1 Methane biofiltration using Autoclaved Aerated Concrete as carrier material
2
3 Giovanni Ganendra a,d , Daniel Mercado Garcia a , Emma Hernandez-Sanabria a , Pascal
4 Boeckx b
, Adrian Ho c
, Nico Boon
a,#
5
6
7
8 a
Laboratory of Microbial Ecology and Technology, Ghent University, Coupure Links
9 653, B-9000, Gent, Belgium
10 b
Laboratory of Applied Physical Chemistry, ISOFYS, Ghent University, Coupure Links
11 653, B-9000 Gent, Belgium
12 c
Department of Microbial Ecology, Netherlands Institute of Ecology (NIOO-KNAW),
13 Droevendaalsesteeg 10, 6708 PB, Wageningen, The Netherlands
14 d SIM vzw, Technologiepark 935, BE-9052, Zwijnaarde, Belgium
15
16 * Correspondence to: Nico Boon, Ghent University; Faculty of Bioscience Engineering;
17 Laboratory of Microbial Ecology and Technology (LabMET); Coupure Links 653; B-
18 9000 Gent, Belgium; phone: +32 (0)9 264 59 76; fax: +32 (0)9 264 62 48; E-mail:
19 Nico.Boon@UGent.be
; Webpage: www.labmet.Ugent.be
.
20
1

1
2 The methane removal capacity of Methane-Oxidizing Bacteria (MOB) in a biofilter
3 setup using Autoclaved Aerated Concrete (AAC) as carrier material was tested. Three
4 biofilters (MBF-A, MBF-B, and MBF-C) were run in series to remove methane at
5 ~1000 ppmv methane concentration for 127 days. Optimum methane removal was
6 obtained when calcium chloride was not added during inoculation step and 10 mm thick
7 specimens were used. Optimum methane removal was exhibited when two biofilters
8 were run in series (average removal efficiency = 35 %). During biofilter test, liquid
9 nutrient feeding was essential to the methane removal capacity of the bacteria. Higher
10 MOB abundance was observed at the bottom of the filter due to the higher methane
11 concentration, and therefore growth, in that part of the filter. Overall, AAC holds an
12 advantage over other carrier material as the material could sequester carbon dioxide
13 production by the bacteria making AAC a sustainable carrier material.
14
15 Keywords: Methane biofilter, Methane-Oxidizing Bacteria, Autoclaved Aerated
16 Concrete, Carbon sequestration
17
18
2

1
2 Atmospheric methane plays a significant role in the global climate warming by
3 contributing to 0.5 W m
-2
out of 2.77 W m
-2 of the total radiative forcing of long lived
4
greenhouse gasses (Dlugokencky et al., 2011). Both natural (347 Tg year
-1 ) and
5 anthropogenic processes (331 Tg year
-1
) contribute with almost equal rates to the total
6
atmospheric methane emission (Kirschke et al., 2013). Among different anthropogenic
7 emission sources, methane emission from both energy (e.g., fossil fuel combustion) and
8 agricultural sectors (e.g., livestock) are the largest contributors. With increasing food
9 and energy demand due to the human population growth, global methane emission is set
10 to increase in the future. Therefore, its mitigation is essential for the regulation of the
11 global methane budget.
12 Methane-Oxidizing Bacteria (MOB) are responsible for ~5 % of the global methane
13
sink (Conrad, 2009). MOB are microorganisms capable of oxidizing methane as both
14 carbon and energy sources. MOB possess the Methane Monooxygenase (MMO)
15 enzyme which enable the bacteria to oxidize methane to methanol. Via a series of
16 oxidation reactions, methanol is subsequently converted to formaldehyde, a central
17
component in the bacterial metabolism (Hanson & Hanson, 1996). Formaldehyde is
18 partly utilized to synthesize biomass and partly to carbon dioxide to generate energy.
19 Due to its versatility and ease of applicability, MOB is central in the biotechnological
20
applications to mitigate methane emission (Semrau et al., 2010).
21 Biofiltration is the typical biotechnological application to mitigate methane emission in
22
places with high methane emission (e.g., coal mines, livestock barns) (Apel et al., 1991;
23
Veillette et al., 2012). Methane biofiltration is the removal of methane by MOB
24 embedded on static carrier material from waste gas flowing through the filter. Biofilter
3

1 performance is influenced by several technical (e.g., reactor dimension) and operational
2 parameters (e.g., inlet load). In biofilter process design, the carrier material selection is
3
essential for optimum methane removal (Huang et al., 2011). Biofilter carrier material
4 can be organic (e.g., compost) and inorganic (e.g., gravel stone) material. The advantage
5 of using organic material includes additional nutrient provision (e.g., N and P sources)
6 for the bacteria from the material whereas inorganic material is more durable as it does
7
not deteriorate with time (Akdeniz et al., 2011; Veillette et al., 2012). For better biofilter
8 performance (e.g., higher elimination capacity), inorganic carrier material is typically
9
preferred (Nikiema et al., 2005).
10
One of the most important properties of the carrier material is porosity (Cohen, 2001;
11
Pratt et al., 2012). Material with high porosity can accommodate a high number of
12
bacteria by providing a vast adsorption site for the bacteria (Cohen, 2001; Samonin &
13
Elikova, 2004). Among different materials, Autoclaved Aerated Concrete (AAC)
14 possesses this beneficial characteristic. AAC is a lightweight porous concrete consisting
15 of calcium silicate hydrate typically used for wall, floor, and roof panels of residential
16 and industrial buildings. AAC possesses a high porosity (i.e., up to 80 % (v/v)) due to
17
the gas entrapment by the aerating agent during manufacturing process (Narayanan &
18
Ramamurthy, 2000).
19 In a proof of concept study, a high methane removal was exhibited by different MOB at
20 low (~100 ppmv) and high (~20 % (v/v)) methane concentration when they were
21
immobilized on AAC (Ganendra et al., 2014). Presently, we aim to investigate the
22 performance of MOB immobilized on AAC to remove methane in a biofilter setup.
23 Firstly, we performed batch tests to optimize MOB immobilization on AAC. Secondly,
24 the methane removal capacity of the immobilized MOB in a biofilter setup was tested.
25
4

1
2 2.1 Materials
2.1.1 Microorganisms and growth medium
3 MOB mixed culture was enriched from circum-neutral agricultural soil (pH ~7.9)
4
originating from Ghent, Belgium (Ganendra et al., 2014). The culture was
5 predominantly composed of Methylocystis spp . as revealed by a diagnostic microarray
6 analysis targeting the pmoA
gene of the MOB (Ganendra et al., 2014). The use of mixed
7 culture may be more feasible (e.g. set up need not be sterile) and effective as methane
8
uptake rates have been shown to be higher in co-cultures (Ho et al., 2014). Prior to
9 experimental set-up, 200 ml culture was sub-cultivated in Nitrate Mineral Salt (NMS)
10
medium (Whittenbury et al., 1970) in a 1 L serum bottle (Schott Duran, USA). For
11 culture sub-cultivation, a 10% (v/v) culture from previous enrichment was added in
12 NMS medium before the bottle was closed using a butyl rubber stopper and sealed with
13 an aperture cap. Methane gas (99.5 % (v/v); Linde Gas, Belgium) was subsequently
14 injected into the headspace until it reached ~20 % (v/v) headspace concentration. The
15 bottle was incubated on a shaker (120 rpm) at 20 °C until the culture was enriched (~2
16 X 10
8
cells ml
-1
).
17
2.1.2 Autoclaved Aerated Concrete
18 AAC ( Ytong , Belgium) was cut into triangular specimens (FIG 1a.) for the batch
19 optimisation test or circular discs for the biofilter test (FIG 1b.). The triangular
20 specimens were 30 mm in radius with varying thickness (see batch optimisation test
21 procedure). The circular discs were 10 mm thick with a diameter of 90 mm with four
22 openings (15 mm x 10 mm). The openings were made for gas passage in the biofilter.
5

1 The specimens were stored at 28 ° C prior to use.
2
3 2.2 Methods
2.2.1 Batch optimisation test
4 Batch optimisation test was aimed to optimize MOB immobilization on AAC specimens
5 for the biofilter test. The test was conducted to investigate the influence of: (1) calcium
6 chloride addition into the bacterial culture and (2) AAC specimens thickness on the
7 methane removal capacity of the immobilized bacteria. Calcium addition to bacterial
8
culture is known to promote bioflocculation (Sobeck & Higgins, 2002). Floc formation
9 inside AAC specimens is anticipated to improve MOB immobilization on the
10 specimens.
11 Forty grams of AAC specimens were inserted into 1 L serum bottles containing 200 ml
12 of enriched MOB culture. The bottles were subsequently closed using a butyl rubber
13 stopper and sealed with an aperture cap. Methane gas was subsequently injected into the
14 headspace until it reached ~20 % (v/v) headspace concentration. The bottles were
15 incubated on a shaker (120 rpm) at 20 °C for 48 hours. Afterwards, the liquid was
16 poured out of the bottles and the bottles were closed before new methane gas was
17 injected into the headspace.
18 The methane removal capacity of the MOB was investigated by analyzing the evolution
19 of the methane concentration in the headspace of the bottles. As oxygen was also a
20 substrate for the bacterial methane oxidation, the oxygen concentration in the headspace
21 was maintained above 5 % (v/v) by injecting new oxygen (99,5 % (v/v); Air Liquide,
22 Belgium) into the bottles. Methane was replenished to maintain a constant concentration
23 level at ~20 % (v/v).
6

1 For incubations with varying calcium chloride addition, different volume of 1 M
2 calcium chloride was added to different bottles containing bacterial culture after the
3 specimens were inserted to reach 30, 50, 70, and 90 mM final calcium chloride
4 concentration. Incubations of AAC specimens in MOB culture without calcium chloride
5 addition served as reference incubations. In this test, 10 mm thick AAC specimens were
6 used. A 10, 15, or 20 mm thick specimen was used for incubations with varying
7 specimen thickness. For simplification, the following symbols were assigned to
8 different incubations with the following calcium chloride addition: 0 mM (MC1,
9 controls), 30 mM (MC2), 50 mM (MC3), 70 mM (MC4), and 90 mM (MC5). The
10 following symbols were assigned for incubations with: 10 mm (MT1), 15 mm (MT2),
11 20 mm (MT3) thick specimens. Each treatment was performed in triplicate.
12
2.2.2 Biofilter configuration and test
13 Three biofilters were made from a hollow transparent polyethylene (PE) tube (ISPA
14 plastic, The Netherlands) with a dimension of: 80 cm (length) X 9 cm (diameter). For
15 each biofilter, 7 holes with 12 mm diameter were made at 9 cm apart along the biofilter
16 bed. The holes, which would be used as gas sampling ports, were closed with butyl
17 rubber stoppers and sealed with epoxy glue (Loctite, USA).
18 Before placing in the biofilter, the disc was innoculated with MOB by immersing it in
19 fully grown culture in 5 L serum bottles (Schott Duran, USA) for 48 hours using
20 method described previously. At the end of the immersion period, the discs were
21 separated from the culture and placed inside the biofilter in spiral trajectory direction
22 (FIG 2a) aiming to prevent gas flow obstruction caused by biofilm clogging. To hold
23 the discs and for gas inlet distribution, a circular plastic with numerous holes was glued
7

1 to the biofilter wall at the base of the bottom disc. Both ends of the biofilter were
2 subsequently closed with PVC flanges (ISPA plastic, The Netherlands) and tightened
3 with 8 screws (M8 x 60 mm; Ijzewaar, Belgium) to make the biofilter gas tight.
4 The biofilter bed was made out from 45 cm tall specimens which started at 15 cm from
5 the bottom of the filter. Five gas sampling ports were situated equally to different
6 specimens’ height whereas the other two ports were located close to the gas inlet and
7 outlet. The sampling ports were numbered sequentially from the bottom to the top part
8 of the filter (i.e., 1 st and 7 th ports were the one adjacent to the gas inlet and outlet,
9 respectively).
10 The biofilter process configuration can be seen in FIG 2b. The biofilter was connected
11 at its base to a gas line coming from compressed air (Air Compact, Belgium) and
12 methane gas (99,5 % (v/v); Air Liquide, Belgium) mixing point. Both gas flows were
13 regulated using mass flow controllers that were connected to a control module (EX-
14 FLOW; Bronkhorst, The Netherlands) so that ~1000 ppmv methane/air mixing ratio
15 was fed into the biofilter. To check for leakage, gas was fed into the hollow filter prior
16 to the addition of carrier material for one week and the methane concentration in the
17 filter was monitored regularly.
18 NMS medium was intermittently fed at the top of the filter from a 10 L nutrient tank to
19 have a gas/liquid counter current flow in the biofilter. The liquid was fed every 6 hours
20 for 1 minute (120 ml min
-1
) using a pump (Cole-Parmer, USA) equipped with a timer
21 (Chacon, Belgium). At the bottom of the filter, the liquid was collected and circulated
22 back to the tank. The liquid nutrient composition was checked regularly to ensure
23 enough nutrient provision for the bacteria. When one of the nutrients was depleted, the
24 liquid was replaced by fresh NMS medium. The biofilter was operated in a temperature
8

1 controlled room (20° C). Summary of the biofilter operating and design parameters can
2 be seen in Table 1.
3 In the final configuration, three biofilters were run in series. Initially, one biofilter
4 (MBF-A) was operated for 36 days. For the first 20 days the gas flow rate was set at
5 approximately half of the design value (Table 1). MBF-B was installed after 36 days
6 and it was placed prior to MBF-A. Both biofilters were subsequently run in series for 37
7 days before the third biofilter (MBF-C) was installed. MBF-C was placed prior to MBF-
8 B and all biofilters were operated for 53 days before the experiment was stopped. The
9 biofilters were arranged to promote higher biomass growth in the newly installed
10 biofilter by feeding it with methane at higher concentration. To simulate the influence of
11 liquid nutrient feeding on the bacterial methane removal, from day 86 to 106, the liquid
12 feeding was stopped.
13 Several parameters were calculated in this study. The Empty Bed Residence Time
14 (EBRT) was calculated using the following equation:
15 EBRT (s) = V f
/Q (1)
16 where V f
is the filter bed volume (m 3 ) and Q is the inlet gas flow rate (m 3 h -1 ). The
17 methane volumetric load is divided into Inlet Load (IL) and Outlet Load (OL) and they
18 were calculated using the following equations:
19 IL (g m
-3
h
-1
) = (C in
X Q)/V f
20 OL (g m
-3
h
-1
) = (C out
X Q)/V f
(2)
(3)
21 where C in
and C out
were the inlet and outlet methane concentrations (g m -3 ),
22 respectively, measured from gas samples taken from the 1 st
(C in
) and 7 th
(C out
). C out
23 always corresponded to the top (7 th ) sampling port of MBF-A. C in
corresponded to the
24 bottom (1 st
) sampling port of MBF-A, MBF-B, and MBF-C in the 1 st
(day 1 to 36), 2 nd
9

1 (day 37 to 72), and 3 rd
(day 73 to 127) period, respectively. To obtain the methane
2 removal capacity of MOB, the Elimination Capacity (EC) was calculated using the
3 following equation:
4 EC (g m
-3
h
-1
) = (Q/V f
) X (C in
– C out
) (4)
5 The EC was calculated for each biofilter with the total EC being the sum of all. The
6 methane Removal Efficiency (RE) of MOB was calculated using the following
7 equation:
(5) 8 RE (%) = ((C in
– C out
)/C in)
X 100%
9
2.2.3 Gas composition analysis
10 Gaseous oxygen, methane, and carbon dioxide were analyzed. For batch optimisation
11 test, 1 ml of gas sample was taken from the headspace of each bottle using a gas tight
12 syringe (Hamilton, Belgium) and directly analyzed. For the biofilter test, duplicate gas
13 samples were taken from each sampling port using a gas tight syringe (Hamilton,
14 Belgium) and transferred to 12 ml vacutainers (Becton Dickinson, Belgium) that had
15 been vacuumed prior to the analyses. Oxygen, methane (in batch optimization test), and
16 carbon dioxide were measured using a Compact Gas Chromatography (GC) (Global
17 Analyser Solution, The Netherlands) equipped with a thermal conductivity detector, a
18 Porabond pre-column, and a Molsieve SA column. Methane in the gas samples was
19 analyzed using a Trace GC Ultra (Thermo Scientific, Belgium) equipped with a flame
20 ionization detector. Gas pressure inside the biofilters and serum bottles was measured
21 using a tensimeter (WIKA, Germany).
22
2.2.4 Nutrient composition analysis
10

1 Nitrate, nitrite, sulphate, and phosphate were determined in the liquid from the nutrient
2 tank. A 1 ml liquid sample was taken from the tank and diluted 10 times prior to
3 analysis. The ions concentration in the sample was analyzed using a 761 Compact Ion
4 Chromatograph (Metrohm, Switzerland) equipped with a thermal conductivity detector.
5
2.2.5 DNA extraction
6 At the end of the biofilters test, triplicate samples were collected from the surface of
7 specimen located beside the sampling port. There were five different types of specimen
8 samples from each biofilter. The samples were taken by means of scrapping the
9 specimen’s surface (1 mm deep). The samples consisted mainly of cell biomass with
10
11 residual AAC specimens. After homogenization, an aliquote of the sample was dried in an oven at 70° C for 24 hours to determine the dry weight. The remaining sample was
12 stored in the -20˚C freezer till DNA extraction. Afterwards, the samples were weighed
13 and stored at -20 °C prior to analyzes.
14 Total DNA was extracted from the samples using a physical disruption method
15 described in detail before (Vilchez-Vargas et al., 2013). Upon lysis and disruption at
16 1,800 rpm for 3 min, phenol-chloroform-isoamyl ethanol (25:24:1) extractions were
17
followed. DNA was precipitated and washed twice with cold ethanol (Hernandez-
18
Sanabria et al., 2010) and resuspended in 50 µl of TE buffer (10 mM Tris-HCl, 1 mM
19 EDTA [pH 8.0]). The amount and quality of DNA were measured using an ND 1000
20 spectrophotometer (NanoDrop Technologies, Wilmington, DE).
21
2.2.6 qPCR analysis
22 A quantitative PCR (qPCR) assay was performed to investigate the MOB abundance on
11

1 AAC specimens at different bed heights. The quantification of the pmoA gene was used
2 as proxy for the total MOB community. The pmoA gene (a subunit of the gene encoding
3 for the particulate methane monooxygenase enzyme) is present in virtually all obligate
4
methanotrophs and is congruent with the 16S rRNA gene phylogeny (Kolb et al., 2003),
5 making the pmoA
gene suitable for the detection of methanotrophs (Ho et al., 2011).
6
7 qPCR targeting the pmoA gene was performed using the A189f/mmb661r primer combination. Briefly, each qPCR reaction (total volume 20 µl) consisted of 10 µl 2X
8
9
10
SensiFAST SYBR (BIOLINE, the Netherlands), 3.5 µl of A189f forward primer (5 pmol/µl), 3.5 µl mmb661r reverse primer (5 pmol/µl), 1 µl Bovine Serum Albumin (5 mg/ml; Invitrogen, the Netherlands), and 2 µl diluted template DNA. In a preliminary
11 qPCR run, DNA template was diluted (10X, 50X, and 100X dilution) to determine the
12 optimal target yield. Henceforth, DNA was diluted 100X to achieve the optimum pmoA
13 gene copy numbers. The PCR program consisted of an initial denaturation step at 95°C
14
15 for 3 min, followed by 45 cycles of 95°C for 10 s, 62°C for 10 s, and 72°C for 25 s.
Fluorescence signal was obtained at 87°C (8 sec) after each cycle, and melt curve
16 obtained from 70°C to 99°C (1°C temperature increase). The qPCR was performed with
17 a Rotor-Gene Q real-time PCR cycler (Qiagen, the Netherlands). Duplicate qPCR
18 reactions were performed for each template DNA giving a total of six replicates per
19 sampling point along the vertical profile.
20
2.2.7 Statistical analysis
21 Except for the biofilter test (duplicate measurements), values are the mean of triplicate
22 measurement values. Error bars represent standard deviations. Comparison of means,
23 assuming normal distribution was done using one-way ANOVA test (p =0.05).
12

1 Subsequent pairwise multiple comparisons tests (Holm–Sidak procedure) were
2 performed to compare the differences between two mean values in the experiment (α =
3 0.05). Statistical analyses were carried out in SigmaPlot v12.0 (Systat Software Inc.,
4 USA).
5
13

1
2 3.1 Batch optimization test
3 Beside in MC5, methane was removed from the headspace of all bottles after each
4 methane addition (Fig 3a.). MC5 were stopped after 32 days due to negligible methane
5 consumption by the MOB. The highest initial methane removal rate was exhibited in
6 MC2 (179.7 ± 3.9 µg C-CH
4
(g AAC d) -1 ) and MC3 (179.8 ± 1.2 µg C-CH
4
(g AAC -
7 d)
-1
). After the fifth methane addition, methane was removed at higher rate in MC2
8 (301.6 ± 1.9 µg C-CH
4
(g AAC
-
d)
-1
) than in MC3 (245.2 ± 0.9 µg C-CH
4
(g AAC
-
d)
-
9
1
). Although having lower initial methane removal rate, similar rate was observed
10 afterwards in M1, M2, and M3. The methane removal rate differences were not
11 significant when specimens with different thickness were used up to 20 days of
12 incubation (p<0.05) (Fig 3b.). This was observed in all incubations up to the fourth
13 methane addition. Afterwards, higher removal rate was examined in MT1 (304.2 ± 1.8
14 µg C-CH
4
(g AAC d) -1 ) than in MT2 (196.2 ± 26.6 µg C-CH
4
(g AAC d) -1 ) or
15 MT3(200.1 ± 11.7 µg C-CH
4
(g AAC
-
d)
-1
).
16 The bacterial culture was supplemented with Calcium chloride to promote floc
17 formation inside AAC specimens. Salt addition has been known as a strategy to induce
18
floc formation in wastewater treatment processes (Moussa et al., 2006; Zita &
19
Hermansson, 1994). Dications salt (e.g., calcium chloride) addition into wastewater
20 stream increases the attractive energy (DLVO theory) between negatively charged
21 bacteria by forming a “cation bridge”. This preludes floc formation and its subsequent
22
settling (Zita & Hermansson, 1994). In this study, we hypothesized that calcium salt
23 addition after specimens immersion in the culture would increase the probability of
14

1 having bacterial aggregate size exceeding the specimens pore size after the bacteria
2 entered the specimen. Thus, when separating the liquid, these aggregates remained in
3 the specimens instead being poured out of the bottles. By incorporating higher number
4 of bacteria inside the specimens, higher methane removal by the bacteria was
5 anticipated.
6 Calcium chloride addition up to 70 mM only slightly increased the initial methane
7 removal rate by the bacteria (Fig 3a.). Although higher number of bacteria might be
8 incorporated inside the specimens, salt addition inhibited the bacterial methane uptake.
9 Salt addition lowers the methane solubility in the medium and its transfer from the gas
10 to the liquid phase (Schnell and King, 1996). Previous researches showed that salinity
11 inhibited the methane removal capacity of both pure (i.e., M. trichosporium OB3b) and
12
mixed MOB culture (King & Schnell, 1998; van der Ha et al., 2010). The methane
13 removal capacity of MOB was lost upon 90 mM calcium chloride addition as a result of
14 plasmolysis caused by hyperosmotic shock on the cells. Plasmolysis inhibits molecular
15
biosynthesis which impair cellular activities (Csonka, 1989).
16 Higher methane removal rate was shown in incubations using thinner specimens beyond
17 20 days of incubation (FIG 3b.). Thinner specimens exhibited higher surface area
18 whereas thicker specimens possess higher pore volume per gram of material. This
19 implies that higher methane removal rate could be obtained using thinner specimens due
20 to the higher contact area between bacteria and methane. On the other hand, higher
21 methane removal rate could also be obtained using thicker specimens as higher number
22 of bacteria can be immobilized on the specimens. In our study, higher methane removal
23 rate was eventually observed in incubations using thinner specimens. We concluded that
24 due to bacterial growth on the specimens surface after 20 days, the bacterial layer had
15

1 increased the methane diffusion resistance to deeper pores and this had lowered the
2 methane conversion rate. Overall, 10 mm thick specimens would be used for the
3 biofilter test and calcium chloride would not be added during bacterial inoculation step.
4
5 3.2 Biofilter test
6 Methane was removed in the biofilter with varying EC and RE over the period of the
7 test (FIG 4a. and FIG 4b.). After 20 days, the EC of MBF-A increased concomitantly
8 when the IL was doubled. After MBF-B was installed, lower total EC (250.3 g-C CH
4
9 m
-3
day
-1
) was observed (day 44) compared to when MBF-A was run individually
10 (425.1 g-C CH
4 m
-3
day
-1
; day 36). When both biofilters were run longer, the total EC
11 increased sharply and this was contributed from the increasing EC in MBF-A (468.5 g-
12 C CH
4 m
-3
day
-1
; day 64). The EC in MBF-B remained at the same level (~100 g-C CH
4
13 m -3 day -1 ) after its start-up. Prior to MBF-C installation on day 72, highest total EC (630
14 g-C CH
4 m
-3
day
-1
) and RE (65.54 %) were reached on day 66.
15 A decrease in the total EC was observed when MBF-C was added. The total EC
16 decreased to 396 g-C CH
4 m -3 day -1 (day 80) after all three biofilters were run in series.
17 When liquid nutrient feeding was stopped on day 86, the total EC decreased further
18 until it reached the lowest capacity (176.7 g-C CH
4 m
-3
day
-1
) on day 106. After liquid
19 nutrient feeding was started again, the total EC subsequently increased to 222.9 g-C
20 CH
4 m
-3
day
-1
(day 117) and it remained approximately at the same level until the end of
21 the test. Carbon dioxide concentration during the experiment always remained at
22 atmospheric level (~300 ppmv; unpublished data). Overall, the average RE during the
23 experiment was: 28.5 % (day 1 to 36), 39.7 % (day 37 to 72), and 27.8 % (day 72 to
24 127).
16

1 The methane removal profile along the filter bed varied over the period of the test (FIG
2 5.). Initially, methane was equally removed in MBF-A (day 12). Afterwards, higher
3 methane removal was observed at the top (port 5 and 6) of the filter (days 27 and 36).
4 When MBF-B was started, the methane removal profile in MBF-A reverted back to the
5 linear trend examined on day 12 (day 44). The high total EC observed after day 44
6 occurred mainly at the bottom part of MBF-A (port 1 and 2; day 64). After MBF-C
7 addition, the high activity shown at the bottom part of MBF-A was reduced and linear
8 methane removal pattern was again observed along the filter length (day 78). As the
9 three filters were continuously operated, higher methane removal was exhibited in the
10 middle (ports 3 and 4) and top (ports 5 and 6) of MBF-A (day 117) whereas a linear
11 methane removal pattern was always shown in MBF-B and MBF-C since they were
12 started.
13 Among the MBFs, significantly higher total MOB abundance was observed in MBF-B
14 (9.5 X 10
8
± 1.6 X 10
8 pmoA copies g dw
-1
; p<0.05) (FIG 6.). MOB abundance were not
15 significantly different in MBF-A and MBF-C (p<0.05). In all MBFs, the highest MOB
16 abundance was detected at the bottom part of the biofilters (sampling ports 1 and 2) ,
17 whereas the lowest abundance was found in the middle region (sampling ports 3 and 4).
18 More specifically, the highest and lowest MOB abundance were detected at sampling
19 port 2 ( 8.8 X 10
8
± 3.5 X 10
8
pmoA copies g dw
-1
) and 4 of MBF-B ( 8.9 X 10
4
± 4.3
20 X 10
4
pmoA copies g dw
-1
), respectively.
21 After enrichment at high methane concentration (~20 % (v/v)), MOB mixed culture
22 could remove methane at significantly lower concentration (~1000 ppmv) in the
23 biofilter for a relatively long period. The initial objective of the biofilter test was to have
24 an MOB enrichment on the filter bed. The culture was predominantly comprised of
17

1 Methylocystis spp . and Methylocystis -like spp.
(Ganendra et al., 2014), proven to
2 oxidize methane at low atmospheric concentration (i.e., ppmv level) and grown under
3
~1000 ppmv methane concentration (Knief & Dunfield, 2005).
4 Bacterial adaptation period, indicated by the lower methane removal capacity, was
5 exhibited as soon as environmental changes were applied in the niche of the bacteria.
6 Previous batch and biotrickling filter studies also showed that MOB experienced an
7
adaptation period as soon as one of the process conditions was altered (Knief &
8
Dunfield, 2005; Yoon et al., 2009). In our study, this was observed after a new biofilter
9 was added on day 36 (MBF-B added) and 72 (MBF-C added) where the bacteria had to
10 adapt to the changing IL into the biofilter (FIG 4.). The adaptation period for MOB in
11 all biofilters lasted approximately 10 days. Afterwards, higher methane removal
12 capacity in the biofilter was examined with the exception being after MBF-C was
13 installed (FIG 4.). We concluded that although MOB in the other biofilters had adapted
14 after MBF-C was started, the liquid nutrient feeding was subsequently stopped. The
15 methane removal capacity was lowered as the bacteria needed nutrients (e.g., copper
16 and iron) to synthesize enzymes to oxidize methane.
17 The bacterial methane removal capacity was dependent on essential nutrients provision
18 (e.g., P and Cu sources) to the bacteria. Higher activity observed at the top part of MBF-
19 A after 36 days was a consequence of the proximity of the filter bed to the liquid feed
20 inlet (FIG 5.). At the start of the test, liquid was not uniformly distributed in the filter.
21 Hence, nutrients were more accessible to the bacteria at the top part of the filter. Higher
22 methane removal capacity observed at the lower bed of MBF-A after 64 days was a
23 consequence of the bed proximity to the gas inlet, hence, exposing the bacteria to a
24 higher methane concentration (FIG 5.). From the kinetic study, immobilized MOB on
18

1 building materials always exhibited higher methane removal rate at higher methane
2
concentration (Ganendra et al., 2014). Higher biomass growth was also anticipated as a
3 consequence of higher methane removal and this was further verified by the qPCR
4 analysis (FIG 6.).
5 Phosphorus islimiting nutrient for the bacterial methane removal (unpublished results).
6 This agrees with Nikiema et al (date) which examined that among other nutrients,
7 phosphorus was central to the bacterial growth and methane removal capacity of MOB
8
(Nikiema et al., 2010). This was further demonstrated when lower MOB activity was
9 observed when liquid feeding was stopped and subsequently increased when liquid
10 feeding was started again (day 86 to 106). Additionally, liquid feeding is important to
11 maintain humidity inside the biofilter.
12 Higher MOB abundance was not found in the biofilter with the longest methane feeding
13 period (i.e., MBF-A) (FIG 6.). Significantly higher MOB abundance was instead found
14 in MBF-B. The MOB abundance difference between MBF-A and MBF-C was not
15 significant although MBF-A was started 72 days before MBF-C. We concluded that a
16 fraction of MOB population on MBF-A had been lysed and degraded at the end of the
17 experiment making the total population similar to the growing biomass in MBF-C.
18
19 3.3 State-of- the -art on methane biofilter
20 The performance of methane biofilter used in this study is compared to the ones found
21 in the literatures (Table 2). The average RE obtained in our study (35 %) is similar to
22 other biofilters although operating at much lower EBRT (0.78 min) (Table 2). As
23
observed in the batch study (Ganendra et al., 2014), carbon dioxide was not produced
24 from the biofilter and this holds an advantage over other biofilters (Table 2). MOB
19

1 produce carbon dioxide, another greenhouse gas, as part of their catabolic reaction
2
(Hanson & Hanson, 1996). When AAC is used as carrier material, carbon dioxide was
3 removed by reacting with the binder material (i.e., 1,1-nm tobermorite) of AAC
4
(Ganendra et al., 2014). Therefore, a complete methane carbon sequestration is possible,
5 making AAC a sustainable carrier material.
6 On the other hand, inorganic material like AAC gives several disadvantages as the
7 biofilter carrier material. Firstly, the material cannot provide additional nutrients to the
8 bacteria. Therefore, external nutrient supply should be provided into the filter, which is
9 not necessarily the case when organic material was used (e.g., compost; Table 2). It was
10 further proven in this study that stopping the liquid nutrient feeding would lower the
11 methane removal capacity of the bacteria. Additional nutrient supply implies higher
12 biofilter operational cost. Secondly, bacterial inoculation of the filter is not needed
13 using organic filter material as MOB can be enriched from the indigenous organic
14 material (e.g., soil) community. The inoculation step is, on the other hand, needed when
15 inorganic material is used.
16
17
18
3.4 outlook
We demonstrated that two biofilters operating in series were sufficient to obtain
19 optimum methane removal capacity. Several considerations are needed for future
20 research. The influence of different EBRT (i.e., varying gas flow rate) on the RE in the
21 biofilter should be investigated to optimize the flow rate. The evolution of MOB
22 abundance during the test should also be analyzed. This could not be performedin the
23 current setup because itis not designed for non-invasive biomass sampling. We suggest
24 to use a mixture of organic and inorganic material (e.g., AAC and soil) as the carrier
20

1 material for application purposes in order tothe advantages of using both types of
2 materials (i.e., nutrient provision from the carrier material and carbon dioxide
3 sequestration) when operating the biofilter.
4 The biofilter application for the removal of methane from in-situ livestock gas waste
5 where lower methane mixing ratio was found (~100 ppmv) will be investigated in our
6 next study. The influence of other waste gas such as ammonia, commonly found in the
7 livestock gas waste to the methane removal capacity of MOB will be examined.
8 Ammonia is in equilibrium with ammonium in water and ammonium is known to bind
9 to MMO enzyme possessed by MOB as it is evolutionary related to the Ammonium
10
Mono-Oxygenase (AMO) in Ammonia Oxidizing Bacteria (Bedard & Knowles, 1989).
11
21

1
2 An optimum methane removal by MOB mixed culture was reached when the bacteria
3 were immobilized on 10 mm AAC specimens without calcium chloride addition prior to
4 the inoculation. The immobilized bacteria could remove methane at low methane
5 mixing ratio (~1000 ppmv) in a biofilter setup for 127 days. Optimum methane removal
6 capacity (RE = 35%) was obtained when two biofilters were operated in series. During
7 biofilter test, liquid feeding containing essential nutrients were necessary to obtain
8 optimum bacterial methane removal. Higher MOB growth, inferred by increased pmoA
9 gene abundance was observed at the bottom of the filter presumably due to the higher
10 methane concentration.
11
22

1
2 The project is fully funded by SIM-SHE SECEMIN project (SIM 2009-1) and the
3 Geconcerteerde Onderzoeksactie (GOA) of Ghent University (BOF09/GOA/005).
4 Special thanks to Ralph Lindeboom, Alberto Scoma, and Amanda Luther for critically
5 reviewing this paper. This publication is Publication no. xxx of the Netherlands Institute
6 of Ecology.
7
8
23

1
2 Akdeniz, N., Janni, K.A., Salnikov, I.A. 2011. Biofilter performance of pine nuggets
3
5 and lava rock as media. Bioresource Technology , 102 (8), 4974-4980.
4 Apel, W.A., Dugan, P.R., Wiebe, M.R. 1991. Use of methanotrophic bacteria in gasphase bioreactors to abate methane in coal-mine atmosphere. Fuel , 70 (8), 1001-
6 1003.
7 Bedard, C., Knowles, R. 1989. Physiology, Biochemistry, and Specific Inhibitors of
8
9
CH4, NH4+, and Co oxidation by Methanotrophs and Nitrifiers.
Microbiological Reviews , 53 (1), 68-84.
10 Cohen, Y. 2001. Biofiltration - the treatment of fluids by microorganisms immobilized
11 into the filter bedding material: a review.
Bioresour Technol , 77 (3), 257-274.
12 Conrad, R. 2009. The global methane cycle: recent advances in understanding the
13
15 microbial processes involved. Environ Microbiol Rep , 1 (5), 285-292.
14 Csonka, L.N. 1989. Physiological and genetic responses of bacteria to osmotic-stress.
Microbiol. Rev.
, 53 (1), 121-147.
16 Dlugokencky, E.J., Nisbet, E.G., Fisher, R., Lowry, D. 2011. Global atmospheric
17 methane: budget, changes and dangers. Phil. Trans. R. Soc. A , 369 (1943), 2058-
18
20
2072.
19 Ganendra, G., De Muynck, W., Ho, A., Hoefman, S., De Vos, P., Boeckx, P., Boon, N.
2014. Atmospheric methane removal by methane-oxidizing bacteria
21
22 immobilized on porous building materials. Applied Microbiology and
Biotechnology , 98 (8), 3791-3800.
23 Girard, M., Viens, P., Ramirez, A.A., Brzezinski, R., Buelna, G., Heitz, M. 2012.
24

1
2
Simultaneous treatment of methane and swine slurry by biofiltration. J Chem
Technol Biotechnol , 87 (5), 697-704.
3 Hanson, R.S., Hanson, T.E. 1996. Methanotrophic bacteria. Microbiol. Rev.
, 60 (2), 439-
4 471.
5 Hernandez-Sanabria, E., Guan, L.L., Goonewardene, L.A., Li, M., Mujibi, D.F.,
6
7
8
Stothard, P., Moore, S.S., Leon-Quintero, M.C. 2010. Correlation of Particular
Bacterial PCR-Denaturing Gradient Gel Electrophoresis Patterns with Bovine
Ruminal Fermentation Parameters and Feed Efficiency Traits. Applied and
9 Environmental Microbiology , 76 (19), 6338-6350.
10 Ho, A., de Roy, K., Thas, O., De Neve, J., Hoefman, S., Vandamme, P., Heylen, K.,
11
12
Boon, N. 2014. The more, the merrier: heterotroph richness stimulates methanotrophic activity. Isme Journal , 8 (9), 1945-1948.
13 Ho, A., Lueke, C., Frenzel, P. 2011. Recovery of methanotrophs from disturbance:
14
16 population dynamics, evenness and functioning. Isme Journal , 5 (4), 750-758.
15 Huang, Q., Zhang, Q., Cicek, N., Mann, D. 2011. Biofilter: a promising tool for mitigating methane emission from manure storage. Journal of Arid Land , 3 (1),
17 61-70.
18 Kim, T.G., Lee, E.-H., Cho, K.-S. 2013. Effects of nonmethane volatile organic
19
20 compounds on microbial community of methanotrophic biofilter. Applied
Microbiology and Biotechnology , 97 (14), 6549-6559.
21 King, G.M., Schnell, S. 1998. Effects of ammonium and non-ammonium salt additions
22
23 on methane oxidation by Methylosinus trichosporium OB3b and Maine forest soils. Applied and Environmental Microbiology , 64 (1), 253-257.
24 Kirschke, S., Bousquet, P., Ciais, P., Saunois, M., Canadell, J.G., Dlugokencky, E.J.,
25

1
2
Bergamaschi, P., Bergmann, D., Blake, D.R., Bruhwiler, L. 2013. Three decades of global methane sources and sinks. Nature Geoscience , 6 (10), 813-823.
3 Knief, C., Dunfield, P.F. 2005. Response and adaptation of different methanotrophic
4 bacteria to low methane mixing ratios. Environ. Microbiol.
, 7 (9), 1307-1317.
5 Kolb, S., Knief, C., Stubner, S., Conrad, R. 2003. Quantitative detection of
9
10
6
7 methanotrophs in soil by novel pmoA-targeted real-time PCR assays. Applied and Environmental Microbiology , 69 (5), 2423-2429.
8 Melse, R.W., Van der Werf, A.W. 2005. Biofiltration for mitigation of methane emission from animal husbandry. Environmental Science & Technology , 39 (14),
5460-5468.
11 Moussa, M.S., Sumanasekera, D.U., Ibrahim, S.H., Lubberding, H.J., Hooijmans, C.M.,
12
13
Gijzen, H.J., van Loosdrecht, M.C.M. 2006. Long term effects of salt on activity, population structure and floc characteristics in enriched bacterial
14
16 cultures of nitrifiers. Water Research , 40 (7), 1377-1388.
15 Narayanan, N., Ramamurthy, K. 2000. Structure and properties of aerated concrete: a review. Cement & Concrete Composites , 22 (5), 321-329.
17 Nikiema, J., Bibeau, L., Lavoie, J., Brzezinski, R., Vigneux, J., Heitz, M. 2005.
18 Biofiltration of methane: An experimental study. Chemical Engineering
19
21
Journal , 113 (2-3), 111-117.
20 Nikiema, J., Brzezinski, R., Heitz, M. 2010. Influence of phosphorus, potassium, and copper on methane biofiltration performance. Canadian Journal of Civil
22 Engineering , 37 (2), 335-345.
23 Pratt, C., Walcroft, A.S., Tate, K.R., Ross, D.J., Roy, R., Reid, M.H., Veiga, P.W. 2012.
24 In Vitro Methane Removal by Volcanic Pumice Soil Biofilter Columns over
26

1
3
One Year. Journal of Environmental Quality , 41 (1), 80-87.
2 Ramirez, A.A., Jones, J.P., Heitz, M. 2012. Methane treatment in biotrickling filters packed with inert materials in presence of a non-ionic surfactant. Journal of
4 Chemical Technology and Biotechnology , 87 (6), 848-853.
5 Samonin, V.V., Elikova, E.E. 2004. A study of the adsorption of bacterial cells on
6
8 porous materials. Microbiology , 73 (6), 696-701.
7 Semrau, J.D., DiSpirito, A.A., Yoon, S. 2010. Methanotrophs and copper. FEMS
Microbiol. Rev.
, 34 (4), 496-531.
9 Sobeck, D.C., Higgins, M.J. 2002. Examination of three theories for mechanisms of
10 cation-induced bioflocculation. Water Research , 36 (3), 527-538.
11 Streese, J., Stegmann, R. 2003. Microbial oxidation of methane from old landfills in
12 biofilters. Waste Management , 23 (7), 573-580.
13 van der Ha, D., Hoefman, S., Boeckx, P., Verstraete, W., Boon, N. 2010. Copper
14
15 enhances the activity and salt resistance of mixed methane-oxidizing communities. Applied Microbiology and Biotechnology , 87 (6), 2355-2363.
16 Veillette, M., Girard, M., Viens, P., Brzezinski, R., Heitz, M. 2012. Function and limits
17
18 of biofilters for the removal of methane in exhaust gases from the pig industry.
Appl Microbiol Biotechnol , 94 (3), 601-611.
19 Whittenbury, R., Phillips, K.C., Wilkinson, J.F. 1970. Enrichment, isolation and some
20 properties of Methane-Utilizing Bacteria. J. Gen. Microbiol.
, 61 , 205-218.
21 Yoon, S., Carey, J.N., Semrau, J.D. 2009. Feasibility of atmospheric methane removal
22
23 using methanotrophic biotrickling filters. Applied Microbiology and
Biotechnology , 83 (5), 949-956.
24 Zita, A., Hermansson, M. 1994. Effects of Ionic-Strength on Bacterial Adhesion and
27

1
2
3
4
5
6
Stability of Flocs in a Waste-Water Activated-Sludge System. Applied and
Environmental Microbiology , 60 (9), 3041-3048.
28

1
2
3 Figure. 1. AAC specimens used to immobilize MOB for the: (a) batch optimization test,
4 and (b) biofilter test.
5
6 Figure. 2. (a) Biofilter bed configuration when placed inside the biofilter and (b) the
7 biofilter process configuration when all three biofilters were run in series. When MBF-
8 A and MBF-B were run in series (day 37 to 72), gas entered firstly in MBF-B. When all
9 biofilters were run in series (day 73 to 127), gas entered firstly in MBF-C.
10
11 Figure. 3. The methane removal in different incubations when: (a) different amount of
12 calcium chloride was added into the culture or when (b) different thickness of AAC
13 specimens were used. The following symbols represent different calcium chloride
14 addition: 0 mM (MC1), 30 mM (MC2), 50 mM (MC3), 70 mM (MC4), and 90 mM
15 (MC5). For varying specimens thickness incubations: 10 mm (MT1), 15 mm (MT2),
16 and 20 mm (MT3).
17
18 Figure. 4. (a) Methane IL, OL, RE, and (b) EC in the biofilters over 127 days of
19 operation. Different area in (b) indicates the EC in each biofilter.
20
21 Figure. 5. Methane removal profile along the biofilter bed. Top three figures depict the
22 methane consumption in MBF-A in the first 36 days. The middle six figures depict the
23 methane removal in MBF-A and B from day 37 to 72. The bottom nine figures depict
24 the methane removal in MBF-A, B, and C from day 73 to 127. Methane always entered
29

1 from port 1 with port 2 to 6 situated equally to different filter bed height. Sampling port
2 7 was placed close to the gas outlet.
3
4 Figure. 6. MOB abundance on different filter bed height obtained from qPCR analyses.
5
30

1
2 Table 1. Biofilter design and operating parameters
Parameter
Empty biofilter volume (m
3
)
Biofilter bed volume (m
3
)
Biofilter bed mass (kg)
Compressed air flow rate (m
3
h
-1
)
Methane flow rate (m
3
h
-1
)
Inlet methane concentration (ppmv)
Temperature (°C)
EBRT (s)
Methane IL (g C-CH4 d
-1
m
-3
)
Nutrient feeding rate a
(ml min
-1
)
3 a
Nutrient feeding rate for 1 minute every 6 hours
Value
960
20
47
944.7
120
0.0051
0.0026
1.02
0.1998
0.0002
31

1 Table 2. Overview of state of the art of methane biofilter
Inoculum Gas source Filter bed
MOB mixed culture
Indigenous soil community
Activated sludge
Indigenous soil
Synthetic gas
Synthetic gas a a
AAC
Compost
Liquid manure storage Compost and Perlite b
Synthetic gas c Landfill soil
Temperature
(° C)
20
30
C in
(%)
V
(m 3 f
)
0.1 0.003
2.5 0.015
Q
(m 3 h-1)
0.2
IL EC
(g m -3 h -1 ) (g m -3 h -1 )
39.4
e 14 f
0.15 165 64
12
20
0.85
0.1
0.16
0.004
0.75
0.003
25
24
8
24
32
100
RE
(%)
35 f
39
EBRT
(min)
0.78
6
13.5
80
CO
2 emitted g
No
Liquid feeding
Yes
No
Reference
This study
(Streese & Stegmann, 2003
)
Yes
Yes
No
No
(Melse & Van der Werf, 2005)
(Pratt et al., 2012)
community
Swine slurry
Old biofilter
Landfill soil
2
3
4
5
6
7
8
9
Synthetic gas a
Synthetic gas a
Synthetic gas a
Gravel stone
Gravel stone
20-25
24
0.43
0.7
0.017
0.017
0.25
0.25
9
61.8
3.9
33
43
53
4.2
4.2
Yes
Yes
Yes
Yes
Pumice stone and
activated carbon d
20-25 1.5 0.005 0.015 30.4 18 59 20 Yes a Synthetic gas is defined as a mixture of methane and compressed air originating from source bottles b Filter bed volume composition is 60 (compost):40 (perlite) % (v/v) c Synthetic gas is defined as a mixture of biogas and compressed air originating from source bottles d
Filter bed weight composition is 10 (pumice stone):1 (activated carbon) (w/w) e
Value is average over the biofilter test period. The initial period where the inlet load was set half the design value was not taken into account f Value is an average taken from day 20 (inlet load set to design value) to 72 (before MBF-C was installed). Measurements from day 72 onwards was not considered because the liquid nutrient feeding
was stopped and this had influenced the elimination capacity of the bacteria. g Carbon dioxide production analyzed in the gas outlet stream
(Girard et al., 2012)
(Ramirez et al., 2012)
(Kim et al., 2013)
32