Content uploaded by Jan B. A. Arends
Author content
All content in this area was uploaded by Jan B. A. Arends on Apr 09, 2014
Content may be subject to copyright.
biblio.ugent.be
The UGent Institutional Repository is the electronic archiving and dissemination platform for all
UGent research publications. Ghent University has implemented a mandate stipulating that all
academic publications of UGent researchers should be deposited and archived in this repository.
Except for items where current copyright restrictions apply, these papers are available in Open
Access.
This item is the archived peer-reviewed author-version of:
Title: Greenhouse gas emissions from rice microcosms amended with a plant microbial fuel cell.
Authors: Jan B. A. Arends, Jonas Speeckaert, Evelyne Blondeel, Jo De Vrieze, Pascal Boeckx, Willy
Verstraete, Korneel Rabaey, Nico Boon.
In: Applied Microbiology and Biotechnology 98 (7) 3205-3217
Link: http://link.springer.com/article/10.1007%2Fs00253-013-5328-5
To refer to or to cite this work, please use the citation to the published version:
Arends, J. B. A., Speeckaert, J., Blondeel, E., Vrieze, J., Boeckx, P., Verstraete, W., Rabaey, K., Boon, N.
2014. Greenhouse gas emissions from rice microcosms amended with a plant microbial fuel
cell. Applied Microbiology and Biotechnology, 98(7), 3205-3217. DOI: 10.1007/s00253-013-
5328-5
Greenhouse gas emissions from rice microcosms amended with a plant microbial fuel cell.
Jan B. A. Arends1, Jonas Speeckaert1, Evelyne Blondeel1#, Jo De Vrieze1, Pascal Boeckx2, Willy
Verstraete1, Korneel Rabaey1, Nico Boon1*
1Laboratory of Microbial Ecology and Technology (LabMET), Ghent University, Coupure Links 653, B-
9000 Gent, Belgium.
2Isotope Bioscience Laboratory (ISOFYS), Ghent University, Coupure Links 653, B-9000 Gent, Belgium.
# Currently at: EnBiChem, HOWEST, Graaf Karel de Goedelaan 5, B-8500 Kortrijk, Belgium
*Corresponding author. Laboratory of Microbial Ecology and Technology (LabMET), Ghent
University, Coupure Links 653, B-9000 Ghent, Belgium;
phone: +32 (0)9 264 59 76;
fax: +32 (0)9 264 62 48;
E-mail: Nico.Boon@ugent.be;
Webpage: www.labmet.Ugent.be.
Key words:
Bioelectrochemical system, methane, microbial ecology, hydrogenotrophic methanogens, electrode
material, nitrous oxide, rhizosphere, Oryza sativa.
Abstract
Methane (CH4) release from wetlands is an important source of greenhouse gas emissions. Gas
exchange occurs mainly through the aerenchyma of plants and production of greenhouse gases is
heavily dependent on rhizosphere biogeochemical conditions (i.e. substrate availability and redox
potential). It is hypothesized that by introducing a biocatalyzed anode electrode in the rhizosphere of
wetland plants, a competition for carbon and electrons can be invoked between electrical current
generating bacteria and methanogenic archaea. The anode electrode is part of a bioelectrochemical
system (BES) capable of harvesting electrical current from microbial metabolism. In this work, the
anode of a BES was introduced in the rhizosphere of rice plants (Oryza sativa) and the impact on
methane emissions was monitored.
Microbial current generation was able to outcompete methanogenic processes when the bulk matrix
contained low concentrations of organic carbon, provided that the electrical circuit with the effective
electro-active microorganisms was in place. When interrupting the electrical circuit or supplying an
excess of organic carbon, methanogenic metabolism was able to outcompete current generating
metabolism. The qPCR results showed hydrogenotrophic methanogens were the most abundant
methanogenic group present, while mixotrophic or acetoclastic methanogens were hardly detected
in the bulk rhizosphere or on the electrodes. Competition for electron donor and acceptor were likely
the main drivers to lower methane emissions. Overall, electrical current generation with BESs is an
interesting option to control CH4 emissions from wetlands but needs to be applied in combination
with other mitigation strategies to be successful and feasible in practice.
1. Introduction
Methane (CH4) emissions to the atmosphere arise from various (a)biotic sources. CH4 has a global
warming potential (GWP) of 25 times the GWP of CO2. Therefore these emissions need to be
carefully managed (Forster et al. 2007). The main anthropogenic sources of methane emissions are
ruminant farming (15-32%) and rice agriculture (9-19%). Taking also natural emissions into account,
wetlands and rice agriculture combined amount to 32-47% of total methane emissions (Denman et
al. 2007).
Anaerobic decomposition of organic matter resulting in the formation of methane is the
consequence of a series of biochemical transformations. Bacteria ensure the formation of monomers
from complex organic polymers. The monomers are subsequently transformed into various organic
acids. In the third step acetic acid, CO2 and H2 are formed which are finally transformed into methane
by the hydrogenotrophic or the acetoclastic methanogens (Appels et al. 2008; Conrad 2002). CH4
formation occurs under specific biogeochemical conditions such as redox potential (Eh), pH and
ammonia concentration (Appels et al. 2008; Conrad 2002; Johnson-Beebout et al. 2009; Majumdar
2003). These parameters give, next to competition for substrate, various handles for control of CH4
production. The main pathway of gas exchange with the rhizosphere under waterlogged conditions is
by transport through the aerenchyma of plants (90%). Other mechanisms include diffusion and
ebullition (Aulakh et al. 2000; Bazhin 2010). Therefore a promising route for mitigation of CH4
emissions is to prevent the formation of CH4 in the rhizosphere. Next to introduction of alternative
electron acceptors, other mitigation strategies include fertilizer management, water management,
biochar addition to the soil and choosing crop varieties with little aerenchyma or little
rhizodeposition (Aulakh et al. 2001; Aulakh et al. 2000; Conrad 2002; Majumdar 2003; Singh et al.
1999; Zhang et al. 2012).
The last decades microbial extra cellular electron transfer has gained attention on the promise of
generating electrical energy directly from various sources of organic matter (Rabaey and Verstraete
2005). During anaerobic respiration of organic carbon in a microbial fuel cell (MFC), microorganisms
are able to use an electrode (the anode) as an electron acceptor. When such a biocatalyzed anode
(bioanode) is connected with a (bio)cathode, electrical current can be directly harvested from the
microbial decomposition of organic matter in a bioelectrochemical system (BES). The typical
substrate for the anodic process is acetate although it has been shown that H2 can also be used
(Logan et al. 2006). Waterlogged soils and sediments containing organic matter have been exploited
for direct electrical current generation in sediment-MFCs (Donovan et al. 2011; Reimers et al. 2001;
Tender et al. 2008). A drawback of sediment-MFCs is the low flux of organic matter towards the
anode, limiting high current production. To overcome this issue, plants have been introduced into
the anode of a sediment-MFC creating a Plant-MFC which enables sustained current generation from
organic matter due to rhizodeposition processes (Strik et al. 2008). An example is the use of rice
plants, as they are able to withstand prolonged inundation of their rhizosphere (De Schamphelaire et
al. 2008; Kaku et al. 2008).
Methane emissions from rice paddies can be estimated to range from 0 up to 60 mgCH4 m -2 h-1
(Gogoi et al. 2005; Singh et al. 1999; Xu et al. 2007). This corresponds to the equivalent of an
electrical current of 0 to 804 mA m -2 (See supplementary information for calculations). Plant-MFCs
have been reported to produce a current up to 120 mA m -2 (De Schamphelaire et al. 2008). These
values are well within the same order of magnitude, therefore it has been hypothesized that current
generating metabolism can mitigate CH4 emissions from rice paddy soils and in more general terms,
waterlogged wetlands (Cabezas da Rosa 2010; De Schamphelaire et al. 2008; Hong et al. 2009; Ishii et
al. 2008; Kaku et al. 2008). Introducing an electrode (anode) in the rhizosphere of waterlogged plants
possibly enables the removal of electrons from the commonly present methanogenic metabolism
(Arends and Verstraete 2012; Cabezas da Rosa 2010; De Schamphelaire et al. 2008; Kaku et al. 2008).
This is analogous to CH4 emission mitigation strategies such as the addition of ferric iron or sulphate
to stimulate iron reducing or sulphate reducing microorganisms in their competition for reducing
equivalents with methanogens (Conrad 2002; Liesack et al. 2000).
Anodic oxidation of organic matter leads to a decrease in pH during current generation (Rozendal et
al. 2006; eq. S2). Methanogenic metabolism can be inhibited by a decrease in pH (Appels et al. 2008).
Acidification due to current generating metabolism can possibly form a second route to inhibition of
methanogenic activity, next to competition for substrate.
Cabezas de Rosa (2010) has observed a 47% decrease in pore water CH4 concentrations in a rice
paddy soil sediment-MFC without plants. Kaku et al. (2008) were not able to show lower CH4
emissions in closed circuit operation of a Plant-MFC during a field trial in a rice paddy. In this work
the results of a microcosm study are presented in which rice plants were grown in vermiculite with a
BES in the rhizosphere. The aim of this work was to investigate 1) the influence of the external
resistance 2) the mechanism and 3) the archaeal pathways involved in possible methane emission
mitigation in Plant-Sediment MFC rhizosphere/anode environments.
2. Materials and Methods
2.1 Microcosm setup
All experiments were carried out in microcosms consisting of Perspex tubes (14 cm inner diameter,
35 cm high). The bottom 15 cm contained 12 sample ports at various heights while the top 10 cm
contained 4 sample ports (Figure 1, S2). The containers were filled with graphite granules (type
00514, Mersen, Belgium) and vermiculite (Nestaan, Belgium) in a volume ratio of 2:1, to a height of
18 cm. This ratio was chosen in order to have a good electrical conductivity in the rhizosphere
(Arends et al. 2012) and to ensure that only added organic carbon (COD) or rhizodeposition were the
sources organic carbon. On top of this mixture, a layer of 3 cm vermiculite was added which
prevented electrical contact between the anode and the cathode. Vermiculite and not rice paddy soil
was chosen to create uniform rhizosphere conditions and to be able to measure and/or control the
amount of organic carbon in the rhizosphere/anode as much as possible. The sides of containers
were covered with black tape to create a dark root environment. At 5 and 15 cm height, a piece of
graphite felt (10*10*0.3 cm; Alfa Aesar, Germany) interwoven with a carbon rod (5 mm diameter;
Morgan, Belgium) was placed as the anodic current collector. The cathode was made of two sections
of 5*10*0.3 cm graphite felt interwoven with a carbon rod. The two sections were placed so that the
plants were able to grow in the middle of the Perspex tube. A watertight connection between the
wire and the carbon rod was established by means of conductive carbon cement (CCC, Leit-C; Fluka,
Germany) and vulcanizing tape. An Ag/AgCl reference electrode (BASi, United Kingdom) was placed
at 10 cm height in the rhizosphere/anode compartment and also in the cathode compartment. To be
able to measure soluble pore liquid components, a rhizon sampler was located at 6.7 and 16.7 cm
height (10 cm porous, Rhizosphere Research Products, The Netherlands). In order to collect gas
emissions, the top of the Perspex container contained a ledge for a water lock. This ledge was able to
support a tube of 12.3 cm inner diameter and 1 m height (Figure S2). The top of this tube contained
four connectors for sampling or other uses. One connection was occupied by the electrical
connections for a fan to mix the contents of the headspace; another was used for temperature
measurements during headspace gas sampling. All sampling i.e. pore liquid, trace gas and
polarization curves was performed at the same time for a microcosm during the whole experimental
period to minimize circadian (day/night) influences on the results. For the exploratory study,
additional square containers were used with the same configuration except for the surface area
(18*18 cm). In experimental setups meant for open circuit conditions, only 1 anode and 1 rhizon
sampler were placed in the container. An overview of all microcosms and experimental conditions
can be found in table S1.
Figure 1: a) Schematic overview of microcosm, not to scale. b) photograph of empty microcosm. 1) anode
current collector, 2) Ag/AgCl reference electrode, 3) port for liquid addition, 4) positions for Rhizon samplers, 5)
vermiculite level, 6) water level, 7) cathode electrode and wire, 8) rim to hold gas cover.
2.2 Rice plants & rhizosphere organic carbon
Rice seeds (Oryza sativa, ssp. indica, cultivar C101 PKT) were germinated in ¼ Hoagland nutrient
solution at 28°C in the dark (De Schamphelaire et al. 2008). The seedlings were subsequently
transplanted into the Perspex containers, which was considered the start of the experiment. Two
seedlings were planted per microcosm. The microcosms were placed in a light (12/12 light/dark using
Osram 400W plantastar and Osram 400W powerstar daylight HQI-BT Lamps) and temperature (30 ±
2 °C) controlled growth room. Development of the plants was scored according to Counce et al.
(Counce et al. (2000). To maintain a constant water level above the vermiculite, plants were regularly
watered with tap water and once per week with ½ Hoagland nutrient solution. Upon good visible
plant development, nitrate was removed from the Hoagland nutrient solution (Helder et al. 2012).
Dried and ground barley straw was added to the rhizosphere as additional organic carbon in the
exploratory study (Table S1). Soluble organic carbon mimicking exudation was added to the
rhizosphere in the main study as an equal weight mixture of acetate, starch, glucose, malate and
succinate. The amount of organic carbon added was calculated to attain approximately 30 mg CH4 m-
2 h -1 emission resulting in an organic carbon addition rate of 2.9 gCOD m -2 d -1 (~ 401 mA m -2; starting
from day 39; See supplementary information S1.).
2.3 Electrochemical analysis
Cell potential over an external resistor (500 Ω, unless stated otherwise), anode and cathode
potentials (Ean respectively Ecath) versus a reference electrode were measured continuously at 5 min
intervals (HP 34970A, Agilent). The potential of the Ag/AgCl reference electrodes were regularly
monitored relative to a calomel electrode (+244 mV vs. Standard Hydrogen Electrode (SHE); QIS, the
Netherlands) for correct conversion of the electrode potentials compared to the SHE. Electrode
potentials are consequently reported versus the standard hydrogen electrode. Polarization and power
curves were recorded on a weekly basis with a Bistat potentiostat (Biologic, France) at a scanrate of 1
mV s -1 following a 20 min stabilization period in open circuit. Electrochemical calculations were
performed according to Rabaey et al. (2005) and are based on hourly averages. Current and power
density are reported normalized to the plant growth area (0.015 m2; Table S1).
2.4 Chemical analysis
Liquid and headspace gas samples were taken twice a week. Anions (PO4
3-, NO3
-, NO2
-, SO4
2-) in the
pore liquid were analysed using a metrosep A Supp 5-150 column after a metrosep A 4/5 guard
column in a 761 Compact IC with a conductivity detector (Metrohm, Switzerland). Pore liquid soluble
chemical oxygen demand (COD) was determined using commercial kits according to the
manufacturer’s instructions (Machery-Nagel, Germany). Volatile fatty acids (VFA) were determined,
after etheric extraction of 2 mL pore liquid, on an EC-1000 Econo-Cap column in a gas chromatograph
(2014, Shimadzu) with a flame ionization detector (FID). Pore liquid and overlying water pH and
conductivity (EC) were determined using a handheld probe (SP10B and SK20, Consort, Belgium).
Headspace air temperature was monitored during gas sampling.
2.5 Trace gas analysis
Twice a week the headspace of the microcosms was closed to determine gas emissions from the
rhizosphere-plant continuum. Every 30 min. duplicate 15 mL gas samples were taken from the
headspace over a period of 2 hours. A duplicate 15 mL background air sample was taken to account
for gas concentrations already present in the rice cultivation room. All gas samples were stored in a
12 mL vacutainer, allowing at least 3 subsamples for analysis. Vacutainers were stored in the dark at
ambient temperature (20 ± 2 °C) until analysis. CH4 and CO2 were analysed using a gas
chromatograph (Trace GC Ultra, Thermo Fisher Scientific, Germany). Methane was determined using
FID while CO2 was determined using a thermal conductivity detector (TCD). Nitrous oxide was
measured using a gas chromatograph (14B, Shimadzu, Japan) with a 63Ni electron capture detector
(ECD).
2.6 PCR-DGGE and qPCR
At the end of the experimental period, samples were taken for microbial community analysis. Sample
positions were: anode current collector, anode felt (top & bottom) and anode granules with and
without rice roots (Figure S2). Anode granules without rice roots are considered as bulk vermiculite.
Total DNA was extracted using the method as described by Boon et al. (Boon et al. 2000). The 16S
rRNA-gene was amplified using PCR (Table 1). The bacterial community structure was visualized by
means of denaturing gradient gel electrophoresis (DGGE) of the obtained PCR amplicons using an 8%
polyacrylamide gel and a denaturing gradient of 40-60% on an INGENY phorU2X2 system for 16 h at
120 V (Goes, The Netherlands). The gel was stained with Sybr green in 1x TAE buffer. The resulting
community structures were analysed using BioNumerics software v5.1 (Applied Maths, Belgium).
DGGE profile similarities were based on band based clustering using the Jaccard coefficient to
minimize the influence of background subtraction on the clustering. Theoretical ecological
parameters, community organization (Co) and richness (Rr), were used to numerically describe the
bacterial communities. In brief, Co describes the species abundance distribution based on the gini
coefficient of the microbial community. Rr describes the species richness of the microbial community
and can be interpreted as the carrying capacity of a certain environment (Read et al. 2011). The
archaeal community structure was analysed by means of qPCR (Table 1 & S2) using the GoTaq qPCR
MasterMix (Promega, Belgium) on a StepOnePlus qPCR machine and accompanying software
(Applied Biosystems, United Kingdom).
Table 1: (q)PCR primers and conditions used for elucidating the microbial communities. D: denaturation, A:
Annealing, E: Elongation phases of the PCR cycle. For the specific methanogenic groups, taxonomic level and
major metabolism are indicated.
Reaction
Target
Primers
Program
Reference
PCR-DGGE
Bacterial 16S rRNA gene
338F-GC
518R
D: 94°C, 300s
30 cycles
D: 95°C, 60s
A:53°C, 60s
E: 72°C, 120s
E: 72°C, 600s
(Boon et al.,
2000; Ovreas et
al., 1997)
qPCR
Total Bacteria 16S rRNA
gene
338F
518R
D: 94°C, 10min
40 cycles
D: 94°C, 15s
A&E: 60°C, 60s
Total Archaea 16S rRNA
gene
ARC787F
ARC1059R
D: 94°C, 10min
40 cycles
D: 94°C, 10s
A&E: 60°C, 60s
(Yu et al., 2005)
Methanosaetaceae
Family (acetoclastic)
Mst 702F
Mst 862R
See total archaea
Methanosarcianceae
Family (mixotroph)
Msc 380F
Msc 828R
See total archaea
Methanobacteriales
Order (H2/CO2)
MBT 857F
MBT 1196R
See total archaea
Methanomicrobiales
Order (H2/CO2)
MMB 282F
MMB 832R
D: 94°C, 10min
40 cycles
D: 94°C, 10s
A&E: 63°C, 60s
3. Results
A first set of exploratory experiments was used to determine the key operational parameters
(electrode position, organic matter content and initial microbial community) relevant for CH4
emissions. A second, more detailed study focused on the effect of placing a sediment MFC in the
rhizosphere of rice plants on greenhouse gas emissions.
3.1 Exploratory study
From the exploratory study no clear difference could be distinguished in methane emissions between
closed and open circuit conditions as well as between planted and sediment conditions. Methane
emissions averaged across the open (n=7) and closed circuit (n=3) configurations (table S1) were 157
± 76 vs. 169 ± 68 mg CH4 m -2 h -1 respectively. These values are high compared to natural field
conditions (0-60 mg CH4 m -2 h-1 (Gogoi et al. 2005; Singh et al. 1999; Xu et al. 2007)). The microcosm
that was not supplemented with straw was the configuration that came close to natural CH4 fluxes in
the order of 35 mgCH4 m -2 h -1. Interestingly, the exploratory study revealed that interrupting the
current flow from a well performing Plant-MFC resulted in a doubling of CH4 emission flux (Figure
S3). This doubling was about 20 times higher than the loss in electrons due to current (See
supplementary information for calculations). This suggests the influence of changing electrochemical
or thermodynamic conditions, possibly related to local acidification close to the anode electrode,
next to direct competition for electrons between the current generating and methanogenic microbial
populations. Based on the results of the exploratory work, it was chosen to work in the detailed
study with no added organic carbon at the start of the experiment and also to not add methanogenic
sludge.
The next sections provide the results of the detailed study.
3.2 Plant growth characteristics
4 Microcosms were set up for the detailed study. One microcosm (M4) lagged behind in
development rate in the first growth phase, probably due to leakage of copper from the current
collector wire at the cathode i.e. at the top of the microcosm. This was amended by replacing the
wire and removing the affected vermiculite, after which the rate of development of the plants in M4
as well as electrochemical and gas emissions were similar to comparable microcosms. Final above
ground dry biomass were M1: 9.5 ± 1.2 gDW plant -1 M2: 10.9 ± 0.1 gDW plant-1 M3: 11.5 ± 1.7gDW
plant-1 M4: 4.3 ± 2.2 gDW plant-1.
3.3 Current, power and electrochemical performance of the Plant-MFCs
Cell potentials of all four plant-MFCs increased during the first 21 days after planting to an average
maximum value of 0.23 ± 0.039 V resulting in a current density of 27.9 ± 4.2 mA m-2 (figure 2b). After
day 21, the cell potentials decreased. Addition of organic carbon (starting at day 39) resulted in an
increase in current density for M3 and M4 (days 45-55) to 45.5 ± 11.8 mA m-2 for M3 and 25.9 ± 13.7
mA m-2 for M4 (Figure 2b). M1 and M2 did not receive any organic carbon (Table S1). Consequently,
no increase in current nor in greenhouse gas emissions were observed. M2 is not plotted in figure 2
as data were similar to M1. The high standard deviations are due to a circadian rhythm with daily
maxima up to 69.6 mA m-2 and 62.4 mA m-2 for M3 and M4 respectively and nightly minima as low as
22.9 mA m-2 and 9.7 mA m-2 for M3 and M4 (Figure 3). After the open circuit period (day 63-72), the
current density did not recover up to these values. Attempts to increase current density by lowering
the external resistance did not result in an increased current flow (Figure 2b). The main cause for the
low current density was a deterioration of the cathode performance as evidenced by the cathode
potential (Ecath). While the cathode potential reached daily maxima of 0.84 V and 0.62 V for M3 and
M4 and nightly minima of 0.07 V and 0.17 V during days 45-55 (Figure 3), daily cathode maxima went
only up to 0.40 V and 0.37 V for M3 and M4 upon re-closing the opened electrical circuit with a 500 Ω
resistor. Decreasing the external resistance to 100 Ω (d 79) resulted in an average (including day &
night) Ecath over the remainder of the experimental period of -0.11 ± 0.13 V for M3 and -0.11 ± 0.07 V
for M4. The anode potential of M3 and M4 remained low after day 39, -0.20 ± 0.16 V for M3 and -
0.19 ± 0.09 V for M4 (Figure 1c and 3). Variations in anode potential are mainly caused by
polarization curves (invasive technique) or sampling. Overall average power production during days
45-55 resulted in 22.2 ± 9.5 mW m-2 for M3 and 6.6 ± 6.6 mW m-2 for M4. Here again high standard
deviations can be seen, due to the circadian rhythm influencing cathode performance. From the
polarization and power curves the maximum attainable current and power can be estimated. The
maximum power output was attained at day 61, 72 mW m-2 for M3 and 56 mW m-2 for M4. The
maximum short circuit current generated during the polarization measurements was 0.58 A m-2 for
M3 at day 53 and 0.20 A m-2 for M4 at day 61. For M1 and M2, the maximum values amounted to 9
mW m-2 and 1 mW m-2 respectively, obtained at day 19
Figure 2: overview of a) chemical oxygen demand, b) current density, c) anode potentials, d) methane flux,
multiply value by 25 to arrive at CO2 equivalents and e)nitrous oxide flux, multiply value by 298 to arrive at CO2
equivalents during the rice growth period. M1: M3: M4: . M2 is not plotted as data are
similar to M1. For continuous measurements, b) and c), symbols are omitted. Experimental stages: 0: initial
plant growth and MFC start-up 1: addition of organic carbon 2: open circuit period 3: closed circuit period, 500
Ω 4: external resistance lowered from 500 Ω to 100 Ω for M3 and M4. No electrochemical data available for day
58-63.
Figure 3: Close up of data between day 45-55 showing the diurnal variation. a) anode and cathode potentials
during day 45-55. Day 48: anode reference electrodes disconnected. Vertical solid lines; 7 am, lights on. Vertical
dotted lines; 7 pm, lights off. M3 anode: M3 cathode: M4 anode: M4 cathode: b)
Current density resulting from the varying electrode potentials day 45-55. M3 daily max.: M3 Nightly
min. : M4 daily max.: M4 Nightly min. :
3.4 Current was generated after addition of organic carbon.
The overall availability of organic carbon in the pore liquid was low during the first part of the
experiment (35.3 ± 10.5 mgCOD L-1 for M3 and 31.8 ± 9.3 mgCOD L-1 for M4 till day 55; Figure 1a). To
compensate for the low reducing power provided by the plants, soluble organic carbon was added to
mimic the rhizodeposition process. The added COD was immediately used by the electroactive
microbial community present on the anode, as can be seen by the almost instantaneous increase in
current (Figure 2b). The organic carbon was efficiently oxidized as only on day 70 it started to appear
in the pore liquid of the bulk (Figure 2a). No short chain fatty acids were detected in the bulk liquid of
the microcosms (data not shown). The efficient use of exogenous organic matter was not only for
current generation, as current generation only accounted for maximum 15% of the organic carbon
([~60 mA m-2] / [401 mA m-2] *100%). The organic carbon that was not used for current generation
accumulated in the system or was metabolised using alternative electron acceptors such as nitrate
and sulphate or O2 due to radial oxygen loss from the rice roots. For M3 and M4 a decrease of 69 %
and 75 % COD equivalents was noticed for nitrite, nitrate and sulphate combined after day 39. The
substrate limited situation of all microcosms could also be derived from the fact that M2 received the
same nutrient solution as M3 & M4 but without organic carbon. The addition of only Hoagland
solution did not result in an increase of current, nor CH4 emissions nor organic carbon concentration
in the pore liquid. M1 did not receive any extra organic carbon or liquid, which also did not result in
an increase of available organic carbon in the pore liquid.
Due to current generating metabolism, a drop in pH could be expected but this was not detected in
the bulk pore liquid. The pH of all pore liquid samples was 7.79 ± 0.44 over the whole experimental
period.
3.5 Anode biofilm redox potentials in relation to greenhouse gas emissions
Methane was produced when favourable redox conditions existed for the producing microbial
community (i.e. < -150 mV vs. SHE; Hou et al., 2000; Yu and Patrick, 2003; Johnson-Beebout et al.,
2009). The potentials measured in the reference electrode/ bioanode combination could be related
with bulk gas emissions for CH4 (Figure 4a). N2O emissions showed a less clear relationship with the
redox potentials as measured with the reference electrode/ bioanode combination (figure 4b).
Figure 4: Greenhouse gas emissions in relation to measured anode potential. a) CH4 multiply value by 25 to
arrive at CO2 equivalents b) nitrous oxide flux, multiply value by 298 to arrive at CO2 equivalents. M3:
M4: .
3.6 Methane emissions lagged behind current generation.
Organic carbon was efficiently metabolised to current, to other sinks or accumulated, but was not
metabolised to methane. Methane only appeared in the headspace measurements of M3 & M4 after
the current flow was interrupted by removing the external resistor from the electrical circuit for 9
days (day 63-72). Towards the end of this period (day 69; Figure 2a), organic carbon concentrations
started to increase in the pore liquid (up to 577 mgCOD L -1 for M3, day 85 and 265 mgCOD L -1 for
M4, day 89). This increase was followed by an increase in CH4 emissions (starting on day 75; Figure
2d). Closing the electrical circuit with a 500 Ω resistor on day 72 did not result in a decrease in CH4
emissions. Conversely, CH4 emissions increased even further. Methane release from the microcosms
did not show any relation with either closing the electrical circuit or lowering the external resistance
to 100Ω (day 79; Figure 2b and 2d). The effect of the electrical current on the CH4 emissions after re-
closing the external circuit remains to be determined as the current was markedly lower compared
to the period before the electrical circuit was opened. This was the result of the low activity of the
(bio)cathode resulting in an overall low current density (Figure 2b and § 3.3). Methane emissions in
these microcosms were indeed facilitated by the aerenchyma of the rice plants. This was established
by harvesting the above-ground biomass at the end of the study period and subsequently
determining the CH4 flux. Cutting the biomass in M3 and exposing the aerenchyma to the air resulted
in a 14% lower emission the next day (32.3 to 27.8 mgCH4 m - 2 h -1). Cutting the biomass in M4 and
raising the water level to above the aerenchyma resulted in a 85% decrease of emissions the next
day (9.5 to 1.4 mgCH4 m - 2 h -1).
3.7 Bacterial community structure is driven by available organic carbon
Based on the DGGE-profile of the 16S rRNA gene of the bacterial community, the main parameter
affecting the enrichment of the various microbial communities was the amount of available electron
donor (organic carbon) present at the various locations. This was confirmed by two methods of
cluster analysis, Jaccard band based clustering (Figure 5) and via a Pearson correlation clustering (not
shown). M3 and M4 were able to support a bacterial community with a higher community
organisation (Co) and a higher richness (R) compared to M1 and M2. The higher richness indicates a
more developed microbial community. It can be reasoned that this is related to a metabolic network
where fermentation of substrates to H2 plays a key role however this could not be validated with the
used technique.
Figure 5: Jaccard bandbased clustering of 16s rRNA-gene based microbial community. Microbial Resource
Management (MRM) parameters community organization (Co) and range weighted richness (Rr) are indicated
(Read et al. 2011).
3.8 Archaeal community structure
The results of the qPCR analysis of the archaeal microbiome at the end of the experiment indicates
that H2 based methanogenesis (orders Methanobacteriales & Methanomicrobiales) was likely the
dominant process compared to acetate based methanogenesis (Family Methanosaetaceae). M1 (low
COD concentration in the pore liquid, and resulting low CH4 emissions) showed a slight enrichment in
H2 consuming Methanomicrobiales compared to acetate consuming methanogens (3.9 vs. 3.8 log
copies g -1, other groups were below detection) at the roots of the rice plants. The microcosms with a
higher COD-load (M3 & M4) showed at all sample locations (with and without roots, top and bottom
felt) a higher abundance of H2-consuming methanogens. Methanosaetaceae (acetate dependent)
were present in a lower abundance compared to Methanosarcinaceae (mixotrophic) at the bottom
anode (5.2 vs. 6.0 log copies g -1 for M3 & 4.7 vs. 5.1 log copies g -1 M4) whereas they were more
abundantly present at the top anode (5.0 vs. 4.7 log copies g -1 for M3 & 4.7 log copies g -1 vs. < 3.7 log
copies g -1 for M4). All archaeal groups were below detection limit at all sample locations for M2,
correlating with the low COD concentrations in the rhizosphere. The same was valid for M1 where
archaea were only detected in samples with roots. The copy numbers of all four groups of
methanogens added up to the copy numbers obtained by means of the total archaea qPCR for all
microcosms. This indicates that no other groups were involved in methanogenesis in these
microcosms.
4. Discussion
4.1 Anode/rhizosphere organic carbon and redox potential
The low current density recorded for M1 and M2 over the whole experimental period and for M3
and M4 for the first 40 days indicate that exudation of low molecular weight organic acids or sugars
were not directly responsible for current generation at the anode. Moreover, no short chain fatty
acids were detected in the bulk liquid, indicating that other organic components were the major
components of the organic material present. These results are in line with previous research where
long start-up times for Plant-MFCs were recorded with pristine anodes (De Schamphelaire et al.
2008; Helder et al. 2010; Helder et al. 2012). The long start-up times indeed suggest that components
derived from rhizodeposition other than small molecules released via direct exudation are more
important in current generation by plant-sediment MFCs. The model put forward by Timmers et al.
(2011) showed that a low amount of reducing equivalents available for current generation was
corroborated by the presence of alternative electron acceptors and could possibly be related to
radial oxygen loss into the rhizosphere (Liesack et al. 2000). In this work, redox potentials as
measured at the anode electrode of M3 and M4 decreased over time, with subsequent increase in
bulk gas emissions of CH4, with increased current generation and a decrease in alternative electron
acceptor concentrations. For both CH4 and N2O, favourable redox potentials are needed, i.e. for N2O
Eh >180 mV vs. SHE via denitrification and for CH4 Eh < -150 mV vs. SHE (Hou et al. 2000; Johnson-
Beebout et al. 2009; Yu and Patrick 2003). Bulk methane emissions can be related to redox potentials
as determined with the bioanodes. A clear relation was not observed for N2O emission and bulk
anode potentials (Figure 4). This can be attributed to the position of the reference electrode in
relation to the location of microbial activity. The top of the microcosm which was exposed to
ambient air, while the bulk microcosm was waterlogged.
4.2 Current producing metabolism precedes methanogenic metabolism.
As stated before, organic carbon was instantaneously metabolised to current and not to CH4.
Methane emissions became apparent after interrupting current flow by removing the external load.
During this period, organic carbon became available in the pore liquid and CH4 emissions started to
increase. An exploratory study already revealed that at higher organic carbon loads, no difference in
methane emissions could be observed between the open and closed circuit conditions. In accordance
with the findings in this work, Freguia et al. (2008) have also shown that methanogenesis is a robust
process largely operating independent of electrode based processes. This was established in a
reactor type anode with a packed bed of similar granular electrode material. Kaku et al. (2008) were
also not able to detect decreased CH4 emissions from Plant-MFCs in closed circuit in a rice field. On
the other hand, it was shown that a closed circuit MFC was able to release 10 times less methane
compared to an open circuit MFC using a reactor type setup and rice paddy soil as anode inoculum
(Ishii et al. 2008). Cabezas de Rosa (2010) showed that it was possible to reduce methane emissions
with electrodes in sediment systems by 47%. These seemingly contradicting observations indicate
that the electrical circuit and the current generating microorganisms were only able to outcompete
methanogenic microorganisms at low organic carbon concentrations (electron-donor) and high
electron acceptor concentrations (anode in closed circuit). This concept is supported by the fact that
the affinity (Ks) of Geobacter sulfurreducens (a known current generating microorganism) for acetate
can be as low as 10 µM whereas the lowest affinity for acetate reported for Methanosaetaceae is
160 µM (Esteve-Núñez et al. 2005; Qu et al. 2009). For Methanosarcinaceae this value is even higher,
3 mM (Qu et al. 2009).
Competition for substrate seemed to be most important as acidification of the bulk pore liquid was
not detected. With an average current density of 35 mA m - 2 (day 45-55 Figure 2b), it can be
calculated that about 0.5 mmol of protons are produced per day. This amount of protons produced
locally at the anode can indeed cause local acidification. The results indicate however that local
buffer capacity, in companion with proton diffusion to the cathode, were able to prevent bulk
acidification in the studied microcosms. Lowering the resistance in the external circuit from 500 Ω to
100 Ω did not result in a higher current. This is in conjunction with the observations of Helder et al.
on an a rooftop MFC (Helder et al. 2013). Although they were not able to indicate a cause for this
phenomena, from this work it could be observed that it was mainly due to a deterioration of cathode
performance. The low cathode potentials were most likely caused by an accumulation of organic
detritus (e.g. degraded leaves, algae, roots) on the cathode electrode, leading to a low O2 availability
for electrochemical reduction.
Overall, a sequence of steps can be distinguished; initially, during early waterlogging conditions and
plant development, a high redox potential due to the presence of O2 and other soluble electron
acceptors is noticed. Gradually organic carbon becomes available in the rhizosphere, electron
acceptors are being depleted except for the electrode and redox potential decreases leading to
current generation. Over time, more organic carbon becomes available than the current generating
bacteria can metabolise and CH4 production starts and cannot be stopped anymore.
4.3 Archaeal communities in the rhizosphere of plant-MFCs
Few microbial communities of anodes of Plant-MFCs have been elucidated (Cabezas da Rosa 2010;
De Schamphelaire et al. 2010; Kaku et al. 2008; Timmers et al. 2012). Of the microbial communities
studied, focus has been placed on the anode bacterial community and not on the archaeal microbial
community. In the current study methanogenic community slightly dominated by H2-dependent
methanogens was detected. This was also the case in several other studies and sometimes even
more pronounced in closed circuit compared to open circuit conditions. De Schamphelaire et al.
(2010) have seen a relative enrichment of H2 utilizing methanogens in Plant-MFCs consisting of a soil
or vermiculite rhizosphere with graphite felt electrodes, which only covered a small part of the
rhizosphere. Timmers et al., (2012) characterized the microbial populations in the anode of a high
current (167 ± 65 mA m - 2) vs. a low current (67 ± 58 mA m - 2) producing Plant-MFC (Glyceria
maxima) with the complete anode compartment filled with electrode material. The relatively high
current indicated that in that case sufficient organic substrate was present in the anode under
reducing conditions. In their study, a selection towards a H2-based metabolism was detected after
225 days of operation where 95% of the archaea could be related to the genus Methanobacterium in
the high current producing Plant-MFC vs. a more diverse archaeal community in the low current
MFC. The bacterial community supported these results as H2 producing clostridia were abundantly
found in the high current producing systems (Timmers et al. 2012).
Using a sediment MFC, Cabezas da Rosa (2010) found that H2-based methanogenesis was more
important compared to acetate based methanogenesis at the archaeal community level, i.e.
Methanosaetaceae were less abundant compared to Methanosarcinaceae and
Methanomicrobiaceae in the current generating systems. This result was corroborated by the stable
isotope ratios of CH4 indicating lower acetoclastic CH4 production in closed circuit systems (Cabezas
da Rosa 2010).
In the microcosms studied in this work, an alternative electron acceptor was introduced in the
rhizosphere aimed at lowering the activity of the methanogenic microbial community. As seen in
other (plant) MFC based anodes, also here hydrogenotrophic methanogens were most abundantly
detected. This indicates that to effectively lower methane emissions by introducing an anode, the
focus should be on stimulating hydrogen to current conversions.
4.4 Outlook & challenges
From this longitudinal study several challenges for the control of methane emissions from rice
paddies or other wetlands on a larger than lab-scale by means of a bioelectrochemical system can be
envisioned. Condensing the limited information that is known on (competitive) inhibition of CH4
formation in anodes of (sediment) bioelectrochemical systems, it shows that contradicting data are
present in the existing literature, because of 1) different inocula, 2) a varying electrode configuration,
3) a varying organic loading rate and 4) a different electrochemical control (Cabezas da Rosa 2010;
Freguia et al. 2008; Ishii et al. 2008; Kaku et al. 2008).
A first challenge is the effective range of the anode and the technical application of the
bioelectrochemical system. In order to achieve an effect, carbon granules were applied in a 2/3 ratio
vs. the amount of vermiculite matrix. Combining these dispersed granules with a suitable current
collector and a reference system poses a great challenge. On top of that, agricultural practices(e.g.
plowing) can interfere with the bioelectrochemical system. A solution to this challenge might be the
use of electron shuttle molecules (phenazines, humic substances etc). However the effectiveness of
these molecules in electron transfer processes in soils and sediments needs to be established.
Moreover, these molecules can also interfere with other processes than electron transfer.
The second challenge is the question of electrochemical control. In this work, electrochemical control
was applied by using an external resistor coupled to an O2-reducing (bio)cathode. Visual inspection,
polarization curves and electrochemical monitoring of the cathode potential indicated deterioration
of cathode performance over time. A probable cause for this deterioration is the amount of organic
carbon that accumulated on the cathode electrode over time due to growth of algae and dead
(plant)biomass. Another option for electrochemical control, is the polarization of the anode
electrode at a certain potential (e.g. 0 vs. SHE). This might however lead to large energy investments
due to the internal resistance of a sediment-BES configuration (see Supplementary information S2,
Figure S1).
A third challenge is the durability of CH4 emissions mitigation over a long period of time. In this work,
it was established that CH4 emissions could be postponed by introducing a bioanode. However, the
effectiveness on a time scale longer than one growing season needs to be determined. A longer
effective working period can possibly be obtained in combination with an anode of conductive
biochar in the rhizosphere. As Feng et al. (2012) has shown, biochar is able to increase the
abundance of methanotropic bacteria. During an anaerobic period, a conductive biochar can act as
an anode, while during an aerobic period it can act as a methanotrophic stimulant. The application of
agricultural practices such as mid-season drainage (Xu et al. 2007) to provide a competitive
advantage for the current generating microbial community (that is able to withstand O2 (Freguia et
al. 2008; Nevin et al. 2011)) in conjunction with methanotropic stimulation is another option.
In conclusion, It was shown that placing an anode of a bioelectrochemical system in a waterlogged
rhizosphere led to current generation before methane emissions started, possible leading to
postponement of methane emissions from these anoxic systems in the short term. However, in the
long-term (i.e. period of complete rice cropping season), low methane emissions could not be
maintained due to an excess of organic matter in the rhizosphere. Based on qPCR analysis of the
archaeal community in the rhizosphere/anode plane, H2 was indicated as the most important
precursor for methane production. This shows that there was an effective competition on the level of
acetoclastic methanogenesis, but that the competition for H2 needs to be enhanced to achieve a
more effective and sustainable mitigation of methane emissions by means of a bioelectrochemical
system. The effect of local acidification due to anodic metabolism on methanogenic metabolism
could not be established in this work. To achieve a practical, large scale application of a
bioelectrochemical system to mitigate methane emissions, a combined strategy with other methane
mitigation approaches is envisioned.
Acknowledgements
J.A. was funded by the European Community’s Seventh Framework Programme FP7/2007-2013
under Grant Agreement No. 226532. W.V. is supported by a Ghent University Multidisciplinary
Research Partnership (MRP): Biotechnology for a sustainable economy (01 MRA 510W). The authors
thank Jan Vermeulen for assistance with trace gas analysis. Simon Decorte, Joachim Desloover and
Giovanny Ganendra are acknowledged for their fruitful discussions on the manuscript.
References
Appels L, Baeyens J, Degrève J, Dewil R (2008) Principles and potential of the anaerobic digestion of
waste-activated sludge. Prog Energ Combust 34(6):755-781
Arends JBA, Blondeel E, Tennison S, Boon N, Verstraete W (2012) Suitability of granular carbon as an
anode material for sediment microbial fuel cells. J Soils Sediments 12(7):1197-1206
Arends JBA, Verstraete W (2012) 100 years of microbial electricity production: three concepts for the
future. Microbial Biotechnology 5(3):333-346
Aulakh MS, Wassmann R, Bueno C, Kreuzwieser J, Rennenberg H (2001) Characterization of Root
Exudates at Different Growth Stages of Ten Rice (Oryza sativa L.) Cultivars. Plant Biology
3(2):139-148
Aulakh MS, Wassmann R, Rennenberg H, Fink S (2000) Pattern and Amount of Aerenchyma Relate to
Variable Methane Transport Capacity of Different Rice Cultivars. Plant Biology 2(2):182-194
Bazhin N (2010) Theory of methane emission from wetlands. Energ Environ Sci 3(8):1057-1072
Boon N, Goris J, De Vos P, Verstraete W, Top EM (2000) Bioaugmentation of Activated Sludge by an
Indigenous 3-Chloroaniline-Degrading Comamonas testosteroni Strain, I2gfp. Appl Environ
Microbiol 66(7):2906-2913
Cabezas da Rosa A (2010) Diversity and function of the microbial community on anodes of sediment
microbial fuel cells fueled by root exudates. Dissertation. Philipps-Universität Marburg
Conrad R (2002) Control of microbial methane production in wetland rice fields. Nutrient Cycling in
Agroecosystems 64(1):59-69
Counce PA, Keisling TC, Mitchell AJ (2000) A Uniform, Objective, and Adaptive System for Expressing
Rice Development. Crop Sci 40(2):436-443
De Schamphelaire L, Cabezas A, Marzorati M, Friedrich MW, Boon N, Verstraete W (2010) Microbial
Community Analysis of Anodes from Sediment Microbial Fuel Cells Powered by Rhizodeposits
of Living Rice Plants. Appl Environ Microbiol 76(6):2002-2008
De Schamphelaire L, van den Bossche L, Dang HS, Hofte M, Boon N, Rabaey K, Verstraete W (2008)
Microbial fuel cells generating electricity from rhizodeposits of rice plants. Environ Sci
Technol 42(8):3053-3058
Denman KL, G. Brasseur, A. Chidthaisong, P. Ciais, P.M. Cox, R.E. Dickinson, D. Hauglustaine, C.
Heinze, E. Holland, D. Jacob, U. Lohmann, S Ramachandran, P.L. da Silva Dias, Wofsy SC,
Zhang X (2007) Couplings Between Changes in the Climate System and Biogeochemistry. In:
Solomon S, D. Qin, M. Manning, Z. Chen, M. Marquis, K.B. Averyt, M.Tignor, Miller HL (eds)
Climate Change 2007: The Physical Science Basis Contribution of Working Group I to the
Fourth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge
University Press, Cambridge United Kingdom and New York, NY, USA
Donovan C, Dewan A, Peng H, Heo D, Beyenal H (2011) Power management system for a 2.5 W
remote sensor powered by a sediment microbial fuel cell. J Power Sources 196(3):1171-1177
Esteve-Núñez A, Rothermich M, Sharma M, Lovley D (2005) Growth of Geobacter sulfurreducens
under nutrient-limiting conditions in continuous culture. Environ Microbiol 7(5):641-648
Feng Y, Xu Y, Yu Y, Xie Z, Lin X (2012) Mechanisms of biochar decreasing methane emission from
Chinese paddy soils. Soil Biology & Biochemistry 46:80-88
Forster P, V. Ramaswamy, P. Artaxo, T. Berntsen, R. Betts, D.W. Fahey, J. Haywood, J. Lean, D.C.
Lowe, G. Myhre, J. Nganga, R. Prinn, G. Raga, Schulz M, Dorland RV (2007) Changes in
Atmospheric Constituents and in Radiative Forcing. In: Solomon S, D. Qin, M. Manning, Z.
Chen, M. Marquis, K.B. Averyt, M.Tignor, Miller HL (eds) Climate Change 2007: The Physical
Science Basis Contribution of Working Group I to the Fourth Assessment Report of the
Intergovernmental Panel on Climate Change. Cambridge University Press, Cambridge United
Kingdom and New York, NY, USA
Freguia S, Rabaey K, Yuan ZG, Keller J (2008) Syntrophic processes drive the conversion of glucose in
microbial fuel cell anodes. Environ Sci Technol 42(21):7937-7943
Gogoi N, Baruah KK, Gogoi B, Gupta PK (2005) Methane emission characteristics and its relations with
plant and soil parameters under irrigated rice ecosystem of northeast India. Chemosphere
59(11):1677-1684
Helder M, Strik DPBTB, Hamelers HVM, Kuhn AJ, Blok C, Buisman CJN (2010) Concurrent bio-
electricity and biomass production in three Plant-Microbial Fuel Cells using Spartina anglica,
Arundinella anomala and Arundo donax. Bioresour Technol 101(10):3541-3547
Helder M, Strik DPBTB, Hamelers HVM, Kuijken RCP, Buisman CJN (2012) New plant-growth medium
for increased power output of the Plant-Microbial Fuel Cell. Bioresour Technol 104:417-423
Helder M, Strik DPBTB, Timmers RA, Raes SMT, Hamelers HVM, Buisman CJN (2013) Resilience of
roof-top Plant-Microbial Fuel Cells during Dutch winter. Biomass and Bioenergy 51(0):1-7
Hong S, Chang I, Choi Y, Kim B, Chung T (2009) Responses from freshwater sediment during
electricity generation using microbial fuel cells. Bioprocess and Biosystems Engineering
32(3):389-395
Hou AX, Chen GX, Wang ZP, Van Cleemput O, Patrick WH (2000) Methane and nitrous oxide
emissions from a rice field in relation to soil redox and microbiological processes. Soil Science
Society of America Journal 64(6):2180-2186
Ishii Si, Hotta Y, Watanabe K (2008) Methanogenesis versus electrogenesis: Morphological and
phylogenetic comparisons of microbial communities. Biosci Biotechnol Biochem 72(2):286-
294
Johnson-Beebout SE, Angeles OR, Alberto MCR, Buresh RJ (2009) Simultaneous minimization of
nitrous oxide and methane emission from rice paddy soils is improbable due to redox
potential changes with depth in a greenhouse experiment without plants. Geoderma 149(1–
2):45-53
Kaku N, Yonezawa N, Kodama Y, Watanabe K (2008) Plant/microbe cooperation for electricity
generation in a rice paddy field. Appl Microbiol Biotechnol 79(1):43-49 doi:10.1007/s00253-
008-1410-9
Liesack W, Schnell S, Revsbech NP (2000) Microbiology of flooded rice paddies. Fems Microbiol Rev
24(5):625-645
Logan BE, Hamelers B, Rozendal R, Schröder U, Keller J, Freguia S, Aelterman P, Verstraete W, Rabaey
K (2006) Microbial Fuel Cells: Methodology and Technology. Environ Sci Technol 40(17):5181-
5192
Majumdar D (2003) Methane and nitrous oxide emission from irrigated rice fields: Proposed
mitigation strategies. Current Science 84(10):1317-1326
Nevin KP, Pei Z, Franks AE, Woodard TL, Lovley DR (2011) Anaerobes unleashed: Aerobic fuel cells of
Geobacter sulfurreducens. J Power Sources 196(18)
Ovreas L, Forney L, Daae FL, Torsvik V (1997) Distribution of bacterioplankton in meromictic Lake
Saelenvannet, as determined by denaturing gradient gel electrophoresis of PCR-amplified
gene fragments coding for 16S rRNA. Appl Environ Microbiol 63(9):3367-3373
Qu X, Vavilin VA, Mazéas L, Lemunier M, Duquennoi C, He PJ, Bouchez T (2009) Anaerobic
biodegradation of cellulosic material: Batch experiments and modelling based on isotopic
data and focusing on aceticlastic and non-aceticlastic methanogenesis. Waste Management
29(6):1828-1837
Rabaey K, Clauwaert P, Aelterman P, Verstraete W (2005) Tubular microbial fuel cells for efficient
electricity generation. Environ Sci Technol 39(20):8077-8082
Rabaey K, Verstraete W (2005) Microbial fuel cells: novel biotechnology for energy generation.
Trends Biotechnol 23(6):291-298
Read S, Marzorati M, Guimaraes BCM, Boon N (2011) Microbial Resource Management revisited:
successful parameters and new concepts. Appl Microbiol Biotechnol 90(3):861-871
Reimers CE, Tender LM, Fertig S, Wang W (2001) Harvesting energy from the marine sediment-water
interface. Environ Sci Technol 35(1):192-195
Rozendal RA, Hamelers HVM, Buisman CJN (2006) Effects of membrane cation transport on pH and
microbial fuel cell performance. Environ Sci Technol 40(17):5206-5211
Rozendal RA, Hamelers HVM, Rabaey K, Keller J, Buisman CJN (2008) Towards practical
implementation of bioelectrochemical wastewater treatment. Trends Biotechnol 26(8):450-
459
Singh S, Singh JS, Kashyap AK (1999) Methane flux from irrigated rice fields in relation to crop growth
and N-fertilization. Soil Biology and Biochemistry 31(9):1219-1228
Strik D, Hamelers HVM, Snel JFH, Buisman CJN (2008) Green electricity production with living plants
and bacteria in a fuel cell. International Journal of Energy Research 32(9):870-876
Tender LM, Gray SA, Groveman E, Lowy DA, Kauffman P, Melhado J, Tyce RC, Flynn D, Petrecca R,
Dobarro J (2008) The first demonstration of a microbial fuel cell as a viable power supply:
Powering a meteorological buoy. J Power Sources 179(2):571-575
Timmers RA, Rothballer M, Strik DPBTB, Engel M, Schulz S, Schloter M, Hartmann A, Hamelers B,
Buisman C (2012) Microbial community structure elucidates performance of Glyceria maxima
plant microbial fuel cell. Appl Microbiol Biotechnol 94(2):537-548
Timmers RA, Strik DPBTB, Arampatzoglou C, Buisman CJN, Hamelers HVM (2011) Rhizosphere anode
model explains high oxygen levels during operation of a Glyceria maxima PMFC. Bioresour
Technol 108 60-67
Xu SP, Jaffe PR, Mauzerall DL (2007) A process-based model for methane emission from flooded rice
paddy systems. Ecol Model 205(3-4):475-491
Yu KW, Patrick WH (2003) Redox range with minimum nitrous oxide and methane production in a
rice soil under different pH. Soil Science Society of America Journal 67(6):1952-1958
Yu Y, Lee C, Kim J, Hwang S (2005) Group-specific primer and probe sets to detect methanogenic
communities using quantitative real-time polymerase chain reaction. Biotechnology and
Bioengineering 89(6):670-679
Zhang A, Bian R, Pan G, Cui L, Hussain Q, Li L, Zheng J, Zheng J, Zhang X, Han X, Yu X (2012) Effects of
biochar amendment on soil quality, crop yield and greenhouse gas emission in a Chinese rice
paddy: A field study of 2 consecutive rice growing cycles. Field Crops Research 127:153-160
Supplementary information: Greenhouse gas emissions from rice microcosms amended with a
plant microbial fuel cell.
Jan B. A. Arends1, Jonas Speeckaert1, Evelyne Blondeel1#, Jo De Vrieze1, Pascal Boeckx2, Willy
Verstraete1, Korneel Rabaey1, Nico Boon1*
1Laboratory of Microbial Ecology and Technology (LabMET), Ghent University, Coupure Links 653, B-
9000 Gent, Belgium.
2Isotope Bioscience Laboratory (ISOFYS), Ghent University, Coupure Links 653, B-9000 Gent, Belgium.
# Currently at: EnBiChem, HOWEST, Graaf Karel de Goedelaan 5, B-8500 Kortrijk, Belgium
*Corresponding author. Laboratory of Microbial Ecology and Technology (LabMET), Ghent
University, Coupure Links 653, B-9000 Ghent, Belgium;
phone: +32 (0)9 264 59 76;
fax: +32 (0)9 264 62 48;
E-mail: Nico.Boon@ugent.be;
Webpage: www.labmet.Ugent.be.
This supplementary information contains:
Supplementary information S1. Example calculation to convert current density, chemical oxygen
demand removal rate and current in the same units.
Supplementary information S2. Estimation of electrical operating cost for a poised anode in a
wetland.
Figure S1: graphical estimation of electrical operating cost.
Figure S2: detailed schematic overview of the rice microcosms.
Figure S3: results of the exploratory study.
Table S1: Overview of all experimental setups.
Table S2: qPCR parameters.
Supplementary information S1. Example calculation to convert current density, chemical oxygen
demand removal rate and current in the same units
Electrical current density (A m-2), chemical oxygen demand removal rate (gCOD d-1) and methane
emissions (mgCH4 m-2 h-1) can all be recalculated into the same units. All three parameters express, in
essence, the movement of electrons. Here some example calculations are presented to clarify these
conversions.
Electrical current density J (A m -2), can be expressed as the amount of charge Q, expressed in
coulomb (C) per area A (m2) per time t (s)
Q/(t*A) = J (A m -2) (Eq. S1)
The amount of charge can be expressed as mole of electrons by conversion with the Faraday
constant (96485 C mol (e-) -1). Thus electrical current density is a measure of the amount of charge
that moves in a given time through a certain area.
Organic matter removal can be expressed as chemical oxygen demand (COD) removal per day. This
can also be expressed as the movement of charge in a given time. For example acetate oxidation
with O2 can be written according to two half reactions, the first is the oxidation of acetate and the
second the reduction of O2 to water.
2HCO3
- + 9H+ + 8e- CH3COO- + 4H2O (Eq. S2)
2O2 + 8e- + 8H+ 2H2O (Eq. S3)
From these two half reactions it can be established that 1.1 g of O2 (32 g mol -1) is needed for the
oxidation of 1 g of acetate (59 g mol -1). Moreover, as can be seen in S2 and S3, electrons are being
displaced. Accordingly, the oxidation of organic matter per area per time can be expressed as a
current density. For example oxidizing 1 g organic matter per m2 per h results in a current density of
3.4 A m -2.
1 (gCOD m - 2)/32 (g mol -1 O2) * 4 (mol (e- )mol -1 O2) *1/3600 (h s -1)*96485 (C mol (e-) -1) = 3.4
A m -2
Methane removal or production can also be expressed as a current density when applying the same
reasoning.
1HCO3
- + 9H+ + 8e- 1CH4 + 3H2O (Eq. S4)
For example producing 10 mgCH4 per m2 per h results in an equivalent current density of 0.13 A m - 2
10*10-3 (gCH4 m - 2)/16 (g mol -1 CH4) * 8 (mol (e-) mol -1 CH4) *1/3600 (h s -1)*96485 (C mol (e-) -
1) = 0.13 A m - 2
Conversely, current density can be expressed in equivalent units such as gCOD d -1 or mg CH4 m – 2 h -1
enabling fast comparison of the various metablic rates. See also Arends et al. (2012)
Supplementary information S2. Estimation of electrical operating costs for a poised anode in a
wetland
To estimate the amount of electrical power and thus the electrical operating cost per m2 of plant-
MFC in case of anode potentiostatic control, several parameters influence the outcome of this
estimation. One important parameter is the distance between the reference electrode and the
anode electrode. This causes an ohmic voltage drop that must be compensated for to achieve a good
potentiostatic control of the anode. This ohmic voltage drop is dependent on the current that flows
through the system (this can be understood as the methane mitigation rate) and the resistivity of the
matrix (in this case rice or wetland soil). Here an estimation is made for the following case, a wetland
soil with 1 reference electrode, a maximum effective anode-reference distance of 1 m and an energy
price of € 0.20 kwh -1. Soil resistivity was varied between 0 and 50 Ω m (Domínguez-Garay et al.,
2013) and current density between 10 and 300 mA m - 2 which is equivalent to 0.13-3.9 mgCH4 m – 2 h -
1 (Eq. S4). Within these parameters 55 scenarios were calculated which are depicted in Figure S1.
From this estimation it can be seen that when aiming for an appreciable methane mitigation rate, a
considerable monthly cost is incurred. Options to reduce this cost are placing more reference
electrodes (extra capital cost) or opting for less sophisticated electrochemical control by using an
external resistance.
Figure S1: Estimation of electrical operating costs for a poised anode in a wetland based on soil resistivity and
current density. Scenario of 3.9 mg CH4 m-2 h-1 ~ 300 mA m-2 is depicted on the right axis.
Other capital costs such as anode and cathode materials, electrical wiring and control periphirals
have already been described elsewhere for BESs (Foley et al. 2010; Rozendal et al. 2008). On the
profit side a small amount of electrical power can be noted (when using MFC mode) and one can
possibly enter in a CO2-certificate trading scheme when emissions are succesfully lowered.
Figure S2: Left 3 schematics: bottom part of the microcosm used for rice growth. The top ledge contained a 2 cm
wide rim (not shown) with a 1 cm wide slot to accommodate the closed chamber for headspace gas
measurements. Right 2 schematics: Chamber to allow headspace gas measurements on the microcosms.
Figure S3: Result from the exploratory study where removing the external resistance resulted in an increased
CH4-flux 22 times higher than can be estimated based on the current in closed circuit. CC: Closed Circuit, OC:
Open Circuit.
Table S1: Overview of all experiments. 1) Composition see main text. 2) amount of granules based on previous
work (Arends et al., 2012) 3) only current collectors present in the anode compartment. 4) non-planted systems
are considered as sediment-MFC 5) liquid was added here without organic carbon. OC: open circuit. CC: closed
circuit over 500 Ω external resistance. A: anodic effluent of a well performing MFC in the lab 10 vol-% total
anode compartment. M: anaerobic methanogenic sludge from WWTP ‘Ossemeersen’ Gent, Belgium 1 vol-% of
total anode compartment.
#
# of
plants
Seedling age at
transplanting (week)
Added organic
carbon (kg m-2)1)
OC/CC
Inoculum
(A and/or M)
Granules in
anode 2)
Plant growth
area (m2)
Exploratory study
E1
2
7
2
OC
A & M
Y
0.015
E2
2
7
2
CC
A & M
Y
0.015
E3
2
7
2
CC
A & M
No 3)
0.015
E4
-4)
-
2
OC
A & M
Y
0.015
E5
-
-
2
CC
A & M
Y
0.015
E6
2
12
2
OC
A & M
Y
0.032
E7
2
12
2
CC
A & M
Y
0.032
E8
2
12
No
CC
A & M
Y
0.032
E9
-
-
2
CC
A & M
Y
0.032
E10
-
-
2
CC
A
Y
0.032
Detailed study
(kg m-2 d-1)1
M1
2
3
No
CC
A
Y
0.015
M2
2
3
No5)
CC
A
Y
0.015
M3
2
3
0.013
CC
A
Y
0.015
M4
2
3
0.013
CC
A
Y
0.015
Table S2: Quality control parameters for qPCR analysis. Parameters as obtained during analysis with
StepOnePlus software V2.2.2. Detection limit calculated to copies/gram wet weight of the sample taking
dilution and extraction efficiency into account.
Amplification plot
Standard Curve
Threshold
(ΔRn)
Baseline
Slope
R2
Eff (%)
Detection limit
(copies g-1)
Total Bacteria
3.1
3
-3.3
0.98
100
3.1x105
Total Archaea
4.5
3
-3.2
0.97
100
4.4 x103
Methanobacteriales
2.9
3
-3.6
0.99
88
3.9 x103
Methanomicrobiales
2.6
3
-4.0
0.99
78
3.6 x103
Methanosarcinaceae
4.5
3
-3.7
1.00
86
4.8 x103
Methanosaetaceae
2.6
3
-3.6
0.99
91
5.1 x103
Supplementary references
Arends, J.B.A., Verstraete, W., 2012. 100 years of microbial electricity production: Three concepts for
the future. Microbial Biotechnology 5, 333-346.
Domínguez-Garay, A., Berná, A., Ortiz-Bernad, I., Esteve-Núñez, A., 2013. Silica colloid formation
enhances performance of sediment microbial fuel cells in a low conductivity soil.
Environmental Science & Technology.
Foley, J.M., Rozendal, R.A., Hertle, C.K., Lant, P.A., Rabaey, K., 2010. Life cycle assessment of high-
rate anaerobic treatment, microbial fuel cells, and microbial electrolysis cells. Environmental
Science & Technology 44, 3629-3637.
Rozendal, R.A., Hamelers, H.V.M., Rabaey, K., Keller, J., Buisman, C.J.N., 2008. Towards practical
implementation of bioelectrochemical wastewater treatment. Trends in Biotechnology 26,
450-459.
