Applied Energy 222 (2018) 369–382

Contents lists available at ScienceDirect

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

Early-stage performance evaluation of flowing microbial fuel cells using T

chemically treated carbon felt and yeast biocatalyst
Marcelinus Christwardanaa,1, Domenico Frattinia,1, Grazia Accardob, Sung Pil Yoonb,
⁎
Yongchai Kwona,
a
Graduate School of Energy and Environment, Seoul National University of Science and Technology 232 Gongneung-ro, Nowon-gu, Seoul 01811, Republic of Korea
b
Fuel Cell Research Center, KIST – Korea Institute of Science and Technology, Hwarang-ro 14-gil 5, Seongbuk-gu, Seoul 02792, Republic of Korea

H I GH L IG H T S G R A P H I C A L A B S T R A C T

• Yeast and chemically treated CF ef-

fects on MFC performance are in-
vestigated.
• CdbndO/CsbndN dangled on CF-PEI
are key bonds for performance en-
hancement.
• Pi-pi bond conjugation and lone elec-
tron pair induce performance en-
hancement.
• High yeast growth rate and optimal
yeast growing time are determined.
• MPD of MFC using CF-PEI is
256.3 mW m . 2

A R T I C LE I N FO A B S T R A C T

Keywords: The performance of closed-loop flowing-type microbial fuel cells using differently pretreated carbon felts is
Carbon felt measured. Yeast cultivated from S. cerevisiae is used as biocatalyst, while glucose is the substrate. For the pre-
Flowing-type microbial fuel cell treatment of felt, acetone, nitric acid, and polyethyleneimine are employed. First the optimal conditions for yeast
Glucose cultivation are quantitatively determined. As a result, a high yeast growth rate (1.083 h−1) and the optimal yeast
Pretreatment
growing time (48 h) for cell tests are obtained. The differently pretreated felts are analyzed by X-ray photo-
Yeast
electron spectroscopy, electrochemical impedance spectroscopy and optical microscopy. Conductivity, charge
transfer resistance, and CdbndO and CsbndN groups dangled on the felt are crucial parameters determining the
performance of the microbial fuel cell. Particularly, the conjugation effects of pi-pi bonds and lone pairs facil-
itating the attachment of yeast to the CdbndO and CsbndN groups on the carbon felt promote (i) mutual adhesion
between them and (ii) growth of yeast on CF-PEI. This correlation is confirmed by optical analysis of the felts
after the cell tests. To evaluate the early-stage performance of the microbial fuel cells using the different felts,
polarization curves are measured. In the measurements, the maximum power density of the cells depends on the
superficial state of felts, while the performance of the cell using the PEI-treated felt is best, at
256.3 ± 11.5 mW·m−2. These data match other results attained by pretreatments.

⁎
Corresponding author.
E-mail address: kwony@seoultech.ac.kr (Y. Kwon).
1
These authors contributed equally to this work.

https://doi.org/10.1016/j.apenergy.2018.03.193
Received 23 November 2017; Received in revised form 11 March 2018; Accepted 31 March 2018
0306-2619/ © 2018 Elsevier Ltd. All rights reserved.
M. Christwardana et al.
1. Introduction
Microbial fuel cells (MFCs) are a fascinating technology that can
achieve both energy extraction from low-grade biomasses and re-
mediation of wastewaters, derived from a large variety of human and
industrial activities [1], and the solid fraction of organic waste [2].
Over the last decade, MFCs have been used to harvest energy not only
from derivatives of sugar [3], oilfield waters [4], urban wastewaters
[5], and rice mill wastewater [6] but also dairy and slaughterhouse
wastewaters [7], swine [8], azo dye residues [9] and composite food
waste [10]. Moreover, these bio-electrochemical systems (BESs) can
efficiently exploit sunlight [11,12], be incorporated in constructed
ecosystems [13] and detect glucose levels [14] due to the strict corre-
lation between their electrochemical and electrical behaviors [15]. An
interesting and recent trend is using microbial electrolysis cells (MECs)
for biohydrogen generation [16], solid waste valorization [17], bior-
emediation [4], and production of other valuable chemicals from solid
waste and food waste [18–20].
However, in spite of such remarkable progress, there are still lim-
itations to be addressed, such as reproducibility, standardized perfor-
mance and real economic feasibility [21–24]. Particularly, the eco-
nomic profitability of MFCs depends on (i) finding cost-effective
electrode and membrane materials to promote the generation of elec-
tricity without using precious and expensive metals and (ii) minimizing
the time-to-power production, i.e., the time required from electrode
pretreatments, start-up, to electricity generation. These points play a
key role in determining whether the MFCs can replace other competi-
tive wastewater treatment technologies because long-term operations
can be achieved by periodic substitution of the substrate and anolyte,
but reducing the time-to-power of MFCs is currently underestimated or
reported to a limited extend [25,26].
For example, Alatrakcthi et al. [27] achieved a power density of
374.9 mW·m−2 after two consecutive batch operations of several days
using an expensive anode with gold nanoparticles. LaBarge et al. [28]
studied the pre-acclimation steps of bog sediments in MECs to produce
methane and a useful gas evolution was achieved only after 4–7 cycles
(total time of 60–80 days), with granular activated carbon and an ex-
pensive ruthenium foil as electrodes. Liang et al. [29] reported power
densities of 600–1200 mW·m−2 when an intermittent switch on/off
regulation procedure was applied to the MFCs using a reduced gra-
phene oxide (rGO)/MnO2 anode and a high concentration of sodium
acetate as substrate. Zhang et al. [30] demonstrated that the current
obtained at the start-up of their MFC was dependent on even low ap-
plied resistances to biofilm formation at an early stage. The time-to-
power was always approximately 5–6 days and they did not mention
the effects of different anode materials. Later, they [31] reported a si-
milar study using a continuous-flow MFC and various external re-
sistances and concluded that biofilm distribution due to anolyte flow
could improve the start-up performance. Recently, Paitier et al. [32]
demonstrated that another limiting factor for the start-up and biofilm
establishment is the presence of competing microbial communities and
with different strains the time-to-stable-power was 10–20 days and
power density of these MFCs could be only maintained for a short in-
terval without a fresh substrate.
Therefore, reducing the start-up and acclimation time is required,
and the parameters that determine them should be effectively managed
to improve the feasibility in real scale applications. To alleviate the
slow procurement issue of starting inoculum, yeast can be considered as
a fast-growing inoculum. Differently from the bacteria used in the
above-mentioned works, yeasts have short growing periods, are time-
saving and in dried form can be stored under soft conditions for a very
long time. In prospect, for real and large scale applications, this is a
competitive advantage because it makes yeasts “ready-to-use”, easy to
stock and transport, unlike some bacteria strains, giving to yeasts a high
potential for practical exploitation in biotechnology processes [18].
In fact, yeasts have been already used in many real applications as
370
Applied Energy 222 (2018) 369–382
for example in biofuel production from residual biomass [20], bio-hy-
drogen evolution from fermentation of agro and industrial wastes and
sugar oxidation [33–35]. Even for MFCs, yeast has facilitated extra-
cellular electron transfer, and in particular, Saccharomyces cerevisiae is a
robust, fast-growing, facultative anaerobe, non-pathogen and tem-
perature tolerant biocatalyst that is mainly employed in mediated MFCs
[36–39] or membraneless MFCs [40]. It is also notable that in case of
oxygen infiltration, biological contamination or accidental increase in
temperature, S. cerevisiae can survive at a higher rate without sudden
stoppage, although it is not the most powerful exo-electrogenic bioca-
talyst due to its low open circuit voltage (OCV) of approximately
0.3–0.4 V [41]. It is therefore thought that its electron transfer me-
chanism should be promoted by using genetic engineering [42] or
surface modification of electrodes [43], with further improvements in
power and electricity production.
Recently, the use of layer-by-layer and physical immobilization
techniques of the biocatalyst has been an effective approach to promote
the interaction between electrode surface and biocatalyst [44–46].
According to Zhu et al. [47], a multistep CF treatment of 6–7 days was
proposed to functionalize carbon felt fibers. Despite the long pretreat-
ment, the start-up data were not clearly mentioned and after the start-
up the time needed for acquiring stable power was 200–250 h. Hidalgo
et al. [48] proposed two surface modifications based on (i) a nitric acid
treatment and (ii) in situ polyaniline deposition. However, even in this
case, the benefits attained by the use of nitric acid and polyaniline
should be better clarified because the glucose concentration was too
high (60 g·L−1) and the OCV of MFC was still low, and its power density
decreased after 24 h. Consequently, electrode modification is another
critical factor to improve the early-stage performance of MFCs and
carbon felt (CF) is mainly used as a supporting material of the electrode
for MFCs, but pretreatments are necessary before use. Therefore, more
efforts are required to further develop effective and easy surface mod-
ification methods to advance the feasibility and convenience of yeasts
as biocatalyst in MFCs.
Unlike the abovementioned studies, in this work, we report a
complete analysis including results from preparation of treated felts,
physical and chemical characterization using scanning electron micro-
scopy (SEM), X-ray photoelectron spectroscopy (XPS), electrochemical
impedance spectroscopy (EIS), optical contact angle (OCA) measure-
ment, and pH-grow curve of biocatalyst to quantify the optimal con-
ditions for yeast cultivation. Afterward, real application of the proposed
treatments is demonstrated with a new closed-loop flowing-type MFC
system using the three different one-step pretreated CF electrodes and
polarization, power curves, OCV, maximum power density (MPD) are
reported to compare the performances. Yeast from S. cerevisiae and
glucose are the biocatalyst and substrate, respectively. The effects of
modifications of the carbon felt electrode on surface chemistry after full
cell tests are characterized using XPS and digital optical microscopy
(DOM), while the increase in immobilized yeast due to the treatments is
measured in terms of dry biomass before and after use in anodes.
2. Materials and methods
2.1. Carbon felt treatments
The carbon felt (XF30A–3.5T) was purchased from Toyobo (Osaka,
Japan). Polyethyleneimine (PEI) (50% w/v) was purchased from Sigma
Aldrich (St. Louis, USA). Acetone (C3H6O, 99.7%) was from Samchun
Chemicals (Gyeonggi-do, Korea). Nitric acid (HNO3, 60% w/v) was
purchased from Matsunoen Chemicals (Osaka, Japan).
Three different surface treatments of CF were initially implemented.
The CF was modified using acetone, nitric acid, or PEI. An untreated
carbon felt was also used as control. The CF samples that had an active
area of 2 × 2 cm2 were immersed in acetone (99.7% w/w), nitric acid
(10% v/v), or PEI (5 mg·mL−1) for 3 h at room temperature. Afterward,
all carbon felts were washed with deionized (DI) water until a neutral
M. Christwardana et al. Applied Energy 222 (2018) 369–382

pH was detected in the water. Finally, samples were vacuum-dried in an

oven at 80 °C for 12 h before being used for characterizations and
single-cell tests.

2.2. In vitro cultivation of yeast Saccharomyces cerevisiae

To estimate the growth of the biocatalyst, 0.7 mg·mL−1 of dried

Saccharomyces cerevisiae was cultivated in modified yeast extract, pep-
tone, and D-glucose (YPD) medium [49], which consisted of 5 mg·mL−1
of yeast extract, 2.5 mg·mL−1 of peptone, and 5 mg·mL−1 of D-glucose
in 0.1 M PBS (pH 7.4). The yeast culture was cultivated under anaerobic
conditions for three days at room temperature (23 ± 3 °C), while it
was stirred at 110 rpm. Every 6 h, a 2 mL sample of culture was col-
lected, centrifuged, dried and then weighed to obtain the dry biomass
during cultivation. The dry biomass was measured to calculate the yeast
growth rate (µ) according to [50]. The growth curve of yeast cultivated
in vitro was used as a control to estimate the behavior of yeast alone in
advance and the appropriate duration for later MFC tests.

2.3. MFC configuration

The performances of the different electrodes were tested in the

flowing-type MFC, which was produced by CNL Energy (Seoul, Korea).
It was made up of PVC end plates with flow channels, collector plates, a
bipolar plate, and a rubber gasket. It had an electrode active area of
4 cm2 and was assembled according to the design shown in Fig. 1.
Differently from standard fuel cell fed with gas mixtures, in which the
pattern of channel is important for reactant distribution [51,52], in our
work two liquid solutions are fed and the anode solution contains the
solid biocatalyst. Therefore, to avoid clogging, we preferred simple flat
plates to highlight the dependency of yeast cell’s immobilization from
the hydrophobic/hydrophilic nature of the surface of the felt.
The pretreated carbon felts were used as electrode at the anode side, Fig. 2. In vitro cultivation of S. cerevisiae: (a) the effect of growth time of yeast
while the cathode was untreated. Nafion 117 treated with 3% w/w on dry weight concentration; (b) the effects of growth time of yeast on both the
H2O2, 0.5 M H2SO4, and DI water was used for the membrane. The pH of the electrolyte and the dry weight concentration.
anode and cathode reservoirs were connected to a Schott’s glass bottle
(Mainz, Germany) and were respectively filled with 50 mL of yeast in
YPD medium as anolyte and 0.1 M PBS as catholyte. For the anode, the

Fig. 1. Flowing-type MFC: (a) single-cell components and assembly; (b) and (c) lateral and front views of assembled cell.

371
M. Christwardana et al.
Fig. 3. SEM images of four different CFs: (a) CF-Untreated; (b) CF-Acetone; (c) CF-Nitric Acid; and (d) CF-PEI.
glass bottle was jacketed to prevent temperature changes (23 ± 3 °C),
and anaerobic conditions were maintained during the process by cap-
ping the bottle, while open-air conditions were used for the cathode.
2.4. Electrochemical characterizations
A WonaTech Zive SP-2 potentiostat (Seoul, Korea) that was
equipped with a frequency response analyzer (FRA) was connected to a
computer and was used for electrochemical measurements. For mea-
suring EIS in half-cell tests, a Pt wire acted as counter-electrode, while a
saturated calomel electrode (SCE) (soaked in 3.0 M KCl) acted as re-
ference electrode. The corresponding felt was loaded on a glass kit,
acting as working electrode and YPD medium in 0.1 M PBS (pH 7.4)
(without yeast) was used as the electrolyte, while N2 gas was fed to
create anaerobic conditions. On the other side, for measuring the po-
larization curves of the MFC full cell, a potentiostat was connected to
the MFC cell device. Then, a centrifugal pump was used to flow anolyte
and catholyte from glass bottles into the anode and cathode reservoirs
with a flow rate of 15 mL·min−1. All experiments were conducted at
room temperature (23 ± 3 °C) within a certain period of time. For the
tests, substrate was not replaced, although replacement was needed to
maintain long-term operation. The MFC full-cell tests were done after
the specific growth curve of the S. cerevisiae was fixed and the OCV
evolution recorded. For evaluating the performance of MFC full cells,
current density-voltage (I-V) and power curves were also measured.
2.5. Optical and chemical characterizations
To investigate surface structure and morphology of untreated and
treated carbon felts, SEM (InspectF, FEI Co.) was used. The adhesion
372
Applied Energy 222 (2018) 369–382
and entrapment of the biocatalyst layer on carbon felt and Nafion 117
membrane after MFC full-cell tests was inspected by digital optical
microscopy (Xi-Cam, Bestechvision Co.). To evaluate their chemical
structure, type and distribution of chemical bonds on the surface of CFs,
before and after full-cell tests, XPS analysis (K-alpha+, Thermo
Scientific Co.) was conducted. The hydrophobic/hydrophilic behavior
of the treated felts and wettability were determined on square speci-
mens (2 × 2 cm2) using an on-line optical contact angle device (OCA,
Dataphysics) with a quartz stage, an adjustable support for alignment, a
CC camera with a 2× max optical zoom and images were processed
with an imaging analyzer (SCA 20, Dataphysics) to calculate the angle.
3. Results and discussion
3.1. In vitro yeast cultivation results
To accelerate the start-up time of an MFC with yeast, (i) the max-
imum growth time, (ii) the growth profile, and (iii) the replacement
time of exhausted substrate with fresh substrate for the yeast biocata-
lyst should be determined. To these ends, the growth of yeast biomass,
calculated as dry weight concentration, was evaluated as a function of
growing time in vitro (Fig. 2a). The average dry weight concentration
during the growing time is represented in Fig. 2b. To ensure reprodu-
cibility, the test was performed three times (Tubes 1–3).
According to Fig. 2, S. cerevisiae yeast generally grew well. As shown
in Fig. 2b, the growth rate of yeast was low for the first 6 h (lag phase),
and then it grew very fast until 12 h, with a growth rate of 1.083 h−1
(exponential growth phase). Its growth stayed almost constant for 30 h,
and then it decreased (death phase).
There are two noticeable findings about the yeast growth. First, the
M. Christwardana et al.
Fig. 4. EIS measurements: (a) Half-cell spectra; (b) Rct vs pH for the CF samples.
growth pattern of yeast (lag – exponential – stationary – death) pro-
ceeded in a standard way, and its growth rate was better than previous
results. For example, our yeast growth rate was 1.083 h−1, while that
reported by Klein et al. was 0.18 h−1 [53], that reported by Lee and Jin
was 0.25 h−1 [54], and that reported by Swinnen et al. was 0.08 h−1
[55]. The better yeast growth rate means that the YPD medium and
growth conditions were favorable for yeast growth. If needed, substrate
can be replaced at the optimal time determined by this growth profile,
prolonging the stationary growth phase and keeping the OCV stable.
From Fig. 2b, the duration of the stationary phase was ∼30 h, and dry
biomass started to decrease after 48 h due to the starvation of yeast and
depletion of substrate.
Second, the pH of the electrolyte plays a critical role in growth
because a low pH can either promote faster proton transfer from anode
to cathode due to the higher proton concentration or induce the bio-
logical death of the microorganisms because they are not able to survive
in a strongly acidic environment [56] and performances are limited
[57]. According to Fig. 2b, yeast grew well under a slightly basic or
neutral pH. When the pH was shifted from neutral to slightly acidic, in
the middle of the stationary phase, yeast biomass decreased slowly with
time and the pH value was not optimal for yeast. The change in pH from
neutral to acidic was attributed to the depletion of glucose and the
increasing concentration of protons or of other acidic species, as re-
ported in previous studies [58].
Table 1
Parameters obtained from Nyquist plots of the CF samples.
Sample Rs (Ω·cm2) Rct (Ω·cm2)
CF-Untreated 27.968 ± 0.066 1.984
CF-Acetone 29.603 ± 0.056 1.234
CF-Nitric acid 29.837 ± 0.020 0.879
CF-PEI 29.210 ± 0.086 1.603
Applied Energy 222 (2018) 369–382
From these two findings, we think that the increase in OCV was
limited during the lag phase (first 8 h), and then it rapidly increased
during the exponential growth phase (next 12 h). Afterward, the OCV
was stable during the stationary phase until it started to decline after
approximately 48 h, corresponding to the yeast starvation time.
3.2. Microstructural, surface, and electrochemical characterization of
treated CFs
Since the OCV evolution is closely related to the metabolism of
yeast, the preliminary characterization of the yeast growth curve is
required to understand how OCV and MFC performance are affected by
yeast growth and to ensure the yeast is always active during the MFC
tests, even without the replacement of substrate or medium. Similarly,
the properties and behavior of the pretreated CFs alone, without yeast,
must be characterized to measure the effectiveness of the applied pre-
treatments. Hence, after revealing the growth characteristics of yeast,
how CF fibers are affected by the three different chemical treatments
was investigated. We denoted the CF samples as CF-Untreated (control
sample), CF-Acetone, CF-Nitric Acid and CF-PEI. Initially, the effects of
these treatments on the microstructure of CF were examined using SEM
micrographs (Fig. 3).
The CF materials were essentially an entangled web of carbon fibers
held together by a self-standing textile-like structure. From these SEM
images, noticeable modifications between the four different CF fibers
were observed. The untreated CF fiber (Fig. 3a) had a plain surface. In
contrast, the CF fibers treated by acetone or nitric acid showed some
eroded traces on their surface (Fig. 3b and c), while the CF fiber treated
by PEI was mostly covered by a PEI layer (Fig. 3d).
Fig. 3 shows that nitric acid and acetone physically eroded the CF
fibers, causing their internal layers to be exposed. Such erosion is prone
to induce (i) elimination of impurities or inhomogeneities of pristine CF
fibers and (ii) an increase in the number of potential weak-interaction
sites [47,49]. In contrast, in the CF fiber immersed in PEI solution,
additional PEI layers are deposited, and such layers play a role in en-
hancing the physical entrapment or electrostatic interactions between
CF, yeast and substrate [59,60].
These qualitative hypotheses were confirmed quantitatively by
graphing half-cell Nyquist plots using EIS. Fig. 4a shows the Nyquist
plots for all CF samples. The measurements were performed three times
in a frequency range of 1 MHz to 0.1 Hz, an AC amplitude perturbation
of 10 mV, in YPD medium (without yeast). The equivalent Randle’s
circuit used is showed in Fig. 4a.
All spectra showed semi-circles representing the charge transfer
resistance (Rct) and the pseudo-capacitance (C) contribution. Here, the
X axis intercept indicates the resistance of solution (Rs). The complete
list of these values is reported in Table 1.
In the YPD medium, Rs was always approximately 28–30 Ω·cm2.
This narrow ohmic resistance variation was random and not associated
to a specific property of felts, and therefore might be neglected. The
charge transfer resistance was more important and was ascribed to the
intrinsic resistance of pretreated felts against the transfer of electrons
from YPD medium to the solid electrode. As there was no yeast, no
electrochemical reactions occurred at anode and only electrostatic
phenomena took place (e.g. deposition of glucose on fibers). These
phenomena were optically inspected after EIS and the samples
C (F·cm−2) W (s5·Ω−1·cm−2)
± 0.018 1.63 · 10−5 ± 6.06 · 10−6 0.017 ± 0.001
± 0.042 3.07 · 10−5 ± 2.62 · 10−6 0.016 ± 0.001
± 0.021 4.86 · 10−5 ± 3.08 · 10−6 0.019 ± 0.006
± 0.103 1.69 · 10−5 ± 3.11 · 10−7 0.018 ± 0.002
373
M. Christwardana et al.
Fig. 5. Optical contact angle images of four different CFs: (a) CF-Untreated; (b) CF-Acetone; (c) CF-Nitric Acid; and (d) CF-PEI.
contacting the YPD medium were inspected by SEM. The micrographs
are presented in Supplementary Fig. S2. In these photographs, the
physical interaction between the surface of samples and glucose was
observed; giving additional evidence that different weak interactions
occurred. Furthermore, all the samples showed the Warburg (W) ele-
ment, i.e. unrestricted diffusion to a planar electrode due to the inert
nitrogen in the system and the absence of a chemical process. These
evidences confirm that the EIS measurements were able to determine
the physical interaction at the felts.
According to Fig. 4a, the Rct of the CF-Nitric Acid sample was
lowest, while that of the CF-Untreated sample was highest. The dif-
ference in Rct was due to the different chemical treatments. This ex-
periment yielded two interesting observations. First, all treatments re-
duced the Rct of CF. This was probably attributable to the generic
removal of matter, such as the elimination of any impurities or other
residues present on the surface of the as-received CF. For this reason,
the carbon sites on the fiber surface were free to undergo further che-
mical modifications caused by the introduction of oxygen atoms or
other charged groups, depending on the type of treatment [47,48].
Second, the trend in Rct agrees well with the pH of the treating solution
(Fig. 4b). Specifically, the Rct decreased as the pH of the treating so-
lution became more acidic because when a carbon fiber contacts a
strong acidic solution, the fiber of CF is chemically attacked by the
solution, and then its surface is modified with an increase in surface
area and the addition of functionalized oxygen bonds [49,61]. How-
ever, if the exposure to an aggressive solution is too long, the CF surface
erosion may result in a rupture of carbon fibers, leading to the dete-
rioration of the CF pore structure and the formation of large holes in-
side the CF. In this case, the CF is damaged so that it cannot be used in
MFCs. This partial erosion due to the low pH of the treating solution is
visible in Fig. 3b and c.
Another importance consequence of the pretreatments is the change
in hydrophobic/hydrophilic behavior of the CF surface. The behavior of
the corresponding CF samples can be observed by contact angle mea-
surements and the results are presented in Fig. 5. This behavior is
374
Applied Energy 222 (2018) 369–382
critical for establishing a contact between substrate, yeast and electrode
and evaluating whether the surface of CF affect the adhesion and im-
mobilization of yeast’s cells. The results obtained from the calculation
of contact angle clearly demonstrated that CF-Untreated and CF-
Acetone had a very high contact angle, > 90° (hydrophobic) while CF-
Nitric Acid was partially hydrophilic with an incomplete wetting be-
havior and a contact angle close to 90°. The best wettability was ob-
tained by the CF-PEI sample that showed complete wetting and hy-
drophilic behavior.
These results suggest that the CF surface is affected by the pre-
treatment. To further evaluate effects of the CF treatments on the
chemical structure of CF surface, XPS analysis was performed.
The XPS spectra of C 1s, N 1s and O 1s elements for all the samples
are represented in Fig. 6. According to Fig. 6, the XPS data clearly
confirmed that the different treatments led to noticeable effects on the
surface of the CF. First, oxygen bonds increased from 5.1% in the un-
treated CF to 15.7% in the acetone- and nitric acid-treated CFs. On the
other hand, PEI treatment slightly increased the oxygen sites from 5.1%
to 9.2%. Second, regarding carbon and nitrogen bonds, CF-PEI had a
very dissimilar C 1s spectrum compared to the remaining samples. This
was because PEI contains nitrogen-based bonds, so new nitrogen sites
were observed on surface of this CF. In fact, in samples that were
treated by acetone or nitric acid, the N 1s spectra were flat, even though
nitric acid has the nitrate group –NO3−. The possible reason for these
results is that PEI is made up of abundant amino groups, and during the
immersion treatment, they attached to the carbon fiber, forming the
functional layer on the surface of the CF, as shown in Fig. 3d. This
mechanism is completely different from the activation by erosion due to
acid attack in the nitric acid and acetone cases.
To distinguish the specific chemical bonds, the spectra of Fig. 6 were
deconvoluted and fitted by Lorentz/Gauss functions. The deconvoluted
spectra for all the samples are presented in Fig. 7 and the peak height
data are summarized in the supplementary section (Tables S1–S4).
According to these calculations, the spectra of CF-Untreated, CF-
Acetone and CF-Nitric Acid were qualitatively similar: the dominating
M. Christwardana et al.
Fig. 6. Comparison of XPS spectra for treated CF: (a) C 1s spectrum; (b) O 1s
spectrum; and (c) N 1s spectrum.
bond was CeC, with some secondary CeOH bonds with very few C]O.
In addition, in CF-Acetone and CF-Nitric Acid, the peak height and peak
area in O 1s for CeOH and C]O bonds increased. These results show
that the hydrophobic behavior of untreated felt was turned into less
hydrophobic behavior with the acidic treatments because of the two
lone electron pairs on O atoms.
This result is linked to MFC performance because the distribution of
YPD aqueous medium in the anode was better and more surface area
was effectively used for YPD-Yeast-electrode electron transfer in the CF
375
Applied Energy 222 (2018) 369–382
[31,47].
In contrast, the deconvoluted spectra of CF-PEI were different from
those of other samples due to two new chemical bonds observed in the
N 1s spectra. Namely, the N 1s spectrum was deconvoluted into two
major peaks: CeNH2, i.e., primary amine, and CeNR2, i.e., secondary
or tertiary amine, in which R2 = C, C or C, H. The O 1s spectrum of CF-
PEI was similar to that of pristine CF with a small increase in C]O
bonds, probably due to the generic cleaning effect due to the immersion
in a treating solution. The chemical structure of PEI (see Fig. S1) has
many branched arms, including the primary amines as terminal groups
and the backbone of secondary amines, whereas there are few tertiary
amines. From the deconvolution result, these peaks appeared in the N
1s spectrum of CF-PEI, indicating that the treatment using PEI was ef-
fective to introduce different types of N-groups on the CF surface. The
presence of these groups was critical because CF is hydrophobic, and
yeast is a facultative anaerobe. Therefore, the adhesion, good attach-
ment and electron transfer between yeast and CF in an anaerobic en-
vironment are not always easy without a mediator [38,39]. However, if
the physical behavior of CF is more hydrophilic, the contact between
yeast, YPD medium and the entire CF surface can be increased. In that
case, it will be possible for yeast to secrete extracellular polymeric
substances (EPS) such as rDNA, proteins and polysaccharides to attach
them to the support (in this case, CF), attaching a biocatalyst layer on
the surface [62,63]. The EPS already had a large number of peptide
bonds and other CeN bonds, and this is why the presence of N-groups
on carbon fibers potentially increased the biocompatibility of CF-PEI
and induced a better growth than CF-Acetone and CF-Nitric Acid [64].
The optical results of Figs. 3 and 5 also support these observations.
3.3. Performance of MFC employing S. cerevisiae as biocatalyst with treated
felts
After the preliminary characterizations of a yeast biocatalyst and
the four different CFs, the performance of the MFCs using both the yeast
biocatalyst and the four different CF electrodes was investigated by
measuring their OCV and polarization curves. We measured the OCV
and polarization curves three times per case to guarantee the re-
producibility of the MFC tests. Initially, the effect of time on the OCV of
all the MFCs was measured, and the results are reported in Fig. 8. In
Fig. 8a–d, there are three noticeable results. First, the OCV of the MFC
employing CF-PEI was highest; the OCV of the MFCs using CF-Un-
treated, CF-Acetone, CF-Nitric Acid, and CF-PEI was 0.297 ± 0.029,
0.277 ± 0.036, 0.342 ± 0.006, and 0.362 ± 0.036 V, respectively.
The treatments using nitric acid and PEI changed pristine CF to hy-
drophilic, so that physical bonds such as hydrogen bonds and Van der
Waals forces occurred easily [65,66]. Moreover, there were electro-
static interactions and physical entrapment between the CF-PEI and
polysaccharides of the yeast that improved the CF-PEI performance due
to the generally opposite charges between PEI groups on the carbon felt
and the yeast cells [45,49,60,67]. As a result, the electrons produced by
yeast were promptly transferred to the electrode by these bonds.
Compared to CF-PEI, the CF-Nitric Acid performance was slightly
lower because nitric acid treatment caused erosion in the CF, as shown
in Fig. 3c, and due to the absence of abundant C-N bonds on the anode,
decreasing the performance of the MFC. Second, the OCV margin of the
MFC using CF-Acetone was widest. This can be attributed to the mild
aggressiveness of acetone. Therefore, the surface of the CF was not
uniformly eroded, causing its performance and stability to vary. The
MFC using CF-Nitric Acid had the minimum error in OCV, although
nitric acid showed the most aggressive behavior in terms of pH. Unlike
the acetone treatment, the use of nitric acid made CF more homo-
genous, i.e., all portions of the carbon felt had similar surface properties
and conductivity. Certainly, the adaptation and growth of yeast are
influenced by various external conditions, and any such effects play an
important role in the reproducibility of the MFC’s performance, espe-
cially during the start-up phase of the MFC tests, and this is the most
M. Christwardana et al.
Fig. 7. XPS deconvolution of C 1s, O 1s and N 1s for all treated CFs.
common source of uncertainty in the MFC tests. However, biocompat-
ibility and surface interactions between CF and medium can be con-
trolled by these pretreatments to alleviate this uncertainty. The third
notable aspect of Fig. 7 is the change in OCV during the first 10–20 h of
operation. The OCV curve also represents the growth of yeast during
the MFC test because the number of electrons affecting the OCV is
proportional to the number of cultivated yeast cells attached to the felt
for the direct electron transfer.
The OCV of the MFCs using differently treated CFs was measured,
and it was found that the MFCs had different lag phase patterns. In the
case of the MFC using CF-PEI, two out of three MFC tests had lag phase
time of 8–10 h, and then the OCV increased, indicating that the yeast
started to grow accordingly to growth curve in Fig. 2, they formed a
diffused biocatalyst layer, and more electrons were generated. On the
other hand, in the MFC using CF-Nitric Acid, two out of three MFC tests
had the shortest lag phase time of 3–7 h, and then the OCV significantly
increased, as expected. This means that the yeast easily adapted to the
environment and formed an active biocatalyst layer rapidly, generating
more electrons. As previously described, the hydrophilicity of CF-PEI
376
Applied Energy 222 (2018) 369–382
and CF-Nitric Acid allowed for the yeast to adhere more easily, and in
that environment, the yeast grew fast. Compared to CF-Nitric Acid, the
lag time of CF-PEI was slower because more time was needed for the
yeast to adapt to an environment where both physical entrapment
(between PEI and polysaccharides of yeast) and electrostatic interac-
tions (between CF-PEI and yeast cells) occurred. However, the slower
lag time of CF-PEI did not preclude a high OCV, stable operation and
fast start-up, while these did not occur at all in the CF-Nitric Acid MFC
because even though the lag time was faster with CF-Nitric Acid, the
OCV overshoot right after the short lag time (Fig. 7c) was not a stable
condition, probably due to missing N bonds on fiber surface from the
beginning. In the MFC using CF-Untreated and the MFC using CF-
Acetone, two out of three MFC tests had the longest lag phase time of
13–14 h. These results imply that the yeast needed a longer time to
adapt to the environment because CF-Untreated and CF-Acetone are not
hydrophilic and don’t have nitrogen groups (needed to form the bio-
catalyst layer). The MPD of the MFCs was also measured from the po-
larization curves, and the results are shown in Fig. 9.
According to these results, the MPD of the MFC using CF-PEI is
M. Christwardana et al.
Fig. 8. Open circuit voltage (OCV) of yeast MFC adopting: (a) CF-Untreated; (b) CF-Acetone; (c) CF-Nitric Acid; and (d) CF-PEI.
256.3 ± 11.5 mW·m−2, while that of the MFCs using CF-Untreated,
CF-Acetone, and CF-Nitric Acid is 123 ± 6.8, 68 ± 7.21, and
217 ± 8.5 mW·m−2, respectively. By adoption of CF-PEI, the perfor-
mance of the MFC increases 108% compared to that of the MFC using
CF-Untreated within a short time period, saving considerable start-up
time. These MPDs are comparable with the previously reported ones
(Table 2).
The MPD of a MFC using yeast immobilized on a reticulated vitreous
carbon electrode with a neutral red mediator [69] was 100 mW·m−2,
while the value in another MFC using yeast immobilized on a glassy
carbon electrode with a resourfin mediator was 155 mW·m−2 [70]. The
comparison in Table 2 shows that there are not only many different
treatments, as proposed in literature, but also many different config-
urations for testing them. As a result, it is found that an exact com-
parison is difficult but the relative effectiveness of the treatments on
MPD can be compared. Although many benchmarks claim higher MPD,
it is due to the convenient use of ferricyanide as terminal electron ac-
ceptor (TEA), instead of the oxygen in air, to facilitate cathodic process.
In fact, in these cases, the relative increase of MPD is lower than that
obtained in this work. For instance, the good results obtained by Hi-
dalgo et al. were ascribed to the anode pre-treatment as well as the use
of methylene blue as mediator and a very high concentration of glucose
(60 g·L−1) and yeast (50 g·L−1). In spite of that, after the operation for
two days, its MPD rapidly decreased, proving that the effectiveness
cannot be maintained and the growth conditions might be optimized.
Similarly, Liang et al. used a flowing type MFC with the uses of acetate,
ferricyanide and an untreated stainless steel anode as reference. Ac-
cording to the results, the extremely high increase in MPD is due to a
poor baseline performance of the reference steel electrode.
377
Applied Energy 222 (2018) 369–382
Here, it is worth mentioning that no mediators and no substrate
replacement were considered in our MFC tests. This means that in a
very short start-up time of 48 h, a marked increase in MPD was
achieved in our system adopting CF pretreatments, and this result is
better than those of other studies.
Finally, a brief chemical explanation of these results is given by XPS
deconvolution of the CF samples after full-cell tests in Fig. 10.
As can be observed in Fig. 10, the types of chemical bonds did not
change, but a different distribution was noticed. After full-cell tests,
even in CF-Untreated, many CeN bonds were detected due to the
presence of the entrapped yeast biocatalyst. Specifically, peptide
(O]CeN), primary amine (CeNH2) and secondary/tertiary amine
bonds (CeNR2) were identified. Since the N 1s spectrum of CF-Un-
treated before tests in Fig. 6 shows no evidence of N groups, the bonds
identified in Fig. 10 can be attributed to the EPS mentioned above, as a
proof of the biocatalyst activity and adaptation. However, the N 1s
spectrum of CF-PEI showed a small difference in the distribution of
amine groups because the primary amine peak was missing. This means
that all CeN functional groups in the systems were employed in other
bonds, and this is evidence of the higher number of interactions at-
tained by CF-PEI between yeast cells and the electrode surface.
The origin of OCV in a MFC when a mixed substrate, deriving from
wastewater or food waste, is used in MFCs, the description of the
chemical reactions, especially at anode side, is not obvious because
many compounds can be degraded and a deep chemical analysis is
needed. In this work, with the yeast as biocatalyst, the glucose as
substrate, and the oxygen from air as TEA, the origin of OCV can be
described with the following anodic and cathodic reactions:
C6 H12 O6 + 6H2 O→ 6CO2 + 24H+ + 24e− (1)
M. Christwardana et al. Applied Energy 222 (2018) 369–382

Fig. 9. Polarization and power density curves of yeast MFC adopting: (a) CF-Untreated; (b) CF-Acetone; (c) CF-Nitric Acid; d) and CF-PEI.

Table 2
Comparison of anode treatments with literature benchmarks.
Anode material/treatment Cathode material/TEAa Substrate/Biocatalyst MFC type, mediator (Y/N) MPD (mW·m−2) Relative increase in MPD Ref.
(%)

Carbon felt/PEI Carbon felt/aerated PBS Glucose/S. cerevisiae Double chamber, flowing- 256.3 ± 11.5 108.3 This work
type, N
Graphite plate/thionine Graphite plate/aerated Glucose/S. cerevisiae Double chamber H-type, Y 60 106.9 [68]
PBS
Carbon felt/nitric acid + PANI Carbon paper/ Glucose/S. cerevisiae Double chamber PEM-type, Y 460 187.5 [48]
ferricyanide
Carbon felt/nano nickel Carbon felt/ferricyanide Fructose/C. Double chamber tube-type, N 700 189.2 [43]
melibiosica
Carbon paper/nano gold Carbon paper/ Acetate/wastewater Double chamber, H-type, N 461.6 27.7 [27]
ferricyanide
Activated carbon/polydopamine Activated carbon/air (O2) Acetate/wastewater Single chamber, cube-type, N 803 ± 6 30.8 [59]
Stainless steel/rGOb + MnO2 Graphite brush/ Acetate/mixedc Double chamber, Flowing 1045 2171 [29]
ferricyanide type, N

a
Terminal electron acceptor.
b
Reduced graphene oxide.
c
Not specified inoculum source from a 6-months culture.

1 For our scope, in a separated experiment, in the same conditions of

O2 + 2H+ + 2e− → 2H2 O
2 (2)
The net reaction, after balance, is the complete oxidation of one
molecule of glucose. The reaction (1) is a typical anaerobic hydrolysis
done by the yeast as a result of its catabolic activity, although this re-
action is the sum of several cascade reactions occurred by the metabolic
cycles inside yeast cells. We did not treat with the description of yeast’s
metabolic pathways because they are beyond the scope of this work and
the two basic reactions above model the interaction between modified
electrode, yeast population and the electrochemical process inside the
MFCs.
in vitro cultivation tests, we evaluated the increase of dry biomass at-
tached to the pretreated felts (2 × 2 × 0.3 cm3) after 2 days to quantify
the amount of entrapped yeast related to each pretreatment and the
results are showed in Table 3.
The biomass of immobilized yeast increased in the pretreated felts
and showed that a pretreatment is needed to improve the properties of
untreated felt. This trend is coherent with contact angle measurements,
XPS analysis and MPD values confirming that wettability, entrapment
and surface bonds of felt are strictly correlated in yeast-MFC perfor-
mance when mediators are not used. Moreover for the CF-Untreated

378
M. Christwardana et al. Applied Energy 222 (2018) 369–382

Fig. 10. XPS deconvolution of C 1s and N 1s for CF-Untreated and CF-PEI.

Table 3 electrode to correlate OCV, pretreatment effectiveness and yeast po-

Amount of entrapped yeast on pretreated felts after 2 days. pulation increase. These potentials are showed in Fig. 11.
Sample Initial dry Final dry Dry yeast Entrapped yeast
The better entrapment and growth of yeast cells on CF-PEI is the
weight (mg) weight (mg) biomass (mg) (%) cause of the gradual increase of anode potential, leading to a larger
OCV with respect to the untreated felt used as control. The origin of
CF-Untreated 178 223 45 25.3 OCV is therefore ascribed to the formation of a proper yeast layer on the
CF-Acetone 176 229 53 30.1
CF-Nitric Acid 155 214 59 38.1
treated felt because the potential at cathode is almost constant and does
CF-PEI 177 245 68 38.4 not affect too much OCV.
Proper formation of a yeast layer is important in the early stage of
an MFC because it is the place where electron transfer occurs. To in-
spect the yeast layer formed on the electrode surface during the MFC
tests, micrographs of carbon felt electrodes were taken, and the images
are reported in Fig. 12.
Fig. 12a–d shows clear evidence that the interaction between fibers,
yeast, and substrate formed a thin layer (or biofilm) between fibers of
carbon felts. In CF-Untreated and CF-Acetone (Fig. 12a–b), the formed
film was poorly distributed between single fibers, with a just-sufficient
coverage, and was not homogenous. In contrast, in CF-Nitric Acid, the
layers formed were smaller but more homogenous and were well dis-
tributed, as shown in Fig. 10c. We hypothesized that the semi-hydro-
philic nature of CF-Nitric Acid, demonstrated in the contact angle
measurement, contributed to this result. In contrast to other samples, in
CF-PEI, the yeast layers formed were mostly well distributed. As dis-
cussed above, this behavior can allow for better contact and transfer
between anolytes, yeast cells and fibers due to the more hydrophilic
nature of the material, the establishment of weak chemical interactions
such as hydrogen or van der Waals bonds and physical immobilization.
Detailed images of Fig. 12 are also presented in Supplementary
Fig. 11. Anode potentials of CF-Untreated and CF-PEI vs time. Information (Fig. S3).

and CF-PEI, as representative samples, we monitored also the evolution

of anode and cathode potentials by employing a reference SCE

379
M. Christwardana et al.
Fig. 12. DOM (500×) of treated CFs in contact with yeast and YPD medium after MFC tests: (a) CF-Untreated; (b) CF-Acetone; (c) CF-Nitric Acid; (d) and CF-PEI.
4. Conclusions
The use of yeasts in microbial fuel cells as biocatalyst is a stimu-
lating topic because many advantages are established but many chal-
lenges still remains on the profitability of their use in real applications.
From an industrial point of view, yeast Saccharomyces cerevisiae is
commercially available, cheap, easy to cultivate and stock but the
growing conditions need to be optimized, according to the medium
employed. The major issue is to reduce the needs for an external
mediator and improve the interaction between yeast and supporting
electrode, mainly carbon felts, with an effective surface modification
pretreatment because in large scale systems the amount of yeast im-
mobilized directly on the anode is critical to obtain high voltage and
efficient electron transfer.
In this work a new flowing-type MFC system using a yeast-based
biocatalyst and modified CF as anodic electrode was established for
future applications and development in the valorization of wastewater
and food waste containing sugars and their derivatives. To establish
basic performance on model substrate, experiments using differently
pretreated CFs were designed, and their performance was evaluated
with S. cerevisiae yeast as the biocatalyst, and glucose as the substrate.
The work includes electrode treatments, characterization, single cell
tests and post-analysis of used anodes to provide a complete overview
of the implications and real application of the suggested pretreatments.
For the pretreatment of carbon felts, acetone, nitric acid, and
polyethyleneimine were used. As an initial step, the result of the in vitro
cultivation tests showed that a high yeast growth rate (1.083 h−1) for
cell tests were obtained in a growing time of 48 h. The X-ray photo-
electron and electrochemical impedance spectroscopy techniques re-
vealed that conductivity, charge transfer resistance, and C]O and CeN
groups dangled on the carbon felts were crucial parameters determining
the performance of the microbial fuel cell. Particularly, the conjugation
effects of π-π bonds and lone pairs promoted (i) the mutual adhesion
between them and (ii) the growth of yeast on CF-PEI. This correlation
was confirmed by optical analysis, wettability measurements and yeast
dry biomass growth experiment on the carbon felts disentangled after
380
Applied Energy 222 (2018) 369–382
the cell tests. To evaluate the early-stage performance of the microbial
fuel cells using the different carbon felts, polarization curves were
plotted. The maximum power density of the MFCs depended on the
superficial state of the CFs, while the performance of the MFC using the
PEI-treated carbon felt was best as 256.3 ± 11.5 mW·m−2.
In conclusion, this work reported the qualitative and quantitative
effectiveness of three different pretreatments by using several techni-
ques and tests. These findings were compared and explained with other
results attained in previous studies on carbon felts and we believe that
the CF-PEI treatment and results reported in this work could be of in-
terest for future developing of lab and pilot scale microbial fuel cells
employing yeasts or mixed substrate from wastewater or food waste
containing non-edible and non-reusable sugars or derivatives.
Declaration of interest
The authors declare no conflicts of interests.
Acknowledgements
Dr. Domenico Frattini was supported by the Korea Research
Fellowship through the National Research Foundation of Korea (NRF)
funded by the Ministry of Science and ICT of the Republic of Korea (No.
2017H1D3A1A01013887) and by the Korea Institute of Energy
Technology Evaluation and Planning (KETEP) and the Ministry of
Trade, Industry, and Energy (MOTIE) of the Republic of Korea (No.
20164030201060). This work was also supported by the In-house
Program (2E28272) of the Korea Institute of Science & Technology
(KIST), Republic of Korea.
Appendix A. Supplementary material
Supplementary data associated with this article can be found, in the
online version, at http://dx.doi.org/10.1016/j.apenergy.2018.03.193.
M. Christwardana et al.
References
[1] Rabaey K, Verstraete W. Microbial fuel cells: novel biotechnology for energy gen-
eration. Trends Biotechnol 2005;23:291–8. http://dx.doi.org/10.1016/j.tibtech.
2005.04.008.
[2] Meric S, Selcuk H, Onat B, Ongen A. Sustainable technologies for recycling and
reuse: an overview. Environ Sci Pollut Res 2018;25:2993–5. http://dx.doi.org/10.
1007/s11356-017-0770-z.
[3] Pandey P, Shinde VN, Deopurkar RL, Kale SP, Patil SA, Pant D. Recent advances in
the use of different substrates in microbial fuel cells toward wastewater treatment
and simultaneous energy recovery. Appl Energy 2016;168:706–23. http://dx.doi.
org/10.1016/j.apenergy.2016.01.056.
[4] Roustazadeh Sheikhyousefi P, Nasr Esfahany M, Colombo A, Franzetti A, Trasatti
SP, Cristiani P. Investigation of different configurations of microbial fuel cells for
the treatment of oilfield produced water. Appl Energy 2017;192:457–65. http://dx.
doi.org/10.1016/j.apenergy.2016.10.057.
[5] Rodrigo MA, Cañizares P, Lobato J, Paz R, Sáez C, Linares JJ. Production of elec-
tricity from the treatment of urban waste water using a microbial fuel cell. J Power
Sources 2007;169:198–204. http://dx.doi.org/10.1016/j.jpowsour.2007.01.054.
[6] Behera M, Jana PS, More TT, Ghangrekar MM. Rice mill wastewater treatment in
microbial fuel cells fabricated using proton exchange membrane and earthen pot at
different pH. Bioelectrochemistry 2010;79:228–33. http://dx.doi.org/10.1016/j.
bioelechem.2010.06.002.
[7] Prabowo AK, Tiarasukma AP, Christwardana M, Ariyanti D. Microbial fuel cells for
simultaneous electricity generation and organic degradation from slaughterhouse
wastewater. Int J Renew Energy Dev 2016;5:107–12. http://dx.doi.org/10.14710/
ijred.5.2.107-112.
[8] Jeon Y, Park CH, Kim S. Electricity generation from swine wastewater in media-
torless single-chamber microbial fuel cells. Bull Korean Chem Soc
2016;37:1148–51. http://dx.doi.org/10.1002/bkcs.10821.
[9] Lai C-Y, Wu C-H, Meng C-T, Lin C-W. Decolorization of azo dye and generation of
electricity by microbial fuel cell with laccase-producing white-rot fungus on
cathode. Appl Energy 2017;188:392–8. http://dx.doi.org/10.1016/j.apenergy.
2016.12.044.
[10] Frattini D, Falcucci G, Minutillo M, Ferone C, Cioffi R, Jannelli E. On the effect of
different configurations in air-cathode MFCs fed by composite food waste for energy
harvesting. Chem Eng Trans 2016;49:85–90. http://dx.doi.org/10.3303/
CET1649015.
[11] Najafgholi Z, Rahimnejad M. Improvement of sediment microbial fuel cell perfor-
mance by application of sun light and biocathode. Korean J Chem Eng
2016;33:154–8. http://dx.doi.org/10.1007/s11814-015-0123-x.
[12] Han HX, Shi C, Yuan L, Sheng GP. Enhancement of methyl orange degradation and
power generation in a photoelectrocatalytic microbial fuel cell. Appl Energy
2017;204:382–9. http://dx.doi.org/10.1016/j.apenergy.2017.07.032.
[13] Xu L, Wang B, Liu X, Yu W, Zhao Y. Maximizing the energy harvest from a microbial
fuel cell embedded in a constructed wetland. Appl Energy 2018;214:83–91. http://
dx.doi.org/10.1016/j.apenergy.2018.01.071.
[14] Christwardana M, Ji J, Chung Y, Kwon Y. Highly sensitive glucose biosensor using
new glucose oxidase based biocatalyst. Korean J Chem Eng 2017;34:2916–21.
http://dx.doi.org/10.1007/s11814-017-0224-9.
[15] Christwardana M, Chung Y, Kwon Y. A correlation of results measured by cyclic
voltammogram and impedance spectroscopy in glucose oxidase based biocatalysts.
Korean J Chem Eng 2017;34:3009–16. http://dx.doi.org/10.1007/s11814-017-
0213-z.
[16] Zhen G, Kobayashi T, Lu X, Kumar G, Hu Y, Bakonyi P, et al. Recovery of biohy-
drogen in a single-chamber microbial electrohydrogenesis cell using liquid fraction
of pressed municipal solid waste (LPW) as substrate. Int J Hydrogen Energy
2016;41:17896–906. http://dx.doi.org/10.1016/j.ijhydene.2016.07.112.
[17] Sun R, Xing D, Jia J, Liu Q, Zhou A, Bai S, et al. Optimization of high-solid waste
activated sludge concentration for hydrogen production in microbial electrolysis
cells and microbial community diversity analysis. Int J Hydrogen Energy
2014;39:19912–20. http://dx.doi.org/10.1016/j.ijhydene.2014.09.163.
[18] Frenkel VS, Cummings GA, Maillacheruvu KY, Tang WZ. Food-processing wastes.
Water Environ Res 2017;89:1360–83. http://dx.doi.org/10.2175/
106143017X15023776270368.
[19] Beyene HD, Werkneh AA, Ambaye TG. Current updates on waste to energy (WtE)
technologies: a review. Renew Energy Focus 2018;24:1–11. http://dx.doi.org/10.
1016/j.ref.2017.11.001.
[20] Bi Z, Zhang J, Zhu Z, Liang Y, Wiltowski T. Generating biocrude from partially
defatted Cryptococcus curvatus yeast residues through catalytic hydrothermal li-
quefaction. Appl Energy 2017;209:435–44. http://dx.doi.org/10.1016/j.apenergy.
2017.11.031.
[21] Weng HL, Lee DJ. Performance of sulfate reducing bacteria-microbial fuel cells:
reproducibility. J Taiwan Inst Chem Eng 2015;56:148–53. http://dx.doi.org/10.
1016/j.jtice.2015.04.028.
[22] Yang W, Kim K-Y, Saikaly PE, Logan BE. The impact of new cathode materials
relative to baseline performance of microbial fuel cells all with the same archi-
tecture and solution chemistry. Energy Environ Sci 2017;10:1025–33. http://dx.
doi.org/10.1039/C7EE00910K.
[23] Trapero JR, Horcajada L, Linares JJ, Lobato J. Is microbial fuel cell technology
ready? An economic answer towards industrial commercialization. Appl Energy
2017;185:698–707. http://dx.doi.org/10.1016/j.apenergy.2016.10.109.
[24] Frattini D, Accardo G, Ferone C, Cioffi R. Fabrication and characterization of gra-
phite-cement composites for microbial fuel cells applications. Mater Res Bull
2017;88:188–99. http://dx.doi.org/10.1016/j.materresbull.2016.12.037.
381
Applied Energy 222 (2018) 369–382
[25] Gude VG. Wastewater treatment in microbial fuel cells – an overview. J Clean Prod
2016;122:287–307. http://dx.doi.org/10.1016/j.jclepro.2016.02.022.
[26] Rodrigo MA, Cañizares P, García H, Linares JJ, Lobato J. Study of the acclimation
stage and of the effect of the biodegradability on the performance of a microbial
fuel cell. Bioresour Technol 2009;100:4704–10. http://dx.doi.org/10.1016/j.
biortech.2009.04.073.
[27] Alatraktchi FA, Zhang Y, Angelidaki I. Nanomodification of the electrodes in mi-
crobial fuel cell: impact of nanoparticle density on electricity production and mi-
crobial community. Appl Energy 2014;116:216–22. http://dx.doi.org/10.1016/j.
apenergy.2013.11.058.
[28] LaBarge N, Yilmazel YD, Hong PY, Logan BE. Effect of pre-acclimation of granular
activated carbon on microbial electrolysis cell startup and performance.
Bioelectrochemistry 2017;113:20–5. http://dx.doi.org/10.1016/j.bioelechem.
2016.08.003.
[29] Liang P, Zhang C, Jiang Y, Bian Y, Zhang H, Sun X, et al. Performance enhancement
of microbial fuel cell by applying transient-state regulation. Appl Energy
2017;185:582–8. http://dx.doi.org/10.1016/j.apenergy.2016.10.130.
[30] Zhang L, Zhu X, Li J, Liao Q, Ye D. Biofilm formation and electricity generation of a
microbial fuel cell started up under different external resistances. J Power Sources
2011;196:6029–35. http://dx.doi.org/10.1016/j.jpowsour.2011.04.013.
[31] Zhang L, Li J, Zhu X, Ye D, Fu Q, Liao Q. Startup performance and anodic biofilm
distribution in continuous-flow microbial fuel cells with serpentine flow fields: ef-
fects of external resistance. Ind Eng Chem Res 2017:acs.iecr.6b04619. http://doi.
org/10.1021/acs.iecr.6b04619.
[32] Paitier A, Godain A, Lyon D, Haddour N, Vogel TM, Monier JM. Microbial fuel cell
anodic microbial population dynamics during MFC start-up. Biosens Bioelectron
2017;92:357–63. http://dx.doi.org/10.1016/j.bios.2016.10.096.
[33] Hubenova Y, Mitov M. Potential application of Candida melibiosica in biofuel cells.
Bioelectrochemistry 2010;78:57–61. http://dx.doi.org/10.1016/j.bioelechem.
2009.07.005.
[34] Qi G, Zhang H, Huang C, Guo H, Xiong L, Wang C, et al. Liquefaction and char-
acterization of residue of oleaginous yeast in polyhydric alcohols. Korean J Chem
Eng 2016;33:2858–62. http://dx.doi.org/10.1007/s11814-016-0122-6.
[35] Soltan M, Elsamadony M, Tawfik A. Biological hydrogen promotion via integrated
fermentation of complex agro-industrial wastes. Appl Energy 2017;185:929–38.
http://dx.doi.org/10.1016/j.apenergy.2016.10.002.
[36] Hubenova Y, Mitov M. Extracellular electron transfer in yeast-based biofuel cells: a
review. Bioelectrochemistry 2015;106:177–85. http://dx.doi.org/10.1016/j.
bioelechem.2015.04.001.
[37] Schaetzle O, Barrière F, Baronian K. Bacteria and yeasts as catalysts in microbial
fuel cells: electron transfer from micro-organisms to electrodes for green electricity.
Energy Environ Sci 2008;1:607. http://dx.doi.org/10.1039/b810642h.
[38] Rossi R, Cavina M, Setti L. Characterization of electron transfer mechanism in
mediated microbial fuel cell by entrapped electron mediator in Saccharomyces
cerevisiae. Chem Eng Trans 2016;49:559–64. http://dx.doi.org/10.3303/
CET1649094.
[39] Rossi R, Fedrigucci A, Setti L. Characterization of electron mediated microbial fuel
cell by Saccharomyces cerevisiae. Chem Eng Trans 2015;43:337–42. http://dx.doi.
org/10.3303/CET1543057.
[40] Christwardana M, Kim KJ, Kwon Y. Fabrication of mediatorless/membraneless
glucose/oxygen based biofuel cell using biocatalysts including glucose oxidase and
laccase enzymes. Sci Rep 2016;6:30128. http://dx.doi.org/10.1038/srep30128.
[41] Raghavulu SV, Goud RK, Sarma PN, Mohan SV. Saccharomyces cerevisiae as anodic
biocatalyst for power generation in biofuel cell: influence of redox condition and
substrate load. Bioresour Technol 2011;102:2751–7. http://dx.doi.org/10.1016/j.
biortech.2010.11.048.
[42] Valle-Rodríguez JO, Shi S, Siewers V, Nielsen J. Metabolic engineering of
Saccharomyces cerevisiae for production of fatty acid ethyl esters, an advanced
biofuel, by eliminating non-essential fatty acid utilization pathways. Appl Energy
2014;115:226–32. http://dx.doi.org/10.1016/j.apenergy.2013.10.003.
[43] Hubenova YV, Rashkov RS, Buchvarov VD, Arnaudova MH, Babanova SM, Mitov
MY. Improvement of yeast−biofuel cell output by electrode modifications. Ind Eng
Chem Res 2011;50:557–64. http://dx.doi.org/10.1021/ie1000949.
[44] Chung Y, Hyun KH, Kwon Y. Fabrication of a biofuel cell improved by the [small
pi]-conjugated electron pathway effect induced from a new enzyme catalyst em-
ploying terephthalaldehyde. Nanoscale 2016;8:1161–8. http://dx.doi.org/10.
1039/C5NR06703K.
[45] Christwardana M, Chung Y, Kwon Y. Co-immobilization of glucose oxidase and
catalase for enhancing the performance of a membraneless glucose biofuel cell
operated under physiological conditions. Nanoscale 2017;9:1993–2002. http://dx.
doi.org/10.1039/C6NR09103B.
[46] Hyun KH, Han SW, Koh WG, Kwon Y. Fabrication of biofuel cell containing enzyme
catalyst immobilized by layer-by-layer method. J Power Sources
2015;286:197–203. http://dx.doi.org/10.1016/j.jpowsour.2015.03.136.
[47] Zhu N, Chen X, Zhang T, Wu P, Li P, Wu J. Improved performance of membrane free
single-chamber air-cathode microbial fuel cells with nitric acid and ethylenedia-
mine surface modified activated carbon fiber felt anodes. Bioresour Technol
2011;102:422–6. http://dx.doi.org/10.1016/j.biortech.2010.06.046.
[48] Hidalgo D, Tommasi T, Bocchini S, Chiolerio A, Chiodoni A, Mazzarino I, et al.
Surface modification of commercial carbon felt used as anode for microbial fuel
cells. Energy 2016;99:193–201. http://dx.doi.org/10.1016/j.energy.2016.01.039.
[49] Christwardana M, Kwon Y. Yeast and carbon nanotube based biocatalyst developed
by synergetic effects of covalent bonding and hydrophobic interaction for perfor-
mance enhancement of membraneless microbial fuel cell. Bioresour Technol
2017;225:175–82. http://dx.doi.org/10.1016/j.biortech.2016.11.051.
[50] Zakhartsev M, Yang X, Reuss M, Pörtner HO. Metabolic efficiency in yeast
M. Christwardana et al.
Saccharomyces cerevisiae in relation to temperature dependent growth and biomass
yield. J Therm Biol 2015;52:117–29. http://dx.doi.org/10.1016/j.jtherbio.2015.
05.008.
[51] Falcucci G, Jannelli E, Minutillo M, Ubertini S. Fluid dynamic investigation of
channel design in high temperature pem fuel cells. J Fuel Cell Sci Technol
2012;9:21014. http://dx.doi.org/10.1115/1.4005628.
[52] Krastev VK, Falcucci G, Jannelli E, Minutillo M, Cozzolino R. 3D CFD modeling and
experimental characterization of HT PEM fuel cells at different anode gas compo-
sitions. Int J Hydrogen Energy 2014;39:21663–72. http://dx.doi.org/10.1016/j.
ijhydene.2014.09.015.
[53] Klein M, ul Islam Z, Knudsen PB, Carrillo M, Swinnen S, Workman M, et al. The
expression of glycerol facilitators from various yeast species improves growth on
glycerol of Saccharomyces cerevisiae. Metab Eng Commun 2016;3:252–7. http://dx.
doi.org/10.1016/j.meteno.2016.09.001.
[54] Lee WH, Jin YS. Improved ethanol production by engineered Saccharomyces cere-
visiae expressing a mutated cellobiose transporter during simultaneous sacchar-
ification and fermentation. J Biotechnol 2017;245:1–8. http://dx.doi.org/10.1016/
j.jbiotec.2017.01.018.
[55] Swinnen S, Ho PW, Klein M, Nevoigt E. Genetic determinants for enhanced glycerol
growth of Saccharomyces cerevisiae. Metab Eng 2016;36:68–79. http://dx.doi.org/
10.1016/j.ymben.2016.03.003.
[56] Puig S, Serra M, Coma M, Cabré M, Balaguer MD, Colprim J. Effect of pH on nu-
trient dynamics and electricity production using microbial fuel cells. Bioresour
Technol 2010;101:9594–9. http://dx.doi.org/10.1016/j.biortech.2010.07.082.
[57] Jannelli N, Nastro RA, Cigolotti V, Minutillo M, Falcucci G. Low pH, high salinity:
too much for microbial fuel cells? Appl Energy 2017;192:543–50. http://dx.doi.
org/10.1016/j.apenergy.2016.07.079.
[58] Karthikeyan R, Selvam A, Cheng KY. Wong JW-C. Influence of ionic conductivity in
bioelectricity production from saline domestic sewage sludge in microbial fuel cells.
Bioresour Technol 2016;200:845–52. http://dx.doi.org/10.1016/j.biortech.2015.
10.101.
[59] Du Q, An J, Li J, Zhou L, Li N, Wang X. Polydopamine as a new modification
material to accelerate startup and promote anode performance in microbial fuel
cells. J Power Sources 2017;343:477–82. http://dx.doi.org/10.1016/j.jpowsour.
2017.01.093.
[60] Christwardana M, Kwon Y. Effects of multiple polyaniline layers immobilized on
382
Applied Energy 222 (2018) 369–382
carbon nanotube and glutaraldehyde on performance and stability of biofuel cell. J
Power Sources 2015;299:604–10. http://dx.doi.org/10.1016/j.jpowsour.2015.08.
107.
[61] Rinaldi A, Mecheri B, Garavaglia V, Licoccia S, Nardo D, Traversa E. Engineering
materials and biology to boost performance of microbial fuel cells: a critical review.
Energy Environ Sci 2008:417–29. http://dx.doi.org/10.1039/b806498a.
[62] Hall-Stoodley L, Costerton JW, Stoodley P. Bacterial biofilms: from the natural
environment to infectious diseases. Nat Rev Microbiol 2004;2:95–108. http://dx.
doi.org/10.1038/nrmicro821.
[63] Lopez D, Vlamakis H, Kolter R. Biofilms. a000398–a000398 Cold Spring Harb
Perspect Biol 2010;2. http://dx.doi.org/10.1101/cshperspect.a000398.
[64] Morozan A, Stamatin L, Nastase F, Dumitru A, Vulpe S, Nastase C, et al. The bio-
compatibility microorganisms-carbon nanostructures for applications in microbial
fuel cells. Phys. Status Solidi Appl. Mater. Sci. 2007;204:1797–803. http://dx.doi.
org/10.1002/pssa.200675344.
[65] Briandet R, Herry J-M, Bellon-Fontaine M-N. Determination of the van der Waals,
electron donor and electron acceptor surface tension components of static gram-
positive microbial biofilms. Colloids Surfaces B Biointerfaces 2001;21:299–310.
http://dx.doi.org/10.1016/S0927-7765(00)00213-7.
[66] Takahashi H, Suda T, Tanaka Y, Kimura B. Cellular hydrophobicity of Listeria
monocytogenes involves initial attachment and biofilm formation on the surface of
polyvinyl chloride. Lett Appl Microbiol 2010;50:618–25. http://dx.doi.org/10.
1111/j.1472-765X.2010.02842.x.
[67] Prindle A, Liu J, Asally M, Ly S, Garcia-Ojalvo J, Süel GM. Ion channels enable
electrical communication in bacterial communities. Nature 2015;527:59–63.
http://dx.doi.org/10.1038/nature15709.
[68] Rahimnejad M, Najafpour GD, Ghoreyshi AA, Talebnia F, Premier GC, Bakeri G,
et al. Thionine increases electricity generation from microbial fuel cell using
Saccharomyces cerevisiae and exoelectrogenic mixed culture. J Microbiol
2012;50:575–80. http://dx.doi.org/10.1007/s12275-012-2135-0.
[69] Wilkinson S, Klar J, Applegarth S. Optimizing biofuel cell performance using a
targeted mixed mediator combination. Electroanalysis 2006;18:2001–7. http://dx.
doi.org/10.1002/elan.200603621.
[70] Bennetto HP, Stirling JL, Tanaka K, Vega CA. Anodic reactions in microbial fuel
cells. Biotechnol Bioeng 1983;25:559–68. http://dx.doi.org/10.1002/bit.
260250219.