Coupling Dark Fermentation and Microbial Electrolysis To Enhance Bio-Hydrogen Production From Agro-Industrial Wastewaters and By-Products in A Bio-Refinery Framework

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 6 ) 1 e1 3
Available online at www.sciencedirect.com
ScienceDirect
journal homepage: www.elsevier.com/locate/he
Coupling dark fermentation and microbial

electrolysis to enhance bio-hydrogen production
from agro-industrial wastewaters and by-products
in a bio-refinery framework
Antonella Marone a,*, Olga R. Ayala-Campos b, Eric Trably a,

Alessandro A. Carmona-Martı́nez a,1, Roman Moscoviz a, Eric Latrille a,
Jean-Philippe Steyer a, Vı́ctor Alcaraz-Gonzalez b, Nicolas Bernet a
a
LBE, INRA, 102 Avenue des Etangs, 11100, Narbonne, France
b
n
Universidad de Guadalajara-CUCEI, Departamento de Ingenierı́a Quı́mica, Blvrd. Gral. Marcelino Garcı́a Barraga
1451, 44430, Guadalajara Jalisco, Mexico
article info abstract
Article history: The aim of this work is to evaluate biohydrogen production from agro-industrial waste-
Received 4 April 2016 waters and by-products, by combining dark fermentation and microbial electrolysis in a
Received in revised form two-step cascade process. Such coupling of both technologies constitutes a technological
13 September 2016 building block within a concept of environmental biorefinery where sustainable production
Accepted 24 September 2016 of renewable energy is expected.
Available online xxx Six different wastewaters and industrial by-products coming from cheese, fruit juice,
paper, sugar, fruit processing and spirits factories were evaluated for the feasibility of
Keywords: hydrogen production in a two-step process. The overall hydrogen production when
Biohydrogen coupling dark fermentation and microbial electrolysis was increased up to 13 times when
Dark fermentation (DF) compared to fermentation alone, achieving a maximum overall hydrogen yield of
Microbial electrolysis cell (MEC) 1608.6 ± 266.2 mLH2/gCODconsumed and a maximum of 78.5 ± 5.7% of COD removal.
Agro-industrial wastewaters and These results show that dark fermentation coupled with microbial electrolysis is a
by-products highly promising option to maximize the conversion of agro-industrial wastewaters and
Two-stage process by-products into bio-hydrogen.
© 2016 Hydrogen Energy Publications LLC. Published by Elsevier Ltd. All rights reserved.
processes (open-cycles) toward natural closed ecological

Introduction cycles. Such alternative production chains need a replace-
ment of current raw materials, including energy, with
Nowadays alternative energy production chains are renewable resources.
increasingly studied with the objective to reduce the Currently, more than 60 Million Tons of hydrogen per year
dependence on oil, mitigate climate change originated by are produced in the World with about 10 MTH2/y in the EU.
human activities and redirect the current production
* Corresponding author.
E-mail address: marone.ant@gmail.com (A. Marone).
1
, Alcala
Current address: IMDEA Water Institute, Technological Park of the University of Alcala de Henares, Spain.
http://dx.doi.org/10.1016/j.ijhydene.2016.09.166
0360-3199/© 2016 Hydrogen Energy Publications LLC. Published by Elsevier Ltd. All rights reserved.
Please cite this article in press as: Marone A, et al., Coupling dark fermentation and microbial electrolysis to enhance bio-hydrogen
production from agro-industrial wastewaters and by-products in a bio-refinery framework, International Journal of Hydrogen Energy
(2016), http://dx.doi.org/10.1016/j.ijhydene.2016.09.166
2 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 6 ) 1 e1 3
Although H2 is mostly used as chemical reactant in various such as COD removal, material and energy recovery into a
industrial sectors, such oil refining (44%) or ammonia pro- single cascade process increases the system overall efficiency
duction (38%), hydrogen is considered as one of the most and potential for deployment. This provides dual environ-
serious alternatives to fossil fuels for future transportation mental benefits in the direction of wastewater treatment
vehicles [1]. Indeed, the major advantage of energy from along with sustainable bioenergy generation [7]. In the field of
hydrogen is the absence of polluting emissions since the ME cells for the production of hydrogen as a value-added
utilization of hydrogen, either via combustion or via fuel product researches have explored all the parameters that
cells, results in pure water [2]. It has also the highest energy might influence their performance, e.g.: electrode materials,
content per unit weight of any known fuel (142 kJ g1) and pH, temperature, type of substrate, among many others
can be transported for domestic/industrial consumption [13e15].
through conventional means [3]. However, hydrogen pro- The present manuscript tackles one pending issue in the
duction at industrial scale is mainly based on natural gas field of MECs: the use of dead-end fermentation products as
reforming with high environmental impacts through the substrates for microbial electrolysis thus significantly
release of huge amounts of fossil carbon dioxide [4]. The two increasing the overall production of hydrogen by coupling
most important challenges are to generate sustainably both technologies.
enough H2 from renewable sources and to reduce the pro- To date, only a few studies have dealt with H2 production
duction costs [5]. The development of green technologies is in DF process coupled with ME. Mostly, single carbon sour-
therefore crucial for the sustainability of a future H2 trans- ces were used as model substrates [8,16,17]. Nonetheless,
portation sector. more complex (fermentable) substrates, such as cellulose
Currently, besides hydrogen production by water electrol- [18,19] have also been investigated. Lalaurette et al. [20]
ysis based on renewable resources, the promising sectors of were the first to propose a two-stage system (DFeME)
hydrogen production are those of biomass biorefinery. using organic waste (acid-pretreated corn stover) as sub-
Among diverse renewable biomass based H2-producing strate, but they observed methane generation in the ME.
technologies, Dark Fermentation (DF) using anaerobic micro- Recently, different kinds of organic waste and by-products
bial communities, has gained increased attention not only have been investigated for H2 production by coupling DF
because it achieves high H2 production rates, but also for its and ME in a cascade process: acid-pretreated corn stalk [21];
ability to produce H2 at low costs when degrading complex crude glycerol [22] cheese whey [23] and waste activated
and unsterilized substrates such as waste [6]. sludge [24,25]. All these studies reported relatively high H2
Among various organic-rich substrates, wastewaters and yields, with both overall H2 yield and energy efficiency being
industrial by-products have gained a considerable attention much higher than the one obtained when only dark
due to their advantages such as high organic loads, low fermentation was used. Although all these works deal with
nutrient requirements and positive net energy gain [7]. bench-scale reactors, they clearly demonstrated the suit-
Particularly, wastewaters represent abundant, cheap and ability of the use of dead-end fermentation products as
widely available sources of biodegradable substrates not yet substrates for microbial electrolysis, significantly increasing
exploited to produce bio-hydrogen by dark fermentation (DF) the overall production of hydrogen by coupling both
[2]. The use of mixed consortia as biocatalyst for H2 produc- technologies.
tion, can be a practical and promising option for scaling up the The aim of this research was to evaluate and compare in
DF technology, especially when wastewaters are used as the same conditions the potential of different agro-industrial
substrates [8]. However, fermentative H2 production provides effluents, wastewaters and by-products, for the application
only partial oxidation of the organic substrate. Therefore, it is of a two-step biohydrogen production process, combining DF
likely to be industrially viable only if the fermentative bio- and ME in a biorefinery framework.
processes are integrated within a stream that can utilize the Six different substrates coming from cheese (CW), fruit
fermentation by-products [9]. juice (FJW), paper (PW), sugar (SW), fruit processing (FPW) and
Bio-electrochemical systems have been proposed as a spirits (VB2) factories were selected to evaluate the suitability
technology that can be coupled with fermentative hydrogen of two-step process at bench-scale. These substrates were
production [10]. More specifically, microbial electrolysis (ME) representative of different agro-industrial sectors, with
that is a promising candidate for the augmentation of clas- wastewater treatment systems (or by-products current use)
sical, single-stage dark fermentation to generate hydrogen gas that could be upgraded by producing biohydrogen. The efflu-
with a better efficiency [11]. Indeed, microbial electrolysis is ents from the fermentation of these substrates were tested in
accomplished in an electrochemical reactor, with an anode at two chambers ME cells to further recover H2 and COD removal.
which organic substrates are microbially oxidized to provide
an electric current that is then used at the cathode to elec-
trochemically produce hydrogen from water. The anode and Materials and methods
cathode may be or not separated by an ionic membrane. This
process requires a small external power supply in order to Agro-industrial wastewaters and by-products
make the hydrogen production thermodynamically favorable
[12]. Combining DF with ME in a two-steps cascade process Samples of 20 different agro-industrial wastewaters and by-
can result in a complete exploitation of biodegradable products were collected from a wide range of industrial sec-
wastewaters, which maximizes at the same time both energy tors (i.e. agroindustry, food, manufacturing, biofuel, waste-
recovery and effluent treatment [10]. Combining necessities water treatment). They were evaluated in terms of H2
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 6 ) 1 e1 3 3
potential in DF and fully characterized before and after dark
1102.72
(mg/L)
Ca2þ
210.50
354.63
203.49
282.93
(FJW); fruit processing wastewater (FPW); paper mill wastewater (PW); sugar production wastewater (SW) and vinasse residues from the distillation of sugar beets for
fermentation [26]. Six different wastewaters and by-products,
Table 1 e Chemical and biochemical characteristics of raw agro-industrial wastewaters and by-products: ricotta cheese whey (CW); fruit juice production wastewater
32.58
representative of different agro-industrial sectors, with
wastewater treatment systems (or by-products current use)
that could be upgraded by producing biohydrogen, were
(mg/L)
217.58
Mg2þ
83.12
13.98
83.64
11.79
32.05
selected on the basis of fermentative conversion efficiency
(i.e., both H2 and metabolic by-product accumulation) to be
used as substrate for the present study. They were: deprotei-
(mg/L)
270.07
NH4 þ
nized ricotta cheese whey (CW) namely “scotta”, mixed sheep
20.73
0.00
0.00
0.00
0.00
and cattle, supplied from a dairy factory in central Italy; fruit
processing wastewater (FPW) from a food manufacturing in-
dustry from south of France; sugar production wastewater
3243.77
(mg/L)
NO3 2
65.16
56.55
(SW) provided by a French sugar factory; wastewater from an
8.83
5.76
6.28
industrial fruit juice production plant (FJW) located in south
France; concentrated vinasse residues from the distillation of
(mg/L)
NO2
sugar beets for spirits production, supplied by industrial plant
13.03
37.91
0.00
0.00
0.00
9.39
(VB2) of a French cooperative and paper mill wastewater (PW)
collected from a French paper factory.
It's worth to precise that, whilst FPW, SW, FJW and PW are
N Tot
23.99
(g/L)
0.94
0.01
0.12
0.02
0.02
typical industrial wastewaters which undergone a classical
(aerobic or anaerobic) wastewater treatment process, CW and
VB2 should be considered as industrial by-products. CW,
49.92
(g/L)
Kþ
namely “scotta”, is the effluent from the Ricotta cheese pro-
0.51
0.03
0.36
0.03
0.10
duction, unlike “classical” whey, which is rich in proteins and
is currently used as a livestock feed, scotta is poor in proteins
24.73
(g/L)
Naþ
9.57
0.92
7.01
0.82
4.08
[27]. This makes scotta unattractive for further processing;
therefore the liquid itself has been poorly studied so far and is
actually considered as a waste [28]. VB2, instead, is the result
SO4 2
23.57
(g/L)
9.52
0.84
9.45
1.02
1.02
of six fold concentration of the original distillation effluent
and currently it is valued as fertilizer.
As shown in Table 1, the substrates composition varied
PO4 3
(g/L)
significantly both in terms of total organic matter content, and
0.44
0.00
0.00
0.00
0.00
1.09
biochemical or chemical composition.
For CW, FPW and FJW most of the determined COD
(g/L)
(Chemical Oxygen Demand) was explained by its sugars con-

7.98
0.60
3.45
0.64
6.39
7.92
Cl
tent (from 60 to 80%). Proteins were only present in a minor

extent in CW where they represented max. 10% of the total
Lactate
COD. Surprisingly, in SW which came from a sugar factory,

66.60
(g/L)
0.60
0.00
0.00
0.25
0.17
sugars represented only 5% of VSS (Volatile Suspended Solids)

while most of the organic matter content was related to
organic acids and alcohol, mainly acetic acid, lactic acid and
Glycerol
ethanol (data not shown). This suggests that this wastewater

94.47
(g/L)
0.00
0.00
0.00
0.00
0.00
had already undergone a natural fermentation process. In VB2

the glycerol represented a significant component of COD,
around 30%, the remaining COD was explained by approxi-
Proteins
mately 30% of proteins and around 30% of organic acids (in

149.96
(g/L)
5.88
0.08
0.75
0.15
0.12
part lactate) and alcohol, sugars represented only a minor

content (3% of VSS). PW is a kind of wastewater which is
difficult to treat by conventional aerobic and anaerobic
Sugars
methods cause of its contents in complex molecules and the

39.36
16.96
13.19
(g/L)
1.03
0.07
0.09
presence of toxic compounds. In PW the contents in reduced

spirits production (VB2).
metabolites exceeded the 20% of total COD, including a little

content of lactic acid. It is possible to assume that substantial
455.67
COD
(g/L)
57.80
39.10
amount of the organic matter is represented by complex

1.80
4.24
2.35
molecules as cellulose.
Substrate
Dark fermentation experiments

FPW
FJW
VB2
CW
Dark fermentation batch experiments were carried out in

SW
PW
500 mL serum flasks with a working volume of 200 mL. To
uniform the experimental conditions, raw substrate samples acetate as unique carbon and energy source. During these two
were diluted to reach an organic content 3 gCOD L1. SW chronoamperometric cycles a medium with the following
and FJW which were characterized by a lower organic matter composition was used: 2 g/L NH4Cl, 0.5 g/L K2HPO4, 9.76 g/L
content (2.35 and 1.80 gCOD L1 respectively) were used MES (2-[N-morpholino] ethane sulfonic acid) buffer, trace vi-
without dilution. MES buffer 50 mM and 1 mL of oligoele- tamins/minerals solutions [32] and 2-brome-ethane sulfonate
ments solution [29] were added to each flask. Anaerobically (BES) 10 mM to inhibit methanogens. At the beginning of each
digested sludge pretreated by heat/shock treatment (90 , enrichment cycle, the ME reactors were inoculated with
10 min) was used as inoculum at a substrate-to-biomass ratio anaerobic sediments (10% v/v) as described in Ref. [33].
(S/X) of 10. The initial pH was adjusted to 5.5 with NaOH 2N or Duration of a chronoamperometric cycle during enrichment
37% HCl. All batch tests were performed in triplicate and procedure varied from 6 to 7 days. Cyclic voltammetry (CV)
standard deviations are presented through all the manu- was performed at the beginning of each MEC enrichment
script. After inoculation, each bottle was flushed with ni- batch cycle before reactor inoculation (control-CV) to assure
trogen gas (>99.995%) for 10 min to ensure anaerobic that the electrochemical system was working adequately and
conditions. The bottles were then capped with a rubber at the maximum of current density production during chro-
stopper and incubated at 37 C. Reactors headspace were noamperometry. An illustrative example of chronoampero-
sampled at 6 h intervals and gas pressure and composition metric cycles and relatives CVs performed during the
(O2, H2, N2 and CH4) were measured using an automated enrichment procedure of electroactive biofilm is reported in
multiplexed micro-gas chromatograph (lGC R3000, SRA in- Fig. S1 of Supplementary Information (SI).
strument) equipped of a thermal conductivity detector using After the enrichment cycles, the medium in the anode
12e20 flasks in parallel, as previously described [30]. All chamber was replaced with the effluents from the fermenta-
volumes were expressed under temperature and pressure tion as carbon and energy source, neither inoculum nor
standard conditions. Cumulative hydrogen production was additional nutrient components were added to the effluent
determined by mass balance as previously described in Ref. during the test. Prior to use dark fermentation effluents, these
[31]. The experimental procedure ended when the pressure were pretreated by removing microbial biomass through
in the flask headspace started to drop, indicating hydrogen centrifugation (at 8000 rpm for 20 min) to avoid any possible
consumption. After this, liquid samples were taken at the influence. The pH was adjusted to 7.0 and effluents were
end of batch operation for pH and metabolite analyses. The flushed with nitrogen gas (purity 99.995%, Linde France S.A.)
effluents of dark fermentation were biochemically charac- for 30 min and 10 min more within the sealed reactor to
terized, and stored in sealed bottles at 20 C to preserve the establish anaerobic conditions. The total duration of a chro-
final composition to be tested in the second stage of micro- noamperometric cycle during experiments varied from 7 to 20
bial electrolysis cells. days. The volume of biogas production in ME reactors was
measured using water displacement method. The current,
Microbial electrolysis cell set up and operation biogas composition, soluble organic molecules and volatile
fatty acids (VFAs) concentration were monitored as described
The electrochemical system used for hydrogen production below.
from fermentation effluents was a cylindrical two-chamber
reactor. Each chamber had a working volume of 400 mL. The Analytical measurements
anode was a 7.0 cm 5.5 cm 1.0 cm carbon felt (RVG4000,
Mersen S.A., France), screwed onto a 2 mm diameter titanium The analysis of Total Suspended Solids (TSS), Volatile Sus-
road (TI007910/13, Goodfellow SARL, France) as electron col- pended Solids (VSS) and Chemical Oxygen Demand (COD)
lector. The cathode was a 90% Platinum-10% Iridium mesh were performed according to standard methods [34]. Total
(Heraeus PSP S.A.S., France). The reactor was hermetically Nitrogen content (TKN) was analyzed using a Buchi digestion
sealed with a silicone seal and crimping ring of stainless steel unit K438 and a Buchi 370-K distillator/titrator. Proteins were
at each chamber. An anion exchange membrane (AEM, determined by multiplying Total Kjeldhal Nitrogen (TKN) by
Fumasep FAA, FuMA-Tech BWT GmbH, Germany) separated 6.25. The volatile fatty acids (VFA) composition was deter-
both chambers. All ME tests were conducted at a set temper- mined with a gas chromatograph (GC-580, Clarus) equipped
ature of 37 C. A magnetic stirrer rotating at 250 rpm gently with an automatic sampler, an Elite-FFAP column (Perki-
homogenized the mixture during all the experiments. MEC nElmer) and a flame ionization detector (FID) according to
tests were carried out in triplicate for FPW, PW and SW and in Que me
neur et al. [30]. The concentrations of soluble sugars
duplicate for FJW, CW and VB2; standard deviations are pre- (glucose, fructose, sucrose, galactose, arabinose, lactose and
sented through all the manuscript. xylose), glycerol and other soluble organic molecules such as
All the bioelectrochemical experiments were conducted ethanol, lactate, succinate and formate were measured by
using a saturated calomel electrode (SCE, Hg/Hg2Cl2 þ240 mV HPLC analysis and refractometric detection, as previously
vs. SHE) (Origalys Electrochem SAS, France). The anode po- described in Ref. [35]. Ionic content was analyzed with ion
tential was fixed at þ200.0 mV vs. SCE (þ444.0 mV vs. SHE) chromatograph (ICS 3000, Dionex, USA) equipped with pre-
using a Potentiostat VMP3 (BioLogic Science Instruments, columns NGI 2 mm and CG 11 2 mm followed by separation
France) controlled by using the EC-Laboratory v.10.1 software. columns CS 16 3 mm and AG 15 2 mm for cations and anions,
For all experiments, the current density was recorded every respectively. The eluents used for this analysis were HMSA
60 s. The electroactive biofilm in the anode was previously (25e40 mM) pumped at 0.3 mL/min for cations and KOH
enriched by using two chronoamperometric cycles with 2 g/L (10e74 mM) pumped at 0.35 mL/min for anions [36].
For ME experiments, the gas composition (CH4, CO2, H2 and measured during the experiment and the ideal gas law. Cal-
N2) of the anodic and cathodic chambers was analyzed daily culations for hydrogen yield (YH2, mLH2/gCOD) were based on
from a sample taken directly in the gaseous headspace of the the hydrogen volume produced for unit of consumed COD as
reactor with a syringe which was analyzed by a gas chro- YH2 ¼ VH2/DCOD.
matograph (Clarus 580 GC, Perkin Elmer, USA) coupled to The performances in terms of energy recovery (YE, %) were
thermal catharometer detector as described in Ref. [37]. All based on the energy content of the hydrogen produced,
volumes were expressed under temperature and pressure compared to the electric energy input added into the system
standard conditions. Volatile fatty acids (VFA), residual sugars by the potentiostat (We, kJ), as YE ¼ WH2/We. The amount of
and non-VFA metabolic byproducts such as organic acids energy recovered by the hydrogen produced over a batch cycle
(lactate or formate) and ethanol in the liquid phase were (WH2, kJ) was calculated as WH2 ¼ nH2 DHH2 on the basis of
analyzed every 2e3 days as described before. the hydrogen heat combustion energy, DHH2 ¼ 285.8 kJ/mol.
The net electric energy yield has been calculated as electric
Microbial community analysis of anodic biofilm energy input (Wh) consumed per mole of H2 produced as We/
nH2. The overall hydrogen yield was calculated as the sum of
At the end of ME experiment, the anodic biofilm was sampled the cumulated volume of hydrogen per unit of COD consumed
and genomic DNA was extracted and purified as previously in each stage (Dark Fermentation þ Microbial Electrolysis).
described [32]. The V4 and V5 regions of the 16S rRNA genes Analysis of variance (ANOVA) was performed for compar-
were amplified and sequenced by Illumina MiSeq (get.geno- ison of mean with R software (version 3.3.0, available in CRAN
toul.fr) and sequences were analyzed as described elsewhere website) by using Rcmdr library.
[38]. Sequences have been submitted to GenBank, with
accession No. KU998264eKU999067.
Results and discussion
Calculations
Dark fermentation
A modified Gompertz equation was used to model hydrogen
production in dark fermentation experiments [39], Excel Add- Hydrogen production from fermentation of different agro-
in Solver was used to fit this model to the experimental data as industrial wastewaters and by-products
previously described in Ref. [40]. The performance of the ME Table 2 summarizes the results of the fermentation of
reactors were characterized and compared based on the different agro-industrial wastewaters and by-products,
following parameters: Coulombic efficiency (CE), cathodic including H2 production performance parameters and meta-
hydrogen recovery (rcat), hydrogen yield (YH2), maximum bolic end products accumulation.
current density (A/m2) and electrical energy recovery yield Depending on the substrate, the maximum cumulative
(YE). All calculations were based on COD removal because of hydrogen production yield varied from 18.8 ± 4.8 to 95.1 ± 26.4
the complexity of the substrate used. The coulombic effi- mLH2 gCODinitial 1 (Fig. 1). The highest yield was obtained with
ciency (CE, %) was calculated as CE ¼ nce/nth, where nce is the CW, which was the richest substrate in soluble sugars. Inter-
number of hydrogen moles that could be recovered based on estingly, CW also showed the highest variability between
the measured current and nth is the theoretical maximal pro- replicates (Fig. 1 and Table 2). This result is consistent with the
duction based on COD removal. The cathodic hydrogen re- yield of 102 ± 33 mLH2 gVS1 previously reported for the same
covery (rcat, %) was calculated as rcat ¼ nH2/nce, where nH2 is kind of substrate [41]. The wastewater issued from a sugar
number of moles of hydrogen actually produced at the cath- factory (SW), showed the lowest yield at 18.8 ± 4.8 mLH2
ode, which were calculated from the volume of hydrogen gCODinitial 1 , which is not surprising when considering the low
Table 2 e Fermentation parameters of raw agro-industrial wastewaters and by-products: ricotta cheese whey (CW); fruit
juice production wastewater (FJW); fruit processing wastewater (FPW); paper mill wastewater (PW); sugar production
wastewater (SW) and vinasse residues from the distillation of sugar beets for spirits production (VB2).
Substrate Time pHF condF H2 Yield Gompertz parameters ASuc EtOH AAce APro ABut 13Prop
(h) (mS cm1) (% mol/mol) (mLH2 $ gCOD1
initial)
pH2 r l R2
1
(mL) (mLH2 h ) (h)
CW 37 5.5 7.4 13.8 95.1 54.6 2.00 12 1.0 12.4 101 35.3 0.0 81.1 0.0
WC 34 4.8 5.9 12.1 57.6 39.5 2.48 12 1.0 0 235 105 0 92.1 0
SW 37 5.4 15.8 3.2 18.8 8.5 1.08 21 1.0 35.7 76.7 279 34.3 209 0
FJW 35 5.1 6.1 8.5 74.9 25.6 2.73 13 0.99 0 0 147 0 261 0
VB2 59 5.6 6.6 5.1 35.4 15.5 0.81 26 1.0 19.2 0 94.1 8.95 286 154
PW 37 5.5 7.4 4.7 19.3 14.2 0.82 24 1.0 127 62.4 74.1 56.7 117 0
Notes: time required to reach maximum production of hydrogen (Time), final pH (pHF), final conductivity (condF), H2% and H2 production yields
(mLH2 gCOD1 2
initial); H2 Gompertz equation parameters: production potential (PH2), rate (r), lag phase (l) and Pearson correlation coefficients (R )
between observed and predicted data; and metabolic end-products accumulation in mgCOD gCOD1L1 (ASuc ¼ succinic acid, EtOH ¼ ethanol,
AAce ¼ acetic acid, AProp ¼ propionic acid, ABut ¼ butyric acid,13Prop ¼ 1,3 propanediol) when the maximum hydrogen was reached. (nd ¼ not
determined).
Fig. 1 e Maximum Bio-Hydrogen production yield (mLH2/

gCODinitial) obtained for various organic agro-industrial
wastewaters and by-products: deproteinized ricotta cheese
whey (CW); wastewater from an industrial fruit juice
production plant (FJW); fruit processing wastewater (FPW); Fig. 2 e Cumulative hydrogen production (mL) from dark
paper mill wastewater (PW); sugar production wastewater fermentation of different agro-industrial wastewaters and
(SW) and vinasse residues from the distillation of sugar by-products: deproteinized ricotta cheese whey (CW);
beets for spirits production (VB2). Error bars represent wastewater from an industrial fruit juice production plant
standard deviation of triplicates. (FJW); fruit processing wastewater (FPW); paper mill
wastewater (PW); sugar production wastewater (SW) and
vinasse residues from the distillation of sugar beets for
amount of sugars content present (i.e. 5% of organic content) spirits production (VB2). Points represent experimental
in the raw wastewater. Indeed, carbohydrate-rich substrates data, while the curves are calculated by Gompertz
are mostly referred in the literature to favor biohydrogen equation.
production through fermentative pathways [6,8,42]. Our re-
sults are consistent with these findings since a significant
linear correlation (R2 ¼ 0.89) was found between the cumu- inoculum which favorably selected spore-forming anaerobic
lated volume of H2 and the initial soluble sugars content in bacteria, belonging mainly to the Clostridium genus. Indeed,
substrates. fermentative metabolism of Clostridium species produces
As shown in Table 2, the overall fermentation behaviors, mainly butyrate and acetate as primary soluble metabolites
including the time required to reach maximum production of [9]. Interestingly, ethanol was the main metabolite produced
hydrogen (Time), the final pH (pHF), the final conductivity during CW fermentation, which was the substrate which
(condF) and the metabolic end-products patterns (succinic yielded the max H2. Such difference in metabolic pathways
acid, ethanol, propionic acid, butyric acid and 1,3 propane- could be explained from the different initial composition of
diol), varied in a large extent according to the origin of the substrates. For example, CW was the sole substrate contain-
substrate. Interestingly, the Gompertz model was highly ing lactose. As reported elsewhere, the fermentation of lactose
suitable to describe the kinetics of the BHP tests for all sub- by a Clostridium strain (Clostridium thermolacticum) produces
strates (Fig. 2), as shown by the R2 coefficients (>0.99) gener- mainly ethanol and acetate as end-products [43].
ated after fitting the Gompertz model to experimental data The amounts and the relative abundances of acetate,
(Table 2). ethanol or butyrate varied depending on the type of substrate;
propionate was produced in some cases but only in a poor
Metabolic end products from fermentation of different agro- quantity. Although lactate was not detected as end-product, it
industrial wastewaters and by-products was initially detected in some substrates (Table 1). Probably,
At the end of fermentation experiments, no sugars were lactate was converted into butyrate with concomitant H2
detected, nor lactate, initially contained in VB2 and CW. This production. This former metabolic pathway has been widely
shows that sugars and lactate were effectively consumed as observed in mixed culture hydrogen fermentative systems
substrates in the fermentative metabolism. Meanwhile, other [44e46]. Guo et al. [47] reported that lactate had even a positive
reduced organic compounds, such as succinic acid, acetic acid effect on bioH2 production for carbohydrate-poor substrates,
and ethanol, which were detected in some of the raw waste- while it showed a strong negative effect for carbohydrate-rich
waters (FPW, SW and PW), were not degraded and their con- substrates. The authors proposed an interaction term be-
centrations remained unchanged at the end of fermentation tween carbohydrates and lactate concentrations in a model
or were slightly accumulated during the process. Values re- explaining the microbial pathways for hydrogen production
ported in Table 2 are the ones observed in the samples at the from a large spectrum of organic waste [47]. In addition, 1,3
end of the batch fermentation. Butyrate was the main end- propanediol was also detected as end-product during the
product, followed by acetate and/or ethanol depending on fermentation of VB2, which is not surprising since this sub-
the origin of the substrate. This main metabolic pathway was strate initially contained glycerol. Such metabolite, 1,3 pro-
likely a result of initial heat-shock treatment of the microbial panediol, could be produced as by-product of glycerol
acetate consumption was observed and current started to

increase.
Regarding the coulombic efficiency (CE) obtained from the
diverse effluents tested (Fig. 3), it is important to point out the
high range observed between the substrates. Such variability
might be due to the diverse metabolic pathways that bacteria
used to degrade the organic matter. Lower values of CE% mean
an important loss of electrons from the substrate. A possible
explanation of such loss is that the fed substrate, as in the
case of PW, might have been partly used for biosynthesis of
cellular components, e.g. biofilm formation, instead of direct
substrate oxidation providing electrons to the anode. The best
result in terms of coulombic efficiency was reached when the
ME reactor was fed with fruit processing wastewater (FPW)
with a value of 81.4 ± 4.7%. Moreover, the average CE from the
Fig. 3 e ME stage performance in terms of coulombic diverse effluents was 62.7 ± 10.4%, which is a high value
efficiency (CE, %) and maximum current density (A/m2) compared with CEs reported in previous studies dealing with
from diverse fermented agro-industrial wastewaters and pre-fermented effluents [20,22,23].
by-products: ricotta cheese whey (CW); fruit juice Chronoamperometric batch cycle performed with diverse
production wastewater (FJW); fruit processing wastewater fermented effluents are showed in Fig. S2 of Supplementary
(FPW); paper mill wastewater (PW); sugar production Information.
wastewater (SW) and vinasse residues from the distillation
of sugar beets for spirits production (VB2). Hydrogen generation and yield in the MEC
The results of the hydrogen production in MEC are shown in
fermentation, but this metabolic pathway is not linked to the Fig. 4 and Table 3. The highest hydrogen yield (YH2) was
production of hydrogen [48]. observed with fruit juice wastewater (FJW), i.e.,
1478.2 ± 267.6 mLH2 gCODconsumed 1 . The best hydrogen
Further H2 production from fermented agro-industrial cathodic recovery (rcat) at 101.4 ± 0.9% was also found with
wastewaters and by-products in microbial electrolysis cells fruit juice wastewater (FJW).
As reported in Table 3, a high variability of the perfor-
Coulombic efficiency and current density generation mances was observed between substrates. These results show
Overall, ME results showed satisfactory performances in that carbon degradability and substrate conversion mainly
terms of electrical current production for almost all of the depended on the different metabolites fractions fed to the ME
effluents when comparing to other works using a similar reactors. Interestingly, the FJW, which showed the best per-
experimental system, ie. a carbon felt electrode colonized by formances in MEC, had the simplest initial composition of all
an electroactive biofilm fed with an easy degradable substrate the effluents, ie. only acetate and butyrate. Meanwhile, the
such as acetate [49]. The highest value of 7.46 ± 1.76 Am2 was PW, showing the worst result, had the most complex initial
obtained with deproteinized ricotta cheese whey (CW) composition. Finally, a successful substrate to hydrogen
effluent. Interestingly, a complete conversion of metabolic by- conversion was obtained when the initial composition of the
products to acetate was first observed before significant effluent had a significant quantity of ethanol (see Table 2).
Fig. 4 e Hydrogen yield (YH2, mLH2/gCODconsumed) and cathodic hydrogen recovery (rcat, %) obtained in ME cells from
fermentation effluents of diverse agro-industrial wastewaters and by-products: ricotta cheese whey (CW); fruit juice
production wastewater (FJW); fruit processing wastewater (FPW); paper mill wastewater (PW); sugar production wastewater
(SW) and vinasse residues from the distillation of sugar beets for spirits production (VB2).
Table 3 e MEC hydrogen production parameters of raw agro-industrial wastewaters and by-products.
Substrate Time (days) H2 (mL) H2 yield (mLH2 , gCODconsumed 1 ) Electric energy
yield (Wh/H2 mol)
CW 14 ± 0 128.9 ± 4.8 714.7 ± 58.0 18.1 ± 0.8
FJW 16 ± 6 158.1 ± 14.4 1478.2 ± 267.6 11.0 ± 0.2
FPW 13 ± 7 56.9 ± 22.3 652.4 ± 38.5 21.5 ± 0.5
PW 9±3 20.1 ± 20.3 219.4 ± 139.9 23.3 ± 3.0
SW 12 ± 5 51.3 ± 53.4 344.3 ± 175.7 25.2 ± 8.2
VB2 16 ± 0 176.7 ± 73.0 1399.57 ± 388.4 18.4 ± 1.8
CW ¼ ricotta cheese whey; FJW ¼ fruit juice production wastewater; FPW ¼ fruit processing wastewater; PW ¼ paper mill wastewater;
SW ¼ sugar production wastewater and VB2 ¼ vinasse residues from the distillation of sugar beets for spirits production.
Values of cathodic hydrogen recovery ranged between process. Commercial low temperature electrolyzers have
52.1 ± 0.1% and 101.4 ± 0.9%. In all tested effluents, at least half system efficiencies of 56e73% (70.1e53.4 kWh/kg H2 at 1 atm
of the current generated by the anode was used to produce and 25 C) [1] it means that the net electric energy (Wh)
hydrogen at the cathode. The lowest values of rcat (Fig. 4) consumed per mole of H2 produced ranges from 107 to 140. As
implied an important loss of electrons. As the cathodic reported in Table 3, the net energy yields of this study ranged
chamber was abiotic, a low rcat means that electrochemical from 11.0 ± 0.2 to 25.2 ± 8.2 Wh/mol of produced H2, which is
side-reactions, such as reduction of metal ions, occurred in considered low. Thus, although these results are not repre-
the cathodic chamber. It is likely that some ions of the sub- sentative of an industrial process, they support the use of MEC
strates present in the anodic chamber diffused to the cathode technologies as an alternative for H2 production, especially
chamber through the membrane. It is worth mentioning that when combined with organic waste valorization.
no methane was detected in the anode or in the cathode Anyway, an effective comparison should be done only with
chambers and therefore could not explain electrons loss. systems in comparable conditions. Batlle-Vilanova et al. [50],
compared biotic and an abiotic (pure electrochemical re-
COD removal and energy yield in MEC actions) MEC for hydrogen production in the identical opera-
Regarding the energy yield (Fig. 5), the total electrical input tive conditions [50]. In that study, hydrogen was produced in a
energy that was recovered from all effluents was always more efficiently way in the biotic MEC and the energy demand
higher than 100%. In other words, the energy that could be of the abiotic MEC (170.1e995.5 kWh kg1 H2) was higher than
recovered by combustion of H2 was always higher that the the biotic MEC (33.2e117.0 kWh kg1 H2).
supplied electrical needs. The minimum value was COD removal efficiency observed in MEC stage ranged from
336.3 ± 16.0% meaning that the energy produced was at least 35 to 63% in average and showed a great variability between
three times higher than the supplied energy. The maximum replicates (Fig. 4). High variability of the energy yields and COD
energy recovery was 725.0 ± 7.8% from the fruit juice waste- removals in MEC were likely resulting from the initial effluent
water (FJW). Both high hydrogen yield and low electricity COD concentration and composition. As shown in Fig. 4, the best
supply contributed to an increase of the overall energy yield. results were obtained for CW, FJW and VB2. When comparing
Although the results presented in this study are obtained at the data in Table 2 and Fig. 4, a relationship was found between
bench scale, therefore they are not representative of an in- high energy recovery and simple initial composition or a high
dustrial process, it is interesting to compare them to the initial amount of butyrate. The removal sequence of by-
consumption of abiotic electrochemical electrolysis, as actual metabolites was as follows: ethanol > butyrate > acetate,
industrial alternative to the proposed bio-based H2 production which supports the previous conclusion.
Fig. 5 e Energy yield (YE, %) and COD removal (%) in MEC stage from diverse fermented effluents: ricotta cheese whey (CW);
fruit juice production wastewater (FJW); fruit processing wastewater (FPW); paper mill wastewater (PW); sugar production
wastewater (SW) and vinasse residues from the distillation of sugar beets for spirits production (VB2).
Table 4 e Compilation of published works that report the production of hydrogen from complex substrates by coupling dark fermentation and microbial electrolysis.
Substrate Type of DF reactor Electrodes materials used in ME Applied YH2 DF-I stage YH2 ME-II stage YH2 coupling Electric efficiency COD Ref.
source stage volt. (V) (mLH2/gCOD) (mLH2/gCOD) (mLH2/gCOD) (Wh/mol H2) removal (%)
Anode Cathode
Domestic Batch Graphite brush Carbon cloth with a 0.8 135.15* 1200.00 1335.15* 51.56* 50 [24]
wastewater 35 C Pt catalyst layer.
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 6 ) 1 e1 3
Cellulose CSTR Brush carbon Carbon cloth with a 0.44 57.58* 812.72* 870.30* 78.91* 15 [18]
60 C fibers Pt catalyst layer
Corn stalk Batch Two pieces of Carbon cloth with a 0.8 168.16* 1000.00 1168.16* 47.83* 44 ± 2 [21]
36 C square graphite Pt catalyst layer
felts
Crude glycerol Batch Heat-treated Carbon cloth with 1.0 124.10* 106.15 230.25* NA 10.4 ± 1.4/40.6 ± 4.9% [22]
35 C graphite brush stainless steel mesh
Lignocellulose Batch Graphite fiber Flat carbon cloth 0.5 225.92* 900.00 1125.92* 36.09* 50 [20]
and cellobiose 50 C brush pretreated containing a Pt
using an NH3 catalyst
Ricotta cheese Batch Carbon felt 90% Platinum-10% 0.2 122.27 714.69 836.96 18.12 78.5 ± 5.7 This work
whey (CW) 37 C Iridium mesh
Fruit juice Batch Carbon felt 90% Platinum-10% 0.2 130.41 1478.22 1608.63 10.95 71.8 ± 1.6 This work
production 37 C Iridium mesh
wastewater
(FJW)
Fruit processing Batch Carbon felt 90% Platinum-10% 0.2 101.85 652.42 754.27 21.50 73.8 ± 3.6 This work
wastewater 37 C Iridium mesh
(FPW)
Paper mill Batch Carbon felt 90% Platinum-10% 0.2 37.37 219.39 256.76 23.26 57.1 ± 6.1 This work
(PW)
Sugar production Batch Carbon felt 90% Platinum-10% 0.2 51.91 344.33 396.24 25.16 67.0 ± 13.1 This work
(SW)
Vinasse from Batch Carbon felt 90% Platinum-10% 0.2 87.70 1399.57 1487.27 18.37 47.3 ± 6.4 This work
spirits 37 C Iridium mesh
production
(VB2)
YH2 ¼ Hydrogen yield; DF ¼ dark fermentation; ME ¼ microbial electrolysis ; * ¼ Calculated; NA ¼ not available.
9
Microbial community analysis of the anodic biofilms Table 4 reports a compilation of published works that
MiSeq sequencing was performed on anodic biofilm samples report the production of hydrogen from complex substrates by
after the end of the ME experiments. All the communities coupling dark fermentation and microbial electrolysis.
found on the anodes were highly diverse, with no clear Although it is a difficult task to compare among groups due to
dominance of any species (Simpson diversity indexes ranging the lack of standardization regarding the way of performing
from 0.91 to 0.93) except for the community found on the experiments and reporting date extracted from them [13],
anode fed with FJW wastewater (Simpson diversity index of from Table 4 we can observe that there is a clear tendency to
0.73). This anode was the one exhibiting the best perfor- use raw sources of substrates such as real wastewaters, by-
mances (see Table 3). ME reactor anodic biofilm fed with FJW products and lingo-cellulose materials coming from
was dominated by an OTU related to Geovibrio ferrireducens different industrial processes. In spite of the fact that all these
(99% sequence similarity) representing 50.3% of the total works deal with bench-scale reactors, they clearly demon-
abundance. This OTU is found in most of the anodic biofilm strated the suitability of the use of dead-end fermentation
microbial communities but at a lower abundance. Similarly to products as substrates for microbial electrolysis, significantly
other electroactive bacteria belonging to the Geobacteracea increasing the overall production of hydrogen by coupling
family, G. ferrireducens is known to be a Fe(III)-reducing bacte- both technologies. Another pending task to tackle in the near
rium able to consume acetate and propionate, but unable to future is the scaling-up of both coupled technologies. So far,
degrade ethanol and butyrate [51] which is consistent with the the works compiled in Table 4 deal with bench-scale reactors
fact that the anodes were initially fed with acetate only. that although provide useful information about the func-
However, at this stage, the G. ferrireducens electrochemical tioning of the coupled technologies, it does not produce in-
activity cannot be definitively confirmed [52,53]. A detailed formation about their success in the industrial scale.
investigation of the electrochemical activity of G. ferrireducens However, there is an increase in interest in developing
should be specifically carried out as usually performed for these bio-based technologies and when considered as single
other electrochemically active microorganisms [54]. Statistical process there currently are tests carried out at least at pilot
analysis revealed that G. ferrireducens OTU abundance was scale.
correlated with a good overall H2 production yield (r ¼ 0.76, Tapia et al. [55], reviewed the existing pilot scale hydrogen
p 0.05). Therefore this suggests that the bacteria corre- production dark fermentation experiences and concluded
sponding to this OTU was able to perform an efficient con- that improvement of energetic yields of dark fermentation is
version of the organic substrates into electricity. Butyrate and required to reduce system costs to make this technology
ethanol degradation and conversion into electricity was competitive in the current market. Extensive research has
probably carried out by bacteria able to convert these sub- been conducted to overcome these limits and several strate-
strates into acetate. Such conversion, as observed during the gies have been proposed, such as pretreatment of the seed
experiments, could have been supported by species related to sludge, design/retrofitting of bioreactors, optimization of bio-
the Clostridiales order, since their total abundance was posi- process parameters (e.g. pH, temperature, organic loading
tively correlated to a good overall hydrogen yield (r ¼ 0.75, rate, hydraulic retention time, effluent recycling ratio, etc.),
p 0.05). Another OTU of interest was revealed by statistical reduction of H2 partial pressure, bioaugmentation, among
analysis and was related to an uncultured bacterium issued others [56].
from the Coriobacteriaceae family. This OTU was present on all
the anodes (1e10% of the total abundance) and its abundance
positively correlated to a good CE (r ¼ 0.92, p 0.01) and a high Conclusions
current density (r ¼ 0.81, p 0.05). However, this OTU func-
tions are still unknown. Six different industrial wastewaters and by-products coming
from cheese, fruit juice, paper, sugar, fruit processing and
spirits factories were evaluated for their potential to generate
Energy recovery from tested agro-industrial wastewaters hydrogen by coupling dark fermentation and microbial elec-
and by-products by coupling dark fermentation and trolysis in a two-step process. By considering both processes
microbial electrolysis concomitantly (and not separately), a maximum total
hydrogen yield of 1608.6 ± 266.2 mLH2/gCODconsumed and a
Overall, combining dark fermentation and microbial elec- maximum of 78.5 ± 5.7% of COD removal were achieved. An
trolysis into a two-stage process resulted in an good overall increase up to 13 times of the total H2 recovered from real
hydrogen yield for the six tested effluents (Table 4). There was wastewater was reached when considering both processes in
a significant effect of the substrate type on the overall comparison to the DF step alone. Fruit Juice Wastewater (FPW)
hydrogen yield at the p < 0.05 level for the six tested effluents provided the best results which could be due to the compo-
[F(5,9) ¼ 4.959, p ¼ 0.0186]. A maximum value of 1608.6 ± 266.2 sition of the effluent in terms of easy degradable metabolites.
mLH2 gCODconsumed 1 , around 9.82 mol H2 mol glucose1 was Interestingly, this raw wastewater could be treated without
obtained for fruit juice wastewater, which is one of the best the need of any dilution, in contrast with other wastewaters
results when compared with literature values (See Table 4). and by-products, and high H2 yield could be recovered.
Such hydrogen yield is very close to the theoretically Therefore the proposed biorefinery scheme could offer a
maximum value of 12 mol H2 mol glucose1. Additionally, a captivating alternative to the current expensive wastewaters
satisfactory COD removal (47.3%e78.5%) was obtained when treatment technologies, by transform what is currently
coupling both processes for all tested substrates. considered as a waste in a value added product.
Meanwhile, if considering the initial COD content of raw [7] Lin CY, Lay CH, Sen B, Chu CY, Kumar G, Chen CC, et al.
substrates, the most interesting substrates to be exploited for Fermentative hydrogen production from wastewaters: a
H2 production in such two-stage system could be vinasse review and prognosis. Int J Hydrogen Energy
2012;37:15632e42. http://dx.doi.org/10.1016/
residue from spirits production (VB2) followed by deprotei-
j.ijhydene.2012.02.072.
nized ricotta cheese whey (CW). Indeed, up to 28 ± 5 L of H2 [8] Hawkes F, Hussy I, Kyazze G, Dinsdale R, Hawkes D.
could be produced per liter of raw VB2 and 8.1 ± 1.4 L of H2 Continuous dark fermentative hydrogen production by
could be produced per liter of raw CW. However, in the case of mesophilic microflora: principles and progress. Int J
VB2, it is already available as a value-added chain (i.e. it is Hydrogen Energy 2007;32:172e84. http://dx.doi.org/10.1016/
used as fertilizer), it should be considered whether it is j.ijhydene.2006.08.014.
convenient to replace it with the proposed one. In this case, H2 [9] Hallenbeck PC. Fermentative hydrogen production:
principles, progress, and prognosis. Int J Hydrogen Energy
production potential of VB2 should be evaluated before this
2009;34:7379e89. http://dx.doi.org/10.1016/
spirits production effluent undergoes concentration proced- j.ijhydene.2008.12.080.
ure. On the other hand, there are currently no exploitation [10] Rozendal RA. Hydrogen production through biocatalyzed
systems for deproteinized ricotta cheese whey (CW), namely electrolysis. 2007.
“scotta”, therefore the results presented here could open an [11] Kumar G, Bakonyi P, Kobayashi T, Xu KQ, Sivagurunathan P,
interesting field of investigation for the valorization of this Kim SH, et al. Enhancement of biofuel production via
microbial augmentation: the case of dark fermentative
dairy by-product.
hydrogen. Renew Sustain Energy Rev 2016;57:879e91. http://
Overall, these results showed that coupling DF and MEC for
dx.doi.org/10.1016/j.rser.2015.12.107.
organic wastewater treatment and/or by-products trans- [12] Logan BE, Regan JM. Electricity-producing bacterial
formation constitutes not only a suitable but also a highly communities in microbial fuel cells. Trends Microbiol
promising route for producing bio-H2 within the scheme of an 2006;14:512e8. http://dx.doi.org/10.1016/j.tim.2006.10.003.
environmental biorefinery. [13] Escapa A, Mateos R, Martı́nez EJ, Blanes J. Microbial
electrolysis cells: an emerging technology for wastewater
treatment and energy recovery. From laboratory to pilot
plant and beyond. Renew Sustain Energy Rev 2016;55:942e56.
Acknowledgments http://dx.doi.org/10.1016/j.rser.2015.11.029.
[14] Lu L, Ren ZJ. Microbial electrolysis cells for waste biorefinery:
a state of the art review. Bioresour Technol 2016;215:254e64.
This research was supported by Waste2bioHy project (FP7-
http://dx.doi.org/10.1016/j.biortech.2016.03.034.
MC-IEF e326974) and BITA Project (Marie Curie IRSES; PIR-
[15] Kadier A, Simayi Y, Abdeshahian P, Azman NF,
SES-GA-2011-295170) within the EU FP7. A.A.C.M. and R.M. Chandrasekhar K, Kalil MS. A comprehensive review of
were supported by the French National Research Agency microbial electrolysis cells (MEC) reactor designs and
Projects: ANR-09-BioE-10 DefiH12 and ANR-10-BTBR-02, configurations for sustainable hydrogen gas production. Alex
respectively. Eng J 2014;55:427e43. http://dx.doi.org/10.1016/
j.aej.2015.10.008.
[16] Lu L, Ren N, Xing D, Logan BE. Hydrogen production with
effluent from an ethanol-H2-coproducing fermentation
Appendix A. Supplementary data reactor using a single-chamber microbial electrolysis cell.
Biosens Bioelectron 2009;24:3055e60. http://dx.doi.org/
Supplementary data related to this article can be found at 10.1016/j.bios.2009.03.024.
http://dx.doi.org/10.1016/j.ijhydene.2016.09.166. [17] Rivera I, Buitro n G, Bakonyi P, Nemesto thy N, Be
lafi-Bako
K.
Hydrogen production in a microbial electrolysis cell fed with
a dark fermentation effluent. J Appl Electrochem
2015;45:1223e9. http://dx.doi.org/10.1007/s10800-015-0864-6.
references
[18] Wang A, Sun D, Cao G, Wang H, Ren N, Wu W-M, et al.
Integrated hydrogen production process from cellulose by
combining dark fermentation, microbial fuel cells, and a
[1] Holladay JD, Hu J, King DL, Wang Y. An overview of hydrogen microbial electrolysis cell. Bioresour Technol
production technologies. Catal Today 2009;139:244e60. 2011;102:4137e43. http://dx.doi.org/10.1016/
http://dx.doi.org/10.1016/j.cattod.2008.08.039. j.biortech.2010.10.137.
[2] Kapdan IK, Kargi F. Bio-hydrogen production from waste [19] Ullery ML, Logan BE. Anode acclimation methods and their
materials. Enzyme Microb Technol 2006;38:569e82. http:// impact on microbial electrolysis cells treating fermentation
dx.doi.org/10.1016/j.enzmictec.2005.09.015. effluent. Int J Hydrogen Energy 2015;40:6782e91. http://
[3] Benemann J. Hydrogen biotechnology: progress and dx.doi.org/10.1016/j.ijhydene.2015.03.101.
prospects. Nat Biotechnol 1996;14:1101e3. [20] Lalaurette E, Thammannagowda S, Mohagheghi A, Maness P-
[4] Megret O, Hubert L, Calbry M, Trably E, Carrere H, Bernet N. C, Logan BE. Hydrogen production from cellulose in a two-
Production d ’ hydroge ne a
partir de de
chets. 2015. stage process combining fermentation and
[5] Westermann P, Jorgensen B, Lange L, Ahring B, electrohydrogenesis. Int J Hydrogen Energy 2009;34:6201e10.
Christensen C. Maximizing renewable hydrogen production http://dx.doi.org/10.1016/j.ijhydene.2009.05.112.
from biomass in a bio/catalytic refinery. Int J Hydrogen [21] Li X-HH, Liang D-WW, Bai Y-XX, Fan Y-TT, Hou H-WW.
Energy 2007;32:4135e41. http://dx.doi.org/10.1016/ Enhanced H2 production from corn stalk by integrating dark
j.ijhydene.2007.06.018. fermentation and single chamber microbial electrolysis cells
[6] Guo XM, Trably E, Latrille E, Carre re H, Steyer J-P. Hydrogen with double anode arrangement. Int J Hydrogen Energy
production from agricultural waste by dark fermentation: a 2014;39:8977e82. http://dx.doi.org/10.1016/
review. Int J Hydrogen Energy 2010;35:10660e73. http:// j.ijhydene.2014.03.065.
dx.doi.org/10.1016/j.ijhydene.2010.03.008.
[22] Chookaew T, Prasertsan P, Ren ZJ. Two-stage conversion of [37] Cazier EAA, Trably E, Steyer JPP, Escudie R. Biomass
crude glycerol to energy using dark fermentation linked with hydrolysis inhibition at high hydrogen partial pressure in
microbial fuel cell or microbial electrolysis cell. N Biotechnol solid-state anaerobic digestion. Bioresour Technol
2014;31:179e84. http://dx.doi.org/10.1016/j.nbt.2013.12.004. 2015;190:106e13. http://dx.doi.org/10.1016/
[23] Moreno R, Escapa A, Cara J, Carracedo B, Go mez X. A two- j.biortech.2015.04.055.
stage process for hydrogen production from cheese whey: [38] Moscoviz R, Trably E, Bernet N. Consistent 1,3-propanediol
integration of dark fermentation and biocatalyzed production from glycerol in mixed culture fermentation over
electrolysis. Int J Hydrogen Energy 2015;40:168e75. http:// a wide range of pH. Biotechnol Biofuels 2016;9:32. http://
dx.doi.org/10.1016/j.ijhydene.2014.10.120. dx.doi.org/10.1186/s13068-016-0447-8.
[24] Liu W, Huang S, Zhou A, Zhou G, Ren N, Wang A, et al. [39] Chen W-H, Chen S-Y, Khanal SK, Sung S. Kinetic study of
Hydrogen generation in microbial electrolysis cell feeding biological hydrogen production by anaerobic fermentation.
with fermentation liquid of waste activated sludge. Int J Int J Hydrogen Energy 2006;31:2170e8. http://dx.doi.org/
Hydrogen Energy 2012;37:13859e64. http://dx.doi.org/ 10.1016/j.ijhydene.2006.02.020.
10.1016/j.ijhydene.2012.04.090. [40] Varrone C, Giussani B, Izzo G, Massini G, Marone A,
[25] Lu L, Xing D, Liu B, Ren N. Enhanced hydrogen production Signorini A, et al. Statistical optimization of biohydrogen and
from waste activated sludge by cascade utilization of organic ethanol production from crude glycerol by microbial mixed
matter in microbial electrolysis cells. Water Res culture. Int J Hydrogen Energy 2012;37:16479e88. http://
2012;46:1015e26. http://dx.doi.org/10.1016/ dx.doi.org/10.1016/j.ijhydene.2012.02.106.
j.watres.2011.11.073. [41] Marone A, Varrone C, Fiocchetti F, Giussani B, Izzo G,
[26] Marone A, Latrille E, Steyer JP. Trably E. Dark fermentation of Mentuccia L, et al. Optimization of substrate composition for
organic wastewaters: effect of wastewater composition on biohydrogen production from buffalo slurry co-fermented
hydrogen production. [n.d.]. with cheese whey and crude glycerol, using microbial mixed
[27] Sansonetti S, Curcio S, Calabro V, Iorio G. Bio-ethanol culture. Int J Hydrogen Energy 2015;40:209e18. http://
production by fermentation of ricotta cheese whey as an dx.doi.org/10.1016/j.ijhydene.2014.11.008.
effective alternative non-vegetable source. Biomass [42] Sinha P, Pandey A. An evaluative report and challenges for
Bioenergy 2009;33:1687e92. http://dx.doi.org/10.1016/ fermentative biohydrogen production. Int J Hydrogen Energy
j.biombioe.2009.09.002. 2011;36:7460e78. http://dx.doi.org/10.1016/
[28] Zoppellari F, Bardi L. Production of bioethanol from effluents j.ijhydene.2011.03.077.
of the dairy industry by Kluyveromyces marxianus. N [43] Christophe C, Adler N, Schwitzgue bel J-P, Pe
ringer Paul.
Biotechnol 2013;30:607e13. http://dx.doi.org/10.1016/ Hydrogen production by Clostridium thermolacticum during
j.nbt.2012.11.017. continuous fermentation of lactose. Int J Hydrogen Energy
[29] Pierra M, Trably E, Godon J-J, Bernet N. Fermentative 2004;29:1479e85. http://dx.doi.org/10.1016/
hydrogen production under moderate halophilic conditions. j.ijhydene.2004.02.009.
Int J Hydrogen Energy 2014;39:7508e17. http://dx.doi.org/ [44] Marone A, Massini G, Patriarca C, Signorini A, Varrone C,
10.1016/j.ijhydene.2013.08.035. Izzo G. Hydrogen production from vegetable waste by
[30] Que meneur M, Hamelin J, Barakat A, Steyer J-P, Carre re H, bioaugmentation of indigenous fermentative communities.
Trably E. Inhibition of fermentative hydrogen production by Int J Hydrogen Energy 2012;37:5612e22. http://dx.doi.org/
lignocellulose-derived compounds in mixed cultures. Int J 10.1016/j.ijhydene.2011.12.159.
Hydrogen Energy 2012;37:3150e9. http://dx.doi.org/10.1016/ [45] Hashsham SA, Fernandez AS, Dollhopf SL, Dazzo FB,
j.ijhydene.2011.11.033. Hickey RF, Tiedje JM, et al. Parallel processing of substrate
[31] Marone A, Izzo G, Mentuccia L, Massini G, Paganin P, Rosa S, correlates with greater functional stability in methanogenic
et al. Vegetable waste as substrate and source of suitable bioreactor communities perturbed by glucose. Appl Environ
microflora for bio-hydrogen production. Renew Energy Microbiol 2000;66:4050e7. http://dx.doi.org/10.1128/
2014;68:6e13. http://dx.doi.org/10.1016/j.renene.2014.01.013. AEM.66.9.4050-4057.2000.
[32] Carmona-Martı́nez AA, Trably E, Milferstedt K, Lacroix R, [46] Lee Z-K, Li S-L, Lin J-S, Wang Y-H, Kuo P-C, Cheng S-S. Effect
Etcheverry L, Bernet N. Long-term continuous production of of pH in fermentation of vegetable kitchen wastes on
H2 in a microbial electrolysis cell (MEC) treating saline hydrogen production under a thermophilic condition. Int J
wastewater. Water Res 2015;81:149e56. http://dx.doi.org/ Hydrogen Energy 2008;33:5234e41. http://dx.doi.org/10.1016/
10.1016/j.watres.2015.05.041. j.ijhydene.2008.05.006.
[33] Marone A, Carmona-Martı́nez AA, Sire Y, Meudec E, [47] Guo XM, Trably E, Latrille E, Carrere H, Steyer J-P. Predictive
Steyer JP, Bernet N, et al. Bioelectrochemical treatment of and explicative models of fermentative hydrogen production
table olive brine processing wastewater for biogas from solid organic waste: role of butyrate and lactate
production and phenolic compounds removal. Water Res pathways. Int J Hydrogen Energy 2014;39:7476e85. http://
2016;100:316e25. http://dx.doi.org/10.1016/ dx.doi.org/10.1016/j.ijhydene.2013.08.079.
j.watres.2016.05.008. [48] Zeng A-P, Biebl H, Schlieker H, Deckwer W-D. Pathway
[34] APHA. Standard methods for the examination of water and analysis of glycerol fermentation by Klebsiella pneumoniae:
wastewater. 2005. http://dx.doi.org/10.2105/AJPH.51.6.940-a. regulation of reducing equivalent balance and product
[35] Monlau F, Sambusiti C, Barakat A, Guo XM, Latrille E, formation. Enzyme Microb Technol 1993;15:770e9. http://
Trably E, et al. Predictive models of biohydrogen and dx.doi.org/10.1016/0141-0229(93)90008-P.
biomethane production based on the compositional and [49] Chen S, Hou H, Harnisch F, Patil SA, Carmona-Martinez AA,
structural features of lignocellulosic materials. Environ Sci Agarwal S, et al. Electrospun and solution blown three-
Technol 2012;46:12217e25. http://dx.doi.org/10.1021/ dimensional carbon fiber nonwovens for application as
es303132t. electrodes in microbial fuel cells. Energy Environ Sci
[36] Uggetti E, Sialve B, Latrille E, Steyer JP. Anaerobic digestate as 2011;4:1417. http://dx.doi.org/10.1039/c0ee00446d.
substrate for microalgae culture: the role of ammonium [50] Batlle-Vilanova P, Puig S, Gonzalez-Olmos R, Vilajeliu-
concentration on the microalgae productivity. Bioresour Pons A, Ba??eras L, Balaguer MD, et al. Assessment of biotic
Technol 2014;152:437e43. http://dx.doi.org/10.1016/ and abiotic graphite cathodes for hydrogen production in
j.biortech.2013.11.036. microbial electrolysis cells. Int J Hydrogen Energy
2014;39:1297e305. http://dx.doi.org/10.1016/ [54] Badalamenti JP, Krajmalnik-Brown R, Torres CI. Generation

j.ijhydene.2013.11.017. of high current densities by pure cultures of anode-respiring
[51] Caccavo F, Coates JD, Rossello-Mora RA, Ludwig W, Geoalkalibacter spp. under alkaline and saline conditions in
Schleifer KH, Lovley DR, et al. Geovibrio ferrireducens, a microbial electrochemical cells. MBio 2013;4. http://
phylogenetically distinct dissimilatory Fe(III)-reducing dx.doi.org/10.1128/mBio.00144-13. e00144e13.
bacterium. Arch Microbiol 1996;165:370e6. [55] Tapia-Venegas E, Ramirez-Morales JE, Silva-Illanes F, Toledo-
[52] Ortega-Martı́nez AC, Jua rez-Lo
pez K, Solorza-Feria O, Ponce- Alarco n J, Paillet F, Escudie R, et al. Biohydrogen production
Noyola MT, Galindez-Mayer J, Rinderknecht-Seijas N, et al. by dark fermentation: scaling-up and technologies
Analysis of microbial diversity of inocula used in a five-face integration for a sustainable system. Rev Environ Sci
parallelepiped and standard microbial fuel cells. Int J Biotechnol 2015;14:761e85. http://dx.doi.org/10.1007/s11157-
Hydrogen Energy 2013;38:12589e99. http://dx.doi.org/ 015-9383-5.
10.1016/j.ijhydene.2013.02.023. [56] Sivagurunathan P, Kumar G, Bakonyi P, Kim SH, Kobayashi T,
[53] Katuri KP, Enright A-M, O'Flaherty V, Leech D. Microbial Xu KQ, et al. A critical review on issues and overcoming
analysis of anodic biofilm in a microbial fuel cell using strategies for the enhancement of dark fermentative
slaughterhouse wastewater. Bioelectrochemistry hydrogen production in continuous systems. Int J Hydrogen
2012;87:164e71. http://dx.doi.org/10.1016/ Energy 2016;41:3820e36. http://dx.doi.org/10.1016/
j.bioelechem.2011.12.002. j.ijhydene.2015.12.081.

Coupling Dark Fermentation and Microbial Electrolysis To Enhance Bio-Hydrogen Production From Agro-Industrial Wastewaters and By-Products in A Bio-Refinery Framework

Uploaded by

Copyright:

Available Formats

Coupling Dark Fermentation and Microbial Electrolysis To Enhance Bio-Hydrogen Production From Agro-Industrial Wastewaters and By-Products in A Bio-Refinery Framework

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Coupling Dark Fermentation and Microbial Electrolysis To Enhance Bio-Hydrogen Production From Agro-Industrial Wastewaters and By-Products in A Bio-Refinery Framework

Uploaded by

Copyright:

Available Formats

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 6 ) 1 e1 3

Available online at www.sciencedirect.com

journal homepage: www.elsevier.com/locate/he

Coupling dark fermentation and microbial

Antonella Marone a,*, Olga R. Ayala-Campos b, Eric Trably a,

article info abstract

processes (open-cycles) toward natural closed ecological

potential in DF and fully characterized before and after dark

(Chemical Oxygen Demand) was explained by its sugars con-

tent (from 60 to 80%). Proteins were only present in a minor

COD. Surprisingly, in SW which came from a sugar factory,

sugars represented only 5% of VSS (Volatile Suspended Solids)

ethanol (data not shown). This suggests that this wastewater

had already undergone a natural fermentation process. In VB2

mately 30% of proteins and around 30% of organic acids (in

part lactate) and alcohol, sugars represented only a minor

methods cause of its contents in complex molecules and the

presence of toxic compounds. In PW the contents in reduced

metabolites exceeded the 20% of total COD, including a little

amount of the organic matter is represented by complex

Dark fermentation experiments

Dark fermentation batch experiments were carried out in

500 mL serum flasks with a working volume of 200 mL. To

Fig. 1 e Maximum Bio-Hydrogen production yield (mLH2/

acetate consumption was observed and current started to

2014;39:1297e305. http://dx.doi.org/10.1016/ [54] Badalamenti JP, Krajmalnik-Brown R, Torres CI. Generation

You might also like