for CH 4 removal from anaerobic effluents. 878
The most common technique is the downflow hanging sponge (DHS). This method has proven 879
highly effective by numerous studies, as previously reviewed [28]. In comparison, another 880
potentially promising treatment system that has yet to be fully investigated is known as 881
denitrifying anaerobic CH 4 oxidation (DAMO), and is reviewed herein. 882
2.3.3.1 Effective CH 4 removal by the downflow hanging sponge (DHS) 883
Several recent studies have utilized DHS bioreactors for the aerobic oxidation of dissolved CH 4 884
with relatively consistent removal results [112, 123-125]. Through the optimization of 885
operational parameters such as wastewater composition and air flowrates, systems have 886
achieved removal of multiple residuals (e.g., CH 4, ammonium, sulfur, etc.). While some have 887
employed varying HRTs and aeration rates to achieve removals of up to 97% of dissolved CH 4 888
using single stage DHS reactors [123], other work has shown that two-stage DHS systems are 889
capable of both recovering CH 4 in off-gas at high concentrations (>30%) and oxidizing the 890
remaining content to achieve near complete removal of dissolved CH 4 (>99%) from effluents 891
[124]. Air flowrate is a critical operational parameter, as varying oxygen affinity and growth 892
rates among different microbial communities significantly affect removal [125]. 893
2.3.3.2 The case for denitrifying anaerobic CH 4 oxidation (DAMO) 894
A more recently proposed method, known as DAMO, provides a potential solution for CH 4 895
removal through its use as an electron donor. DAMO archaea are capable of reducing nitrate to 896
nitrite while DAMO bacteria convert nitrite to nitrogen gas. Combining DAMO and anammox 897
has recently been proposed as a means for simultaneous nitrogen and CH 4 removal from 898
39

anaerobic effluents [126]. Nitrate reduction by DAMO archaea and nitrite reduction by DAMO 899
bacteria with CH 4 oxidation are achieved through the sequential CH 4 oxidation processes below, 900
CH 4 + 4NO 3
−
→ CO 2 + 4NO 2
−
+ 2H 2O 901
3CH 4 + 8NO 2
−
+ 8H
+
→ 3CO 2 + 4N 2 + 10H 2O 902
while nitrite reduction/ammonium oxidation by anammox concurrently produces nitrate: 903
NH 4
+
+ 1.3NO 2
−
→ N 2 + 0.3NO 3
−
+ 2H 2O 904
Recent studies by Chen et al. [127, 128] have developed a system based on this model using a 905
membrane biofilm reactor (MBfR). Multiple lab-scale investigations have shown that biofilms 906
containing cocultures of DAMO and anammox microorganisms can achieve sufficient nitrate and 907
nitrite reduction and ammonia oxidation [129, 130]. Ultimately, the application of a combined 908
anammox and DAMO process could offer significant economic and practical advantages over 909
conventional practices if successfully combined with anaerobic systems. The implementation of 910
this process for treatment of anaerobic bioreactor effluents, however, is highly dependent on 911
the co-enrichment of specific DAMO and anammox organisms and the supplementation of 912
nitrite to the system. Although magnetically stirred gas lift reactors (MSGLRs), MBfRs, and 913
granular sludge reactors have all been identified as capable of supporting growth of DAMO 914
microorganisms and retaining biomass effectively [127-129, 131-134], the most feasible option 915
thus far for integration of anammox and DAMO is MBfRs. 916
Although such applications are still in their infancy, there are multiple practical advantages to 917
applying DAMO as part of anaerobic effluent treatment processes [135, 136]. With CH 4 as the 918
sole electron donor for DAMO microorganisms, no additional organic carbon sources would be 919
needed. Further, the slow growth rates of DAMO microorganisms such as M. oxyfera (doubling 920
40

time of 1–2 weeks), and low yields of DAMO microorganisms in general [127, 133], alleviate the 921
necessity of sludge disposal. Oxygen delivery via hollow fiber membrane units [128, 132] or 922
granular-based optimization of oxygen levels [133] require further investigation to practically 923
alleviate the negative impacts of aeration on anammox/DAMO. Nonetheless, research thus far 924
on MBfRs and granular sludge reactors suggests that they may soon be a feasible basis for post- 925
treatment of anaerobic effluents. 926
2.3.3.3 MFCs as an alternative biological process 927
Microbial fuel cells (MFCs) have also been considered for effluent dissolved CH 4 management 928
[137-140]. MFCs are bioelectrochemical systems where exoelectrogenic microorganisms oxidize 929
organics and directly deposit electrons onto an anode [8-10]. Methane, as an organic substrate, 930
can be used as an energy source to drive MFCs, converting it directly to electricity [11, 137-139, 931
141, 142]. For example, a study by McAnulty et al. [138] manipulated engineered archaeal 932
strains to produce acetate from CH 4 anaerobically via methyl coenzyme M reductase, 933
subsequently generating electricity in a two-chamber MFC. Chen et al. [139] also reported 934
electricity generation from CH 4 using a single-chamber MFC while observing microbial 935
interactions between aerobic methanotrophs and exoelectrogenic Geobacter. It should be 936
noted that these emerging biological processes, and specifically MFCs, require further 937
optimization in terms of capital cost reduction and achieving consistent treatment performance 938
before scaling up to pilot- and full-scale application [143, 144]. 939
2.3.4 Implication of physical vs. biological systems for CH 4 mitigation 940
It should be noted that the aforementioned biological approaches, while potentially requiring 941
less energy input than membrane contactors and still mitigating GHG emissions, do not capture 942
CH 4 for energy recovery. Although this is a significant limitation for DHS systems, DAMO's use of 943
41

CH 4 as an electron donor for nitrate reduction provides an alternative route to its utilization 944
when nitrogen removal is required (e.g., effluent discharge to nitrogen sensitive waterways). 945
MFCs, although only recently demonstrated for methane, could be advantageous over both 946
physical CH 4 recovery (using membrane contactors) and other biological approaches due to their 947
ability to directly recover energy. 948
2.4 Future GHG management perspectives 949
2.4.1 N 2O as an energy recovery oxidant 950
N 2O, as a powerful oxidant, has the potential to be selectively produced in wastewater 951
treatment processes (e.g., via coupled aerobic-anoxic side-stream nitrogen removal) and serve 952
as a combustion oxidant in combination with CH 4 [145, 146]. In conventional WWTPs, N 2O off- 953
gas collection could be accomplished via the installation of covers on treatment unit processes, 954
however this may be impractical for nitrification due to the large volume of gas produced by 955
aeration. Although selective reduction of N 2O has been practiced industrially, similar processes 956
may not be cost-effective in large application to low N 2O containing gases as they require 957
catalysts and high reaction temperatures [147]. Nevertheless, with consistent N 2O production 958
and improved collection efficiency in newly designed processes [92], this could be a worthwhile 959
future research topic. 960
2.4.2 Technologies for targeting CO 2 capture 961
Given the potential significance of non-biogenic CO 2 sources in municipal wastewater influents, 962
it is important to also consider possible means for direct CO 2 sequestration or capture. However, 963
considering the relatively high solubility of CO 2 in water and its potential cost of recovery, the 964
treatment of dissolved CO 2 in-situ using emerging technologies has recently become a topic of 965
interest. For example, phototrophic technologies relying on algae and/or phototrophic bacteria 966
42

could promote carbon fixation while simultaneously achieving nutrient removal [148]. 967
Alternatively, carbonic anhydrase, a ubiquitous enzyme capable of catalyzing the hydration of 968
CO 2 into bicarbonate and hydrogen at high rates, could potentially be incorporated into 969
engineered systems to sequester carbon directly [148]. Microbially assisted electrolytic systems 970
also have the potential to sequester and convert CO 2 to bicarbonate using either wastewater or 971
seawater as the electrolyte while producing beneficial products such as H 2 [148, 149]. 972
2.4.3 Methanotroph-based recovery of high-value end products 973
Methanotrophs can be metabolically engineered to synthesize a range of high-value products 974
including single cell proteins, biopolymers (e.g., polyhydroxyalkanoates, PHB), soluble 975
metabolites (e.g., methanol, formaldehyde, formate), lipids, lycopene, C30 carotenoid, lactic 976
acid and exopolysaccharides [150-153]. Methane oxidation is a multi-step process in which CH 4 977
is oxidized to methanol, formaldehyde, formate, and CO 2 sequentially. Given that methanol 978
dehydrogenase is located in the periplasmic membrane of methanotrophs, methanol must be 979
transported out of the cell membrane in order for subsequent processes to ensue. Based on 980
this, methanotrophs can be genetically engineered or supplemented with inhibitors to suppress 981
this dehydrogenase and stop CH 4 oxidation at methanol [154], which can then be collected and 982
enriched for use in MFCs (currently commercially available and used as portable electricity 983
sources). The application of these processes, among others, could be especially useful when 984
remote electricity generation is needed, the energy for which could be supplied exclusively from 985
treated wastewater. 986
2.4.4 Methane adsorbent-related technology 987
Recent studies have demonstrated that structures with high adsorption capacity and packing 988
density can be used to adsorb and store CH 4. For example, an investigation by Bagheri et al. 989
43

[155] demonstrated that microporous activated carbon generated from corn cobs was capable 990
of high levels of CH 4 adsorption (150 v/v). Other materials such as constructed multilayer 991
graphene nanostructures (MGNs) with optimized layer distances were able to satisfy the U.S. 992
department of engineering target for adsorbents (180 v/v) [156]. Although the aforementioned 993
materials require pressures of >100 psi to effectively sorb CH 4 into their structures (reducing 994
their viability from an energy and GHG footprint perspective), the recent synthesis of a 995
monolithic metal-organic framework has proven capable of reaching a CH 4 packing density of 996
259 v/v at pressures previously comparable to those of half of its capacity [157]. The continuous 997
improvement of adsorptive materials and the potential for their exploration at lower sorption 998
pressures could lead to viable use for CH 4 capture, purification, and transport from wastewater 999
effluents. 1000
2.5 Perspectives on the direct comparison of CAS- and anaerobic-based GHG 1001
emissions 1002
Based on a normalized analysis of existing literature, herein we provide a parallel assessment of 1003
CAS-based and anaerobic-based mainstream wastewater treatment using their dominant GHG 1004
sources and assuming equivalent levels of treatment for effluents (i.e., nitrogen and COD 1005
removal). Recent studies have shown that approximately 0.298 kg CO 2e/m
3
of GHG emissions 1006
come from electricity usage in conventional WWTPs (based on 0.472 kg CO 2/kWh energy carbon 1007
footprint) [32, 33, 48], with 50% of that energy demand being consumed by aeration [17]. It is 1008
estimated that roughly 25% of plant electricity use can be offset by energy produced from 1009
sludge digestion [17]. Taking these values into account, total fossil fuel-generated emissions 1010
from conventional treatment would be on the order of 0.224 kg CO 2e/m
3
DWW. Total fugitive 1011
CH 4 emissions were also included at an average of 0.095 kg CO 2e/m
3
[59]. 1012
44

Assuming full-scale mainstream anaerobic treatment energy demands are comparable to those 1013
of conventional WWTPs before considering aeration requirements, it can be anticipated that 1014
their electricity consumption accounts for approximately 0.149 kg CO 2e/m
3
DWW. The amount 1015
of energy achievable from direct biogas recovery (headspace) was further calculated based on 1016
methane loss values extracted from Figure 3 at 25° C and 10° C (15 and 50 mg CH 4/L, 1017
respectively). Assuming a 95% conversion of incoming COD (430 mg/L) to CH 4, energy density of 1018
55.5 MJ/kg, conversion efficiency to electricity of 35%, and a CO 2 emission factor from electricity 1019
usage of 0.472 kg CO 2/kWh [158], it is estimated that anaerobic mainstream treatment 1020
electricity-associated GHG footprints could be reduced to below 0.02 kg CO 2e/m
3
at 10° C, while 1021
actually achieving energy positive operation at 25° C (-0.073 kg CO 2e/m
3
). Based on these 1022
calculations, it can be concluded that mainstream anaerobic treatment has the potential to 1023
more significantly offset electricity-associated GHG emissions when compared to conventional 1024
WWTPs with anaerobic digestion (0.224 kg CO 2e/m
3
). However, without downstream CH 4 1025
recovery, anaerobic effluents would contribute GHG emissions of approximately 0.51 and 1.70 1026
kg CO 2e/m
3
at 25° C and 10° C, respectively (using CH 4 GWP of 34 and dissolved methane- 1027
temperature relationship obtained from Figure 3). 1028
N 2O emissions from the nitrogen removal process in WWTPs have been identified as the most 1029
widely varying and least predictable of GHG sources (ranging from 0 to 14.6% of incoming 1030
nitrogen). Nonetheless, an analysis of several representative full-scale studies of conventional 1031
anoxic-oxic processes revealed an average emission factor of 1.5% [45, 46, 50, 159-161], 1032
resulting in N 2O emissions of 0.281 kg CO 2e/m
3
DWW for conventional WWTPs (assuming 1033
influent of 20 mg N/L and N 2O GWP of 298). If nitritation coupled with anammox is employed as 1034
the nitrogen removal process for mainstream anaerobic treatment and an average nitrogen to 1035
N 2O ratio of 2.8% is used (estimated from full-scale nitration-anammox studies) [84, 85, 162, 1036
45

163], N 2O emissions for anaerobic treatment can be calculated as 0.529 kg CO 2e/m
3
WW. The 1037
relatively higher emissions observed for nitration-anammox have been attributed to a lack of 1038
process optimization for N 2O mitigation [25], which can likely be improved upon significantly in 1039
future research. 1040
Therefore, total GHG emissions from CAS WWTPs are significantly lower than mainstream 1041
anaerobic systems, even at 25° C (0.599 vs. 0.966 kg CO 2e/m
3
WW). Anaerobic treatment GHG 1042
footprints would likely be exacerbated at lower ambient temperatures, reaching up to 2.25 1043
CO 2e/m
3
at 10° C, if no effluent recovery was employed. As outlined in this review, however, 1044
emerging techniques for both nitrogen and dissolved CH 4 removal/recovery could effectively 1045
negate these outstanding issues. If, for example, energy-efficient dissolved CH 4 recovery is 1046
employed and comparable N 2O emissions are achieved through nitrogen removal processes 1047
optimization, mainstream anaerobic system GHG footprints would easily drop below those 1048
calculated for current CAS WWTP processes. Further, recent research has implicated 1049
mainstream anaerobic effluents as likely to become acceptable for direct irrigation reuse from a 1050
microbial safety perspective [164, 165]. This application of nutrient-rich treated effluents could 1051
negate the necessity of nitrogen removal, essentially allowing for the elimination of direct N 2O 1052
emissions from mainstream anaerobic treatment and such systems approaching carbon 1053
neutrality. 1054
2.6 Conclusions 1055
Existing literature on WWTP GHGs has reported broadly varying total emissions ranging from 1056
0.243 to 2.4 kg CO 2e/m
3
WW. A unified and comprehensive plant-wide approach inclusive of all 1057
direct and indirect emissions is necessary for accurate WWTP carbon footprint interpretation. 1058
Overall, the most significant obstacle facing GHG mitigation in CAS WWTPs is associated with 1059
46

understanding N 2O generation, whereas for the sustainability of mainstream anaerobic 1060
wastewater treatment, dissolved CH 4 emissions are of greatest concern. Other specific 1061
observations of this review are summarized as follows: 1062
• N 2O emissions are both dominant and highly variable in conventional aerobic-based 1063
WWTPs (0 to 95% of N for lab-scale and 0 to 14.6% of N for full-scale), with several 1064
critical factors influencing this variability including: DO, pH, nitrite, carbon source 1065
availability and ammonium loading. 1066
• More research is specifically needed in elucidating the pathways involved in N 2O 1067
formation (i.e., nitrifier denitrification, heterotrophic denitrification, and hydroxylamine 1068
oxidation) at different operational conditions, which can be then used to correlate 1069
practical mitigation strategies with specific processes and configurations. 1070
• Dissolved CH 4 contributions in mainstream anaerobic treatment account for the 1071
majority of GHG emissions. Anaerobic system GHG emissions are inversely correlated 1072
with operational temperature due to increasing CH 4 solubility and supersaturation 1073
(Section 3.1). 1074
• Analysis of membrane contactors for physical dissolved CH 4 removal showed that for 1075
sweep gas systems, gas to liquid (G/L) ratio is a critical parameter influencing CH 4 1076
removal efficiency and off-gas CH 4 concentration. Vacuum driven membrane contactors, 1077
although capable of high-quality gas recovery, require operation at TMPs below 50 kPa 1078
to achieve energy neutrality. 1079
• Several emerging methods for dissolved CH 4 recovery are likely to play significant roles 1080
in future management of dissolved CH 4. For example, DAMO combined with anammox 1081
could allow for the simultaneous removal of both nitrogen and CH 4 from anaerobic 1082
effluents. 1083
47

To significantly reduce WWTP GHG emissions, future research on CAS must focus on N 2O 1084
management strategies to minimize emissions. For anaerobic systems, both efficient CH 4 and 1085
nitrogen resource recovery must be achieved without introducing incidental increases in N 2O 1086
generation. The accomplishment of this goal appears to be within reach, given the prospects of 1087
emerging CH 4 recovery processes and likelihood of effluent reuse. 1088
1089
1090
1091
1092
1093
1094
1095
1096
1097
1098
1099
1100
48

Chapter 3. Methane-driven microbial fuel cells recover energy and 1101
mitigate dissolved methane emissions from anaerobic effluents 1102
Abstract 1103
The effluents of mainstream anaerobic treatment processes such as anaerobic membrane 1104
bioreactors (AnMBRs) contain dissolved methane that represents a large fraction of the 1105
available energy (approximately 50% at 15° C) and a significant greenhouse gas (GHG) emission if 1106
released to the atmosphere. Microbial fuel cells (MFCs), which rely on exoelectrogenic 1107
microorganisms to generate electricity from organic or inorganic matter, could be used to 1108
recover energy and prevent GHG emissions from dissolved methane. Two replicate air-cathode, 1109
single-chamber MFCs and one dual-chamber MFC were constructed and operated in continuous 1110
mode on a synthetic, methane-saturated media at 20° C and hydraulic retention times of 4, 8, 1111
and 16 h. Up to 85% dissolved methane removal was achieved resulting in generation of 0.55 ± 1112
0.06 V. Geobacter, a common exoelectrogen, and methanotrophs were identified in anode 1113
biofilm samples by Illumina sequencing targeting the 16S ribosomal RNA (rRNA) and 16S rRNA 1114
gene. Activity quantification of key microbial populations via reverse transcription-quantitative 1115
polymerase chain reaction (RT-qPCR) indicated that methane removal and voltage production 1116
result from a consortium of aerobic methanotrophs enriched in a cathode biofilm that produce 1117
intermediate metabolites (e.g., formate and acetate) that serve as substrates for Geobacter in 1118
the anode biofilm. 1119
3.1 Introduction 1120
Anaerobic membrane bioreactors (AnMBRs) are a promising low-strength wastewater 1121
treatment technology combining anaerobic biological treatment with membrane separation in a 1122
49

single unit process. Wastewater organics are effectively converted into methane-rich biogas in 1123
AnMBRs which can subsequently be converted to electricity and heat. Further, AnMBRs have 1124
low sludge production, limited energy requirement assuming membrane fouling control is done 1125
efficiently, and similar carbon removal as aerobic processes at a range of operational 1126
temperatures [6, 166]. However, AnMBRs produce an effluent saturated or supersaturated with 1127
dissolved methane [114] dramatically increasing greenhouse gas (GHG) emissions relative to 1128
aerobic treatment processes if this methane is released to the atmosphere [1]. For example, 1129
approximately 50% of total produced methane remained in the dissolved form at 15° C when 1130
treating a synthetic wastewater representative of domestic wastewater [5] and at temperatures 1131
<5° C, essentially all produced methane was dissolved in the effluent [6]. Although wastewater 1132
treatment plants are not regulated on GHG emissions, preventing these emissions and ideally 1133
recovering the embedded energy is necessary before AnMBRs or other mainstream anaerobic 1134
treatment processes can be responsibly implemented at the full scale. 1135
Several studies have reported on recovery or removal of dissolved methane from anaerobic 1136
effluents with varying success [102, 105, 106, 115, 116, 119, 124, 167]. Giménez et al. [106] 1137
applied biogas-assisted mixing to sparge anaerobic effluents to avoid supersaturation of 1138
methane. However, methane losses of 42.6% and 46.6% at 30° C and 20° C, respectively, were 1139
still observed as biogas-assisted mixing only created equilibrium between gaseous and dissolved 1140
forms of methane. Limiting supersaturation may enhance energy recovery via gaseous methane 1141
production but is insufficient to mitigate GHG emissions. To achieve similar GHG emissions to 1142
aerobic treatment, approximately 90% or more of dissolved methane needs to be removed from 1143
AnMBR effluent [1]. To achieve dissolved methane removal beyond limiting supersaturation, 1144
membrane contactors in sweep gas desorption or vacuum degassing mode have been proposed 1145
[102, 105, 115, 116, 119, 167]. However, energy demands for these proposed processes remain 1146
50

high, often an order of magnitude greater than energy available in recovered gaseous methane. 1147
Biological treatment is another strategy for dissolved methane management with aerobic 1148
methanotrophy and anaerobic oxidation of methane (AOM) being the most prevalent microbial 1149
pathways. AOM coupled with sulfate reduction or denitrification, referred to as denitrifying 1150
anaerobic methane oxidation (DAMO), has been reported extensively [168-172]. In DAMO, 1151
dissolved methane serves as the electron donor for denitrification via reduction of nitrate or 1152
nitrite and could thus potentially be used in a downstream nitrogen removal process [170, 171]. 1153
It is important to note that AnMBRs and other mainstream anaerobic processes do not provide 1154
direct nutrient removal. Downstream DAMO processes would first require nitritation or 1155
nitrification of effluent ammonia which could be challenging without stripping dissolved 1156
methane into the gas phase or oxidizing it via aerobic methanotrophs, which may have higher 1157
oxygen affinity than DAMO microorganisms [173]. Aerobic methanotrophy is a multi-step 1158
metabolic pathway requiring oxygen to initially oxidize methane to methanol mediated by 1159
soluble/particulate monooxygenase. After, methanol is either oxidized to formaldehyde, 1160
formate, and carbon dioxide, or absorbed in the form of formaldehyde for biosynthesis of 1161
cellular components [150, 174, 175]. Under low oxygen availability, aerobic methanotrophs 1162
have been reported to release intermediate metabolites such as methanol, formaldehyde, and 1163
formate rather than completely mineralizing methane to carbon dioxide [176-178]. Matsuura et 1164
al. [124] reported using aerobic methanotrophy to oxidize dissolved methane in down-flow 1165
hanging sponge reactors, similar to trickling filters. Although these systems efficiently prevent 1166
GHG emissions - removal efficiencies of 70% in a one-stage system and up to 99% in a two-stage 1167
system - they fail to recover energy and require energy input to supply oxygen. Potentially more 1168
promising, methanotrophs could be used to recover biomaterials such as polyhydroxyalkanoate 1169
(PHA) [151], but this line of research requires further investigation before it is a realistic strategy 1170
51

for dissolved methane management in anaerobic effluents. Despite the various physical and 1171
biological approaches evaluated to date for dissolved methane management, feasibility remains 1172
questionable and effluent dissolved methane persists as a barrier to mainstream anaerobic 1173
treatment. 1174
Microbial fuel cells (MFCs), bioelectrochemical systems where microorganisms directly deposit 1175
electrons to an anode during oxidation of organic or inorganic compounds [8-10], have 1176
traditionally been studied for energy recovery directly from domestic wastewater and thus can 1177
be seen as a direct competitor to mainstream anaerobic treatment [8, 179]. Single-chamber, air- 1178
cathode MFC designs allow for a system with essentially no energy demands due to oxygen 1179
being provided passively at the atmosphere-exposed cathode [9]. MFCs have been 1180
demonstrated with anaerobic post-treatment [109, 180, 181] as a coupled approach for 1181
domestic wastewater treatment. For example, Ren et al. [109] demonstrated an MFC coupled 1182
with an anaerobic fluidized bed membrane bioreactor and were able to produce a high-quality 1183
effluent with minimal energy requirements. However, in such configurations, residual chemical 1184
oxygen demand (COD) from the MFC does not provide sufficient organics for the anaerobic 1185
treatment system to produce gaseous methane and therefore recover energy. The anaerobic 1186
treatment system essentially acts to convert residual COD from the MFC into dissolved 1187
methane. We propose to invert this configuration and operate an MFC downstream of an 1188
anaerobic treatment system to recover energy and prevent GHG emissions from dissolved 1189
methane. This configuration eliminates any risk of dissolved methane being stripped and 1190
released from anaerobic effluents. Further, MFCs are an attached growth process and thus 1191
biological post-treatment of AnMBR effluents may be possible without subsequent solids 1192
removal. A methane-driven MFC could also be used to power underwater sensors in marine 1193
52

applications [141, 182] or to convert gaseous methane (i.e., natural gas) into electricity to 1194
reduce or eliminate methane leaks that occur during transportation and storage [142]. 1195
An early 1965 study by van Hees [11] reported using methane as the only organic in a two- 1196
chamber MFC that produced 0.5-0.6v using a pure culture of Pseudomonas methanica. This 1197
work suggested that a methanotroph could act as an exoelectrogen similar to Geobacter, a 1198
common exoelectrogen in MFCs. However, no reports of exoelectrogenic methanotrophs exist 1199
outside of this study. It is important to note that Geobacter does not have the metabolic 1200
pathways for direct methane oxidation. A recent study reported a two-chamber MFC using 1201
methane as the electron donor at the anode inoculated with DAMO-archaea [137]. A relatively 1202
low voltage production was reported but there was a correlation between voltage and methane 1203
addition to the anode chamber. The anode was enriched with DAMO-archaea and Geobacter 1204
suggesting that these populations may work together to deposit electrons on the anode. 1205
Recently, McAnulty et al. [138] demonstrated electricity production from methane in a two- 1206
chamber MFC by constructing an engineered archaeal strain capable of producing methyl- 1207
coenzyme M reductase to convert methane to acetate (i.e., reverse methanogenesis). With the 1208
aid of electron shuttles and methane-acclimated sludge containing Paracoccus denitrificans, 1209
Geobacter sulfurreducens was shown to generate electricity from methane via this synthetic 1210
microbial consortium. 1211
This study is the first to investigate the potential for air-cathode MFCs to be used as a post- 1212
treatment biotechnology for energy recovery and mitigation of GHG emissions from anaerobic 1213
effluents. The performance of bench-scale, methane-driven MFCs was evaluated using a 1214
synthetic media representative of AnMBR effluent produced during domestic wastewater 1215
treatment. Process performance was evaluated at hydraulic retention times (HRTs) of 16, 8, and 1216
4 h by measuring voltage production, dissolved methane removal efficiency, and other water 1217
53

quality parameters. High-throughput sequencing targeting 16S ribosomal RNA (rRNA) and 16 1218
rRNA genes and reverse transcription-quantitative polymerase chain reaction (RT-qPCR) were 1219
used to evaluate microbial community structure and activity of anode and cathode biofilms. 1220
3.2 Materials and methods 1221
3.2.1 MFC configurations 1222
Two replicate single-chamber, air-cathode MFCs with a 240 mL working volume were 1223
constructed based on the design of those reported in [183]. Three carbon brushes (Zoltek PX 35 1224
carbon fiber, Mill-Rose Company, Mentor, OH) with 2.5 cm diameter and 5 cm length were used 1225
as the anode. Carbon nanofibers were twisted on a titanium rod which provided current 1226
collection. The anodes were pretreated by soaking the brushes in a solution of ammonium 1227
peroxydisulfate (200 g /L) and concentrated sulfuric acid (100mL/L) for 15 min according to 1228
[184]. A carbon cloth (30% wet-proofing, carbon cloth CC4 wet proofed, Fuel Cell Earth, 1229
Woburn, MA) was pretreated as described in [185] for use as the cathode. Two pieces of fabric 1230
cloth (Amplitude EcoCloth, Contect, Inc., Spartanburg, SC) were integrated between the anode 1231
and cathode as a separator to prevent excess oxygen diffusion and potential short-circuiting. 1232
Reactors were operated with 1,000-ohm resistance in the external circuit. A data acquisition 1233
device (DI245, DATAQ, Akron, OH) was used to measure voltage every 16 s and subsequent data 1234
was trimmed to every 20 min for analysis. Reactors were located in a temperature-controlled 1235
incubator (Drosophila Incubator, Genesee Scientific, CA) and the synthetic wastewater media (SI 1236
Table 1) was placed in a refrigerated water bath (Lindberg/Blue M Shaking Water bath, Thermo 1237
Fisher Scientific, Waltham, MA) such that both maintained a temperature of 20° C. Initially, both 1238
replicate MFCs were filled with a media comprised of half primary effluent (120mL) from the 1239
Hyperion Wastewater Treatment Plant (Los Angeles, CA) and half synthetic media (120mL) 1240
54

consisting of acetate and nutrients to inoculate the MFCs. The inoculum was screened for the 1241
presence of Geobacter and methanotrophs via PCR and gel electrophoresis (methodology 1242
described below; SI Figure 2). After demonstrating stable and reproducible voltage production, 1243
both MFCs were operated on acetate containing synthetic media in batch mode to benchmark 1244
performance of our system against similar systems fed acetate reported on in the literature. 1245
Afterwards, the MFCs were operated in continuous mode on a synthetic anaerobic methane- 1246
saturated media with dissolved methane as the only organic (no acetate was in the media). 1247
Methane was dissolved in the media by vigorously purging with an 80% methane and 20% 1248
carbon dioxide gas mixture for 15 min. Media was kept in gas-tight containers and constantly 1249
bubbled with the gas mixture at 10-20 mL/min to assure influent dissolved methane 1250
concentration and prevent intrusion of atmospheric oxygen. The system was initially operated 1251
at an 8 h HRT which was later varied to evaluate performance at 16, 8, and 4 h HRTs. 1252
A two-chamber MFC was constructed using the same electrode materials as the single-chamber 1253
MFCs with glass reaction chambers (MFC 250.40.0, Adams & Chittenden Scientific Glass, 1254
Berkeley, CA). Inoculation and batch mode operation was performed similarly to the single- 1255
chamber MFCs. The anode chamber was continuously fed with methane containing media, 1256
while the cathode chamber was continuously sparged with air. A cation exchange membrane 1257
(CMI-7000, Membranes International Inc., Ringwood, NJ ) was placed between the anode and 1258
cathode chambers for proton exchange. 1259
Power density, 𝑃 𝑐𝑎𝑡
, defined as power generation normalized by the cathode area, was used to 1260
characterize MFC power production and compare with relevant MFCs reported in the literature: 1261
𝑃 𝑐𝑎𝑡
=
𝑉 2
𝑅 ⁄
𝐴 𝑐𝑎 𝑡 1262
Where V is the voltage production, R is the resistance, and 𝐴 𝑐𝑎𝑡
is the cathode area (60 cm
2
). 1263
55

Coulombic efficiency (CE) defines the percentage of electrons recovered in the form of 1264
electricity over total electrons originating from available organics. Dissolved methane can thus 1265
be converted to theoretical maximum current (TMC) and converted to CE with the actual 1266
current (AC). TMC and CE were calculated as follows: 1267
TMC =
𝐷 𝐶𝐻
4
𝐶 𝑏𝐹𝑄 𝑀𝑅
; CE =
𝐴𝐶
𝑇𝑀𝐶 1268
Where 𝐷 𝐶𝐻
4
the dissolved methane concentration (mg/L), C is the COD equivalent of methane (4 1269
g O 2/ g CH 4), b is the number of electrons transferred per mole of oxygen, F is Faraday’s 1270
constant, Q is the incoming flowrate, M is the molecular weight of oxygen, and R is the 1271
resistance in the external circuit (1000 Ω). 1272
3.2.2 Chemical assays 1273
pH and dissolved oxygen (DO) were measured by pH probe (SevenCompact pH/Ion S220, 1274
Mettler Toledo) and DO meter (SevenGo Duo pro SG98, Mettler Toledo), respectively. Influent 1275
and effluent dissolved methane concentrations were measured daily and influent DO was 1276
measured periodically to confirm that the influent remained anaerobic. Dissolved methane in 1277
the synthetic media and effluent was stripped out of liquid by vigorously shaking for 1 min in a 1278
gas-tight syringe in the presence of an equal volume of nitrogen. Then, 1 mL of the resulting 1279
methane and nitrogen mixture was injected into a Trace 1310 Gas Chromatograph (Thermo 1280
Fisher Scientific, Waltham, MA) equipped with a 30m x 0.53mm x 20µm TracePLOT TG-BOND Q 1281
column and flame ionization detector to quantify methane content. The instrument was 1282
operated in split injection mode, with the inlet temperature at 250° C, oven at a constant 150° C, 1283
and detector at 250° C. Dissolved methane concentration was then calculated based on the 1284
methane content using the ideal gas law: 1285
56

𝑐 𝐶𝐻
4
(mg / L )=
𝑃𝑉
𝑅𝑇
𝑀 𝐶𝐻
4
𝑉 𝑙 1286
Where 𝑉 𝑙 is the volume of influent or effluent sample, P is the atmospheric pressure (1 atm), V is 1287
the volume of methane stripped out of the liquid sample, R is the ideal gas constant, T is the 1288
operational temperature (20 ℃), and M
CH
4
is the molecular weight of methane. 1289
Volatile fatty acids (acetate, formate, propionate, butyrate, and valerate) and inorganic ions 1290
(chloride, phosphate, sulfate, and nitrate) were quantified by ion chromatrography (ICS 2100, 1291
Thermo Fisher Scientific, Waltham, MA) with a 2mm AS-11HC column (Dionex, Sunnyvale, CA). 1292
Samples were filtered by 0.2 µm Whatman filters (GE Healthcare Life Sciences, Pittsburgh, PA) 1293
and loaded into a temperature controlled autosampler at 4° C before being injected. The 1294
flowrate was set at 0.50 mL/min and KOH was used as eluent. The instrument was operated at 1295
an eluent flowrate of 0.5 mL/min with eluent KOH concentration at 1mM during the first 15 min 1296
and then linearly ramped to 60 mM until the end of the run (total run time of 28 min). Standards 1297
containing VFAs and inorganic ions were prepared and ran in triplicate at 1, 5, 10, and 50 mg/L. 1298
3.2.3 Microbial community structure and activity 1299
Anode and cathode biomass samples were collected periodically from bench-scale MFCs and 1300
immediately stored at -80° C. RNA samples were preserved in DNA/RNA Shield (Zymo Research, 1301
Irvine, CA). DNA was extracted from biomass by initially mixing with prebaked 0.1 mm diameter 1302
zirconium beads and lysis buffer, followed by three, 2-min bead beating steps (Mini-Beadbeater- 1303
24, BioSpec Products, Bartlesville, OK). Supernatant was taken and digested with proteinase K, 1304
followed by automated extraction via the Maxwell® Magnetic Particle Processor (Promega, 1305
Madison, WI) using Maxwell 16 LEV Blood DNA kits according to manufacturer’s instructions. 1306
RNA was extracted by three, 1-minute zirconium bead beating steps using lysis buffer and 1- 1307
thiolyglycerol from the Maxwell 16 LEV simplyRNA blood kit, followed by automatic extraction 1308
57

according to manufacturer’s instruction. DNA and RNA extracts were quantified by 1309
spectrophotometry (BioSpectrometer Fluorescence, Eppendorf, Germany). After that, DNA 1310
concentration was further quantified via Quant-iT
TM
PicoGreen dsDNA Assay (Thermo Fisher 1311
Scientific, Waltham, MA). RNA extracts went through additional treatment to remove DNA 1312
contamination using the Invitrogen DNA-free DNA removal kit (Thermo Fisher Scientific, 1313
Waltham, MA). PCR amplification of contaminating 16S rRNA genes and gel electrophoresis 1314
were performed to assure RNA purity (Supplementary Information (SI) Figure 2). RNA was then 1315
quantified using the Quant-iT
TM
RiboGreen RNA Assay kit (Thermo Fisher Scientific, Waltham, 1316
MA). After that, RNA samples were reverse transcribed to generate complementary single- 1317
stranded DNA (cDNA) using the GoScript Reverse Transcription system (Promega, Madison, WI). 1318
Illumina MiSeq sequencing targeting the V4 region of Bacteria and Archaea was performed by 1319
the Host Microbiome Initiative (University of Michigan, Ann Arbor, MI). A set of barcoded 1320
primers described in Kozich et al. [186] were used to sequence 16S rRNA and 16S rRNA genes. 1321
PCR reactions were conducted by the following reagent composition: primers at 500 nM, 10µL 1322
2X Accuprime buffer 11 (Invitrogen, Waltham, MA), 0.15 µL Accuprime HiFi TAQ, 0.5 ng 1323
template, and nuclease-free water in a total volume of 20 µL. Thermocycling included an initial 2 1324
min denaturation at 95° C, followed by 30 cycles of denaturing at 95° C for 20 s, annealing at 55° C 1325
for 15 s, and extension at 72° C for 5 min, with a final extension at 72° C for 5 min. SequalPrep 1326
Normalization Plate Kits (Life Technologies, Grand Island, NY) were used to pool amplicons by 1327
equal mass. Multiplexed amplicons were sequenced via Illumina MiSeq using the MiSeq Reagent 1328
Kit V2 (2x250 bp reads) at the University of Michigan. Mothur [186] was used to analyze 1329
sequencing results following the Schloss MiSeq SOP. The UCHIME algorithm was applied to 1330
perform chimera removal. After quality filtering, an average of 23,431 ± 3,364 paired-end 1331
sequences per sample were obtained, with minimum and maximum sequences of 21,627 and 1332
58

28,474. Sequences were aligned with the SILVA reference database [187], and were subsampled 1333
to 21,627 sequences before conducting further analysis with operational taxonomic unit (OTU)- 1334
based clustering (average neighbor algorithm at 3% cutoff) of subsampled sequences. 1335
Primers for RT-qPCR were used to target Geobacter (Geo564F (AAGCGTTGTTCGGAWTTAT) and 1336
Geo840R (GGCACTGCAGGGGTCAATA)) [188], particulate methane monooxygenase (pmoA; a 1337
functional gene of methanotrophs, A189F (GGNGACTGGGACTTCTGG) and mb661R 1338
(CCGGMGCAACGTCYTTACC)) [189], and the 16S rRNA V4 region of Bacteria and Archaea (515F 1339
(GTGCCAGCMGCCGCGGTAA) and 806R (GGACTACHVGGGTWTCTAAT)) [186] in RNA extracts 1340
from cathode and anode biofilm biomass. RT-qPCR was conducted in 15uL reactions on a 1341
LightCycler® 96 (Roche Molecular Systems, Inc.) with 7.5 µL qPCR master mix (Fast Plus 1342
EvaGreen® qPCR master mix, Biotium), 0.3µM for forward and reverse primers, 1µL cDNA 1343
template, and DNase/RNase-free water. RT-qPCR standards were generated using the 1344
aforementioned primers in 10 µL reactions on a Mastercycler
®
Nexus thermocycler (Eppendorf, 1345
Germany) with 5 µL PCR master mix (NEBNext® Q5® Hot Start HiFi PCR Master Mix, New 1346
England BioLabs® Inc.), 1 µM forward and reverse primers, and 2 µL cDNA template pooled 1347
from reverse transcribed RNA extracts of samples from the bench-scale MFCs. Isolated PCR 1348
products were run on an agarose gel, purified with the Wizard® SV Gel and PCR Clean-Up 1349
System (Promega, Madison, WI), and quantified with Quant-iT™ PicoGreen (Thermo Fisher 1350
Scientific, Waltham, MA), after which serial dilutions of 10
8
to 10
1
copies of each gene were 1351
prepared. PCR and RT-qPCR temperature cycling programs are shown in SI Table 2. 1352
59

3.3 Results and discussion 1353
3.3.1 Preliminary findings suggest methane removal and voltage production driven by co- 1354
culture of methanotrophs and Geobacter 1355
The replicate bench-scale MFCs (Reactor A and B) were first fed a 1,000 mg/L acetate containing 1356
media to enrich for exoelectrogens, particularly Geobacter, from the primary effluent inoculum. 1357
Stable and repeatable voltage production was observed during two batch runs (SI Figure 3) 1358
suggesting successful acclimation of Geobacter on the anode. The replicate MFCs were then 1359
switched to continuous operation with saturated methane containing media (no acetate 1360
present). A voltage plateau of 0.5 to 0.6 V was recorded which is in agreement with results 1361
commonly obtained for similar MFCs operated on acetate or domestic wastewater [183, 190]. 1362
During 50 days of operation, voltage production averaged 0.476 ± 0.122 V and 0.334 ± 0.117 V 1363
for Reactor A and B, respectively (Figure 4a). Dissolved methane removal efficiency averaged 1364
28.3 ± 9.7% and 24.6 ± 8.6%, which resulted in a CE of 4.53 ± 1.23% and 2.85 ± 0.17% for Reactor 1365
A and B, respectively (Figure 4b). Low CE was attributed to low dissolved methane removal 1366
efficiency and relatively low voltage production coupled with a high dissolved methane loading. 1367
60

1368
Figure 4. 1st run while single-chamber MFCs were operated in continuous mode on methane 1369
containing media (a) Preliminary data on voltage production (b) dissolved methane removal 1370
efficiency and Coulombic efficiency (%) from replicate air-cathode single-chamber MFCs 1371
1372
Microbial community structure and activity data derived from high-throughput sequencing 1373
indicated high activity of Geobacter and aerobic methanotrophs in the anode biofilm, suggesting 1374
enrichment of these populations was necessary to convert dissolved methane to electrons. 1375
Relative abundance of Geobacter at the anode was 23.1% and 20.9% in Reactor A and B, 1376
61

respectively (Figure 5). Relative activity of Geobacter was more variable at 41.9% and 6.31% in 1377
Reactor A and B, respectively. The high variability in relative activity may have resulted in 1378
differences in performance immediately prior to sampling or our sampling protocols since RNA 1379
has a relatively short half-life. An OTU classifying within the family Bradyhizobiaceae comprised 1380
1.00% and 2.65% relative abundance and 1.51% and 7.28% relative activity in Reactor A and B, 1381
respectively. A representative sequence from this OTU demonstrated high identity (94%) with 1382
Rhodoseudomonas palustris, a population isolated from an MFC capable of producing a higher 1383
power density than mixed microbial communities [191]. Aerobic methanotrophs in Reactor A 1384
and B anode biofilms comprised 2.50% and 2.96% of relative abundance and 4.10% and 5.11% 1385
of relative activity, respectively. Aerobic methanotrophs classified as Methylomonas methanica, 1386
a species previously classified as Pseudomonas methanica and used extensively as a model 1387
methanotroph [192]. This same population was used in the pure culture methane-driven MFC 1388
reported previously [11]. These preliminary sequencing results suggest that oxygen diffusion 1389
through the cathode fueled initial methane oxidation to intermediate metabolites, likely 1390
formate, that could be converted to electrons by Geobacter. Thus far, there have been no 1391
reports of Geobacter directly metabolizing other potential intermediates in aerobic 1392
methanotrophy (i.e., methanol or formaldehyde). However, it is also possible that acetogens 1393
converted intermediate metabolites of aerobic methanotrophy to acetate via acetogenesis. 1394
Therefore, Geobacter at the anode may have converted both formate and acetate to electrons. 1395
Trace amounts of acetate in the reactor effluent were occasionally detected (data not shown), 1396
suggesting acetogenesis from intermediate metabolites of aerobic methanotrophy did indeed 1397
occur. Despite the previous findings by van Hees [11], it is unlikely that a methanotroph was 1398
acting as an exoelectrogen in our MFCs given the relatively low activity of methanotrophs at the 1399
anode and high activity of Geobacter. A limitation of this preliminary sequencing work is that 1400
62

only the anode biofilm was sampled. A cathode biofilm was also likely present and may have 1401
contained significant activity of methanotrophs given that DO concentrations within the reactor 1402
are highest at the cathode due to diffusion from the atmosphere. In later work, both the anode 1403
and cathode biofilms were characterized. 1404
1405
Figure 5. Relative abundance and relative activity based on 16S rRNA sequencing of single- 1406
chamber MFCs anodic biomass sample identified to genus level 1407
1408
To evaluate the impact of DO on MFC performance in more controlled experimentation, we 1409
operated a two-chamber MFC with influent containing varying DO concentrations of 0, 0.5, 1, 1410
and 2 mg/L by purging influent media with pure methane until the desired DO was achieved (SI 1411
Figure 4). We observed a similar positive correlation between voltage production and influent 1412
DO concentration as was reported in a U.S. patent on methane powered MFCs [141] and a 1413
recent study on DAMO MFC [137]. Initially, voltage at 0 mg/L and 0.5 mg/L DO stabilized around 1414
0.0957 ± 0.0356 V and 0.0784 ± 0.0503 V over 20 days of operation. After that, the MFC was 1415
63

switched to 1 mg/L-DO media for 2 days and 2 mg/L-DO media for more than 1 week. Voltage 1416
increased to 0.272 ± 0.100 V and 0.370 ± 0.026 V with peak voltage surging to approximately 0.4 1417
V. DO was then reduced to 0.5 mg/L and voltage decreased to 0.199 ± 0.067 V. The correlation 1418
between influent DO concentration with voltage production confirmed our hypothesis that DO 1419
impacts aerobic methanotrophy and intermediate metabolite availability for exoelectrogens. 1420
This motivated further evaluation of HRT in the single-chamber MFCs, which directly controls 1421
oxygen availability from diffusion through the cathode. 1422
3.3.2 Dissolved methane removal efficiency strongly correlated with HRT 1423
Both replicate MFCs were operated at HRTs of 4, 8, and 16 h to evaluate a potential correlation 1424
with dissolved methane removal due to increased oxygen diffusion relative to dissolved 1425
methane loading at higher HRTs. Oxygen diffusion through the cathode is constant over time, 1426
and thus increasing HRT increases oxygen availability relative to influent methane loading. At an 1427
HRT of 16 h, Reactor A and B produced 0.61 ± 0.01 and 0.51 ± 0.06 V, respectively (Figure 6a). 1428
When both MFCs were reduced to an HRT of 8 h, Reactor A voltage production remained almost 1429
the same at 0.59 ± 0.07 V, whereas Reactor B voltage deceased to 0.33 ± 0.08 V. At an HRT of 4 1430
h, Reactor A and B voltage production decreased significantly to 0.11 ± 0.04 V and 0.30 ± 0.05 V, 1431
respectively. Occasional performance irregularities occurred (e.g., voltage decrease or high 1432
variability), which were attributed to factors such as biofouling on the cathode, pump system 1433
malfunction, and air intrusion into the anaerobic media container during media replacement, 1434
etc. One concern is that dissolved methane could be removed from the system via diffusion out 1435
of the reactor chamber through the separator and cathode. To evaluate this concern, a small 1436
chamber adjacent to the cathode was sealed from the outside environment to prevent gas 1437
exchange. Gaseous methane concentrations within this chamber were quantified for Reactor A 1438
and B over time (4, 8, and 16 h after sealing the chamber). Gaseous methane concentrations in 1439
64

the chamber correlated linearly with time (R
2
=0.944) with a maximum concentration after 16 h 1440
representing 1.61 and 1.71% of the total methane loading for Reactor A and B, respectively. This 1441
indicates that fugitive methane loss from diffusion through the cathode was insignificant during 1442
operation and that the vast majority of removal was via biotic pathways. 1443
1444
Figure 6. (a) Voltage production while single-chamber MFCs were operated in continuous mode 1445
on methane containing medium under varying HRTs; (b) Dissolved methane removal efficiency 1446
and Coulombic efficiency (%) while single-chamber MFCs were operated in continuous mode on 1447
methane containing medium under varying HRTs 1448
1449
65

Dissolved methane removal efficiency at a 16 h HRT was high and consistent for Reactor A and 1450
B, 84.6 ± 0.1% and 83.9 ± 0.03%, respectively (Figure 6b). Removal decreased but remained 1451
consistent for Reactor A and B at an 8 h HRT, 61.8 ± 0.1% & 62.7 ± 0.1%, respectively. Removal 1452
efficiency was higher than in preliminary operation (Figure 4b) presumably due to adaptation of 1453
the microbial community over time. The consistency in dissolved methane removal efficiency 1454
but not voltage production at this HRT potentially suggests more instability in Geobacter activity 1455
between reactors. Dissolved methane removal efficiency plummeted at a 4 h HRT to 28.2 ± 0.1% 1456
and 33.8 ± 0.04% for Reactor A and B, respectively. It is likely that oxygen diffusion into the MFC 1457
chamber was too low at such a short HRT to maintain performance given the high dissolved 1458
methane loading. According to Cheng et al. [185], oxygen diffusion through the cathode of MFCs 1459
is governed by a diffusion coefficient (D = 0.000322 cm2/s), the gradient difference in oxygen 1460
concentration outside and inside MFCs (7.8 mg/L, assumed saturation concentration of oxygen 1461
in water), and cathode area (A = 60 cm2), resulting in 0.258, 0.129, and 0.0646 mg O 2/mL 1462
diffusion for 16, 8, and 4 h HRTs, respectively (SI Table 3). Therefore, oxygen availability per 1463
influent flow scales linearly with HRT. Dissolved methane removal on a mass/time basis 1464
increased as HRT decreased but was relatively similar between 4 and 8 h HRTs (Table 3). 1465
CE and power density for Reactor A and B, calculated based on dissolved methane removal 1466
efficiency, voltage production, flowrate, and cathode area, decreased as HRT decreased, ranging 1467
from 17.7% to 0.888% and 62.0 mW/m
2
to 2.11 mW/m
2
, respectively (Table 3). Low power 1468
density was likely attributable to insufficient cathode area (60 cm
2
). Further, nonoptimal 1469
distance between the cathode and anode may have increased internal resistance [143]. 1470
Therefore, at the longest HRT, relatively low methane loading (1.03 g/d) and high removal 1471
efficiency (84.6%) enabled the greatest electrical energy recovery efficiency (17.7%). This CE is 1472
similar to those observed in similar single-chamber, air-cathode MFCs operated on other 1473
66

substrates (e.g., glucose or domestic wastewater) [190, 193]. CE for two-chamber systems, 1474
where oxygen diffusion into the anode chamber is limited, are typically significantly higher than 1475
single-chamber systems. However, this benefit is offset by energy requirements for aeration in 1476
the cathode chamber. It is important to note that CE here is also likely reduced by electrons lost 1477
in the initial steps of methanotrophy required to generate metabolites for Geobacter. 1478
1479
1480
Table 3. Average and standard deviation of influent dissolved methane, average methane 1481
loading per cathode area, average and standard deviation of methane removal, average 1482
methane removal per cathode area, average and standard deviation of voltage production, 1483
average power density, and Coulombic efficiency at HRTs of 16, 8, and 4 h for Reactor A and B. 1484
1485
HRT
(hour)
Influent dissolved
methane (mg/L)
Methane
loading per
cathode area
(g/d/m
2
)
Methane
removal
Methane
removal per
cathode area
(g/d/m
2
)
Voltage
production
(V)
Power
density
(mW/m
2
)
Coulombic
efficiency
16
Reactor A 17.1 ± 1.0 1.03 84.6 ± 0.06% 0.868 0.610 ± 0.006 62.0 17.7%
Reactor B 17.1 ± 1.0 1.03 83.9 ± 0.03% 0.861 0.506 ± 0.063 42.7 14.7%
8
Reactor A 17.1 ± 0.9 2.05 61.8 ± 0.05% 1.27 0.591 ± 0.074 58.2 8.60%
Reactor B 17.1 ± 0.9 2.05 62.7 ± 0.07% 1.29 0.331 ± 0.080 18.3 4.81%
4
Reactor A 15.8 ± 0.5 3.78 28.2 ± 0.05% 1.07 0.112 ± 0.037 2.11 0.888%
Reactor B 15.0 ± 0.7 3.60 33.8 ± 0.04% 1.22 0.301 ± 0.052 15.1 2.49%
67

3.3.3 Geobacter and methanotroph activity was spatially distinct in MFCs 1486
Geobacter 16S rRNA and pmoA transcript copy number were normalized to 16S rRNA copy 1487
number to quantify relative activity of each population at anode and cathode sites of the bench- 1488
scale MFCs. At all HRTs, relative activity of Geobacter in the anode biofilm was approximately 1489
10
-1
Geobacter 16S rRNA copies/ total 16S rRNA copies, which was 2 to 3 magnitudes greater 1490
than Geobacter activity in the cathode biofilm (Figure 7a). Provided that Geobacter rely on the 1491
anode to deposit electrons, it is unsurprising that their activity was spatially distributed in this 1492
way. Conversely, methanotroph activity profiles indicated significantly greater activity at the 1493
cathode relative to the anode in most samples (3 to 4 magnitudes higher; Figure 7b). Unlike 1494
Geobacter, methanotrophs can theoretically be active at both the cathode and anode sites; 1495
however, their activity is likely highest at the cathode where DO concentrations are high. 1496
Recalling our preliminary sequencing data, Geobacter relative activity was 41.9% and 6.31% at 1497
the Reactor A and B anode, respectively, whereas methanotroph relative activity was 4.1% and 1498
5.11% at the Reactor A and B anode, respectively, corroborating the spatial distribution of these 1499
populations derived from RT-qPCR results. Relative activity of methanotrophs increased as HRT 1500
decreased likely due to the increased methane loading, providing additional substrate to 1501
methanotrophs. As mentioned above, dissolved methane removal on a mass/time basis 1502
increased at lower HRTs, confirming activity data provided via RT-qPCR. 1503
68

1504
Figure 7. (a) relative activity of Geobacter 16S rRNA gene copy number /16S rRNA gene copy 1505
number and (b) relative activity of pmoA transcript copy number /16S rRNA gene copy number at 1506
anode and cathode from single-chamber MFCs under 3 different HRTs 1507
1508
It is important to note that the use of 16S rRNA to infer microbial activity has limitations: (1) 16S 1509
rRNA copy number does not always perfectly correlate with microbial activity because it 1510
includes both grow and non-growth activities, (2) dormant microbes may be present that 1511
69

develop more 16S rRNA to reserve higher protein synthesis potential, and (3) the relationship 1512
between 16S rRNA copy number and microbial activity varies among different taxa [194]. 1513
Further, normalization of RT-qPCR results to 16S rRNA copy number only provides relative data. 1514
Normalizing to biomass extraction weight can circumvent this concern somewhat, but requires 1515
quantitative RNA extraction which can be challenging given matrix effects that influence 1516
extraction efficiency and can vary widely based on biomass source. Alternative approaches such 1517
as cell quantification via flow cytometry [195], use of internal standards of marker genes prior to 1518
extraction [196], or other methods could be useful to provide more accurate data regarding 1519
microbial activity profiles. The lack of a positive correlation between Geobacter and 1520
methanotroph activity via RT-qPCR as a function of HRT was presumably due to shifts in activity 1521
of other microorganisms (e.g., heterotrophic bacteria) which could have a large impact on 1522
activity ratios. It was challenging to normalize RT-qPCR results to biomass extraction weight for 1523
anode biofilm samples because anode biofilm biomass was mixed with carbon fibers and 1524
difficult to separate. Cathode biofilm biomass was unevenly distributed spatially with varying 1525
biofilm thickness apparent from visual observations. Therefore, it was challenging to normalize 1526
to cathode area. Despite these methodological limitations, the RT-qPCR strongly indicates a 1527
distinct spatial activity profile for Geobacter and methanotrophs in the methane-driven MFC. 1528
3.3.4 Geobacter scavenge methanotrophic metabolites enabling electron recovery from 1529
methane 1530
Several studies have reported aerobic methanotrophs excreting intermediate metabolites under 1531
oxygen limited conditions (e.g., methanol, formaldehyde, and formate) [176-178, 197, 198]. 1532
Methanotrophic intermediate metabolites may also be anaerobically fermented to produce 1533
organics such as acetate, lactate, and succinate [177, 178]. Based on previous studies [177, 198], 1534
aerobic methanotrophs likely yield 50% COD as intermediate metabolites under oxygen limited 1535
70

conditions and use the remaining COD derived from methane for cell synthesis and maintenance 1536
activities. Excreted metabolites in the MFCs may then be transported via diffusion to Geobacter 1537
in the anode biofilm. However, metabolites could also be scavenged by heterotrophic bacteria 1538
and oxidized to carbon dioxide via trace dissolved oxygen unconsumed by methanotrophs. 1539
Assuming formate and acetate are substrates for Geobacter, a theoretical energy balance can be 1540
derived using bioenergetics (SI Table 4). The portion of electrons transferred to the electron 1541
acceptor (anode) is thus 0.290 and 0.409 for formate and acetate, respectively. Combined with 1542
the observed dissolved methane removal efficiencies at each HRT, approximately 85%, 60%, and 1543
30% at 16 h, 8 h and 4 h, respectively, a theoretical CE of 14.9%, 10.5% and 5.25% for HRTs of 16 1544
h, 8 h, and 4 h, respectively, can be calculated (Figure 8). This value is comparable to our 1545
obtained experimental results (Table 3). 1546
1547
1548
Figure 8. Dissolved methane consumption pathways/ mass balance in single-chamber MFCs 1549
under 8 h HRTs 1550
71

1551
Three likely intermediate metabolites, methanol, formaldehyde, and formate, were added 1552
sequentially to the single-chamber MFC with spiked concentrations based on complete 1553
conversion of influent dissolved methane to each metabolite on a COD basis (SI Table 5). We 1554
elected to not add acetate during these experiments as we had previously demonstrated 1555
voltage production on acetate during inoculation. Prior to sequential addition of each 1556
metabolite, MFC influent flow was stopped to provide a baseline voltage when no organics were 1557
present. After methanol was spiked into the MFC, no significant change in voltage from the 1558
baseline was observed (Figure 9). After, formaldehyde was injected resulting in a slight voltage 1559
increase (from 0.0451 ± 0.0011 to 0.0876 ± 0.0204) lasting for approximately 2 days before 1560
voltage decreased to baseline. Finally, formate was added causing a surge in voltage to 0.546 V, 1561
respectively, which surpassed the peak voltage when previously operated on dissolved 1562
methane. Voltage decreased back to baseline after 2 days. The CE on formate was 76.1%. 1563
Therefore, formate and/or acetate was the likely intermediate metabolite consumed by 1564
Geobacter in our bench-scale systems. We propose using multiple lines of evidence that air- 1565
cathode MFCs can be powered solely on methane via a methanotroph-Geobacter interaction; 1566
methanotrophs in the cathode biofilm oxidize methane to formate which is transported via 1567
diffusion to the anode biofilm where Geobacter converts formate to electrons. 1568
72

1569
Figure 9. Voltage production over time with sequential addition of methanol, formaldehyde and 1570
formate into single-chamber MFC operated in batch mode 1571
1572
3.3.5 Methane-driven MFCs outcompete existing approaches for dissolved methane 1573
management 1574
MFCs are an emerging biotechnology with promise for energy recovery from organic waste 1575
streams. Single-chamber, air-cathode MFCs are particularly attractive given that they do not 1576
require energy intensive aeration or costly proton exchange membranes. The primary drawback 1577
is lower CE due to unavoidable oxygen diffusion into the reactor chamber via the cathode and 1578
issues with scalability as reviewed by [143]. Despite these limitations, MFCs may be attractive to 1579
manage dissolved methane in anaerobic effluents or to power sensors in marine applications 1580
(i.e., benthic microbial fuel cells [182]). Current approaches for dissolved methane management 1581
in anaerobic effluents are either energy intensive or fail to recover energy because they oxidize 1582
73

dissolved methane to carbon dioxide or recover a gas of insufficient methane content for energy 1583
recovery via cogeneration. Matsuura et al. [124] recovered 30% of influent dissolved methane 1584
using a two-stage down-flow hanging sponge system. Bandara et al. [102] achieved 22 ± 13% 1585
methane recovery using a vacuum degassing membrane module. Cookney et al. [115] had the 1586
highest methane recovery, 53%, via a sweep gas membrane contactor. However, these 1587
approaches have high energy demands due to relatively high air/liquid ratio requirements for 1588
hanging sponge systems and a high vacuum pressure requirement for degassing membranes. 1589
Moreover, dissolved methane that is stripped out of liquid is diluted with air in the hanging 1590
sponge system making the collected gas mixture non-reusable for energy recovery. With 1591
degassing membranes, coexisting dissolved gases such as nitrogen and carbon dioxide are also 1592
recovered significantly diluting methane in the off-gas [119]. Subsequent purification of 1593
recovered gas would further increase energy requirements and costs of dissolved methane 1594
energy recovery [199]. Considering that methane conversion efficiency to electricity via 1595
cogeneration is relatively low, less than 40%, electricity recovery from dissolved methane in the 1596
aforementioned studies was at most 12.0%, 8.80%, and 21.2%, respectively. Therefore, little 1597
energy was recovered and significant GHG emissions remain a concern. Further, this does not 1598
consider energy demands for these systems which likely exceed energy recovery. Although our 1599
experimental work indicated a rather low maximum of 17.7% conversion of dissolved methane 1600
to electricity, removal was significantly greater than previous studies (up to 85%; Table 4) 1601
resulting in a more substantial decrease in GHG emissions. Therefore, MFCs generally 1602
outperform existing technologies when considering both energy recovery and GHG emissions. 1603
74

1604
1605
1606
Table 4. Comparison of dissolved methane management approaches including dissolved 1607
methane removal, composition of recovered gas, and energy recovery as a percentage based on 1608
dissolved methane loading, recovery gas volume, and methane content. An efficiency of 40% was 1609
assumed for electricity recovery from collected methane using cogeneration. 1610
1611
System Operational condition
Dissolved
methane removal
Methane content
of recovered gas
Energy
conversion
efficiency
Hatamoto et al.
2010
Down-flow
hanging sponge
(DHS)
3.8 m
3
air/m
3
, 2 h HRT 95.0% None N/A
Matsuura et al. 2010 Two-stage DHS
0.25-0.375 m
3

air/m
3
/day first stage,
2.5 m
3
/ air/m
3
/day
second stage
76.8% (single-
stage); >99% (dual-
stage)
over 30% methane
in the recovered
gas
12.0%
Cookney et al. 2012
Sweep gas
membrane
contactor
Lowest liquid velocity
0.0033 m/s, coupled
with 0.85 L/min gas flow
72.0%
0.028 vol % in the
recovered gas
N/A
Cookney et al. 2016
Sweep gas
membrane
contactor
Sweep gas to liquid flow
ratio 0.034
98.0% 53.0% 21.2%
Bandara et al. 2011
Vaccum degassing
membrane module
Vaccum maintained at
50 kPa
68.0 ± 7.0% 22.0 ± 13.0% 8.80%
75

3.4 Conclusions 1612
Bench-scale MFCs treating a synthetic anaerobic effluent demonstrated up to 85% dissolved 1613
methane removal, 0.5 to 0.6 V generation, and a maximum CE of 17.7%. High-throughput 1614
sequencing of anode biofilm samples indicated high activity of Geobacter and methanotrophs, 1615
substantiating voltage production and suggesting a methanotroph-exoelectrogen interaction, 1616
with formate/acetate the likely intermediate metabolites. RT-qPCR results suggested that 1617
methane oxidation and Geobacter extracellular electron transfer occurred primarily at the 1618
cathode biofilm and anode biofilm, respectively. Therefore, oxygen diffusion and HRT 1619
(interconnected parameters) were correlated with methane removal efficiency and voltage 1620
production indicating that longer HRTs improve methane removal due to additional oxygen 1621
availability for methanotrophs. Future research using advanced methods (e.g.,
13
C labeled 1622
methane using RNA-stable isotopic probing or fluorescent in situ hybridization targeting 1623
methanotrophs and Geobacter) are necessary to elucidate the methanotroph-exoelectrogen 1624
interaction suggested here in a methane-driven MFC. This study demonstrated that MFCs are 1625
able to be solely powered by dissolved methane, which presents MFCs as a potentially 1626
promising technology for post-treatment of anaerobic effluents. 1627
1628
1629
1630
76

Chapter 4. Performance and microbial ecology of methane-driven 1631
microbial fuel cells at temperatures ranging from 25 to 5° C 1632
Abstract 1633
The effluent of mainstream anaerobic processes is saturated with dissolved methane, 1634
representing a lost energy source and potent greenhouse gas emission if left unmanaged. This 1635
study investigated the impact of operational temperature on methane-driven microbial fuel cells 1636
(MFCs) designed for continuous operation to mitigate dissolved methane emissions in anaerobic 1637
effluents. Two bench-scale, single-chamber MFCs were operated sequentially at 25, 20, 15, 10 1638
and 5° C. Voltage production from both MFCs ranged from approximately 0.463 to 0.512 V over 1639
1 kΩ resistance at temperatures 15° C, but abruptly dropped as temperature decreased to 10 1640
and 5° C, averaging just 0.156 and 0.190 V for the replicate systems. Dissolved methane removal 1641
efficiency remained relatively stable across all operational temperatures, ranging from 53.0% to 1642
63.6%. High-throughput sequencing of 16S rRNA genes and reverse transcription quantitative 1643
polymerase chain reaction indicated distinct distribution of methanotrophs (e.g., 1644
Methylomonas) and exoelectrogens (e.g., Geobacter) on the cathode and anode, respectively. 1645
Spearman’s rank correlation suggested that an indirect interaction between methanotrophs and 1646
exoelectrogens via fermentative bacteria (e.g., Acetobacterium) may play a role in system 1647
function. Notably, diversity of the anode microbial community was positively correlated with 1648
both voltage production and Coulombic efficiency, suggesting overall diversity, as opposed to 1649
abundance or activity of exoelectrogens, was the primary factor governing performance at 1650
varying temperatures. 1651
77

4.1 Introduction 1652
High-rate anaerobic biotechnologies such as anaerobic membrane bioreactors (AnMBRs) 1653
convert organic matter in wastewater to methane-rich biogas, a renewable energy resource, 1654
while significantly reducing sludge production and producing an effluent comparable to 1655
activated sludge processes [19]. However, anaerobic effluents, enriched with dissolved methane 1656
and other constituents [4, 173, 200], require additional treatment before being discharged into 1657
receiving waters or reused. If released to the environment without additional treatment, 1658
effluent dissolved methane could be released to the atmosphere as a potent greenhouse gas 1659
[4]. Microbial fuel cells (MFCs), bioelectrochemical systems that use electroactive microbes (i.e., 1660
exoelectrogens) to oxidize organic/inorganic matter and transfer electrons extracellularly [8-10], 1661
have recently been demonstrated to convert methane into electricity [137-140, 201, 202]. 1662
Systems reported in the literature to date feature Geobacter, an exoelectrogen, interacting with 1663
a microbial population able to utilize methane as an electron donor: (1) anaerobic 1664
methanotrophic archaea ANME-2d possibly via direct electron transfer [137]; (2) genetically 1665
engineered archaea performing reverse methanogenesis to acetate [138]; (3) non-ANME 1666
anaerobic oxidizers of methane producing acetate [201] or formate/hydrogen [202]; or (4) 1667
aerobic methanotrophic bacteria performing incomplete methane oxidation to formate [139] or 1668
methanol which is subsequently converted to acetate by fermentative bacteria [140]. In all 1669
cases, produced acetate and/or formate can then be used directly by Geobacter or other 1670
exoelectrogens, resulting in voltage production across an external circuit with applied 1671
resistance. 1672
Temperature is a critical operational parameter influencing performance of high-rate anaerobic 1673
processes [6]. Although the vast majority of engineered anaerobic systems are operated at 1674
mesophilic and thermophilic temperatures, mainstream anaerobic processes must be operated 1675
78

at ambient wastewater temperatures given the immense energy required to heat large 1676
volumetric flows of domestic wastewater relative to the energy recoverable via biogas 1677
production [108]. Although high-rate anaerobic processes have been implemented at the full 1678
scale in hot climates, substantial research efforts are now underway to develop processes 1679
suitable for temperate climates that feature seasonally low psychrophilic temperatures [19]. 1680
However, low temperatures exacerbate dissolved methane losses by increasing methane 1681
solubility, and potentially methane supersaturation, in anaerobic effluents [4, 6]. For example, 1682
Smith et al. evaluated AnMBRs at multiple psychrophilic temperatures ranging from 15 to 3° C 1683
and observed a strong negative correlation between temperature and effluent dissolved 1684
methane [6]. Though high COD removal was accomplished even at 3° C (86 ±4.0%), 7-fold 1685
supersaturation of dissolved methane was reported in the effluent, accounting for 72 ± 10% of 1686
influent COD. Such high methane losses not only originated from increased methane solubility 1687
at low temperature but were also attributed to an increase in membrane biofilm treatment. 1688
Given the strong negative correlation between operational temperature and effluent dissolved 1689
methane, downstream treatment to mitigate methane release is especially critical at 1690
psychrophilic conditions. 1691
Methane-driven MFCs reported to date have been operated at room temperature [202], 20° C 1692
[139], 30° C [138, 140, 203], and 37° C [137, 201]. Lower and temporally variable temperatures 1693
could significantly influence MFC performance metrics such as energy recovery [204-207] and 1694
startup time [208-210]. Multiple studies investigating MFC performance with non-methane 1695
substrates (e.g., acetate) have reported reduced power density at decreasing temperature, 1696
which ranged from a relatively minor decrease of 8.33% when switching from 32 to 20° C [211] 1697
to a significant decrease of 92.1% when switching from 35 to 4° C [206]. For instance, Cheng et 1698
al. reported a linear power density reduction of 33 ± 4 mW/° C in MFCs initially operated at 30° C 1699
79

and subsequently decreased to 4° C [209]. Conversely, some studies have demonstrated an 1700
increase in current [212, 213] and Coulombic efficiency (CE) [209, 210, 212] at lower 1701
temperatures. The leading hypothesis for higher CE is that greater organic matter is diverted to 1702
exoelectrogens at low temperatures, as competing heterotrophic bacteria (i.e., non- 1703
exoelectrogens) are severely suppressed by the temperature decrease [212]. Liu et al. observed 1704
an increase in power density in a two-chamber MFC transitioned from 25° C to 15° C and 1705
suggested that lower anodic resistance at 25° C reduced power output [213]. Several studies 1706
have also identified a negative correlation between temperature and startup time: 30 hours to 1707
144 hours at 30° C, 60 hours to 504 hours at 15 to 22° C [208-210], and no appreciable voltage 1708
over 80 days at 4 and 6° C [209, 210]. However, MFCs have been successfully transitioned to low 1709
temperature after initial startup at higher temperature [209]. Such phenomenon infer that 1710
exoelectrogens can sustain activity at low temperature if initially acclimated to elevated 1711
temperatures. Although our mechanistic understanding of extracellular electron transfer is still 1712
evolving, transfer rates are likely affected by temperature-dependent parameters such as 1713
cathode potential [211], anode potential/impedance due to anodic biofilm development [210], 1714
microbial community structure [207], electrolyte conductivity [214], electrochemical reaction 1715
rates [205], and others. To date, no study has investigated the impact of operational 1716
temperature on performance of methane-driven MFCs. 1717
We previously reported a microbial interaction between aerobic methanotroph Methylomonas 1718
in a cathode biofilm and Geobacter in an anode biofilm in methane-driven, air-cathode MFCs at 1719
20° C, where formate served as the electron shuttle between the two microbial populations 1720
[139]. Up to 85% dissolved methane removal was achieved, resulting in the generation of 0.55 ± 1721
0.06 V. Though aerobic methanotrophs have been reported as active in various cold ecosystems 1722
[215], exoelectrogens such as Geobacter sulfurreducens have optimal growth rate at mesophilic 1723
80

temperatures [216]. To evaluate methane-driven MFC performance and microbial community 1724
adaptation to varying operational temperatures, two replicate air-cathode MFCs were 1725
sequentially operated at 25, 20, 15, 10, and 5° C. Anode and cathode biofilm samples were 1726
obtained at the end of each operational temperature. Activity of methanotrophs and 1727
exoelectrogens were quantified via reverse transcription quantitative polymerase chain reaction 1728
(RT-qPCR), while microbial community structure was evaluated via high-throughput 16S rRNA 1729
gene sequencing. 1730
4.2 Materials and methods 1731
4.2.1 MFC configuration, operation, and monitoring 1732
Two replicate air-cathode MFCs with a 240 mL working volume were operated previously on a 1733
synthetic anaerobic effluent at 20° C [139]. The single-chamber design included three carbon 1734
fiber brushes with titanium rods (Zoltek PX 35 carbon fiber, Mill-Rose Company, Mentor, OH) 1735
serving as the anode. The carbon fibers were pretreated as described in [139]. The cathode was 1736
a carbon cloth (Fuel Cell Earth, Woburn, MA) with two layers of fabric cloth (Amplitude 1737
EcoCloth, Contect, Inc., Spartanburg, SC) applied between the cathode and anode chamber to 1738
limit excess oxygen diffusion and potential short-circuiting. Resistance was applied to the 1739
external circuit at 1000 ohm. 1740
The MFCs were initially inoculated with primary clarifier effluent obtained from the Hyperion 1741
Wastewater Treatment Plant (Los Angeles, CA), operated in batch mode with acetate as the sole 1742
electron donor, and then transitioned to continuous mode operation at a hydraulic retention 1743
time (HRT) of 8 h on a synthetic anaerobic effluent containing dissolved methane at 20° C [139]. 1744
We previously evaluated HRTs of 4, 8, and 16 h, but elected to operate both MFCs at 8 h here. 1745
The MFCs were operated in a temperature-controlled incubator (Drosophila Incubator, Genesee 1746
81

Scientific, CA). A synthetic anaerobic effluent (SI Table 8) was temperature-controlled by 1747
immersion in a water bath connected to a chiller (Lindberg/Blue M Shaking Water Bath, Thermo 1748
Fisher Scientific, Waltham, MA). The synthetic anaerobic effluent was first purged vigorously 1749
with a gaseous blend of 80% methane and 20% carbon dioxide by volume for 15 minutes to 1750
remove dissolved oxygen and saturate the media. During operation, the media was continuously 1751
sparged with the same gaseous blend of 80% methane and 20% carbon dioxide at 10 to 20 1752
mL/min to ensure anaerobic conditions were maintained. Both MFCs were initially operated at 1753
20° C until reaching stable performance (i.e., voltage production of 0.5 ± 0.1 V and dissolve 1754
methane removal efficiency of 60 ± 5%). The MFCs were subsequently subjected to five 1755
consecutive temperatures of 25, 20, 15, 10 and 5° C for at least 1 week at each temperature 1756
from day 54. 1757
Voltage was measured for both MFCs via a data acquisition device (DI245, DATAQ, Akron, OH). 1758
Dissolved methane concentrations were quantified by collecting influent and effluent samples in 1759
gas-tight syringes and equilibrating dissolved gases into an equal volume nitrogen headspace by 1760
vigorously shaking for 1 minute. The gas phase of the syringe was then injected into a 10 mL 1761
water filled sampling vial displacing the water. The resultant gas samples were analyzed by a gas 1762
chromatograph (Trace 1310, Thermo Fisher Scientific, Waltham, MA) equipped with a 1763
TracePLOT TG-BOND Q column (30m x 0.53mm x 20 µm) and a flame ionization detector. The 1764
instrument was operated at the following temperatures: inlet at 250° C, oven at 150° C, and 1765
detector at 250° C. One mL of sample was manually injected. Dissolved methane concentration 1766
was calculated as described previously [139]. Influent and effluent dissolved methane were 1767
monitored on a daily basis. 1768
82

4.2.2 Sequencing and RT-qPCR 1769
Anode and cathode biofilm samples were collected at the end of each operational temperature. 1770
Specifically, at least 1 mg anode/cathode biomass sample was collected at each temperature 1771
except at 20° C. Anode biofilm biomass together with the carbon brush was cut from the anode, 1772
while cathode biofilm biomass was scraped from the cathode/separator cloth on the surface 1773
facing the anode chamber. DNA and RNA were extracted using pre-baked 1 mm diameter 1774
zirconium beads with a bead beater (Mini-Beadbeater-24, BioSpec Products, Bartlesville, OK). 1775
The Maxwell LEV 16 DNA blood kit and simply RNA blood kit (Promega, Madison, WI) were used 1776
according to manufacturer’s instructions. For the RNA extracts, potential residual contaminating 1777
DNA was removed via the Invitrogen DNA-free DNA removal kit following manufacturer’s 1778
instruction (Thermo Fisher Scientific, Waltham, MA). DNA and RNA extracts were quantified via 1779
the Quant-iT
TM
PicoGreen dsDNA and RiboGreen RNA assay kits (Thermo Fisher Scientific, 1780
Waltham, MA) using spectrophotometry (BioSpectrometer Fluorescence, Eppendorf, Germany). 1781
After that, RNA was reverse transcribed to complementary DNA (Promega GoScript Reverse 1782
transcription system, Madison, WI) before conducting reverse transcription quantitative PCR in 1783
which Geobacter spp. 16S rRNA, particulate methane monooxygenase (pmoA, a functional gene 1784
of methanotrophs), and universal 16S rRNA copy number were quantified [139]. Specifically, 1785
three sets of primers: Geo564F (AAGCGTTGTTCGGAWTTAT) and Geo840R 1786
(GGCACTGCAGGGGTCAATA) [188], A189F (GGNGACTGGGACTTCTGG) and 1787
mb661R( CCGGMGCAACGTCYTTACC) [189], and 515F (GTGCCAGCMGCCGCGGTAA) and 806R 1788
(GGACTACHVGGGTWTCTAAT) [186] were used. Each qPCR reaction assay was 15 µL consisting 1789
of 7.5 µL qPCR master mix (Fast Plus EvaGreen® qPCR master mix, Biotium), 0.3 µM for forward 1790
and reverse primers, 1 µL cDNA template, and DNase/RNase-free water. Amplifications were 1791
performed on a LightCycler® 96 instrument (Roche Molecular Systems, Inc.). RT-qPCR standards 1792
83

were prepared as described previously [139]. Microbial community structure was evaluated via 1793
Illumina MiSeq paired-end sequencing targeting the V4 region of the 16S rRNA genes. Library 1794
preparation and sequencing was done by the Microbial Systems Molecular Biology Laboratory 1795
(MSMBL; University of Michigan, Ann Arbor, MI). 1796
4.2.3 Bioinformatics 1797
Mothur was used to process Illumina Miseq sequencing data to generate operational taxonomic 1798
units (OTUs) [186]. After quality filtering, alignment with SILVA reference database [187], and 1799
chimera removal by the VSEARCH algorithm [217], an average of 54796 ± 19408 sequence reads 1800
per sample were generated, with minimum and maximum sequences of 22931 and 100106. 1801
Next, 22931 reads were randomly subsampled from each sample before conducting OTU-based 1802
clustering analysis using the 3% average neighbor cutoff resulting in 1990 distinct OTUs across 1803
all samples and an average of 340 ± 31 OTUs per sample. Microbial community dissimilarity 1804
represented by thetaYC [218] was visualized via non-metric multidimensional scaling (NMDS) . 1805
OTUs were binned into phylotypes based on their taxonomic classification [186] at the genus 1806
level, resulting in 311 phylotypes across all samples. To characterize microbial propensity for 1807
colonization on the anode or cathode, the LEfSe tool was applied to the phylotype data, where a 1808
linear discriminant analysis (LDA) score >2 was deemed significant [219]. Further, Spearman's 1809
rank correlation [220] was calculated between phylotypes (1) within the anode microbial 1810
community, (2) within the cathode microbial community, and (3) between the anode and 1811
cathode communities to evaluate interactions between populations in both biofilms. Microbial 1812
network analysis plots were generated based on Spearman’s rank correlation coefficients of 1813
probability < 0.05. 1814
84

4.3 Results and Discussion 1815
4.3.1 Voltage production abruptly decreased at 10° C 1816
After both replicate MFCs demonstrated stable voltage production on the synthetic anaerobic 1817
effluent at 20° C, operational temperature for both MFCs was increased to 25° C. MFC A 1818
generated a higher voltage, averaging 0.603 ± 0.017 V, relative to MFC B, averaging 0.463 ± 1819
0.093 V (Figure 10B). MFC A voltage decreased by 11.4% when operational temperature was 1820
decreased to 20° C (0.534 ± 0.038 V) and decreased further by 4.12% at 15° C (0.512 ± 0.108V). 1821
MFC B exhibited higher variability in voltage production at 25° C, but voltage production became 1822
more stable at 20 and 15° C resulting in an average voltage of 0.536 ± 0.073 and 0.569 ± 0.054 V, 1823
respectively. Therefore, MFC B voltage production increased by 15.8% when temperature was 1824
decreased from 25 to 20° C, and by an additional 6.16% from 20 to 15° C. Variability in MFC B 1825
voltage production may have originated from instability in the microbial community. Notably, 1826
both MFCs experienced an abrupt decrease in voltage at 10° C, decreasing to 0.196 ± 0.040 and 1827
0.261 ± 0.093 V for MFC A and B, respectively. As temperature was further decreased to 5° C, 1828
voltage production continued to decrease to 0.156 ± 0.039 V and 0.190 ± 0.066 V for MFC A and 1829
B, respectively. Overall, voltage production decreased by 51.2% to 64.3% at 10° C, and further 1830
decreased by 19.2% to 24.6% at 5° C when compared with 10° C. Although other studies have 1831
demonstrated increases in voltage production at lower operational temperatures [209, 210, 1832
212], the methane-driven MFCs investigated here were operated at significantly lower organic 1833
loading rates. Higher organic loading rates likely favor competition between exoelectrogens and 1834
heterotrophic bacteria, resulting in higher voltage production at lower temperatures if 1835
heterotrophs are inhibited to a greater extent than exoelectrogens [212]. The observed abrupt 1836
decrease in voltage production at 10 and 5° C in our study strongly suggests that exoelectrogenic 1837
activity was suppressed at such low temperatures. 1838
85

1839
1840
1841
Figure 10. (A) Influent dissolved methane concentration (bars) and theoretical dissolved methane 1842
derived from Henry’s law (line) across different operational temperatures (25, 20, 15, 10 and 1843
5° C). (B) Voltage production and Coulombic efficiency from continuous operation of replicate air- 1844
cathode MFCs. (C) Dissolved methane removal efficiency (open markers) and average absolute 1845
dissolved methane removal (filled markers). A negative correlation between average absolute 1846
methane removal and temperature was observed with R
2
values equating to 0.8864 and 0.7749 1847
for MFC A and B, respectively (solid lines). Dashed lines represent the linear trend line between 1848
dissolved methane removal efficiency and temperature (R
2
values were 0.3234 and 0.3509 for 1849
MFC A and B, respectively). 1850
86

1851
At the highest voltage production, 0.603 ± 0.017 V, power density was 60.6 mW/m
2
if 1852
normalized to cathode area (60 cm
2
), and is consistent with power density reported in our 1853
previous work [139]. Higher power densities have been reported in two-chamber MFCs, 419.5 ± 1854
5.9 via employing a gas diffusion electrode [201] and as high as 4700 ± 800 when using a 1855
genetically engineered archaeal strain performing reverse methanogenesis [203]. Given that 1856
the MFCs operated here produced similar voltage to previous studies [138, 139, 201, 203], 1857
cathode area is likely the design parameter restricting power density. Power density decreased 1858
to 6.40 and 11.4 mW/m
2
at 10° C in MFC A and B, respectively. CE also decreased significantly for 1859
MFC A and B as temperature decreased. CE at 25, 20, and 15° C averaged 6.90 ± 0.94% and 6.19 1860
± 1.45% for MFC A and B, respectively, whereas it significantly decreased at 10 and 5° C to an 1861
average of 2.10 ± 0.70% and 2.75 ± 1.27% for MFC A and B, respectively. CE at 25, 20, and 15° C 1862
is similar to what we previously reported [139]. Higher CE (65.9 ± 13.2 to 90 ± 10%) has been 1863
reported in other methane-driven MFCs relying on interactions between Geobacter and 1864
anaerobic methanotrophic archaea (ANME) [138] or methanogenic archaea performing reverse 1865
methanogenesis [201]. The lower CE reported here is due to electrons being lost during 1866
conversion of methane to intermediate metabolites by aerobic methanotrophs and scavenging 1867
of metabolites by heterotrophs [139, 140, 174]. It is important to note that studies reporting 1868
higher CE and power density rely on dual-chamber MFCs, in which oxygen or another electron 1869
accepter (e.g., ferriciyanide) [221] is provided in the cathode chamber and oxygen diffusion into 1870
the anode chamber is limited. Dual-chamber systems have substantial energy input if oxygen is 1871
provided, or necessitate unsustainable chemical regeneration of an alternative electron 1872
accepter. Relative to dual-chamber systems, air-cathode MFCs offer a simpler configuration and 1873
alleviate many scale-up issues, despite the lower CE and power density [222]. 1874
87

1875
4.3.2 Dissolved methane removal efficiency was relatively stable across operational 1876
temperature 1877
As operational temperature decreased, influent dissolved methane increased from 14.5 ± 0.8 1878
mg/L to 21.7 ± 0.9 mg/L (Figure 10A) due to increased methane solubility. Relative to theoretical 1879
dissolved methane concentrations derived from Henry’s law, measured influent concentrations 1880
were 19.0 ± 4.1% lower on average, possibly due to inefficiencies in sample collection. Across all 1881
operational temperatures, dissolved methane removal efficiencies were fairly stable, although a 1882
slight decreasing trend was observed: from 66.4% to 58.8% and 63.3% to 53.0% for MFC A and 1883
B, respectively, across the temperature decrease from 25 to 5° C (Figure 10C). Conversely, 1884
absolute dissolved methane removal increased from 9.57 and 9.07 mg/L at 25° C to 13.8 and 1885
11.6 mg/L at 5° C for MFC A and B, respectively, exhibiting a negative linear correlation with 1886
temperature (R
2
of 0.8864 and 0.7749 for MFC A and B, respectively, Figure 10C). This 1887
correlation was due to increased dissolved methane loading and relatively stable dissolved 1888
methane removal efficiency. Notably, dissolved methane removal efficiency exhibited higher 1889
variability at 5° C (SI Table 9), likely originating from instability in aerobic methanotrophic activity 1890
at such low operational temperature [174]. In contrast to literature investigating MFCs for 1891
energy recovery from gaseous methane, our objective here is to both mitigate fugitive dissolved 1892
methane emissions to the atmosphere and recover embedded energy from anaerobic effluents. 1893
In our previous work, we observed a positive correlation between HRT and dissolved methane 1894
removal efficiency, likely due to increased oxygen diffusion through the cathode per methane 1895
loading at longer HRTs. Future research should evaluate alternative strategies to improve 1896
removal efficiency, as >70% dissolved methane removal is likely needed to achieve comparable 1897
greenhouse gas emissions footprint relative to activated sludge processes [1]. 1898
88

4.3.3 Distinct microbial communities were observed on the anode and cathode 1899
A distinct spatial distribution of key microbial populations was observed, with aerobic 1900
methanotrophs colonizing the oxygen-rich cathode/separator and exoelectrogens colonizing the 1901
anode. Although voltage production decreased at the lower operational temperatures (10 and 1902
5° C), relative activity of Geobacter spp. normalized to 16S rRNA copies in the anode biomass did 1903
not demonstrate a corresponding decrease for either MFC (Figure 11A). Relative activity of 1904
methanotrophs, inferred by the ratio of pmoA transcripts to 16S rRNA copies in cathode 1905
biomass, exhibited a plateau corroborating the observed stable dissolved methane removal. 1906
(Figure 11B). It is important to note that normalizing to 16S rRNA has limitations in that activity 1907
ratios are relative to total community activity, which was certainly affected by operational 1908
temperature. Stable Geobacter spp. relative activity, particularly for MFC A, despite the 1909
observed decrease in voltage production, may indicate that other microbial populations in the 1910
anode community were similarly impacted by the decreased temperature. Absolute, as opposed 1911
to relative, activity measurements would likely provide more granularity in future research. 1912
89

1913
Figure 11. Relative activity of Geobacter 16S rRNA copy number normalized to total 16S rRNA 1914
copy number on anode (solid bar) and cathode (pattern filled bar). (B) Relative activity of pmoA 1915
transcripts copy number normalized to total 16S rRNA copy number on anode (solid bar) and 1916
cathode (pattern filled bar). ND = not detected. 1917
1918
Sequencing of 16S rRNA genes in anode and cathode biofilm communities corroborated the 1919
distinct colonization of Geobacter spp. and methanotrophic bacteria on the anode and cathode, 1920
respectively, evident from RT-qPCR data. Exoelectrogens in anode samples were identified 1921
90

belonging to Geobacter (3 OTUs) and Ferribacterium [223] (Figure 12). Across sequencing data 1922
from four operational temperatures (25, 15, 10 and 5° C), total relative abundance of 1923
exoelectrogens ranged from 13.3% to 46.1% in anode samples. Exoelectrogens were also 1924
detected in cathode samples, but at less than 2.29% relative abundance, with the highest 1925
relative abundance being observed at 25° C in MFC B. Geobacter has been observed as an 1926
electrotroph on the cathode of MFCs [224] and it may have performed a similar function in our 1927
system at elevated temperature. Total relative abundance of aerobic methanotrophs 1928
(Methylomonas and Methylobacillus) ranged from 5.86% to 21.6% in cathode samples. 1929
Methanotrophs were also detected in select anode samples and were as high as 10.0% relative 1930
abundance (10° C in MFC A). A recent study observed aerobic methanotrophs acting as 1931
exoelectrogens [225], but that seems unlikely to be occurring here given the inconsistency in 1932
methanotroph detection across anode samples. Distinct distribution of these key microbial 1933
populations implicates specific micro-environments inside methane-driven MFCs and agrees 1934
with our previous findings: an interaction between cathodic methanotrophs and anodic 1935
exoelectrogens powers methane-driven MFCs [139]. 1936
91

1937
Figure 12. Relative abundance of the top 30 abundant OTUs based on 16S rRNA gene (Bacteria 1938
and Archaea) sequencing identified to the genus level at 25, 15, 10, and 5° C. Results are 1939
expressed as a percentage normalized using total of 16S rRNA gene sequences. 1940
1941
Distinct colonization was also evident for other microbial populations within the top 30 most 1942
abundant OTUs at the genus level (Figure 12). Several fermentative populations were identified 1943
primarily in anode samples including Acetobacteirum, Psychrosinus, and Propionivibrio, 1944
comprising up to 15.5% relative abundance but only 1.36% relative abundance in cathode 1945
samples. Spearman’s rank correlation revealed a positive relationship between fermentative 1946
bacteria and exoelectrogens (P= 0.8095, ρ=0.015), suggesting that fermentative bacteria may 1947
play a role in system performance by converting methanotrophic intermediate metabolites (i.e., 1948
methanol, formaldehyde, and formate) to readily biodegradable organics for exoelectrogens as 1949
described previously [139, 140]. Conversely, several OTUs were observed consistently at similar 1950
92

relative abundance in both anode and cathode samples, such as Simplicispia, Flavobacterium, 1951
Bradyhizobium, Diaphorobacter, and Ferruginibacter (Figure 12). Their potential role in 1952
bioelectrochemical systems requires further investigation. 1953
Non-metric multidimensional scaling (NMDS) plots were generated based on the thetaYC 1954
distance matrix to visualize dissimilarity between microbial community structure across 1955
different operational temperatures. Anode and cathode samples formed distinct clusters given 1956
their unique metabolic properties (SI Figure 6). In general, the cathode microbial communities in 1957
MFC A and B displayed a higher similarity across the four operational temperatures than the 1958
anode microbial communities, suggesting greater microbial adaptation to temperature by the 1959
anode community. The anode communities were also distinctly clustered for MFC A and B, likely 1960
resulting in the observed differences in voltage production between the replicate systems. 1961
4.3.4 Network analysis revealed anode and cathode microbial community interactions 1962
LDA scores were determined at the genus level to identify populations showing discriminative 1963
distribution on either the anode or cathode. (SI Figure 7&8) Spearman’s rank correlations were 1964
performed on genera with LDA scores > 2 to investigate potential microbial interactions 1965
amongst genera found on the anode, cathode, and between both communities (SI Table 10). 1966
Within the anode community (Figure 13A), several genera positively correlated with 1967
exoelectrogens. For example, Acetobacterium [226], capable of converting methanol, formate, 1968
and lactate to acetate, positively correlated with Geobacter. Two additional fermentative 1969
bacteria, Psychrosinus and Propionivibrio, were both positively correlated with Geobacter and 1970
another potential exoelectrogen, Desulfovibrio [227]. Kalyuzhnaya et al. reported a 1971
fermentation-based methanotrophy process that produced lactate as a methanotrophic product 1972
[177]. The fermenters positively correlated with Geobacter on the anode may have been 1973
93

performing a similar function by converting produced lactate to acetate. Such findings 1974
implicates another indirect interaction between methanotrophs and exoelectrogens via 1975
fermentative bacteria in addition to their direct interaction [139]. 1976
1977
1978
1979
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
(A)
(B)
94

2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
Figure 13. (A) Correlation network analysis between genera preferentially colonizing on the 2011
anode based on LDA scores >2. Solid lines represent positive correlations and dashed lines 2012
represent negative correlations. Color coding denotes distinct metabolisms: potential 2013
exoelectrogens (red), fermentative bacteria (brown), and others (e.g., heterotrophs; black); (B) 2014
Correlation network analysis between genera preferentially colonizing on the cathode based on 2015
LDA scores >2. Solid lines represent positive correlations and dashed lines represent negative 2016
correlations. Color coding denotes distinct metabolisms: potential methane utilizers (green) and 2017
others (e.g., heterotrophs; black); (C) Correlation network analysis between genera preferentially 2018
colonizing on the anode (○) and genera preferentially colonizing on the cathode ( □) based on 2019
LDA scores >2. Solid lines represent positive correlations and dashed lines represent negative 2020
correlations. Color coding denotes distinct metabolisms: potential methane utilizers (green), 2021
potential exoelectrogens (red), fermentative bacteria (brown), and others (e.g., heterotrophs; 2022
black). 2023
2024
Within the cathode community (Figure 13B), all identified significant correlations were negative. 2025
Most genera identified here were heterotrophic bacteria that have been identified in methane- 2026
fed environments: Sphingomonas, Devosia [228], Stenotrophomonas [229], and 2027
Phenylobacterium [230]. Though methanotrophs can provide organics to cathodic heterotrophs, 2028
the negative correlations between Methylomonas and Terrimonas [231], a potential methane 2029
oxidizer, with other heterotrophs might be due to the complex methanotroph-heterotroph 2030
(C)
95

interactions involving mutual competition for oxygen. Further, heterotrophs rely on 2031
methanotrophic metabolites or microbial decay products [232]. 2032
The highest number of significant correlations were observed between genera on the anode and 2033
cathode (Figure 13C). Positive correlations were observed between Terrimonas and both 2034
Acetobacterium and Psychrosinus, providing further evidence that methanotrophic metabolites 2035
produced on the cathode were transported to the anode and consumed by fermentative 2036
bacteria. In addition, positive correlation between Methylomonas and Anaerosinus suggests that 2037
readily biodegradable organics generated by cathode methanotrophs were likely also scavenged 2038
by heterotrophic bacteria on the anode, rather than exoelectrogens, thereby reducing CE. 2039
4.3.5 Voltage production and Coulombic efficiency correlated with microbial diversity on 2040
the anode 2041
Relative abundance of Geobacter spp. on the anode did not significantly correlate with voltage 2042
production in our study or in another study of an MFC operated on acetate and wastewater 2043
[233]. Operational temperatures have been observed to influence microbial diversity in MFCs 2044
[207], and therefore, we evaluated if a similar relationship occurred in our systems. It is 2045
important to note that substrate type can influence MFC power density [204] and microbial 2046
diversity. For example, in comparable MFCs operated on acetate and wastewater, although 2047
diversity was higher for the acetate fed MFC, performance correlated with microbial diversity 2048
for only the wastewater fed MFC [233]. Unlike acetate and wastewater, methane conversion to 2049
electricity in MFCs requires a more complex microbial community involving diverse metabolic 2050
processes [138-140, 202]. In our work, microbial diversity of the anode for both MFC A and B (SI 2051
Table 11) correlated with voltage production and CE across the operational temperature range 2052
(P>0.7988, ρ<0.05, Table 5), whereas microbial diversity of the cathode did not correlate with 2053
96

either performance metric. Therefore, the anode microbial community is likely more influential 2054
in governing process performance. Thus, the observations here corroborate other work [233] 2055
suggesting that anode microbial diversity, rather than exoelectrogen abundance/activity, is a 2056
significant factor in MFC performance. 2057
2058
Table 5. Statistically significant Spearman’s rank correlations (ρ) between diversity metrics for 2059
the anode microbial community (inverse Simpson index and Shannon index) and reactor 2060
performance (CE, average daily voltage production, methane removal efficiency, and absolute 2061
dissolved methane removal). ** represents P<0.0001 and * represents P<0.05. No significant 2062
correlations were observed between the cathode microbial community and diversity metrics. 2063
Performance Metric Inverse Simpson Index Shannon Index
MFC A MFC B MFC A MFC B
CE 0.953** 0.830** 0.953** 0.799**
Average daily voltage production 0.942** 0.841** 0.942** 0.856**
Methane removal efficiency 0.680* 0.507* 0.608** 0.627*
Absolute dissolved methane removal -0.798** -0.484* -0.798** -0.654*
2064
4.4. Conclusions 2065
• Voltage production was relatively stable across the three highest operational 2066
temperatures (25, 20, and 15° C) for both MFCs (0.540 ± 0.087 V and 0.516 ± 0.069V), 2067
but abruptly decreased at 10°C. 2068
97

• Dissolved methane removal efficiency remained relatively stable, ranging from 66.4% to 2069
58.8% and 63.6% to 53.0% in MFC A and B, respectively. However, average absolute 2070
dissolved methane removal increased from 9.57 to 13.8 and 9.07 to 11.6 mg/L as 2071
temperature decreased due to increased methane solubility. 2072
• RT-qPCR results and 16S rRNA gene sequencing revealed distinct colonization of 2073
methanotrophs and exoelectrogens on the cathode and anode, respectively. 2074
Spearman’s rank correlation suggested that fermentative bacteria play a role by 2075
facilitating interactions between methanotrophs and exoelectrogens. 2076
• Anode microbial diversity in both MFCs was positively correlated with voltage 2077
production and CE, suggesting that anode diversity is an important parameter in 2078
performance at varying temperatures. 2079
• Overall, this work indicates that MFCs can provide relatively stable methane removal 2080
across varying operational temperature, but that energy recovery is severely 2081
compromised at 10°C and below. 2082
2083
2084
2085
2086
98

Chapter 5. Impact of sulfide on methane-driven microbial fuel cells 2087
during treatment of anaerobic effluents 2088
Abstract 2089
Mainstream anaerobic treatment processes produce effluents containing dissolved methane and 2090
sulfide, requiring downstream management to prevent greenhouse gas emissions and other 2091
environmental impacts. Multiple studies have demonstrated methane-driven microbial fuel cells 2092
(MFCs) as a potential biotechnology for energy recovery from methane. However, the impact of 2093
sulfide on performance and potential for additional energy recovery from sulfide (i.e., as another 2094
electron donor) has yet to be evaluated. Therefore, in this study, increasing sulfide concentrations 2095
of 1, 5, 10, and 20 mg/L were supplemented into an air-cathode MFC operated on dissolved 2096
methane at a hydraulic retention time of 16 h. The MFC supplemented with sulfide demonstrated 2097
stable voltage production (0.494 ± 0.008 to 0.452 ± 0.017 V), excellent dissolved methane removal 2098
(97.9 ± 0.6% to 93.6 ± 0.8%), and moderate sulfide removal (46.8 ± 8.9% at 20 mg sulfide/L). 2099
Elemental surface characterization indicated that sulfide was partially converted to elemental 2100
sulfur on the anode, but this did not have any apparent negative impact on electron transfer. 2101
Further, more than 52.4% of removed sulfide was converted to sulfate in the effluent. Accordingly, 2102
we speculate that sulfide serves as an additional electron donor to methane in our systems. 2103
2104
2105
2106
2107
99

5.1 Introduction 2108
Sulfate ranges from 20 to 50 mg/L in domestic wastewater (low to high strength) [234], although 2109
significantly higher concentrations have been observed [5]. High sulfate likely originates from 2110
industrial waste flows: petrochemical, pulp and paper, and tanning industries [235-237]. With 2111
increasing adoption of mainstream anaerobic processes for domestic wastewater management, 2112
influent sulfate becomes problematic by reducing biogas production and generating odorous and 2113
corrosive sulfides [238]. Sulfate reducing bacteria compete readily with methanogens for organic 2114
and inorganic electron donors (e.g., acetate and hydrogen) converting sulfate to sulfide and in the 2115
process consuming 0.67 g COD/g sulfate reduced [239]. Vela et al. [173] reviewed characteristics 2116
of anaerobic wastewater treatment effluent and reported sulfide ranging from 46 to 201 mg 2117
COD/L. Presence of sulfide in the effluent introduces severe operational issues such as toxic and 2118
unpleasant odor, pipe corrosion, etc [240]. Further, hydrogen sulfide in biogas lowers quality and 2119
typically necessitates removal via scrubbing in a side-stream unit process [241] . More importantly, 2120
sulfide exerts toxicity towards key microbial populations in the anaerobic food web, especially 2121
methanogens [242, 243]. Such effects could potentially deteriorate system performance to the 2122
point of reactor upset. Additionally, sulfides have been observed to negatively affect nitrifiers [7] 2123
which could be problematic if a downstream nutrient removal process is coupled with mainstream 2124
anaerobic treatment. 2125
Chemical, biological, and physical methods have been demonstrated for sulfide removal from 2126
wastewater and biogas [10, 240, 241, 244-247]. However, conventional oxygen injection and 2127
chemical addition require transportation and storage entailing additional costs [240]. To improve 2128
cost-effectiveness of effluent sulfide control, several technologies have been proposed. Sahinkaya 2129
et al. [244] delivered oxygen via hollow fiber membranes and demonstrated 37-99% sulfide 2130
oxidation resulting in 64-89% recovery of oxidized sulfide in the membrane biofilm reactor. 2131
100

Similarly, Pikaar et al. [245] generated oxygen in-situ using Ir/Ta mixed metal oxide coated 2132
electrodes and accomplished maximum sulfide removal at 11.8 ± 1.7g S/m
2
projected anode 2133
surface/h. Under micro-aerobic condition, Chen et al. [246] demonstrated enhanced denitrifying 2134
sulfide removal achieving co-management of sulfide and nitrate. In addition, electrochemical 2135
approaches have been demonstrated that recover energy from sulfide oxidation [247], and 2136
recover sulfur by switching electrode polarity [236]. However, elemental sulfur derived from 2137
sulfide oxidation will deposit on the anode requiring periodic cleaning [247]. Conversely, in 2138
microbe-assisted electrochemical systems, anode attached elemental sulfur likely mediates 2139
electron transfer [248], and microbes could facilitate formation of higher sulfur species such as 2140
sulfate [249]. Notably, Rabaey et al. [10] have demonstrated up to 98% sulfide and 46% acetate 2141
removal from anaerobic upflow sludge blanket reactor effluent via MFCs. Further, dissolved 2142
methane, another constituent posing environmental impacts in anaerobic effluents [1, 4, 173], 2143
has recently been demonstrated to power MFCs [138-140]. 2144
MFCs have potential to remove both dissolved methane and sulfide from anaerobic effluents 2145
while recovering energy. In the present study, sulfide was supplemented into a methane-driven 2146
MFC to simulate treatment of anaerobic effluents. Voltage, dissolved methane removal, and 2147
sulfide removal were evaluated. In addition, elemental surface characterization was performed 2148
to investigate elemental sulfur deposition on the anode. To date, this study is the first to evaluate 2149
MFCs co-managing dissolved methane and sulfide. 2150
5.2 Materials and Methods 2151
Two replicate bench-scale MFCs were inoculated and operated at a hydraulic retention time (HRT) 2152
of 16 h in a temperature controlled chamber (Drosophila Incubator, Genesee Scientific, CA) 2153
maintained at 20° C [139]. A synthetic feed was prepared to simulate anaerobic effluents [139]. 2154
101

Throughout operation, MFC A received synthetic anaerobic methane-containing medium which 2155
was supplemented with a concentrated sulfide solution via a syringe pump (New Era, Farmingdale, 2156
NY), while MFC B was operated as a control without sulfide addition. Influent containing sulfide 2157
at four incremental concentrations of 1, 5, 10, and 20 mg sulfide/L was sequentially fed into MFC 2158
A for 10 days at each concentration. 2159
A sulfide solution was prepared using sodium sulfide nonahydrate (Sigma-Aldrich, St. Louis, MO) 2160
and deoxygenated DI water. Influent and effluent samples were collected from the MFC in a 2161
syringe and subsequently stabilized using alkaline antioxidant buffer solution before being 2162
analyzed with a Silver/Sulfide Ion-Selective electrode (Cole Parmer, Vernon Hills, IL) according to 2163
method 4500-S
2-
G [250]. The probe was calibrated via 4500-S
2-
F iodometric method [250] where 2164
a standard potassium bi-iodate solution (VWR, Radnor, PA) was used as a reference. Voltage 2165
generation and dissolved methane were characterized as described previously [139]. Influent and 2166
effluent samples for sulfate quantification were filtered by syringe filter (0.22 µm, Whatman, VWR, 2167
Radnor, PA) and stored at 4° C before being analyzed by ion chromatography (ICS 2100, Thermo 2168
Fisher, Waltham, MA ) as described previously [139]. 2169
Microscopic images of the anode surface were obtained using a JEOL JSM-7001 scanning electron 2170
microscope (SEM; JEOL, MA) at the Core Center of Excellence in Nano Imaging (University of 2171
Southern California, CA). Anode carbon brush was collected at the end of 20 mg sulfide/L addition 2172
from MFC A and the control (MFC B), and immediately stored in 2.5% glutaraldehyde for microbial 2173
fixation at 4° C. Serial dehydrations were performed using ethanol at increasing concentration 2174
(30%, 50%, 70%, 80%, 90%, 95%, and 100%, 10 minutes/wash x 3 washes per concentration). 2175
Samples were further dehydrated in a critical point dryer (Tousimis 815, Tousimis, MD, USA) and 2176
coated with platinum and palladium by a sputter coater (Cressington 108, UK) before imaging. 2177
102

The JEOL JSM-7001 was equipped with energy dispersive analysis X-rays (EDAX) for elemental 2178
analysis. 2179
5.3 Results and Discussion 2180
5.3.1 Dissolved methane removal remained high at all influent sulfide concentrations 2181
As shown in Figure 14, measured influent sulfide concentration was close to the targeted sulfide 2182
addition for the first three concentrations. Measured influent sulfide was greater than the 2183
intended concentration of 20 mg/L and more variable (24.1 ± 5.0 mg/L vs. 21.5 mg/L), but still 2184
comparable to what has been reported for anaerobic effluents [173]. Average sulfide removal 2185
efficiency ranged from 54.1 ± 10.4% to 38.5 ± 9.0% (subset in Figure 14). Variability of sulfide 2186
removal efficiency decreased as influent sulfide concentration increased, possibly due to 2187
temporal microbial community adaptation. At the highest sulfide concentration, MFC A removed 2188
46.8 ± 8.9% of sulfide. Given the relatively stable removal efficiency, sulfide removal rate 2189
increased from 3.62 x 10
-6
to 4.80 x 10
-5
mg/s across operation. Dissolved methane removal 2190
efficiency throughout was stable and higher than has been reported previously: up to 98% in the 2191
present study versus 85% in our previous work at an HRT of 16 h [139]. We hypothesize that 2192
microbial community acclimation over time likely contributed to the improved removal. As sulfide 2193
concentration increased, dissolved methane removal efficiency slightly decreased from 97.9 ± 0.2% 2194
to 93.6 ± 0.8% (Figure 15), negatively correlating with sulfide concentration (R
2
=0.7882). In the 2195
control (MFC B), dissolved methane removal efficiency was somewhat more stable, ranging from 2196
97.4 ± 0.2% to 95.2 ± 0.8% (SI Table 12). A t-test comparing dissolved methane removal efficiency 2197
between MFC A and B across the entire operational period indicated no significant difference ( α 2198
= 0.95, p = 0.1516). Given the similarity in MFC A and B dissolved methane removal efficiency and 2199
data variability, influent sulfide appeared to have limited negative impact on performance. 2200
103

2201
Figure 14. Main figure shows influent and effluent sulfide concentrations. The orange bar 2202
represents effluent sulfide overlaid on a blue bar representing influent sulfide. The red horizontal 2203
lines represent targeted sulfide addition. The inset shows sulfide removal efficiency at each 2204
influent sulfide concentration: 1, 5, 10 and 20 mg/L. The box plot boundaries indicate the standard 2205
deviation while the horizontal line within the box indicates the median. The open diamond marker 2206
in each box represents mean sulfide removal efficiency. The closed diamond markers represent 2207
data from individual measurements. 2208
2209
Visual MinTEQ was used to simulate precipitation of trace metals with sulfide. As shown in SI 2210
Figure 9, Cu
2+
, Pb
2+
, Zn
2+
and Ni
2+
react and precipitate sequentially with increasing sulfide addition 2211
resulting in production of covellite, galena, spharelite, wurtzite, and galena. As sulfide exceeds 5 2212
mg/L, trace metals (Cu
2+
, Ni
2+
, Pb
2+
, Zn
2+
) completely precipitate. Loss of trace metals may 2213
104

potentially exert non-biogenic inhibition towards the microbial community [7]. In addition, sulfide 2214
could precipitate with copper which serves as a vital component of methane monooxygenase [174] 2215
and react with oxygen, negatively impacting methanotrophy. Passive oxygen diffusion across the 2216
air-cathode is assumed to be 1.07 x 10
-3
mg/s [139] which exceeds potential oxygen requirements 2217
for sulfide oxidation at all influent concentrations (SI Table 12). Therefore, competition for 2218
available oxygen between methanotrophs and sulfide oxidation may be limited. In addition, 2219
sulfide transformation into other sulfur species is possible via spontaneous electrochemical 2220
reactions [249], further reducing the likelihood of sulfide inhibition to the microbial community. 2221
Select methanotrophic bacteria have also been observed as highly resistant towards sulfide, such 2222
as Methylocaldum gracile SAD2 [251]. 2223
2224
105

2225
Figure 15. Voltage production and dissolved methane removal efficiency from MFC A (sulfide 2226
addition) and MFC B (control). Solid grey markers represent MFC A dissolved methane removal 2227
and solid orange markers represent MFC B dissolved methane removal. Open green markers 2228
represent MFC A voltage production and open blue marker represent MFC B voltage production. 2229
Microbial samples were collected between different sulfide concentrations creating the gaps in 2230
data. 2231
2232
5.3.2 Voltage production was stable across all influent sulfide concentrations 2233
Prior to sulfide addition, voltage production averaged 0.485 ± 0.027 V and 0.454 ± 0.003 V in MFC 2234
A and B, respectively. Voltage production slightly decreased over sulfide addition, reaching a low 2235
average of 0.452 ± 0.017 V at 10 mg sulfide/L for MFC A. However, this voltage production was 2236
comparable to that of the control. Throughout the experiment, power density normalized to 2237
cathode area (60 cm
2
) ranged from 34.4 ± 2.2 to 40.6 ± 1.2 mW/cm
2
resulting in Coulombic 2238
82%
84%
86%
88%
90%
92%
94%
96%
98%
100%
0.00
0.10
0.20
0.30
0.40
0.50
0.60
0.70
0.80
0.90
1.00
0 5 10 15 20 25 30 35 40 45 50 55 60
Dissolved methane removal efficiency
Voltage production (V)
Time (day)
A voltage production B voltage production
A dissolved methane removal B dissolved methane removal
106

efficiency from 14.1 ± 1.5% to 21.4 ± 3.4% (SI Table 12) in MFC A, which was comparable to our 2239
previous study at the same HRT [139]. 2240
At 10 and 20 mg sulfide/L addition for MFC A, a sudden decrease in voltage was observed upon 2241
initiating operation (Figure 15) followed by a rebound in voltage production. A similar 2242
phenomenon has been reported in another study [249]. Such abrupt decreases in voltage suggest 2243
that spiking sulfide might impose a slight transient shock on exoelectrogenic activity, and may be 2244
an important consideration if influent sulfide is temporally variable. Voltage production in the 2245
control (MFC B) remained relatively stable across operation apart from an operational issue 2246
between days 25 and 35 associated with collecting biomass samples from the anode and cathode. 2247
5.3.3 Elemental surface characterization of the anode revealed elevated sulfur relative to 2248
control 2249
Sulfide oxidation in MFCs has been reported as a combination of electrochemical reactions and 2250
microbial-assisted processes [249]. Visual examination of SEM images on the anode surface prior 2251
to sulfide addition in MFC A and in the control (MFC B) (Figure 16. C1 & D1) relative to MFC A at 2252
the end of 20 mg sulfide/L addition showed substantial granule precipitation. EDAX analysis 2253
showed elevated sulfur elements consisting of 2.05%-5.28% relative weight of total elements 2254
(Figure 16. A1-B2). Such findings suggest that oxidized sulfur species (e.g., elemental sulfur) 2255
precipitated on the anode surface as reported in previous studies [10, 249]. Transformation of 2256
sulfide into elemental sulfur could provide additional electrons resulting in current generation. In 2257
addition, deposited elemental sulfur could play a role in mediating electron transfer between 2258
exoelectrogens and the anode surface [248]. 2259
To further characterize fate of influent sulfide, sulfate was quantified in effluent samples via ion 2260
chromatography. Comparison of removed sulfide and effluent sulfate concentrations indicated 2261
107

that 54.3 ± 24.2%, 52.4 ± 12.4%, and 56.7 ± 13.4% of removed sulfide was converted to sulfate at 2262
sulfide addition of 5, 10 and 20 mg/L, respectively. Although other sulfur species were not 2263
quantified (e.g., thiosulfate), sulfide removal efficiency and effluent sulfate concentrations 2264
suggest that sulfate transformation was likely mediated by microorganisms [248], since formation 2265
of higher sulfur species from elemental sulfur was very limited in abiotic electrochemical systems 2266
[247]. Notably, electrons released during oxidation of sulfide provided extra electrons by 18.4% 2267
relative to MFCs only managing dissolved methane(SI Table 13). 2268
2269
2270
2271
2272
108

2273
(A1)
(B2)
(C1)
(D1)
(A2)
(B2)
(C2)
(D2)
Figure 16. A1 and B1 shows microscopic image on anode surface from MFC A (sulfide
addition), while B1 and B2 shows the corresponding elemental analysis via EDAX. C1
and D1 shows microscopic image on anode surface from MFC B (control), while C2
and D2 shows the corresponding elemental analysis via EDAX.
109

5.4 Conclusions 2274
An air-cathode methane-driven MFC was supplemented with sulfide at 1, 5, 10 and 20 mg/L to 2275
evaluate the efficacy of treating anaerobic process effluents containing both sulfide and dissolved 2276
methane. Dissolved methane removal slightly decreased upon sulfide addition but remained high 2277
(97.9 ± 0.6% to 93.6 ± 0.8%) throughout incremental sulfide increases. At 20 mg/L sulfide addition, 2278
46.8 ± 8.9% sulfide removal was achieved. Voltage production was stable at all sulfide 2279
concentrations, fluctuating <6.8% relative to operation before sulfide addition. SEM imaging and 2280
EDAX analysis of the anode surface revealed elemental sulfur formation. Overall, methane-driven 2281
MFCs are capable of near complete dissolved methane removal from anaerobic effluents and are 2282
not inhibited by sulfide at concentrations as high as 20 mg/L. Further research is necessary to 2283
optimize sulfide removal efficiency and elucidate the mechanism for sulfide transformation in 2284
methane-driven MFCs. 2285
2286
2287
2288
2289
2290
2291
110

Chapter 6. Conclusions and Outlook 2292
6.1 Overview 2293
The objective of this dissertation was to develop an efficient approach to manage anaerobic 2294
process effluents containing dissolved methane and sulfide. This dissertation focused on a 2295
bioelectrochemical system as a downstream biotechnology to treat anaerobic process effluents 2296
prior to discharge or reuse. The study started with a critical review of greenhouse gas emissions 2297
and mitigation strategies from conventional activated sludge processes and mainstream 2298
anaerobic-based treatment processes (Chapter 2). This work then introduced using microbial 2299
fuel cells (MFCs) to treat synthetic anaerobic effluent containing dissolved methane, and 2300
demonstrated up to 85% of dissolved methane removal at 16-hour hydraulic retention time 2301
(HRT) (Chapter 3). Further, we subjected methane-driven MFCs to varying temperatures from 25 2302
to 5° C to investigate the impact of ambient fluctuating temperatures on performance and 2303
microbial ecology (Chapter 4). Finally, we supplemented sulfide of varying concentrations to 2304
synthetic anaerobic effluents and studied performance of MFCs to simultaneously manage both 2305
dissolved methane and sulfide. 2306
6.2 Dissolved methane accounts for the majority of the carbon footprint in 2307
mainstream anaerobic-based processes 2308
Anaerobic technology allows direct energy recovery from domestic wastewater via biogas 2309
production. However, due to effluent dissolved methane, the carbon footprint of mainstream 2310
anaerobic processes is significantly higher than that in conventional activated sludge processes 2311
(Chapter 2). In addition, especially in anaerobic membrane bioreactors (AnMBRs), loss of 2312
methane in the effluent accounts for a large fraction of methane generated, substantially 2313
reducing energy recovery potential of anaerobic technology. Current air-stripping/degassing 2314
111

membrane contactors constrained by high gas/liquid ratio or vacuum pressure are either not 2315
recovering methane in sufficient purity (>5%) or not achieving a neutral or positive net energy 2316
balance. Implementation of anaerobic technology as an alternative approach to manage 2317
domestic wastewater requires investment in new approaches to remove/recover effluent 2318
dissolved methane. 2319
6.3 Methanotrophs-Geobacter interaction drives methane-driven microbial fuel 2320
cells 2321
To efficiently treat effluent dissolved methane, we proposed using MFCs as a downstream 2322
treatment technology. Consistent with the objective to manage domestic wastewater more 2323
sustainably, we elected to use single-chamber air-cathode MFCs where oxygen passively diffuses 2324
across the cathode at a constant rate to provide the electron acceptor at the cathode. Such 2325
bioelectrochemical systems do not require external supply of electron acceptors which entails 2326
energy-intensive aeration or chemical regeneration. Replicate MFCs fed with synthetic 2327
anaerobic effluent where dissolved methane was the sole electron donor demonstrated stable 2328
voltage production and dissolved methane removal (Chapter 3). Sequencing of 16S rRNA of 2329
anode biofilm biomass revealed methanotrophs and high activity of Geobacter spp. suggesting a 2330
synergistic interaction. Reverse transcription quantitative PCR (RT-qPCR) analysis further 2331
revealed distinct distribution of Geobacter spp. and methanotrophs on the anode and cathode, 2332
respectively. Then, sequential addition of three likely methanotrophic intermediate metabolites 2333
and effluent analysis by ion chromatography indicated that formate and acetate serve as the 2334
electron shuttles during methanotroph-Geobacter interactions. Finally, adjusting oxygen 2335
availability relative to influent methane loading via changing HRT resulted in significant impact 2336
to performance. A maximum dissolved methane removal of 84.6% and 83.9% was observed 2337
112

resulting in generation of 0.610 and 0.506 V at a maximum of 17.7% and 14.7% Coulombic 2338
efficiency at 16-hour HRT in replicate systems. 2339
6.4 MFCs demonstrate stable dissolved methane removal at varying temperatures 2340
Anerobic biotechnologies (e.g., AnMBR) have demonstrated excellent treatment of domestic 2341
wastewater at varying temperatures even in the psychrophilic range. However, elevated 2342
effluent dissolved methane was reported across various types of anaerobic bioreactors 2343
especially from AnMBR at low temperature due to increased methane solubility and observed 2344
methane supersaturation (Chapter 2). Therefore, we subjected MFCs to varying temperatures 2345
from 25 to 5° C to evaluate their efficiencies to treat synthetic anaerobic effluents containing 2346
dissolved methane (Chapter 4). Stable dissolved methane removal was observed in replicate 2347
MFCs across all operational temperatures, while voltage production abruptly reduced from 2348
0.463-0.512 V to 0.156-0.190 V at and below 10° C. High-throughput sequencing of 16S rRNA 2349
genes and RT-qPCR data indicated distinct distribution of methanotrophs (e.g., Methylomonas) 2350
and exoelectrogens (e.g., Geobacter) on the cathode and anode, respectively. Further analysis of 2351
microbial community data via linear discriminative analysis and Spearman’s rank correlation 2352
suggested fermentative bacteria (e.g., Acetobacteria) may play a role in facilitating 2353
methanotrophs-exoelectrogen interactions. In addition, anode microbial diversity strongly 2354
correlated with both voltage production and Coulombic efficiency likely serving as the primary 2355
factor influencing performance at varying temperature. 2356
6.5 Sulfide did not negatively impact methane-driven MFC performance 2357
Finally, we supplemented sulfide into one methane-driven MFC to investigate its impact on 2358
system performance (Chapter 5). Dissolved methane removal slightly decreased upon sulfide 2359
addition but remained high (97.9 ± 0.6% to 93.6 ± 0.8%) throughout incremental sulfide addition 2360
113

from 1, to 5, 10 and 20 mg/L. Voltage production was stable at all sulfide concentrations. The 2361
MFC removed 46.8 ± 8.9% sulfide at 20 mg/L addition. SEM imaging and EDAX analysis of the 2362
anode surface revealed elemental sulfur formation. In addition, partial sulfide was converted to 2363
sulfate providing additional electrons but limited voltage contributions due to system 2364
inefficiency. With further optimization of design/operational parameters for sulfide removal, 2365
MFCs appear to be a promising technology to further treat anaerobic process effluents 2366
containing dissolved methane and sulfide. 2367
6.6 Future research directions 2368
Though we have demonstrated that air-cathode single-chamber MFCs provide excellent removal 2369
of dissolved methane and reduce sulfide at varying operational conditions, the underlying 2370
mechanisms, especially microbial community interactions and sulfide transformation, require 2371
further investigation treatment performance can be optimized accordingly. Coulombic efficiency 2372
was relatively low throughout operation, resulting in limited energy recovery. Given the relatively 2373
low price of electricity and high capital costs of MFCs, the economic feasibility of the reported 2374
configuration is questionable. 2375
First, from the perspective of anode exoelectrogens, the extracellular electron transfer pathway 2376
by Geobacter sulfurreducens remains disputed in the literature: electron hopping across redox 2377
cofactors [252] (i.e., hemes of outer membrane or extracellular c-type cytochromes) vs. type IV 2378
pili conferring metallic-like conductivity due to overlapping π – π orbitals of aromatic amino acids 2379
in PilA subunits [253]. Correspondingly, two distinct correlations of conductivity with temperature 2380
were identified: metallic-like conductivity increases as temperature decreases [253], while redox 2381
conductivity increases as temperature increases [254]. Subjecting MFCs to varying temperatures 2382
and targeting biomarkers relevant to extracellular electron transfer (e.g., pilA, omcS, and omcZ) 2383
114

would shed light on the predominant extracellular electron transfer pathway upon temperature 2384
changes. Additionally, in contrast to relative microbial community data obtained from 2385
aforementioned RT-qPCR and 16S rRNA sequencing, adding internal standards prior to 2386
sequencing [196] to obtain absolute microbial structure/activity and employing other omics 2387
approaches (e.g., transcriptomics, proteomics, metabolomics) could provide insightful 2388
information. 2389
Second, from the perspective of cathode methanotrophs, aerobic methanotrophs can oxidize 2390
methane and provide reduced carbon substrates for populations that cannot oxidize methane 2391
directly, which have been reported to support denitrification and reduction of multiple 2392
contaminants [255, 256]. As such, utilizing approaches such as fluorescence in situ hybridization 2393
targeting specific microbial populations could better elucidate spatial distribution of microbial 2394
interactions [202]. In addition, several carbon substrates, such as acetate, formate and lactate, 2395
have been reported from methanotrophic metabolism under oxygen limited condition [177, 257]. 2396
Localizing and engineering methanotrophs on air-diffusing support media in optimal condition 2397
could provide an alternative way to convert effluent dissolved methane into value-added 2398
products for management of other contaminants (e.g., nitrate and bromate). 2399
To advance anaerobic biotechnology for domestic wastewater treatment, nutrients (e.g., 2400
phosphate and ammonium) present in anaerobic process effluents could be recovered via 2401
modifying MFCs into bioelectrochemical systems with capacitive deionization [258]. Voltage 2402
generated from intrinsic dissolved methane and sulfide could offset external electricity demand 2403
in current capacitive deionization systems. Utilizing present electron donors (e.g., dissolved 2404
methane and sulfide) to recover co-existing nutrients not only enhances the wastewater 2405
management energy balance and effluent quality for reuse, but also provides a sustainable 2406
approach to recover nutrients. 2407
115

Further, long-term field tests are required to investigate performance of bench-scale MFCs 2408
treating real anaerobic effluents consisting of complex microbial populations and other 2409
constituents. Presence of other microbial populations and varying ionic strength in the influent 2410
might introduce interruption to the established methanotroph-exoelectrogen interactions. 2411
Possible optimization strategies are proposed here: (1) add more carbon brush anode to enhance 2412
exoelectrogen growth and sulfide oxidation and (2) expand cathode surface area to increase 2413
oxygen diffusion for the growth of cathode methanotrophs. With a more thorough understanding 2414
of the underlying mechanisms and microbial population interactions, coupling anaerobic 2415
treatment processes with MFCs downstream provides a promising and sustainable way to 2416
efficiently manage domestic wastewater. 2417
116

Appendix 2418
A. Supplementary information 2419
SI 1.0 Tables 2420
2421
SI Table 1. Influent media recipe 2422
Substrate Trace metals solution (mg/L)
80% methane saturation /
1 g/L acetate solution
Chromium Nitrate
[Cr(NO 3) 3 ∙9H 2O]
1.00
Chemical compounds (mg/L) Copper Chloride
[CuCl 2]
0.500
Calcium Chloride
[CaCl 2]
75.1 Manganese Sulfate
[MnSO 4]
1.13
Ferric phosphate
[FePO 4]
27.2 Nickel Sulfate
[NiSO 4]
0.230
Phosphate buffer solution (50mM PBS)
(mg/L)
Zinc Chloride
[ZnCl 2]
0.380
Sodium phosphate
dibasic anhydrous
[Na 2HPO 4]

3949
Lead Chloride
[PbCl 2]
0.130
Sodium dihydrogen
phosphate
monohydrate
[NaH 2PO 4•H 2O]

2112

Potassium chloride
[KCl]

112

Ammonium Chloride
[NH 4Cl]

267

2423
117

SI Table 2. PCR and qPCR temperature cycling for primer sets 564F/840R (Geobacter 16S rRNA), A189F/mb661R (pmoA), and 515F/806R (total 2424
16S rRNA) 2425
2426
Primer set
Initial
denaturation
Cycles Denaturation Annealing Extension
Final
extension
High-resolution
melting
Serial dilution
R
2
,
standards
Amplification
efficiency
PCR 564F/840R 99° C 900s 25 94° C 60s 50° C 60s 72° C 60s 72° C 60s
A189F/mb661R 96° C 300s 30 94° C 60s 50° C 60s 72° C 60s 72° C 300s
515F/806R 95° C 300s 30 94° C 60s 58° C 60s 72° C 60s 72° C 600s
qPCR 564F/840R 94° C 240s 45 94° C 30s
Touchdown
65° C, 0 Cyc
-> 55° C (-
0.5° C)
72° C 30s
95° C 20s, 55° C
15s, 72° C 20s
10
0
, 10
-2
, 10
-4
,
10
-6
, 10
-8

1.00 82.0%
A189F/mb661R 95° C 30s 45 95° C 30s 56° C 30s 72° C 30s
95° C 20s, 55° C
15s, 72° C 20s
10
0
, 10
-2
, 10
-4
,
10
-6
, 10
-8

1.00 83.0%
515F/806R 95° C 300s 45 95° C 20s 55° C 15s 72° C 20s
95° C 20s, 55° C
15s, 72° C 20s
10
0
, 10
-2
, 10
-4
,
10
-6
, 10
-8

0.960 99.0%
2427
2428
2429
2430
2431
2432
118

SI Table 3. Oxygen diffusion across cathode 2433
Terms Units Notes
Diffusion coefficient D 0.000322 cm
2
/s k=D/ε
Mass transfer coefficient k 0.0023 cm/s 0.6 to 3.9*10
-3
[Chen et al. 2006]
Membrane thickness ε 0.14 cm [Chen et al. 2006]
Outside oxygen
gradient/concentration
(Saturated)
C s 7.8 mg/L [Chen et al. 2006]
Inside oxygen
gradient/concentration
C 0 mg/L
Membrane area A 60 cm
2

Oxygen
diffusion/wastewater

∆m
0.258
mg O 2/mL
16 h HRT, 0.25 mL/min
0.129 8 h HRT, 0.5 mL/min
0.0646 4 h HRT, 1 mL/min
2434
SI Table 4. Proportion of electrons diverted to electron acceptor and to cell maintenance based 2435
on overall energy balance 2436
2437
Term Formula/value Notes
Formate 1/2HCO 3
-
+ H
+
+ e
-
= 1/2HCOO
-
+ 1/2H 2O ∆G
0'
= 39.19 kJ/e
-
eq
Acetate 1/8CO 2 + 1/8HCO 3
-
+ H
+
+ e
-
= 1/8CH 3COO
-
+
3/8H 2O
∆G
0'
= 27.4 kJ/e
-
eq
Oxygen 1/4O 2 + H
+
+ e
-
= 1/2H 2O ∆G
0'
= -78.72 kJ/e
-
eq
Pyruvate ∆G
0'
= 35.09 kJ/e
-
eq
∆G donor
0'
/∆G c Formate or Acetate Electron donor or carbon source
∆G p ∆G p = ∆G pyr
0'
- ∆G c
0'
Energy to synthesis pyruvate from carbon source

∆G pc

18.8 kJ/e
-
eq when ammonia is nitrogen
source
Energy to make cells from intermediates, based
on empirically derived number of 3.33kJ/g cells,
C 5H 7O 2N for cells, and 20 e
-
transferred/ more
cells yields
∆G r ∆G p = ∆G acceptor
0'
- ∆G donor
0'
Energy available from electron donor to electron
acceptor
ε 0.6 Energy loss including energy transfer efficiency,
ranges from 0.4 to 0.8

A

Energy consumption for cell synthesis/ energy
produced from donor to acceptor

f e
0

Proportion of electrons flow from electron donor
to electron acceptor
119

SI Table 5. Addition volume of methanol, formaldehyde, and formate and corresponding 2438
Coulombic efficiency. 2439

Substrate
Volume
added µL
(per 240
mL)
Theoretical
current from
full substrate
conversion (C)
Total electrons
production (C)

Coulombic efficiency
Reactor A Reactor B Reactor A Reactor B
Methanol 12.1 173.67 1.97 3.91 1.14% 2.25%
Formaldehyde 11.0 115.78 8.33 15.25 7.20% 13.17%
Formate 11.3 57.89 29.66 44.06 51.24% 76.10%
2440
2441
2442
2443
2444
2445
2446
2447
2448
2449
2450
2451
2452
2453
2454
2455
2456
2457
2458
2459
2460
2461
120

SI Table 6. Primer coverage of Archaea and Bacteria for 16S rRNA gene primers F515 (GTGCCAGCMGCCGCGGTAA) and R806 2462
(GGACTACHVGGGTWTCTAAT) targeting the V4 region (Kozich et al. 2013) according to TestPrime 1.0. TestPrime 1.0 evaluates the coverage of 2463
primer pairs by running an in silico PCR using the SILVA databases. Zero primer mismatches were allowed. 2464
2465
Domain Phylum Coverage (%) Domain Phylum

Coverage
(%)
Archaea 52.6
Archaea Aenigmarchaeota 21.5 Archaea Lokiarchaeota 85.1
Archaea Aigarchaeota 1.2 Archaea MSBL1 84.8
Archaea AK8 86.7 Archaea Nanoarchaeota 0
Archaea Altiarchaeales 0 Archaea Nanohaloarchaeota 0
Archaea Bathyarchaeota 0.2 Archaea Parvarchaeota 68.8
Archaea Crenarchaeota 0.2 Archaea pCIRA-13 0
Archaea Diapherotrites 5 Archaea pMC2A209 20
Archaea Euryarchaeota 87.7 Archaea Thaumarchaeota 0.7
Archaea Hadesarchaea 83.4 Archaea TVG8AR30 0
Archaea Korarchaeota 46.3 Archaea WSA2 77
Bacteria 86.5
Bacteria AC1 67.6 Bacteria Gemmatimonadetes 88
Bacteria Acetothermia 26.1 Bacteria GN01 72.4
Bacteria Acidobacteria 91.8 Bacteria Gracilibacteria 74.8
Bacteria Actinobacteria 81.1 Bacteria Hydrogenedentes 88.8
Bacteria Aerophobetes 3.2 Bacteria Ignavibacteriae 89.6
Bacteria Aminicenantes 88.6 Bacteria Latescibacteria 91.1
Bacteria Aquificae 86.9 Bacteria LCP-89 100
Bacteria Armatimonadetes 86.4 Bacteria Lentisphaerae 82.5
Bacteria Atribacteria 93.5 Bacteria MD2896-B216 33.3
Bacteria Bacteroidetes 88.5 Bacteria Microgenomates 1.3
121

Bacteria BJ-169 83.3 Bacteria Nitrospinae 70.4
Bacteria BP4 100 Bacteria Nitrospirae 89.8
Bacteria BRC1 85.6 Bacteria Omnitrophica 62.1
Bacteria Caldiserica 5.3 Bacteria Parcubacteria 4.3
Bacteria Calescamantes 0 Bacteria PAUC34f 60.1
Bacteria Chlamydiae 18.8 Bacteria Peregrinibacteria 57.2
Bacteria Chlorobi 55.8 Bacteria Planctomycetes 80.9
Bacteria Chloroflexi 52.4 Bacteria Poribacteria 11.8
Bacteria Chrysiogenetes 100 Bacteria RsaHF231 85.7
Bacteria Cloacimonetes 82.9 Bacteria Saccharibacteria 3.9
Bacteria CPR2 72.2 Bacteria SBR1093 88.3
Bacteria Cyanobacteria 80.7 Bacteria Spirochaetae 71.9
Bacteria Deferribacteres 87.8 Bacteria Synergistetes 89.2
Bacteria
Deinococcus-
Thermus
92.1
Bacteria TA06
75
Bacteria Dictyoglomi 88.9 Bacteria Tectomicrobia 89
Bacteria Elusimicrobia 88.8 Bacteria Tenericutes 84.2
Bacteria FBP 15 Bacteria Thermodesulfobacteria 93.2
Bacteria FCPU426 75.8 Bacteria Thermotogae 88.5
Bacteria Fervidibacteria 100 Bacteria Verrucomicrobia 84.1
Bacteria Fibrobacteres 85.4 Bacteria WA-aaa01f12 75
Bacteria Firmicutes 88.5 Bacteria WS1 82.2
Bacteria FL0428B-PF49 73.3 Bacteria WS2 84.6
Bacteria Fusobacteria 88.3 Bacteria WS6 0
Bacteria GAL15 88.6 Bacteria WWE3 0
Domain Phylum Class Order Family Genus
Coverage
(%)
Bacteria Proteobacteria 89.8
Bacteria Proteobacteria Alphaproteobacteria 84.8
122

Bacteria Proteobacteria Alphaproteobacteria Rhizobiales 92.9
Bacteria Proteobacteria Alphaproteobacteria Rhizobiales Beijerinckiaceae Methylocapsa 100
Bacteria Proteobacteria Alphaproteobacteria Rhizobiales Beijerinckiaceae Methylocella 100
Bacteria Proteobacteria Alphaproteobacteria Rhizobiales Beijerinckiaceae Methyloferula 100
Bacteria Proteobacteria Alphaproteobacteria Rhizobiales Beijerinckiaceae Methylorosula 88.9
Bacteria Proteobacteria Alphaproteobacteria Rhizobiales Beijerinckiaceae Methylovirgula 94.4
Bacteria Proteobacteria Alphaproteobacteria Rhizobiales Hyphomicrobiaceae 92
Bacteria Proteobacteria Alphaproteobacteria Rhizobiales Hyphomicrobiaceae Methylorhabdus 100
Bacteria Proteobacteria Alphaproteobacteria Rhizobiales Hyphomicrobiaceae Methyloterrigena 100
Bacteria Proteobacteria Alphaproteobacteria Rhizobiales Methylobacteriaceae 89.7
Bacteria Proteobacteria Alphaproteobacteria Rhizobiales Methylobacteriaceae Meganema 94.3
Bacteria Proteobacteria Alphaproteobacteria Rhizobiales Methylobacteriaceae Methylobacterium 89.2
Bacteria Proteobacteria Alphaproteobacteria Rhizobiales Methylobacteriaceae Microvirga 88.9
Bacteria Proteobacteria Alphaproteobacteria Rhizobiales Methylobacteriaceae Psychroglaciecola 89.5
Bacteria Proteobacteria Alphaproteobacteria Rhizobiales Methylobacteriaceae uncultured 92.3
Bacteria Proteobacteria Alphaproteobacteria Rhizobiales Methylocystaceae 94.8
Bacteria Proteobacteria Alphaproteobacteria Rhizobiales Methylocystaceae Albibacter 100
Bacteria Proteobacteria Alphaproteobacteria Rhizobiales Methylocystaceae Hansschlegelia 100
Bacteria Proteobacteria Alphaproteobacteria Rhizobiales Methylocystaceae Hartmannibacter 100
Bacteria Proteobacteria Alphaproteobacteria Rhizobiales Methylocystaceae Methylocystis 94
Bacteria Proteobacteria Alphaproteobacteria Rhizobiales Methylocystaceae Methylopila 100
Bacteria Proteobacteria Alphaproteobacteria Rhizobiales Methylocystaceae Methylosinus 97.1
Bacteria Proteobacteria Alphaproteobacteria Rhizobiales Methylocystaceae Pleomorphomonas 96.4
Bacteria Proteobacteria Alphaproteobacteria Rhizobiales Methylocystaceae uncultured 85.2
Bacteria Proteobacteria Betaproteobacteria 91.4
Bacteria Proteobacteria Betaproteobacteria Methylophilales 89.1
Bacteria Proteobacteria Betaproteobacteria Methylophilales Methylophilaceae 89.1
Bacteria Proteobacteria Betaproteobacteria Methylophilales Methylophilaceae
Candidatus
Methylopumilus
95.2
123

Bacteria Proteobacteria Betaproteobacteria Methylophilales Methylophilaceae Methylobacillus 90
Bacteria Proteobacteria Betaproteobacteria Methylophilales Methylophilaceae Methylophilus 76.9
Bacteria Proteobacteria Betaproteobacteria Methylophilales Methylophilaceae Methylotenera 93
Bacteria Proteobacteria Betaproteobacteria Methylophilales Methylophilaceae Methylovorus 100
Bacteria Proteobacteria Betaproteobacteria Methylophilales Methylophilaceae OM43 clade 91.4
Bacteria Proteobacteria Betaproteobacteria Methylophilales Methylophilaceae PRD01a011B 96.7
Bacteria Proteobacteria Betaproteobacteria Methylophilales Methylophilaceae uncultured 91.2
Bacteria Proteobacteria Betaproteobacteria Rhodocyclales 91.7
Bacteria Proteobacteria Betaproteobacteria Rhodocyclales Rhodocyclaceae 91.7
Bacteria Proteobacteria Betaproteobacteria Rhodocyclales Rhodocyclaceae Methyloversatilis 92.4
Bacteria Proteobacteria Deltaproteobacteria 89.3
Bacteria Proteobacteria Deltaproteobacteria Desulfuromonadales 91.4
Bacteria Proteobacteria Deltaproteobacteria Desulfuromonadales Geobacteraceae 90.9
Bacteria Proteobacteria Deltaproteobacteria Desulfuromonadales Geobacteraceae Geobacter 91.1
Bacteria Proteobacteria Gammaproteobacteria 91.9
Bacteria Proteobacteria Gammaproteobacteria Methylococcales 92.6
Bacteria Proteobacteria Gammaproteobacteria Methylococcales
Marine Methylotrophic
Group 1
95.9
Bacteria Proteobacteria Gammaproteobacteria Methylococcales
Marine Methylotrophic
Group 1 Methyloprofundus
100
Bacteria Proteobacteria Gammaproteobacteria Methylococcales
Marine Methylotrophic
Group 2
92.1
Bacteria Proteobacteria Gammaproteobacteria Methylococcales Methylococcaceae 93.8
Bacteria Proteobacteria Gammaproteobacteria Methylococcales Methylococcaceae
Candidatus
Methylospira
100
Bacteria Proteobacteria Gammaproteobacteria Methylococcales Methylococcaceae Methylobacter 83.1
Bacteria Proteobacteria Gammaproteobacteria Methylococcales Methylococcaceae Methylocaldum 90
Bacteria Proteobacteria Gammaproteobacteria Methylococcales Methylococcaceae Methylococcus 97
Bacteria Proteobacteria Gammaproteobacteria Methylococcales Methylococcaceae Methylogaea 100
124

Bacteria Proteobacteria Gammaproteobacteria Methylococcales Methylococcaceae Methyloglobulus 100
Bacteria Proteobacteria Gammaproteobacteria Methylococcales Methylococcaceae Methylohalobius 100
Bacteria Proteobacteria Gammaproteobacteria Methylococcales Methylococcaceae Methylomarinum 77.8
Bacteria Proteobacteria Gammaproteobacteria Methylococcales Methylococcaceae Methylomicrobium 100
Bacteria Proteobacteria Gammaproteobacteria Methylococcales Methylococcaceae Methylomonas 98.1
Bacteria Proteobacteria Gammaproteobacteria Methylococcales Methylococcaceae Methyloparacoccus 95.6
Bacteria Proteobacteria Gammaproteobacteria Methylococcales Methylococcaceae Methylosarcina 100
Bacteria Proteobacteria Gammaproteobacteria Methylococcales Methylococcaceae Methylosoma 100
Bacteria Proteobacteria Gammaproteobacteria Methylococcales Methylococcaceae Methylosphaera 100
Bacteria Proteobacteria Gammaproteobacteria Methylococcales Methylococcaceae Methylothermus 90
Bacteria Proteobacteria Gammaproteobacteria Methylococcales Methylococcaceae Methylovulum 100
Bacteria Proteobacteria Gammaproteobacteria Methylococcales Methylococcaceae uncultured 91.5
Bacteria Proteobacteria Gammaproteobacteria Methylococcales Milano-WF1B-03 85
Bacteria Proteobacteria Gammaproteobacteria Methylococcales Milano-WF1B-42 100
Bacteria Proteobacteria Gammaproteobacteria Methylococcales pItb-vmat-59 100
Bacteria Proteobacteria Gammaproteobacteria Methylococcales pLW-20 71.4
Bacteria Proteobacteria Gammaproteobacteria Thiotrichales 93.3
Bacteria Proteobacteria Gammaproteobacteria Thiotrichales Thiotrichaceae 93.7
Bacteria Proteobacteria Gammaproteobacteria Thiotrichales Thiotrichaceae Methylohalomonas 82.9
2466
125

SI Table 7. Primer coverage of Archaea for 16S rRNA gene primer 564F (AAGCGTTGTTCGGAWTTAT) and 840R (GGCACTGCAGGGGTCAATA) 2467
targeting Geobacter (Cummings, D.E., et al. 2003) according to TestPrime 1.0. TestPrime 1.0 evaluates the coverage of primer pairs by running an 2468
in silico PCR using the SILVA databases. Zero primer mismatches were allowed. 2469
Domain Phylum Coverage (%) Domain Phylum
Coverage
(%)
Archaea 0 0
Archaea Aenigmarchaeota 0 Archaea Lokiarchaeota 0
Archaea Aigarchaeota 0 Archaea MSBL1 0
Archaea AK8 0 Archaea Nanoarchaeota 0
Archaea Altiarchaeales 0 Archaea Nanohaloarchaeota 0
Archaea Bathyarchaeota 0 Archaea Parvarchaeota 0
Archaea Crenarchaeota 0 Archaea pCIRA-13 0
Archaea Diapherotrites 0 Archaea pMC2A209 0
Archaea Euryarchaeota 0 Archaea Thaumarchaeota 0
Archaea Hadesarchaea 0 Archaea TVG8AR30 0
Archaea Korarchaeota 0 Archaea WSA2 0
Bacteria 0.1
Bacteria AC1 0 Bacteria Gemmatimonadetes 0
Bacteria Acetothermia 0 Bacteria GN01 0
Bacteria Acidobacteria 0 Bacteria Gracilibacteria 0
Bacteria Actinobacteria 0 Bacteria Hydrogenedentes 0
Bacteria Aerophobetes 0 Bacteria Ignavibacteriae 0
Bacteria Aminicenantes 0 Bacteria Latescibacteria 0
Bacteria Aquificae 0 Bacteria LCP-89 0
Bacteria Armatimonadetes 0 Bacteria Lentisphaerae 0
Bacteria Atribacteria 0 Bacteria MD2896-B216 0
Bacteria Bacteroidetes 0 Bacteria Microgenomates 0
Bacteria BJ-169 0 Bacteria Nitrospinae 0
126

Bacteria BP4 0 Bacteria Nitrospirae 0
Bacteria BRC1 0 Bacteria Omnitrophica 0
Bacteria Caldiserica 0 Bacteria Parcubacteria 0
Bacteria Calescamantes 0 Bacteria PAUC34f 0
Bacteria Chlamydiae 0 Bacteria Peregrinibacteria 0
Bacteria Chlorobi 0 Bacteria Planctomycetes 0
Bacteria Chloroflexi 0 Bacteria Poribacteria 0
Bacteria Chrysiogenetes 0 Bacteria RsaHF231 0
Bacteria Cloacimonetes 0 Bacteria Saccharibacteria 0
Bacteria CPR2 0 Bacteria SBR1093 0
Bacteria Cyanobacteria 0 Bacteria Spirochaetae 0
Bacteria Deferribacteres 0 Bacteria Synergistetes 0
Bacteria
Deinococcus-
Thermus
0 Bacteria TA06 0
Bacteria Dictyoglomi 0 Bacteria Tectomicrobia 0
Bacteria Elusimicrobia 0 Bacteria Tenericutes 0
Bacteria FBP 0 Bacteria Thermodesulfobacteria 0
Bacteria FCPU426 0 Bacteria Thermotogae 0
Bacteria Fervidibacteria 0 Bacteria Verrucomicrobia 0
Bacteria Fibrobacteres 0 Bacteria WA-aaa01f12 0
Bacteria Firmicutes 0 Bacteria WS1 0
Bacteria FL0428B-PF49 0 Bacteria WS2 0
Bacteria Fusobacteria 0 Bacteria WS6 0
Bacteria GAL15 0 Bacteria WWE3 0
Domain Phylum Class Order Family Genus
Coverage
(%)
Bacteria Proteobacteria 0.3
Bacteria Proteobacteria Deltaproteobacteria Desulfuromonadales Geobacteraceae 56.3
Bacteria Proteobacteria Deltaproteobacteria Desulfuromonadales Geobacteraceae Geobacter 63.1
2470
127

SI Table 8. Synthetic wastewater medium 2471
Trace metals solution
(mg/L)
Phosphate buffer solution
(mg/L)
Chromium
Nitrate
1
Sodium phosphate
dibasic anhydrous
3949
Copper
Chloride
0.50
Sodium
dihydrogen
phosphate
monohydrate
2112
Manganese
Sulfate
1.13
Potassium
chloride
112
Nickel
Sulfate
0.23
Ammonium
Chloride
267
Zinc
Chloride
0.38

Lead
Chloride
0.13

2472
2473
2474
2475
2476
2477
2478
2479
2480
2481
2482
2483
2484
2485
2486
2487
2488
128

SI Table 9. Average influent dissolved methane, dissolved methane removal efficiency (RE), 2489
absolute dissolved methane removal, voltage production, and CE across 25, 20, 15, 10, and 5° C. 2490
2491
Temperature
(° C)
Influent
dissolved
methane
(mg/L)
Average dissolved
methane RE (%)
Average absolute
methane removal
(mg/L)
Voltage production (V)
CE (%)
MFC A MFC B MFC A MFC B MFCA MFC B MFC A MFC B
25 14.5 ± 0.8
66.4 ±
2.3
63.6 ±
2.6
9.48 ±
0.61
9.03 ±
0.64
0.603 ±
0.017
0.463 ±
0.094
7.61 ±
0.04
5.50 ±
1.77
20 15.1 ± 0.6
57.0 ±
8.7
57.1 ±
1.5
9.35 ±
1.13
8.64 ±
0.56
0.535 ±
0.038
0.536 ±
0.073
6.72 ±
0.44
6.52 ±
0.93
15 17.5 ± 0.6
61.7 ±
4.2
61.1 ±
3.4
11.0 ±
1.1
10.9 ±
1.0
0.512 ±
0.108
0.569 ±
0.054
6.33 ±
1.28
6.59 ±
1.40
10 20.6 ± 0.7
57.8 ±
2.7
58.3 ±
3.6
12.3 ±
1.1
12.9 ±
1.8
0.196 ±
0.040
0.261 ±
0.093
2.53 ±
0.39
3.44 ±
1.01
5 21.7 ± 0.9
58.8 ±
7.3
53.0 ±
5.3
13.9 ±
2.4
12.9 ±
2.4
0.156 ±
0.040
0.190 ±
0.066
1.61 ±
0.66
1.93 ±
1.09
2492
129

SI Table 10. Spearman’s rank correlation were performed on genera that has LDA>2 from both 2493
MFC A and B. (A) Spearman’s correlation coefficients (ρ) and probability values within anode 2494
genera. 2495
Anode Spearman ρ Probability
Propionivibrio Geobacter 0.8571 0.0065
Desulfovibrio Geobacter 0.7619 0.028
Desulfovibrio Propionivibrio 0.9286 0.0009
Psychrosinus Geobacter 0.7143 0.0465
Psychrosinus Propionivibrio 0.7143 0.0465
Psychrosinus Desulfovibrio 0.7857 0.0208
Labilibacter Propionivibrio 0.7381 0.0366
Labilibacter Desulfovibrio 0.7857 0.0208
Acetobacterium Geobacter 0.7619 0.028
Acetobacterium Psychrosinus 0.8095 0.0149
Longilinea Geobacter -0.9048 0.002
Longilinea Propionivibrio -0.7143 0.0465
Longilinea Desulfovibrio -0.7381 0.0366
Longilinea Psychrosinus -0.8095 0.0149
Longilinea Azospira 0.7143 0.0465
Longilinea Acetobacterium -0.9048 0.002
Anaerosinus Desulfovibrio 0.7619 0.028
Anaerosinus Psychrosinus 0.7381 0.0366
Anaerosinus Dysgonomonas 0.8095 0.0149
Anaerosinus Labilibacter 0.7619 0.028
2496
2497
2498
(B) Spearman’s correlation coefficients (ρ) and probability values within cathode genera. 2499
Cathode Spearman ρ Probability
Terrimonas Phenylobacterium -0.881 0.0039
Devosia Stenotrophomonas -0.8571 0.0065
Devosia Leifsonia -0.8333 0.0102
Phenylobacterium Luteibacter -0.8095 0.0149
Terrimonas Sphingomonas -0.8095 0.0149
Methylomonas Stenotrophomonas -0.7381 0.0366
Luteibacter Sphingomonas -0.7381 0.0366
2500
2501
130

(C) Spearman’s correlation coefficients (ρ) and probability values between cathode and anode 2502
genera. 2503
Cathode Anode Spearman ρ Probability
Sphingomonas Propionivibrio -0.7143 0.0465
Sphingomonas Desulfovibrio -0.7143 0.0465
Sphingomonas Psychrosinus -0.9286 0.0009
Sphingomonas Acetobacterium -0.7381 0.0366
Methylobacillus Longilinea -0.8095 0.0149
Stenotrophomonas Propionivibrio -0.7619 0.028
Stenotrophomonas Desulfovibrio -0.7619 0.028
Stenotrophomonas Psychrosinus -0.7143 0.0465
Stenotrophomonas Labilibacter -0.9048 0.002
Stenotrophomonas Anaerosinus -0.8333 0.0102
Methylomonas Anaerosinus 0.8095 0.0149
Hyphomicrobium Azospira -0.7143 0.0465
Nitrosomonas Dysgonomonas -0.7306 0.0396
Leifsonia Geobacter -0.7143 0.0465
Leifsonia Propionivibrio -0.9048 0.002
Leifsonia Psychrosinus -0.7619 0.028
Leifsonia Anaerosinus -0.7381 0.0366
Devosia Geobacter 0.9048 0.002
Devosia Propionivibrio 0.9286 0.0009
Devosia Desulfovibrio 0.9048 0.002
Devosia Labilibacter 0.8571 0.0065
Devosia Acetobacterium 0.7143 0.0465
Devosia Longilinea -0.8333 0.0102
Luteibacter Azospira -0.8333 0.0102
Luteibacter Acetobacterium 0.8095 0.0149
Luteibacter Longilinea -0.7143 0.0465
Phenylobacterium Psychrosinus -0.8095 0.0149
Terrimonas Psychrosinus 0.7381 0.0366
Terrimonas Acetobacterium 0.8571 0.0065
2504
2505
2506
2507
2508
2509
2510
131

SI Table 11. Observed OTUs, Chao1, inverse Shannon index, Shannon index, and alpha diversity 2511
for MFC A and B anode and cathode microbial communities at temperatures of 25, 15, 10, and 2512
5° C. 2513
Reactor

Electrode

Temperature (° C)

Observed OTUs
(observed
richness)

Chao1

Invsimpson

Shannon

Alpha Diversity

MFC A

Anode

25

351

551

28.9

3.97

23.2

15

384

694

22.4

3.85

24.4

10

406

673

21.0

3.81

25.8

5

368

546

20.6

3.77

24.4

Cathode

25

308

454

14.7

3.41

21.3

15

302

445

11.2

3.22

20.2

10

346

498

15.5

3.49

22.1

5

321

470

13.5

3.41

21.2

MFC B

Anode

25

357

600

16.9

3.70

23.7

15

353

586

17.9

3.70

23.4

10

363

540

15.8

3.59

24.3

5

318

665

12.0

3.33

21.2

Cathode

25

333

516

15.9

3.76

21.6

15

317

419

14.0

3.43

21.0

10

295

414

13.7

3.35

19.8

5

324

507

14.0

3.57

21.8

2514
2515
2516
2517
2518
2519
2520
2521
2522
2523
2524
2525
2526
132

SI Table 12. Average voltage production, power density, Coulombic efficiency, dissolved methane 2527
removal efficiency from MFC A (sulfide addition) and MFC B (control). Average influent sulfide, 2528
sulfide removal efficiency, sulfide removal rate (mg/s), proportion of sulfate in removed sulfide 2529
and oxygen diffusion rate at 16-hour HRT from MFC A. 2530

Voltage
production (V)
Power density
(mW/cm
2
)
Coulombic
efficiency (%)
Average
dissolved
methane
removal
efficiency
Average
influent
sulfide,
mg/L
Average
sulfide
removal
efficiency
(%)
Sulfide
removal
rate
(mg/s)
Proportion
of sulfate
(%)
Oxygen
diffusion
rate
(mg/s)
Sulfide
(mg/L)
MFC A MFC B MFC A MFC B MFC A MFC B MFC A MFC B
0
0.485 ±
0.027
0.454 ±
0.003
39.2 ±
1.0
34.4 ±
0.5
21.4 ±
3.4%
20.6 ±
3.5%
97.9 ±
0.6%
97.4 ±
0.2%

1
0.475 ±
0.037
0.432 ±
0.077
37.6 ±
0.9
31.7 ±
7.3
19.5 ±
1.2%
17.7 ±
1.1%
96.4 ±
1.6%
96.4 ±
0.8%
1.58 ±
0.68
50.6 ±
19.4%
3.62E-
06

5
0.494 ±
0.008
0.228 ±
0.118
40.6 ±
1.2
11.1 ±
10.1
16.4 ±
1.3%
10.3 ±
2.4%
94.9 ±
0.5%
96.5 ±
0.1%
5.89 ±
0.74
54.1 ±
10.4%
1.31E-
05
54.3 ±
24.2%
1.07E-
03
10
0.452 ±
0.017
0.457 ±
0.003
34.4 ±
2.2
34.8 ±
0.4
16.6 ±
1.0%
16.1 ±
1.3%
94.7 ±
1.3%
94.1 ±
2.0%
13.0 ±
2.0
38.5 ±
9.0%
2.12E-
05
52.4 ±
12.4%
20
0.474 ±
0.013
0.380 ±
0.065
37.4 ±
1.5
25.6 ±
7.9
14.1 ±
1.5%
12.3 ±
3.4%
93.6 ±
0.8%
95.2 ±
0.8%
24.1 ±
5.0
46.8 ±
8.9%
4.80E-
05
56.7 ±
13.4%
2531
133

SI Table 13. Quantity of electrons represented by influent dissolved methane and oxidation of 2532
sulfide to sulfate over 1 second. All following number were derived under current operational 2533
conditions: 0.25 mL/min and performance parameters including average voltage production, 2534
influent dissolved methane, influent sulfide, sulfide removal efficiency, and proportion of sulfate 2535
in removed sulfide. 2536
Current production
(Coulomb)
Charge embedded in influent
dissolved methane
(Coulomb)
Charge embedded in oxidation of
sulfide to sulfate (Coulomb)
0.000474 0.00243
0.000492 0.00302 0.000174
0.000455 0.00273 0.000264
0.000474 0.00348 0.000643
2537
2538
2539
2540
2541
2542
2543
2544
2545
2546
2547
134

SI 2.0 Figures 2548
2549
2550
2551
2552
2553
2554
2555
2556
2557
2558
2559
2560
SI Figure 1. Gel electrophoresis of 16S rRNA gene PCR products of RNA extracts after being 2561
treated by Invitrogen DNA-free kit. 2562
Lane #

Sample

Lane #

Sample

1

DNA ladder

11

DNA ladder

2

Reactor A Anode biofilm

12

Reactor A Anode biofilm

3

Reactor A Cathode biofilm

13

Reactor A Cathode biofilm

4

Reactor B Anode biofilm

14

Reactor B Anode biofilm

5

Reactor B Cathode biofilm

15

Reactor B Cathode biofilm

6

Reactor A Anode biofilm

16

Primary effluent

7

Reactor A Cathode biofilm

17

Reactor A Anode biofilm

8

Reactor B Anode biofilm

18

Reactor A Cathode biofilm

9

Reactor B Cathode biofilm

19

Positive control

10

DNA ladder

20

DNA ladder

135

2563
2564
SI Figure 2. Gel electrophoresis of Geobacter 16S rRNA gene PCR products (top) and pmoA PCR 2565
products (bottom) of reverse transcribed RNA extracts (cDNA) after being treated by Invitrogen 2566
DNA-free kit. Bands for Geobacter and pmoA (faint) are shown in lane 2 and demonstrate that 2567
the inoculum contained both Geobacter and methanotrophs. 2568

Lane #

Sample
1 DNA ladder
2 Primary effluent
136

2569
SI Figure 3. Two cycles of voltage production while MFCs were operated in fed-batch mode on 2570
acetate containing media. 2571
2572
2573
2574
2575
2576
2577
2578
SI Figure 4. Voltage production while double-chamber MFCs were operated in continuous mode 2579
on methane containing media with different dissolved oxygen concentration 2580
2581
2582
137

2583
2584
2585
2586
2587
2588
2589
2590
2591
2592
2593
2594
2595
2596
2597
2598
2599
2600
2601
2602
2603
2604
2605
2606
2607
2608
SI Figure 5. Gel image on RNA samples after finishing DNA treatment. No band observed in the 2609
gel image indicating no contaminating DNA in the RNA samples. PCR was performed to amplify 2610
16S rDNA. 2611
7 10 1 13 8 2 3 4 5 6 9 11 12 14 15 16 17 18 19 20 21 22
138

2612
2613
2614
SI Figure 6. Non-metric multidimensional scaling plot of anode and cathode communities from 2615
MFC A and B (Stress:0.167223 R
2
= 0.87598). Solid and pattern fills represent MFC A and B, 2616
respectively. ○ and ∆ represent anode and cathode, respectively. 2617
2618
2619
2620
2621
2622
2623
2624
2625
2626
2627
139

2628
2629
SI Figure 7. Taxonomic cladogram for genera that showed significant differential relative 2630
abundance on the anode and cathode for (A) MFC A. Differential relative abundance was defined 2631
as LDA scores ≥ 2. 2632
140

2633
2634
SI Figure 8. Taxonomic cladogram for genera that showed significant differential relative 2635
abundance on the anode and cathode for (B) MFC B. Differential relative abundance was defined 2636
as LDA scores ≥ 2. 2637
2638
141

2639
SI Figure 9. Proportion of individual dissolved ions/ total ions (%), precipitation of Cu
2+
, Ni
2+
, Pb
2+
, 2640
Zn
2+
with addition of increasing sulfide (1, 5, 10 and 20 mg/L) 2641
2642
2643
2644
2645
2646
2647
2648
2649
2650
2651
2652
2653
2654
2655
0
20
40
60
80
100
0 2 4 6 8 10 12 14 16 18 20
Proportion of dissolved ions
/total ions (%)
Sulfide (mg/L)
Cu2+ HS- Ni2+ Pb2+ Zn2+
142

References 2656
2657
1. Smith, A.L., et al., Navigating wastewater energy recovery strategies: a life cycle 2658
comparison of anaerobic membrane bioreactor and conventional treatment systems 2659
with anaerobic digestion. Environmental Science & Technology, 2014. 48(10): p. 5972- 2660
5981. 2661
2. Chernicharo, C.A., et al., Anaerobic sewage treatment in Latin America, in Anaerobic 2662
Biotechnology: Environmental Protection and Resource Recovery. 2015, World Scientific. 2663
p. 263-296. 2664
3. Noyola, A., et al., Typology of municipal wastewater treatment technologies in Latin 2665
America. Clean–Soil, Air, Water, 2012. 40(9): p. 926-932. 2666
4. Chen, S., et al., Emerging investigators series: Revisiting greenhouse gas mitigation from 2667
conventional activated sludge and anaerobic-based wastewater treatment systems. 2668
Environmental Science: Water Research & Technology, 2018. 2669
5. Smith, A.L., S.J. Skerlos, and L. Raskin, Psychrophilic anaerobic membrane bioreactor 2670
treatment of domestic wastewater. Water Research, 2013. 47(4): p. 1655-1665. 2671
6. Smith, A., S. Skerlos, and L. Raskin, Anaerobic membrane bioreactor treatment of 2672
domestic wastewater at psychrophilic temperatures ranging from 15 C to 3 C. 2673
Environmental Science: Water Research & Technology, 2015. 1(1): p. 56-64. 2674
7. Vela, J.D., G.J. Dick, and N.G. Love, Sulfide inhibition of nitrite oxidation in activated 2675
sludge depends on microbial community composition. Water Research, 2018. 2676
8. Rabaey, K. and W. Verstraete, Microbial fuel cells: novel biotechnology for energy 2677
generation. TRENDS in Biotechnology, 2005. 23(6): p. 291-298. 2678
9. Logan, B.E., et al., Microbial fuel cells: methodology and technology. Environmental 2679
Science & Technology, 2006. 40(17): p. 5181-5192. 2680
10. Rabaey, K., et al., Microbial fuel cells for sulfide removal. Environmental Science & 2681
Technology, 2006. 40(17): p. 5218-5224. 2682
11. van Hees, W., A bacterial methane fuel cell. Journal of the Electrochemical Society, 2683
1965. 112(3): p. 258-262. 2684
12. EPA, Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990 - 2016. 2018, U.S. 2685
Environmental Protection Agency. 2686
13. Stocker, T.F., et al., Climate change 2013: the physical science basis. Intergovernmental 2687
panel on climate change, working group I contribution to the IPCC fifth assessment 2688
report (AR5). New York, 2013. 2689
14. Gupta, D. and S.K. Singh, Greenhouse gas emissions from wastewater treatment plants: 2690
a case study of Noida. Journal of Water Sustainability, 2012. 2(2): p. 131-139. 2691
15. Kampschreur, M.J., et al., Nitrous oxide emission during wastewater treatment. Water 2692
Research, 2009. 43(17): p. 4093-4103. 2693
16. Mannina, G., et al., Greenhouse gases from wastewater treatment—A review of 2694
modelling tools. Science of the Total Environment, 2016. 551: p. 254-270. 2695
17. McCarty, P.L., J. Bae, and J. Kim, Domestic wastewater treatment as a net energy 2696
producer–can this be achieved? Environmental Science & Technology, 2011. 45(17): p. 2697
7100-7106. 2698
18. Shoener, B.D., et al., Design of anaerobic membrane bioreactors for the valorization of 2699
dilute organic carbon waste streams. Energy & Environmental Science, 2016. 9(3): p. 2700
1102-1112. 2701
143

19. Smith, A.L., et al., Perspectives on anaerobic membrane bioreactor treatment of 2702
domestic wastewater: a critical review. Bioresource Technology, 2012. 122: p. 149-159. 2703
20. Giraldo, E., et al., Anaerobic sewage treatment technology in Latin-America: A selection 2704
of 20 years of experiences. Proceedings of the Water Environment Federation, 2007. 2705
2007(13): p. 5208-5228. 2706
21. Ozgun, H., et al., A review of anaerobic membrane bioreactors for municipal wastewater 2707
treatment: integration options, limitations and expectations. Separation and Purification 2708
Technology, 2013. 118: p. 89-104. 2709
22. Lin, H., et al., A review on anaerobic membrane bioreactors: applications, membrane 2710
fouling and future perspectives. Desalination, 2013. 314: p. 169-188. 2711
23. Lei, Z., et al., Application of anaerobic membrane bioreactors to municipal wastewater 2712
treatment at ambient temperature: A review of achievements, challenges, and 2713
perspectives. Bioresource Technology, 2018. 267: p. 756-768. 2714
24. Shin, C. and J. Bae, Current status of the pilot-scale anaerobic membrane bioreactor 2715
treatments of domestic wastewaters: A critical review. Bioresource technology, 2018. 2716
247: p. 1038-1046. 2717
25. Massara, T.M., et al., A review on nitrous oxide (N 2O) emissions during biological nutrient 2718
removal from municipal wastewater and sludge reject water. Science of The Total 2719
Environment, 2017. 596: p. 106-123. 2720
26. Law, Y., et al., Nitrous oxide emissions from wastewater treatment processes. Phil. Trans. 2721
R. Soc. B, 2012. 367(1593): p. 1265-1277. 2722
27. Ni, B.-J. and Z. Yuan, Recent advances in mathematical modeling of nitrous oxides 2723
emissions from wastewater treatment processes. Water Research, 2015. 87: p. 336-346. 2724
28. Crone, B.C., et al., Significance of dissolved methane in effluents of anaerobically treated 2725
low strength wastewater and potential for recovery as an energy product: A review. 2726
Water Research, 2016. 104: p. 520-531. 2727
29. Mamais, D., et al., Wastewater treatment process impact on energy savings and 2728
greenhouse gas emissions. Water Science and Technology, 2015. 71(2): p. 303-308. 2729
30. Rosso, D., L.E. Larson, and M.K. Stenstrom, Aeration of large-scale municipal wastewater 2730
treatment plants: state of the art. Water Science and Technology, 2008. 57(7): p. 973- 2731
978. 2732
31. Yan, X., L. Li, and J. Liu, Characteristics of greenhouse gas emission in three full-scale 2733
wastewater treatment processes. Journal of Environmental Sciences, 2014. 26(2): p. 2734
256-263. 2735
32. Masuda, S., et al., The seasonal variation of emission of greenhouse gases from a full- 2736
scale sewage treatment plant. Chemosphere, 2015. 140: p. 167-173. 2737
33. Kyung, D., et al., Estimation of greenhouse gas emissions from a hybrid wastewater 2738
treatment plant. Journal of Cleaner Production, 2015. 95: p. 117-123. 2739
34. EIA, State Electricity Profiles-2016. 2018, U.S. Energy Information Administration. 2740
35. Oshita, K., et al., Methane and nitrous oxide emissions following anaerobic digestion of 2741
sludge in Japanese sewage treatment facilities. Bioresource Technology, 2014. 171: p. 2742
175-181. 2743
36. Chen, Y.-C. and J. Kuo, Potential of greenhouse gas emissions from sewage sludge 2744
management: a case study of Taiwan. Journal of Cleaner Production, 2016. 129: p. 196- 2745
201. 2746
37. Snip, L., Quantifying the greenhouse gas emissions of wastewater treatment plants. 2747
Environmental Sciences Netherlands, 2010: p. 8-13. 2748
144

38. Cadwallader, A. and J.M. VanBriesen, Incorporating uncertainty into future estimates of 2749
nitrous oxide emissions from wastewater treatment. Journal of Environmental 2750
Engineering, 2017. 143(8): p. 04017029. 2751
39. Eggleston, S., et al., 2006 IPCC guidelines for national greenhouse gas inventories. Vol. 5. 2752
2006: Institute for Global Environmental Strategies Hayama, Japan. 2753
40. Tseng, L.Y., et al., Identification of preferential paths of fossil carbon within water 2754
resource recovery facilities via radiocarbon analysis. Environmental Science & 2755
Technology, 2016. 50(22): p. 12166-12178. 2756
41. Griffith, D.R., R.T. Barnes, and P.A. Raymond, Inputs of fossil carbon from wastewater 2757
treatment plants to US rivers and oceans. Environmental Science & Technology, 2009. 2758
43(15): p. 5647-5651. 2759
42. Sweetapple, C., G. Fu, and D. Butler, Identifying key sources of uncertainty in the 2760
modelling of greenhouse gas emissions from wastewater treatment. Water Research, 2761
2013. 47(13): p. 4652-4665. 2762
43. Shahabadi, M.B., L. Yerushalmi, and F. Haghighat, Estimation of greenhouse gas 2763
generation in wastewater treatment plants–Model development and application. 2764
Chemosphere, 2010. 78(9): p. 1085-1092. 2765
44. Keller, J. and K. Hartley, Greenhouse gas production in wastewater treatment: process 2766
selection is the major factor. Water Science and Technology, 2003. 47(12): p. 43-48. 2767
45. Hwang, K.-L., C.-H. Bang, and K.-D. Zoh, Characteristics of methane and nitrous oxide 2768
emissions from the wastewater treatment plant. Bioresource Technology, 2016. 214: p. 2769
881-884. 2770
46. Daelman, M., et al., Methane and nitrous oxide emissions from municipal wastewater 2771
treatment–results from a long-term study. Water Science and technology, 2013. 67(10): 2772
p. 2350-2355. 2773
47. Rodrí guez-Caballero, A., et al., Evaluation of process conditions triggering emissions of 2774
green-house gases from a biological wastewater treatment system. Science of the Total 2775
Environment, 2014. 493: p. 384-391. 2776
48. Rodrí guez-Caballero, A., et al., Minimizing N 2O emissions and carbon footprint on a full- 2777
scale activated sludge sequencing batch reactor. Water Research, 2015. 71: p. 1-10. 2778
49. Daelman, M.R., et al., Seasonal and diurnal variability of N 2O emissions from a full-scale 2779
municipal wastewater treatment plant. Science of the Total Environment, 2015. 536: p. 2780
1-11. 2781
50. Yoshida, H., J. Mønster, and C. Scheutz, Plant-integrated measurement of greenhouse 2782
gas emissions from a municipal wastewater treatment plant. Water Research, 2014. 61: 2783
p. 108-118. 2784
51. Monteith, H.D., et al., A rational procedure for estimation of greenhouse-gas emissions 2785
from municipal wastewater treatment plants. Water Environment Research, 2005. 2786
77(4): p. 390-403. 2787
52. Short, M.D., et al., Dissolved methane in the influent of three Australian wastewater 2788
treatment plants fed by gravity sewers. Science of the Total Environment, 2017. 599: p. 2789
85-93. 2790
53. Liu, Y., et al., CH4 emission and conversion from A2O and SBR processes in full-scale 2791
wastewater treatment plants. Journal of Environmental Sciences, 2014. 26(1): p. 224- 2792
230. 2793
54. Yver Kwok, C., et al., Methane emission estimates using chamber and tracer release 2794
experiments for a municipal waste water treatment plant. 2015. 2795
145

55. Wang, J., et al., Methane emissions from a full-scale A/A/O wastewater treatment plant. 2796
Bioresource Technology, 2011. 102(9): p. 5479-5485. 2797
56. Czepiel, P.M., P.M. Crill, and R.C. Harriss, Methane emissions from municipal wastewater 2798
treatment processes. Environmental science & technology, 1993. 27(12): p. 2472-2477. 2799
57. Aboobakar, A., et al., Methane emissions from aerated zones in a full-scale nitrifying 2800
activated sludge treatment plant. Water, Air, & Soil Pollution, 2014. 225(1): p. 1814. 2801
58. Alshboul, Z., et al., Export of dissolved methane and carbon dioxide with effluents from 2802
municipal wastewater treatment plants. Environmental science & technology, 2016. 2803
50(11): p. 5555-5563. 2804
59. Daelman, M.R., et al., Methane emission during municipal wastewater treatment. Water 2805
research, 2012. 46(11): p. 3657-3670. 2806
60. Schaum, C., T. Fundneider, and P. Cornel, Analysis of methane emissions from digested 2807
sludge. Water Science and Technology, 2016. 73(7): p. 1599-1607. 2808
61. Scheehle, E.A. and M.R. Doorn, Improvements to the US wastewater methane and 2809
nitrous oxide emissions estimates. US EPA, Washington, DC, 2001. 2810
62. Czepiel, P., P. Crill, and R. Harriss, Nitrous oxide emissions from municipal wastewater 2811
treatment. Environmental Science & Technology, 1995. 29(9): p. 2352-2356. 2812
63. Bock, E., et al., Nitrogen loss caused by denitrifying Nitrosomonas cells using ammonium 2813
or hydrogen as electron donors and nitrite as electron acceptor. Archives of 2814
Microbiology, 1995. 163(1): p. 16-20. 2815
64. Poughon, L., C.G. Dussap, and J.B. Gros, Energy model and metabolic flux analysis for 2816
autotrophic nitrifiers. Biotechnology and Bioengineering, 2001. 72(4): p. 416-433. 2817
65. Tallec, G., et al., Nitrous oxide emissions from secondary activated sludge in nitrifying 2818
conditions of urban wastewater treatment plants: effect of oxygenation level. Water 2819
Research, 2006. 40(15): p. 2972-2980. 2820
66. Tallec, G., et al., Nitrous oxide emissions from denitrifying activated sludge of urban 2821
wastewater treatment plants, under anoxia and low oxygenation. Bioresource 2822
Technology, 2008. 99(7): p. 2200-2209. 2823
67. Wang, Y., et al., Nitric oxide and nitrous oxide emissions from a full-scale activated 2824
sludge anaerobic/anoxic/oxic process. Chemical Engineering Journal, 2016. 289: p. 330- 2825
340. 2826
68. Rathnayake, R.M., et al., Effects of dissolved oxygen and pH on nitrous oxide production 2827
rates in autotrophic partial nitrification granules. Bioresource Technology, 2015. 197: p. 2828
15-22. 2829
69. Hynes, R.K. and R. Knowles, Production of nitrous oxide by Nitrosomonas europaea: 2830
effects of acetylene, pH, and oxygen. Canadian Journal of Microbiology, 1984. 30(11): p. 2831
1397-1404. 2832
70. Terada, A., et al., Hybrid nitrous oxide production from a partial nitrifying bioreactor: 2833
hydroxylamine interactions with nitrite. Environmental Science & Technology, 2017. 2834
51(5): p. 2748-2756. 2835
71. Kester, R.A., W. De Boer, and H.J. Laanbroek, Production of NO and N 2O by pure cultures 2836
of nitrifying and denitrifying bacteria during changes in aeration. Applied and 2837
Environmental Microbiology, 1997. 63(10): p. 3872-3877. 2838
72. Peng, L., et al., The combined effect of dissolved oxygen and nitrite on N 2O production by 2839
ammonia oxidizing bacteria in an enriched nitrifying sludge. Water Research, 2015. 73: 2840
p. 29-36. 2841
73. Quan, X., et al., Nitrous oxide emission and nutrient removal in aerobic granular sludge 2842
sequencing batch reactors. Water Research, 2012. 46(16): p. 4981-4990. 2843
146

74. Kampschreur, M.J., et al., Effect of dynamic process conditions on nitrogen oxides 2844
emission from a nitrifying culture. Environmental Science & Technology, 2007. 42(2): p. 2845
429-435. 2846
75. Zheng, H., K. Hanaki, and T. Matsuo, Production of nitrous oxide gas during nitrification 2847
of wastewater. Water Science and Technology, 1994. 30(6): p. 133-141. 2848
76. Burgess, J., et al., Dinitrogen oxide detection for process failure early warning systems. 2849
Water Science and Technology, 2002. 45(4-5): p. 247-254. 2850
77. Ahn, J.H., T. Kwan, and K. Chandran, Comparison of partial and full nitrification 2851
processes applied for treating high-strength nitrogen wastewaters: microbial ecology 2852
through nitrous oxide production. Environmental Science & Technology, 2011. 45(7): p. 2853
2734-2740. 2854
78. Aboobakar, A., et al., Nitrous oxide emissions and dissolved oxygen profiling in a full- 2855
scale nitrifying activated sludge treatment plant. Water Research, 2013. 47(2): p. 524- 2856
534. 2857
79. Goreau, T.J., et al., Production of NO 2
-
and N 2O by nitrifying bacteria at reduced 2858
concentrations of oxygen. Applied and Environmental Microbiology, 1980. 40(3): p. 526- 2859
532. 2860
80. Yu, R., et al., Mechanisms and specific directionality of autotrophic nitrous oxide and 2861
nitric oxide generation during transient anoxia. Environmental Science & Technology, 2862
2010. 44(4): p. 1313-1319. 2863
81. Xie, W.-M., et al., Formation and quantification of soluble microbial products and N 2O 2864
production by ammonia-oxidizing bacteria (AOB)-enriched activated sludge. Chemical 2865
Engineering Science, 2012. 71: p. 67-74. 2866
82. Peng, L., et al., The effect of dissolved oxygen on N2O production by ammonia-oxidizing 2867
bacteria in an enriched nitrifying sludge. Water Research, 2014. 66: p. 12-21. 2868
83. Park, K.Y., et al., Emission and control of nitrous oxide from a biological wastewater 2869
treatment system with intermittent aeration. Journal of Bioscience and Bioengineering, 2870
2000. 90(3): p. 247-252. 2871
84. Castro-Barros, C.M., et al., Effect of aeration regime on N 2O emission from partial 2872
nitritation-anammox in a full-scale granular sludge reactor. Water Research, 2015. 68: p. 2873
793-803. 2874
85. Kampschreur, M.J., et al., Dynamics of nitric oxide and nitrous oxide emission during full- 2875
scale reject water treatment. Water Research, 2008. 42(3): p. 812-826. 2876
86. Law, Y., P. Lant, and Z. Yuan, The effect of pH on N 2O production under aerobic 2877
conditions in a partial nitritation system. Water Research, 2011. 45(18): p. 5934-5944. 2878
87. Hanaki, K., Z. Hong, and T. Matsuo, Production of nitrous oxide gas during denitrification 2879
of wastewater. Water Science and Technology, 1992. 26(5-6): p. 1027-1036. 2880
88. Thörn, M. and F. Sörensson, Variation of nitrous oxide formation in the denitrification 2881
basin in a wastewater treatment plant with nitrogen removal. Water Research, 1996. 2882
30(6): p. 1543-1547. 2883
89. Zhou, Y., et al., Free nitrous acid inhibition on nitrous oxide reduction by a denitrifying- 2884
enhanced biological phosphorus removal sludge. Environmental Science & Technology, 2885
2008. 42(22): p. 8260-8265. 2886
90. Soler-Jofra, A., et al., Importance of abiotic hydroxylamine conversion on nitrous oxide 2887
emissions during nitritation of reject water. Chemical Engineering Journal, 2016. 287: p. 2888
720-726. 2889
91. Schulthess, R., M. Kühni, and W. Gujer, Release of nitric and nitrous oxides from 2890
denitrifying activated sludge. Water Research, 1995. 29(1): p. 215-226. 2891
147

92. Gao, H., et al., Complete nutrient removal coupled to nitrous oxide production as a 2892
bioenergy source by denitrifying polyphosphate-accumulating organisms. Environmental 2893
Science & Technology, 2017. 51(8): p. 4531-4540. 2894
93. Zhu, X. and Y. Chen, Reduction of N2O and NO generation in anaerobic− aerobic (low 2895
dissolved oxygen) biological wastewater treatment process by using sludge alkaline 2896
fermentation liquid. Environmental Science & Technology, 2011. 45(6): p. 2137-2143. 2897
94. Hu, Z., et al., Impact of carbon source on nitrous oxide emission from anoxic/oxic 2898
biological nitrogen removal process and identification of its emission sources. 2899
Environmental Science and Pollution Research, 2013. 20(2): p. 1059-1069. 2900
95. Song, K., et al., Impact of carbon sources on nitrous oxide emission and microbial 2901
community structure in an anoxic/oxic activated sludge system. Clean Technologies and 2902
Environmental Policy, 2015. 17(8): p. 2375-2385. 2903
96. Zhang, X., et al., Reduction of nitrous oxide emissions from partial nitrification process by 2904
using innovative carbon source (mannitol). Bioresource Technology, 2016. 218: p. 789- 2905
795. 2906
97. Zheng, M., et al., Minimization of nitrous oxide emission in a pilot-scale oxidation ditch: 2907
generation, spatial variation and microbial interpretation. Bioresource Technology, 2908
2015. 179: p. 510-517. 2909
98. Burgess, J.E., et al., Dinitrogen oxide production by a mixed culture of nitrifying bacteria 2910
during ammonia shock loading and aeration failure. Journal of Industrial Microbiology 2911
and Biotechnology, 2002. 29(6): p. 309-313. 2912
99. Foley, J., et al., Nitrous oxide generation in full-scale biological nutrient removal 2913
wastewater treatment plants. Water Research, 2010. 44(3): p. 831-844. 2914
100. Cakir, F. and M. Stenstrom, Greenhouse gas production: a comparison between aerobic 2915
and anaerobic wastewater treatment technology. Water Research, 2005. 39(17): p. 2916
4197-4203. 2917
101. Lettinga, G., et al., Anaerobic treatment of domestic sewage and wastewater. Water 2918
Science and Technology, 1993. 27(9): p. 67-73. 2919
102. Bandara, W.M., et al., Removal of residual dissolved methane gas in an upflow anaerobic 2920
sludge blanket reactor treating low-strength wastewater at low temperature with 2921
degassing membrane. Water Research, 2011. 45(11): p. 3533-3540. 2922
103. Souza, C., C. Chernicharo, and S. Aquino, Quantification of dissolved methane in UASB 2923
reactors treating domestic wastewater under different operating conditions. Water 2924
Science and Technology, 2011. 64(11): p. 2259-2264. 2925
104. Bandara, W.M., et al., Anaerobic treatment of municipal wastewater at ambient 2926
temperature: Analysis of archaeal community structure and recovery of dissolved 2927
methane. Water Research, 2012. 46(17): p. 5756-5764. 2928
105. Cookney, J., et al., Recovery of methane from anaerobic process effluent using poly-di- 2929
methyl-siloxane membrane contactors. Water Science and Technology, 2012. 65(4): p. 2930
604-610. 2931
106. Giménez, J., et al., Methane recovery efficiency in a submerged anaerobic membrane 2932
bioreactor (SAnMBR) treating sulphate-rich urban wastewater: evaluation of methane 2933
losses with the effluent. Bioresource Technology, 2012. 118: p. 67-72. 2934
107. Yoo, R., et al., Anaerobic treatment of municipal wastewater with a staged anaerobic 2935
fluidized membrane bioreactor (SAF-MBR) system. Bioresource Technology, 2012. 120: 2936
p. 133-139. 2937
148

108. Pretel, R., et al., Environmental impact of submerged anaerobic MBR (SAnMBR) 2938
technology used to treat urban wastewater at different temperatures. Bioresource 2939
Technology, 2013. 149: p. 532-540. 2940
109. Ren, L., Y. Ahn, and B.E. Logan, A two-stage microbial fuel cell and anaerobic fluidized 2941
bed membrane bioreactor (MFC-AFMBR) system for effective domestic wastewater 2942
treatment. Environmental Science & Technology, 2014. 48(7): p. 4199-4206. 2943
110. Gouveia, J., et al., A novel configuration for an anaerobic submerged membrane 2944
bioreactor (AnSMBR). Long-term treatment of municipal wastewater under psychrophilic 2945
conditions. Bioresource Technology, 2015. 198: p. 510-519. 2946
111. Harb, M., et al., Differences in microbial communities and performance between 2947
suspended and attached growth anaerobic membrane bioreactors treating synthetic 2948
municipal wastewater. Environmental Science: Water Research & Technology, 2015. 2949
1(6): p. 800-813. 2950
112. Matsuura, N., et al., Recovery and biological oxidation of dissolved methane in effluent 2951
from UASB treatment of municipal sewage using a two-stage closed downflow hanging 2952
sponge system. Journal of Environmental Management, 2015. 151: p. 200-209. 2953
113. Shin, C., et al., Pilot-scale temperate-climate treatment of domestic wastewater with a 2954
staged anaerobic fluidized membrane bioreactor (SAF-MBR). Bioresource Technology, 2955
2014. 159: p. 95-103. 2956
114. Smith, A.L., S.J. Skerlos, and L. Raskin, Membrane biofilm development improves COD 2957
removal in anaerobic membrane bioreactor wastewater treatment. Microbial 2958
Biotechnology, 2015. 8(5): p. 883-894. 2959
115. Cookney, J., et al., Dissolved methane recovery from anaerobic effluents using hollow 2960
fibre membrane contactors. Journal of Membrane Science, 2016. 502: p. 141-150. 2961
116. McLeod, A., B. Jefferson, and E.J. McAdam, Toward gas-phase controlled mass transfer 2962
in micro-porous membrane contactors for recovery and concentration of dissolved 2963
methane in the gas phase. Journal of Membrane Science, 2016. 510: p. 466-471. 2964
117. Wongchitphimon, S., et al., Polymer-fluorinated silica composite hollow fiber 2965
membranes for the recovery of biogas dissolved in anaerobic effluent. Journal of 2966
Membrane Science, 2017. 540: p. 146-154. 2967
118. Bandara, W., et al., Introduction of a degassing membrane technology into anaerobic 2968
wastewater treatment. Water Environment Research, 2013. 85(5): p. 387-390. 2969
119. Henares, M., et al., Comparative study of degassing membrane modules for the removal 2970
of methane from Expanded Granular Sludge Bed anaerobic reactor effluent. Separation 2971
and Purification Technology, 2016. 170: p. 22-29. 2972
120. Henares, M., et al., Dissolved methane tecovery from an anaerobic effluent using a 2973
PDMS hollow fiber membrane contactor. Soria, Spain, 2016: p. 30. 2974
121. Rongwong, W., et al., Transport properties of CO 2 and CH 4 in hollow fiber membrane 2975
contactor for the recovery of biogas from anaerobic membrane bioreactor effluent. 2976
Journal of Membrane Science, 2017. 541: p. 62-72. 2977
122. Stolaroff, J.K., et al., Review of methane mitigation technologies with application to 2978
rapid release of methane from the Arctic. Environmental science & technology, 2012. 2979
46(12): p. 6455-6469. 2980
123. Hatamoto, M., et al., Biological oxidation of dissolved methane in effluents from 2981
anaerobic reactors using a down-flow hanging sponge reactor. Water Research, 2010. 2982
44(5): p. 1409-1418. 2983
149

124. Matsuura, N., et al., Closed DHS system to prevent dissolved methane emissions as 2984
greenhouse gas in anaerobic wastewater treatment by its recovery and biological 2985
oxidation. Water Science and Technology, 2010. 61(9): p. 2407-2415. 2986
125. Hatamoto, M., et al., Dissolved methane oxidation and competition for oxygen in down- 2987
flow hanging sponge reactor for post-treatment of anaerobic wastewater treatment. 2988
Bioresource Technology, 2011. 102(22): p. 10299-10304. 2989
126. van Kessel, M.A., et al., Current perspectives on the application of N-damo and 2990
anammox in wastewater treatment. Current Opinion in Biotechnology, 2018. 50: p. 222- 2991
227. 2992
127. Chen, X., et al., Modeling of simultaneous anaerobic methane and ammonium oxidation 2993
in a membrane biofilm reactor. Environmental Science & Technology, 2014. 48(16): p. 2994
9540-9547. 2995
128. Chen, X., et al., A new approach to simultaneous ammonium and dissolved methane 2996
removal from anaerobic digestion liquor: A model-based investigation of feasibility. 2997
Water Research, 2015. 85: p. 295-303. 2998
129. Cai, C., et al., Nitrate reduction by denitrifying anaerobic methane oxidizing 2999
microorganisms can reach a practically useful rate. Water Research, 2015. 87: p. 211- 3000
217. 3001
130. Shi, Y., et al., Nitrogen removal from wastewater by coupling anammox and methane- 3002
dependent denitrification in a membrane biofilm reactor. Environmental science & 3003
technology, 2013. 47(20): p. 11577-11583. 3004
131. Chen, X., et al., Achieving complete nitrogen removal by coupling nitritation ‐anammox 3005
and methane ‐dependent denitrification: A model ‐based study. Biotechnology and 3006
Bioengineering, 2016. 113(5): p. 1035-1045. 3007
132. Chen, X., et al., Model-based feasibility assessment of membrane biofilm reactor to 3008
achieve simultaneous ammonium, dissolved methane, and sulfide removal from 3009
anaerobic digestion liquor. Scientific Reports, 2016. 6. 3010
133. Winkler, M.H., et al., Modelling simultaneous anaerobic methane and ammonium 3011
removal in a granular sludge reactor. Water Research, 2015. 73: p. 323-331. 3012
134. Hu, B., et al., Cultivation of nitrite-dependent anaerobic methane-oxidizing bacteria: 3013
impact of reactor configuration. Applied Microbiology and Biotechnology, 2014. 98(18): 3014
p. 7983-7991. 3015
135. Kampman, C., et al., Enrichment of denitrifying methanotrophic bacteria for application 3016
after direct low-temperature anaerobic sewage treatment. Journal of hazardous 3017
materials, 2012. 227: p. 164-171. 3018
136. Zhu, B., et al., Combined anaerobic ammonium and methane oxidation for nitrogen and 3019
methane removal. 2011, Portland Press Limited. 3020
137. Ding, J., et al., Decoupling of DAMO archaea from DAMO bacteria in a methane-driven 3021
microbial fuel cell. Water Research, 2017. 110: p. 112-119. 3022
138. McAnulty, M.J., et al., Electricity from methane by reversing methanogenesis. Nature 3023
Communications, 2017. 8. 3024
139. Chen, S. and A.L. Smith, Methane-driven microbial fuel cells recover energy and mitigate 3025
dissolved methane emissions from anaerobic effluents. Environmental Science: Water 3026
Research & Technology, 2018. 4(1): p. 67-79. 3027
140. Myung, J., P.E. Saikaly, and B.E. Logan, A two-staged system to generate electricity in 3028
microbial fuel cells using methane. Chemical Engineering Journal, 2018. 352: p. 262-267. 3029
141. Girguis, P. and C.E. Reimers, Methane-powered microbial fuel cells, 3030
PCT/US2009/003209, Editor. 2011, Google Patents. 3031
150

142. Ren, Z.J., Microbial fuel cells: Running on gas. Nature Energy, 2017. 2(17093): p. 1. 3032
143. Logan, B.E., et al., Assessment of microbial fuel cell configurations and power densities. 3033
Environmental Science & Technology Letters, 2015. 2(8): p. 262-267. 3034
144. Ge, Z. and Z. He, Long-term performance of a 200 liter modularized microbial fuel cell 3035
system treating municipal wastewater: treatment, energy, and cost. Environmental 3036
Science: Water Research & Technology, 2016. 2(2): p. 274-281. 3037
145. Scherson, Y.D., S.-G. Woo, and C.S. Criddle, Production of nitrous oxide from anaerobic 3038
digester centrate and its use as a co-oxidant of biogas to enhance energy recovery. 3039
Environmental Science & Technology, 2014. 48(10): p. 5612-5619. 3040
146. Scherson, Y.D., et al., Nitrogen removal with energy recovery through N 2O 3041
decomposition. Energy & Environmental Science, 2013. 6(1): p. 241-248. 3042
147. Jin, Y., M.C. Veiga, and C. Kennes, Bioprocesses for the removal of nitrogen oxides from 3043
polluted air. Journal of Chemical Technology and Biotechnology, 2005. 80(5): p. 483-494. 3044
148. Shoener, B., et al., Energy positive domestic wastewater treatment: the roles of 3045
anaerobic and phototrophic technologies. Environmental Science: Processes & Impacts, 3046
2014. 16(6): p. 1204-1222. 3047
149. Rau, G.H., H.D. Willauer, and Z.J. Ren, The global potential for converting renewable 3048
electricity to negative-CO 2-emissions hydrogen. Nature Climate Change, 2018. 8(7): p. 3049
621. 3050
150. Strong, P.J., S. Xie, and W.P. Clarke, Methane as a resource: can the methanotrophs add 3051
value? Environmental Science & Technology, 2015. 49(7): p. 4001-4018. 3052
151. Strong, P.J., et al., The opportunity for high-performance biomaterials from methane. 3053
Microorganisms, 2016. 4(1): p. 11. 3054
152. Guo, W., et al., Synthesizing value ‐added products from methane by a new 3055
Methylomonas. Journal of Applied Microbiology, 2017. 123(5): p. 1214-1227. 3056
153. Hwang, I.Y., et al., Biological conversion of methane to chemicals and fuels: technical 3057
challenges and issues. Applied Microbiology and Biotechnology, 2018. 102(7): p. 3071- 3058
3080. 3059
154. Taylor, A., et al., Immobilization of Methylosinus trichosporium OB3b for methanol 3060
production. Journal of Industrial Microbiology & Biotechnology, 2018: p. 1-11. 3061
155. Bagheri, N. and J. Abedi, Adsorption of methane on corn cobs based activated carbon. 3062
Chemical Engineering Research and Design, 2011. 89(10): p. 2038-2043. 3063
156. Chen, J.-J., et al., Improving biogas separation and methane storage with multilayer 3064
graphene nanostructure via layer spacing optimization and lithium doping: a molecular 3065
simulation investigation. Environmental Science & Technology, 2012. 46(18): p. 10341- 3066
10348. 3067
157. Tian, T., et al., A sol–gel monolithic metal–organic framework with enhanced methane 3068
uptake. Nature materials, 2018. 17(2): p. 174-179. 3069
158. Conti, J., et al., International energy outlook 2016 with projections to 2040. 2016, USDOE 3070
Energy Information Administration (EIA), Washington, DC (United States …. 3071
159. Sommer, J., et al., Quantification of emitted and retained N 2O in a municipal waste 3072
water treatment plant with activated sludge and nitrifying-denitrifying units. 3073
Agribiological Research (Germany), 1998. 51(1): p. 59-73. 3074
160. Ahn, J.H., et al., N 2O emissions from activated sludge processes, 2008− 2009: results of a 3075
national monitoring survey in the United States. Environmental Science & Technology, 3076
2010. 44(12): p. 4505-4511. 3077
151

161. Sun, S., Z. Bao, and D. Sun, Study on emission characteristics and reduction strategy of 3078
nitrous oxide during wastewater treatment by different processes. Environmental 3079
Science and Pollution Research, 2015. 22(6): p. 4222-4229. 3080
162. Desloover, J., et al., Floc-based sequential partial nitritation and anammox at full scale 3081
with contrasting N2O emissions. Water Research, 2011. 45(9): p. 2811-2821. 3082
163. Kampschreur, M., et al., Emission of nitrous oxide and nitric oxide from a full-scale 3083
single-stage nitritation-anammox reactor. Water Science and Technology, 2009. 60(12): 3084
p. 3211-3217. 3085
164. Harb, M. and P.-Y. Hong, Anaerobic membrane bioreactor effluent reuse: A review of 3086
microbial safety concerns. Fermentation, 2017. 3(3): p. 39. 3087
165. Harb, M. and P.-Y. Hong, Molecular-based detection of potentially pathogenic bacteria in 3088
membrane bioreactor (MBR) systems treating municipal wastewater: a case study. 3089
Environmental Science and Pollution Research, 2017. 24(6): p. 5370-5380. 3090
166. Yoo, R., et al., Effect of temperature on the treatment of domestic wastewater with a 3091
staged anaerobic fluidized membrane bioreactor. Water Science and Technology, 2014. 3092
69(6): p. 1145-1150. 3093
167. Luo, G., W. Wang, and I. Angelidaki, A new degassing membrane coupled upflow 3094
anaerobic sludge blanket (UASB) reactor to achieve in-situ biogas upgrading and 3095
recovery of dissolved CH 4 from the anaerobic effluent. Applied Energy, 2014. 132: p. 3096
536-542. 3097
168. Iversen, N. and B. Jørgensen, Anaerobic methane oxidation rates at the sulfate-methane 3098
transition in marine sediments from Kattegat and Skagerrak (Denmark). Limnol. 3099
Oceanogr, 1985. 30(5): p. 944-955. 3100
169. Boetius, A., et al., A marine microbial consortium apparently mediating anaerobic 3101
oxidation of methane. Nature, 2000. 407(6804): p. 623-626. 3102
170. Islas-Lima, S., F. Thalasso, and J. Gomez-Hernandez, Evidence of anoxic methane 3103
oxidation coupled to denitrification. Water Research, 2004. 38(1): p. 13-16. 3104
171. Raghoebarsing, A.A., et al., A microbial consortium couples anaerobic methane oxidation 3105
to denitrification. Nature, 2006. 440(7086): p. 918-921. 3106
172. Knittel, K. and A. Boetius, Anaerobic oxidation of methane: progress with an unknown 3107
process. Annual Review of Microbiology, 2009. 63: p. 311-334. 3108
173. Delgado Vela, J., et al., Prospects for biological nitrogen removal from anaerobic 3109
effluents during mainstream wastewater treatment. Environmental Science & 3110
Technology Letters, 2015. 2(9): p. 234-244. 3111
174. Hanson, R.S. and T.E. Hanson, Methanotrophic bacteria. Microbiological Reviews, 1996. 3112
60(2): p. 439-471. 3113
175. Kalyuzhnaya, M.G., A.W. Puri, and M.E. Lidstrom, Metabolic engineering in 3114
methanotrophic bacteria. Metabolic Engineering, 2015. 29: p. 142-152. 3115
176. Morinaga, Y., et al., Methane metabolism of the obligate methane-utilizing bacterium, 3116
Methylomonas flagellata, in methane-limited and oxygen-limited chemostat culture. 3117
Agricultural and Biological Chemistry, 1979. 43(12): p. 2453-2458. 3118
177. Kalyuzhnaya, M.G., et al., Highly efficient methane biocatalysis revealed in a 3119
methanotrophic bacterium. Nature Communications, 2013. 4. 3120
178. Gilman, A., et al., Bioreactor performance parameters for an industrially-promising 3121
methanotroph Methylomicrobium buryatense 5GB1. Microbial Cell Factories, 2015. 3122
14(1): p. 182. 3123
152

179. Li, W.-W., H.-Q. Yu, and Z. He, Towards sustainable wastewater treatment by using 3124
microbial fuel cells-centered technologies. Energy & Environmental Science, 2014. 7(3): 3125
p. 911-924. 3126
180. Tian, Y., et al., In-situ integration of microbial fuel cell with hollow-fiber membrane 3127
bioreactor for wastewater treatment and membrane fouling mitigation. Biosensors and 3128
Bioelectronics, 2015. 64: p. 189-195. 3129
181. Kim, K.-Y., et al., Performance of anaerobic fluidized membrane bioreactors using 3130
effluents of microbial fuel cells treating domestic wastewater. Bioresource Technology, 3131
2016. 208: p. 58-63. 3132
182. Donovan, C., et al., Batteryless, wireless sensor powered by a sediment microbial fuel 3133
cell. Environmental Science & Technology, 2008. 42(22): p. 8591-8596. 3134
183. Ahn, Y. and B.E. Logan, A multi-electrode continuous flow microbial fuel cell with 3135
separator electrode assembly design. Applied Microbiology and Biotechnology, 2012. 3136
93(5): p. 2241-2248. 3137
184. Feng, Y., et al., Treatment of carbon fiber brush anodes for improving power generation 3138
in air–cathode microbial fuel cells. Journal of Power Sources, 2010. 195(7): p. 1841- 3139
1844. 3140
185. Cheng, S., H. Liu, and B.E. Logan, Increased performance of single-chamber microbial fuel 3141
cells using an improved cathode structure. Electrochemistry Communications, 2006. 3142
8(3): p. 489-494. 3143
186. Kozich, J.J., et al., Development of a dual-index sequencing strategy and curation 3144
pipeline for analyzing amplicon sequence data on the MiSeq Illumina sequencing 3145
platform. Applied and Environmental Microbiology, 2013. 79(17): p. 5112-5120. 3146
187. Pruesse, E., et al., SILVA: a comprehensive online resource for quality checked and 3147
aligned ribosomal RNA sequence data compatible with ARB. Nucleic Acids Research, 3148
2007. 35(21): p. 7188-7196. 3149
188. Cummings, D.E., et al., Diversity of Geobacteraceae species inhabiting metal-polluted 3150
freshwater lake sediments ascertained by 16S rDNA analyses. Microbial Ecology, 2003. 3151
46(2): p. 257-269. 3152
189. Bourne, D.G., I.R. McDonald, and J.C. Murrell, Comparison of pmoA PCR primer sets as 3153
tools for investigating methanotroph diversity in three Danish soils. Applied and 3154
Environmental Microbiology, 2001. 67(9): p. 3802-3809. 3155
190. Liu, H. and B.E. Logan, Electricity generation using an air-cathode single chamber 3156
microbial fuel cell in the presence and absence of a proton exchange membrane. 3157
Environmental Science & Technology, 2004. 38(14): p. 4040-4046. 3158
191. Xing, D., et al., Electricity generation by Rhodopseudomonas palustris DX-1. 3159
Environmental Science & Technology, 2008. 42(11): p. 4146-4151. 3160
192. Whittenbury, R., K. Phillips, and J. Wilkinson, Enrichment, isolation and some properties 3161
of methane-utilizing bacteria. Microbiology, 1970. 61(2): p. 205-218. 3162
193. Aelterman, P., et al., Continuous electricity generation at high voltages and currents 3163
using stacked microbial fuel cells. Environmental Science & Technology, 2006. 40(10): p. 3164
3388-3394. 3165
194. Blazewicz, S.J., et al., Evaluating rRNA as an indicator of microbial activity in 3166
environmental communities: limitations and uses. The ISME journal, 2013. 7(11): p. 3167
2061-2068. 3168
195. Props, R., et al., Absolute quantification of microbial taxon abundances. The ISME 3169
Journal, 2016. 3170
153

196. Smets, W., et al., A method for simultaneous measurement of soil bacterial abundances 3171
and community composition via 16S rRNA gene sequencing. Soil Biology and 3172
Biochemistry, 2016. 96: p. 145-151. 3173
197. Eshinimaev, B.T., et al., Physiological, biochemical, and cytological characteristics of a 3174
haloalkalitolerant methanotroph grown on methanol. Microbiology, 2002. 71(5): p. 512- 3175
518. 3176
198. Rhee, G.-Y. and G.W. Fuhs, Wastewater denitrification with one-carbon compounds as 3177
energy source. Journal (Water Pollution Control Federation), 1978: p. 2111-2119. 3178
199. Vallieres, C. and E. Favre, Vacuum versus sweeping gas operation for binary mixtures 3179
separation by dense membrane processes. Journal of Membrane Science, 2004. 244(1): 3180
p. 17-23. 3181
200. Zarei-Baygi, A., et al., Evaluating Antibiotic Resistance Gene Correlations with Antibiotic 3182
Exposure Conditions in Anaerobic Membrane Bioreactors. Environmental science & 3183
technology, 2019. 3184
201. Yu, L., et al., Novel Gas Diffusion Cloth Bioanodes for High-Performance Methane- 3185
Powered Microbial Fuel Cells. Environmental Science & Technology, 2018. 53(1): p. 530- 3186
538. 3187
202. Gao, Y., et al., Anaerobic oxidation of methane coupled with extracellular electron 3188
transfer to electrodes. Scientific Reports, 2017. 7(1): p. 5099. 3189
203. Yamasaki, R., T. Maeda, and T.K. Wood, Electron carriers increase electricity production 3190
in methane microbial fuel cells that reverse methanogenesis. Biotechnology for Biofuels, 3191
2018. 11(1): p. 211. 3192
204. Ahn, Y. and B.E. Logan, Effectiveness of domestic wastewater treatment using microbial 3193
fuel cells at ambient and mesophilic temperatures. Bioresource Technology, 2010. 3194
101(2): p. 469-475. 3195
205. Wang, X., Y. Feng, and H. Lee, Electricity production from beer brewery wastewater using 3196
single chamber microbial fuel cell. Water Science and Technology, 2008. 57(7): p. 1117- 3197
1121. 3198
206. Larrosa-Guerrero, A., et al., Effect of temperature on the performance of microbial fuel 3199
cells. Fuel, 2010. 89(12): p. 3985-3994. 3200
207. Mei, X., et al., Adaptation of microbial community of the anode biofilm in microbial fuel 3201
cells to temperature. Bioelectrochemistry, 2017. 3202
208. Min, B., Ó .B. Román, and I. Angelidaki, Importance of temperature and anodic medium 3203
composition on microbial fuel cell (MFC) performance. Biotechnology Letters, 2008. 3204
30(7): p. 1213-1218. 3205
209. Cheng, S., D. Xing, and B.E. Logan, Electricity generation of single-chamber microbial fuel 3206
cells at low temperatures. Biosensors and Bioelectronics, 2011. 26(5): p. 1913-1917. 3207
210. Zhang, Y., et al., Effects of periodically alternating temperatures on performance of 3208
single-chamber microbial fuel cells. International Journal of Hydrogen Energy, 2014. 3209
39(15): p. 8048-8054. 3210
211. Liu, H., S. Cheng, and B.E. Logan, Power generation in fed-batch microbial fuel cells as a 3211
function of ionic strength, temperature, and reactor configuration. Environmental 3212
Science & Technology, 2005. 39(14): p. 5488-5493. 3213
212. Jadhav, G. and M. Ghangrekar, Performance of microbial fuel cell subjected to variation 3214
in pH, temperature, external load and substrate concentration. Bioresource Technology, 3215
2009. 100(2): p. 717-723. 3216
213. Liu, L., et al., Double-chamber microbial fuel cells started up under room and low 3217
temperatures. International Journal of Hydrogen Energy, 2013. 38(35): p. 15574-15579. 3218
154

214. del Campo, A.G., et al., Short-term effects of temperature and COD in a microbial fuel 3219
cell. Applied Energy, 2013. 101: p. 213-217. 3220
215. Trotsenko, Y.A. and V.N. Khmelenina, Aerobic methanotrophic bacteria of cold 3221
ecosystems. FEMS Microbiology Ecology, 2005. 53(1): p. 15-26. 3222
216. Caccavo, F., et al., Geobacter sulfurreducens sp. nov., a hydrogen-and acetate-oxidizing 3223
dissimilatory metal-reducing microorganism. Applied and Environmental Microbiology, 3224
1994. 60(10): p. 3752-3759. 3225
217. Rognes, T., et al., VSEARCH: a versatile open source tool for metagenomics. PeerJ, 2016. 3226
4: p. e2584. 3227
218. Yue, J.C. and M.K. Clayton, A similarity measure based on species proportions. 3228
Communications in Statistics-theory and Methods, 2005. 34(11): p. 2123-2131. 3229
219. Segata, N., et al., Metagenomic biomarker discovery and explanation. Genome Biology, 3230
2011. 12(6): p. R60. 3231
220. Spearman, C., The proof and measurement of association between two things. American 3232
journal of Psychology, 1904. 15(1): p. 72-101. 3233
221. Ucar, D., Y. Zhang, and I. Angelidaki, An overview of electron acceptors in microbial fuel 3234
cells. Frontiers in Microbiology, 2017. 8: p. 643. 3235
222. Liu, H., et al., Scale-up of membrane-free single-chamber microbial fuel cells. Journal of 3236
Power Sources, 2008. 179(1): p. 274-279. 3237
223. Zhang, G., et al., Efficient electricity generation from sewage sludge usingbiocathode 3238
microbial fuel cell. Water Research, 2012. 46(1): p. 43-52. 3239
224. Gregory, K.B., D.R. Bond, and D.R. Lovley, Graphite electrodes as electron donors for 3240
anaerobic respiration. Environmental Microbiology, 2004. 6(6): p. 596-604. 3241
225. Tanaka, K., et al., Extracellular electron transfer via outer membrane cytochromes in a 3242
methanotrophic bacterium Methylococcus capsulatus (Bath). Frontiers in Microbiology, 3243
2018. 9: p. 2905. 3244
226. Ha, P.T., B. Tae, and I.S. Chang, Performance and bacterial consortium of microbial fuel 3245
cell fed with formate. Energy & Fuels, 2007. 22(1): p. 164-168. 3246
227. Zhao, F., et al., Activated carbon cloth as anode for sulfate removal in a microbial fuel 3247
cell. Environmental Science & Technology, 2008. 42(13): p. 4971-4976. 3248
228. Kim, T.G., S.-Y. Jeong, and K.-S. Cho, Functional rigidity of a methane biofilter during the 3249
temporal microbial succession. Applied Microbiology and Biotechnology, 2014. 98(7): p. 3250
3275-3286. 3251
229. Costa, C., et al., Denitrification with methane as electron donor in oxygen-limited 3252
bioreactors. Applied Microbiology and Biotechnology, 2000. 53(6): p. 754-762. 3253
230. Šantl-Temkiv, T., et al., Viable methanotrophic bacteria enriched from air and rain can 3254
oxidize methane at cloud-like conditions. Aerobiologia, 2013. 29(3): p. 373-384. 3255
231. Su, Y., et al., Microbial community and function of enrichment cultures with methane 3256
and toluene. Applied Microbiology and Biotechnology, 2014. 98(7): p. 3121-3131. 3257
232. Hršak, D. and A. Begonja, Possible interactions within a methanotrophic-heterotrophic 3258
groundwater community able to transform linear alkylbenzenesulfonates. Applied and 3259
Environmental Microbiology, 2000. 66(10): p. 4433-4439. 3260
233. Heidrich, E., et al., Temperature, inocula and substrate: Contrasting electroactive 3261
consortia, diversity and performance in microbial fuel cells. Bioelectrochemistry, 2018. 3262
119: p. 43-50. 3263
234. Tchobanoglous, G., F.L. Burton, and H. Stensel, Metcalf & Eddy.(2003). Wastewater 3264
engineering: treatment and reuse. 4. 3265
155

235. Vaiopoulou, E., P. Melidis, and A. Aivasidis, Sulfide removal in wastewater from 3266
petrochemical industries by autotrophic denitrification. Water Research, 2005. 39(17): p. 3267
4101-4109. 3268
236. Dutta, P.K., et al., Electrochemical sulfide removal and recovery from paper mill 3269
anaerobic treatment effluent. Water research, 2010. 44(8): p. 2563-2571. 3270
237. Vannini, C., et al., Sulphide oxidation to elemental sulphur in a membrane bioreactor: 3271
Performance and characterization of the selected microbial sulphur-oxidizing 3272
community. Systematic and applied microbiology, 2008. 31(6-8): p. 461-473. 3273
238. Hao, T.-w., et al., A review of biological sulfate conversions in wastewater treatment. 3274
Water research, 2014. 65: p. 1-21. 3275
239. Grady Jr, C.L., et al., Biological wastewater treatment. 2011: CRC press. 3276
240. Zhang, L., et al., Chemical and biological technologies for hydrogen sulfide emission 3277
control in sewer systems: a review. Water research, 2008. 42(1-2): p. 1-12. 3278
241. Kanjanarong, J., et al., Removal of hydrogen sulfide generated during anaerobic 3279
treatment of sulfate-laden wastewater using biochar: evaluation of efficiency and 3280
mechanisms. Bioresource technology, 2017. 234: p. 115-121. 3281
242. Karhadkar, P., et al., Sulfide and sulfate inhibition of methanogenesis. Water Research, 3282
1987. 21(9): p. 1061-1066. 3283
243. Koster, I., et al., Sulfide inhibition of the methanogenic activity of granular sludge at 3284
various pH-levels. Water Research, 1986. 20(12): p. 1561-1567. 3285
244. Sahinkaya, E., et al., Performance of a sulfide-oxidizing, sulfur-producing membrane 3286
biofilm reactor treating sulfide-containing bioreactor effluent. Environmental Science & 3287
Technology, 2011. 45(9): p. 4080-4087. 3288
245. Pikaar, I., et al., Electrochemical sulfide oxidation from domestic wastewater using mixed 3289
metal-coated titanium electrodes. Water Research, 2011. 45(17): p. 5381-5388. 3290
246. Chen, C., et al., Enhanced performance of denitrifying sulfide removal process under 3291
micro-aerobic condition. Journal of hazardous materials, 2010. 179(1-3): p. 1147-1151. 3292
247. Dutta, P.K., et al., Spontaneous electrochemical removal of aqueous sulfide. Water 3293
research, 2008. 42(20): p. 4965-4975. 3294
248. Dutta, P.K., et al., Role of sulfur during acetate oxidation in biological anodes. 3295
Environmental science & technology, 2009. 43(10): p. 3839-3845. 3296
249. Sun, M., et al., Microbe-assisted sulfide oxidation in the anode of a microbial fuel cell. 3297
Environmental science & technology, 2009. 43(9): p. 3372-3377. 3298
250. E.W. Rice, R.B.B., A.D. Eaton, L.S. Clesceri, Standard methods for the examination of 3299
water and wastewater. 22 ed. 2012: American Public Health Association, American 3300
Water Works Association, Water Environment Federation. 3301
251. Zhang, W., et al., Isolation of a methanotroph from a hydrogen sulfide-rich anaerobic 3302
digester for methanol production from biogas. Process Biochemistry, 2016. 51(7): p. 3303
838-844. 3304
252. Snider, R.M., et al., Long-range electron transport in Geobacter sulfurreducens biofilms is 3305
redox gradient-driven. Proceedings of the National Academy of Sciences, 2012: p. 3306
201209829. 3307
253. Malvankar, N.S., et al., Tunable metallic-like conductivity in microbial nanowire 3308
networks. Nature Nanotechnology, 2011. 6(9): p. 573. 3309
254. Yates, M.D., et al., Thermally activated long range electron transport in living biofilms. 3310
Physical Chemistry Chemical Physics, 2015. 17(48): p. 32564-32570. 3311
255. Alrashed, W., et al., Hypoxic methane oxidation coupled to denitrification in a membrane 3312
biofilm. Chemical Engineering Journal, 2018. 348: p. 745-753. 3313
156

256. Lai, C.-Y., et al., Bromate and nitrate bioreduction coupled with Poly-β-hydroxybutyrate 3314
production in a methane-based membrane biofilm reactor. Environmental science & 3315
technology, 2018. 52(12): p. 7024-7031. 3316
257. Gilman, A., et al., Oxygen-limited metabolism in the methanotroph Methylomicrobium 3317
buryatense 5GB1C. PeerJ, 2017. 5: p. e3945. 3318
258. Ge, Z., et al., Capacitive deionization for nutrient recovery from wastewater with 3319
disinfection capability. Environmental Science: Water Research & Technology, 2018. 3320
4(1): p. 33-39. 3321
3322

Abstract (if available)

Abstract Given the increasing emphasis on resource recovery (e.g., water, energy, and nutrients) from domestic wastewater, high-rate anaerobic processes are drawing attention due to their ability to produce methane-rich biogas, significantly reduce sludge production, and produce an effluent of comparable quality to activated sludge processes. However, anaerobic effluents contain dissolved methane and other constituents (e.g., sulfides and nutrients) that require management prior to discharge or reuse. In particular, dissolved methane poses a risk as a potent greenhouse gas if emitted to the atmosphere. Bioelectrochemical systems such as microbial fuel cells (MFCs) are an attractive technology to recover energy from dissolved methane and sulfide. Specifically, air-cathode MFCs require minimal energy demands due to their ability to passively provide oxygen via diffusion across an atmosphere-exposed cathode. In this dissertation, air-cathode MFCs were investigated to treat dissolved methane and sulfide across a range of operational conditions. First, two replicate bench-scale air-cathode MFCs were operated on a synthetic anaerobic effluent containing dissolved methane as the sole organic electron donor. Up to 85% dissolved methane removal was achieved resulting in 0.5 to 0.6 V and a maximum Coulombic efficiency of 17.7%. Longer hydraulic retention time yielded higher voltage generation and dissolved methane removal due to increased oxygen diffusion. A methanotroph-exoelectrogen interaction was proposed based on distinct colonization of methanotrophs and Geobacter on cathode and anode biofilms, respectively. Sequential addition of likely methanotrophic intermediates suggested that formate served as the electron shuttle between the two populations. Next, the replicate MFCs were subjected to operational temperatures of 25, 20, 15, 10 and 5℃ to investigate performance across seasonal temperature variations common to temperate climates. Though voltage abruptly decreased at and below 10℃, stable dissolved methane removal was achieved at all operational temperatures. High-throughput sequencing revealed methantrophs (e.g., Methylobacillus and Methylomonas) and exoelectrogens (e.g., Geobacter spp. and Ferribacterium) at cathode and anode biofilms, respectively. Spearman rank correlation analysis suggested that fermentative bacteria may play a role mediating interactions between methanotrophs and exoelectrogens. Anode microbial diversity was found to strongly correlate with system performance. Finally, one methane-driven MFC was subjected to sulfide addition at concentrations of 1, 5, 10 and 20 mg/L. Even at 20 mg/L, sulfide addition did not negatively impact dissolved methane removal or voltage production, and sulfide removal was 46.8 ± 8.9%. Future research is necessary to optimize sulfide removal and evaluate long-term performance at the pilot-scale on actual anaerobic effluents.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Advancing energy recovery from food waste using anaerobic biotechnologies: performance and microbial ecology

PDF

Fate of antibiotic resistance in anaerobic membrane bioreactors

PDF

Thermally driven water treatment with membrane distillation: membrane performance, waste heat integration, and cooling analysis

PDF

Catalytic oxidation and precursor identification of disinfection byproducts in recycled water

PDF

Identifying and mitigating the effects of urban heat islands in California

PDF

From cables to biofilms: electronic and electrochemical characterization of electroactive microbial systems

Asset Metadata

Creator Chen, Siming (author)

Core Title Bioelectrochemical treatment of anaerobic process effluents: mitigation of dissolved methane and sulfide

School Viterbi School of Engineering

Degree Doctor of Philosophy

Degree Program Engineering (Environmental Engineering)

Publication Date 08/06/2019

Defense Date 06/07/2019

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag anaerobic effluents,dissolved methane,exoelectrogen,methanotrophs,microbial fuel cell,OAI-PMH Harvest,sulfide

Format application/pdf (imt)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Smith, Adam (committee chair), Boedicker, James (committee member), Childress, Amy (committee member), McCurry, Daniel (committee member)

Creator Email daming6655@gmail.com,simingc@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c89-206561

Unique identifier UC11662692

Identifier etd-ChenSiming-7748.pdf (filename),usctheses-c89-206561 (legacy record id)

Legacy Identifier etd-ChenSiming-7748.pdf

Dmrecord 206561

Document Type Dissertation

Format application/pdf (imt)

Rights Chen, Siming

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA