ww.sciencedirect.com

i n t e rn a t i o n a l j o u r n a l o f h y d r o g e n en e r g y 4 0 ( 2 0 1 5 ) 8 4 9 8e8 5 0 9

Available online at w

ScienceDirect

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

Evaluation of hydrogen and methane productionfrom sugarcane vinasse in an anaerobic fluidizedbed reactor

Cristiane Marques dos Reis a, Mariana Fronja Carosia b,Isabel Kimiko Sakamoto b, Maria Bernadete Amancio Varesche b,Edson Luiz Silva a,*

a Department of Chemical Engineering, Federal University of S~ao Carlos, Rod. Washington Luis, Km 235, 13565-905

S~ao Carlos, SP, Brazilb Department of Hydraulics and Sanitation, School of Engineering of S~ao Carlos, University of S~ao Paulo, Av.

Trabalhador S~ao Carlense, 400, 13566-590 S~ao Carlos, SP, Brazil

a r t i c l e i n f o

Article history:

Received 18 February 2015

Received in revised form

17 April 2015

Accepted 25 April 2015

Available online 23 May 2015

Keywords:

Distillery wastewaters

Co-fermentation

Hydrogen

Hydraulic retention time

Methane

* Corresponding author. Tel.: þ55 16 3351826E-mail address: edsilva@ufscar.br (E.L. Si

http://dx.doi.org/10.1016/j.ijhydene.2015.04.10360-3199/Copyright © 2015, Hydrogen Energ

a b s t r a c t

This study evaluated the hydrogen and methane production from sugarcane vinasse in an

anaerobic fluidized bed reactor. Two reactors were operated with two different substrate

concentrations: R5 (5 g COD L�1) and R10 (10 g COD L�1). During the first stage, glucose was

used as the primary carbon source; vinasse was then added from 0% to 100% of the organic

source in hydraulic retention time (HRT) of 6 h. Later, HRT was changed to 4, 2 and 1 h.

The best hydrogen production rate was 0.57 L h�1 L�1 (R5, HRT ¼ 1 h, 100% vinasse).

Thebest hydrogen yieldwas 3.07mmolH2g�1 CODadded (R5,HRT¼ 6h, vinasse:glucose¼ 1:3).

Main metabolites were ethanol, butyric acid, propionic acid and methanol. Denaturing

gradient gel electrophoresis analysis identifiedPrevotella sp. andMegasphaera sp. belonging to

the Bacteria domain and Methanobacterium sp. and Methanosphaera sp. belonging to the

Archaea domain.

Copyright © 2015, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights

reserved.

Introduction

Sugar and ethanol industry is one of themain segments of the

agribusiness sector in Brazil. In addition to the sugar and

alcohol, the main products, nearly all of the sub-products of

the industry, including bagasse, molasses and sugarcane

vinasse, are processed [1]. Vinasse, one of the major byprod-

ucts of the ethanol production process with nearly 14 L of

4; fax: þ55 312 16 335182lva).36y Publications, LLC. Publ

vinasse produced per liter of ethanol, can cause extensive

pollution due to its high organic load (up to 40 g COD L�1), and

it is a potential anaerobic digestion source [2].

In general, vinasse is a low pH brown-colored residue con-

taining particulate matter and high concentrations of organic

and inorganic compounds [3]. Phenolic compounds (such as

humicacidand tannicacid), themelanoidins (resulting fromthe

reaction of sugars and proteins by the Maillard reaction),

caramel and the furfural components contribute to its color [4].

66.

ished by Elsevier Ltd. All rights reserved.

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 en e r g y 4 0 ( 2 0 1 5 ) 8 4 9 8e8 5 0 9 8499

Thepresenceof thesecompoundsmakesvinasseacomplexand

difficult compound for degradation.

The anaerobic digestion of vinasse has the potential to

reduce the organic load of this residue while generating bio-

gases such as methane (CH4) or hydrogen (H2). Few studies,

however, focus on the energy recovery in the form of H2.

In general, the production of CH4 and H2 is a two-stage pro-

cess involving separation of the acidogenic and the methano-

genic stages. The combination of H2 and CH4 is more efficient

than CH4 alone when used as combustion engine fuel. In addi-

tion, the CH4 and H2 mixture reduces the emission of green-

house gases such as CO, CO2 and unburnt hydrocarbons [5].

Biological H2 production can be accomplished using simple

or complex substrates.Wastewater fromanumerous industries

including thedairy [6], brewery [7], andethanol-sugar industries

[8e10] can be used for biological hydrogen production.

The presence of inhibitory compounds in the distillery

wastewaters canbeabarrier to theanaerobicdigestionprocess

despite the rich amount of organic matter in the wastewater.

Thus, an adjustment period is advisable to promote favorable

consumption of complex substrates by themicroorganisms in

the environment. When added in small quantities, the micro-

bial culture may adapt to the complex residue or to the inhib-

itory compounds [11]. Thus, the use of a simple co-substrate

can improve the microbial degradation of refractory sub-

stances in complex residues such as stillage. Some authors

[8,12,13] who employed distillery residues have also started

adapting or conditioning the reactor to produce hydrogenwith

simplesubstrates suchasglucoseandsucrose.So far, theeffect

of varying proportions of vinasse-glucose in thermophilic

conditions has been tested [14]. However, a study employing

different vinasse-glucose ratios forhydrogenproduction in the

mesophilic range has not been performed. The aim of this

study was to evaluate the production of H2 and CH4 in two

anaerobic fluidized reactors (AFBR) at room temperature

(22 ± 3 �C) with concentrations of 5 and 10 g COD L�1 obtained

from several ratios of diluted vinasse and glucose. Further-

more, the band profiles of the microbial communities were

investigated using polymerase chain reaction-denaturing

gradient gel electrophoresis (PCR-DGGE) in both the AFBRs.

Materials and methods

AFBR

Two acrylic fluidized bed reactors with the following di-

mensions were used: height 151 cm, diameter 3.1 cm, and

total volume of 1452 mL (Fig. 1). The support material used for

immobilizing the microorganisms was expanded clay: grain

sizes between 2.8 and 3.35 mm, a real density of 1.50 g cm�3,

porosity of 23% and minimum fluidization velocity of

1.24 cm s�1. Along the reactors, four samplerswere distributed

in order to collect support material for microbial

characterization.

Reactor start-up procedure

The inoculum was obtained from an upflow anaerobic sludge

blanket (UASB) reactor used for the treatment of swine

wastewater. Swine manure is known for possessing

hydrogen-producing bacteria. The sludge underwent heat

treatment to activate the acidogenic cells, as described in

methods proposed by Ref. [15] and adapted from Ref. [16].

The reactors R5 and R10 worked with distinct substrate

concentrations of 5 g COD L�1 and 10 g COD L�1, respectively.

Some nutrients were also added according to Leite et al. [17] at

the following concentrations (mg L�1): CO(NH2)2 (125);

NiSO4$6H2O (1); FeSO4 $7H2O (5); FeCl3$6H2O (0.5); CaCl2$6H2O

(47.0); CoCl2$2H2O (0.08); SeO (0.07); KH2PO4 (85.0); K2HPO4

(21.7); Na2HPO4$2H2O (33.4). To maintain the pH of the me-

dium between 4 and 5, hydrochloric acid (30%) and sodium

bicarbonate were used as buffer solutions. Both reactors were

operated in batch mode for 96 h in a closed circuit. After this

activation period, the reactors were operated in continuous

mode. Glucose was the only carbon source in the beginning.

However, vinasse was added to the feed gradually (0, 25, 75

and 100% of g COD L�1) until vinasse replaced glucose in the

feedmaking the total concentration of 5 g COD L�1 or 10 g COD

L�1 during the hydraulic retention time (HRT) of 6 h. When

vinasse was the only carbon source, the HRT was varied be-

tween 4, 2 and 1 h. Table 1 illustrates the experimental stages.

To facilitate the discussion of results, each operation phase

was named according to the vinasse content in the substrate

mixture and the HRTs for which reactors were operated, e.g.,

III (75%, 6 h) indicates a phase III where the reactors were fed

with a substrate mixture containing 75% vinasse and were

operated for an HRT of 6 h.

The vinasse used in this experiment was collected at the

Usina S~aoMartinho distillery plant (Pradopolis, SP, Brazil), and

it was stored frozen. Table 2 illustrates the main physical and

chemical properties of vinasse employed. The raw vinasse

had an average concentration of 40 g COD L�1 which was

diluted to achieve the desired concentration in each reactor:

R5 (5 g COD L�1) and R10 (10 g COD L�1). Both reactors were

operated for 364 days and each stage lasted for 40 days in

average.

Analytical methods

The pH, COD, nitrogen, phosphate, sulfate, magnesium, cal-

cium, potassium and the amount of volatile suspended solids

were analyzed according to the Standard Methods for the

Examination of Water and Wastewater [18]. Analyses of total

volatile acids were performed according to the methodology

proposed by Dilalo and Albertson [19], whereas bicarbonate

alkalinity was estimated as described by Ripley et al. [20]. The

total reducing sugars were measured according to Dubois

et al. [21]. The organic acids and alcohol concentrations were

measured by liquid chromatography (HPLC Shimadzu)

equipped with a pump (LC-10ADVP), an autosampler (SIL-20A

HT), a column oven (CTO-20A) at 43 �C, a refractive index

detector (RID-10A), a system Controller (SCL-10AVP) and col-

umn HPX-87H Aminex (300 mm, 7.8 mm, BioRad). The mobile

phase consisted of H2SO4 (0.01 N) at 0.5 mL min�1. The biogas

hydrogen content was determined by gas chromatography

(GC-2010, Shimadzu, Japan) using a thermal conductivity de-

tector (TCD) with argon as the carrier gas and a Supelco Car-

boxen 1010 Plot packed column (30 m � 0.53 mm i.d.).

Volumetric hydrogen production was measured by the Ritter

Fig. 1 e Schematic description of the mesophilic AFBR.

Table 1 e AFBR operating conditions for H2 and CH4

production from glucose and vinasse in R5 and R10.

Phases Substrate mixture HRT(h)

Glucose/Vinasse

variation

I 100% glucose 6

II 75% glucose; 25% vinasse 6

III 25% glucose; 75% vinasse 6

IV 100% vinasse 6

HRT variation V 100% vinasse 4

VI 100% vinasse 2

VII 100% vinasse 1

i n t e rn a t i o n a l j o u r n a l o f h y d r o g e n en e r g y 4 0 ( 2 0 1 5 ) 8 4 9 8e8 5 0 98500

MilliGas-counter (Type: MGC-1). Hydrogen yield and the

hydrogen production rate were measured as it follows:

� Hydrogen production rate (HPR) ¼ {[Biogas production

measured in the MilliGas-counter (L.h�1)]*(% H2)/[Real vol-

ume of reactor based on bed height of particules (L)]}.

� Hydrogen Yield (HYadded) ¼ {[HPR (L h�1 L�1)]/[Organic

Loading Rate (g L�1 h�1)*22.4 L mol�1]}.

Table 2 e Main physical and chemical properties of rawsugarcane vinasse.

Parameters Vinasse

pH 5.0 ± 0.18

COD (mg L�1) 42,818 ± 4391

Total volatile acids (mg L�1) 3012 ± 512

Total alkalinity (mg CaCO3 L�1) 292 ± 38

Volatile suspended solids (mg L�1) 287 ± 72

Total Kjeldahl nitrogen (mg L�1) 244 ± 36

Phosphate (mg L�1) 3796 ± 387

Sulfate (mg L�1) 1400 ± 130

Magnesium (mg L�1) 580 ± 52

Calcium (mg L�1) 757 ± 151

Potassium (mg L�1) 4500 ± 670

Molecular analysis of the microbial community

At the end of each operation stage of the reactors R5 (5 g COD

L�1) and R10 (10 g COD L�1), particle samples were collected

through the samplers distributed laterally along the reactors.

To disengage the biomass attached to the support material,

the particles were washed three times with PBS solution

(NaCl, KCl, Na2POH4, KH2PO4) and centrifuged for 3 min at

3000 rpm.

The extraction of genomic DNA from the samples was

performed by cell lysis [22]. For this procedure we used glass

beads, phenol, chloroform and buffer solution. The 16S rRNA

fragments were amplified by PCR. The primers used were the

following: 968FGC-1401R for the Domain Bacteria [23], and

Parch519fGC-Arch915r for the Domain Archaea [24]. The PCR

amplification product was subjected to electrophoresis in

DGGE for both domains, Bacteria and Archaea, in order to

separate the amplified DNA fragments. The gels contained

45e60% linear denaturing gradient (100% denaturant of 7 M

urea and 40% (v/v) of deionized formamide). Electrophoresis

was run under 75 V and 60 �C for 16 h. The band profiles were

visualized in a chamber with a transilluminator Eagle Eye III

TM (Stratagene) at 254 nm and the images were captured by

Eagle Vista. The BioNumerics 2.5 software was used to

calculate the similarity coefficient (Pearson correlation) from

the band profiles. Clustering was performed using the UPGMA

(unweighted Pair Group Method using Average).

The best condition for biogas production was obtained

from cloning and sequencing the 16s rRNA for which we used

primer 27F-1100R [25] for the Bacteria Domain and primer 21F

Arch958R-for the Archaea Domain [24]. The vector pGEM®

Easy Vector System (Promega) was used and transformed into

competent cells of Escherichia coli. The cloned DNA was then

subjected to PCR using the M13F and M13R primers.

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 en e r g y 4 0 ( 2 0 1 5 ) 8 4 9 8e8 5 0 9 8501

The PCR products were sent to Macrogen® (www.

macrogen.com) for the analysis of the nucleotide sequences.

The sequences were compared with the data from the Na-

tional Center for Biotechnology Information (NCBI) database

Ribosomal Database Project and http://blast.ncbi.nlm.nih.gov/

(http://rdp.cme.msu.edu/seqmatch). The phylogenetic trees

were developed using the MEGA 6.0 software using the

“neighbor-joining” method in a Bootstrap analysis method

with 1000 replications. The sequences from this study were

deposited in GenBank with accession numbers ranging from

KM820898 to KM820906.

Fig. 3 e Biogas content throughout the operation phases in

R10 (10 g COD L¡1).

Results and discussion

Biogas production

Figs. 2 and 3 show the biogas content for the reactors R5 (5 g

COD L�1) and R10 (10 g COD L�1), respectively. Both reactors

produced only H2 and CO2 till phase III (75%, 6 h). However,

CH4 was generated after vinasse became the only carbon

source.

After phase V (100%, 4 h), we observed a reduction in the H2

content along with HRT reduction and reached a minimum

value of 5% for the reactor R5 (5 g COD L�1). At the HRT 1 h, the

reactor stopped producing CH4 in R5 so the content of H2

reestablished and reached 34%. In reactor R10 (10 g COD L�1),

the CH4 production persisted despite the reduction of HRT,

even under turbulent conditions as the HRT 1 h.

Maintaining the pH in acidic conditions and performing

heat treatment of the inoculum are methods of inactivating

the methanogens [26]. In general, methanogenic archaea are

inhibited below pH 6. However, it was verified that H2-con-

sumers are more tolerant to acidic conditions than other CH4-

producers [16].

In this study, there was a variation in biogas composition

when the carbon source was changed from a mixture of

glucose/vinasse (phases I (0%, 6 h), II (25%, 6 h), III (75% 6 h)) to

a composition of vinasse (phase IV (100%, 6 h)).

Xia et al. [27] showed that the mixture cellulose/xylose led

to CH4 production, whereas a mixture cellulose/glucose

Fig. 2 e Biogas content throughout the operation phases in

R5 (5 g COD L¡1).

generated only H2 and CO2 as the biogases. Moreover, Cha-

ganti et al. [28] found that the use of the mixture glucose/

xylose resulted in a better result for H2 production than the

use of separatemediums of glucose and xylose. Oh andMartin

[29] found that the availability of glucose may facilitate or

impede the presence of certain populations in anaerobic

digestion of ethanol stillage. Thus, high values of glucose

decrease the presence of hydrogenotrophic microorganisms,

while facilitating the presence of acetoclastic microorgan-

isms. In contrast, low glucose availability leads to the devel-

opment of hydrogenotrophic archaea. In the present work,

CH4 was produced when glucose was removed from feed.

Some studies using distillery wastewater also verified the

production of CH4 and H2. Espinoza-Escalante et al. [30] re-

ported the production of CH4 with H2 from stillage processing

during the production of tequila with pH below 4.5. Buitron

et al. [8] reported the production of CH4 (35e44%) alongwithH2

at HRT 24 h. CH4 production stopped at HRT 12 h. Possibly, the

methanogenswerewashed reactor hydrodynamics under this

condition. Searmisirimongkol et al. [9] also produced CH4

under an acidogenic condition. It was found that the higher

the applied organic load, the lower the percentage of CH4

generated. In this study, the production of CH4 in reactor R5

ceased when HRT was changed from 2 to 1 h.

Despite being a limiting factor for the production of H2, this

production of CH4 can be valuable. For example, the mixture

of 20% by volume of H2 and CH4, referred to as hythane, is

patented by Hydrogen Consultants Inc [5]. Furthermore, ac-

cording to Bauer and Forest [5], the presence of H2 in the

natural gas used as vehicle fuel helps the engine performance

and reduces the greenhouse gas emissions. Thus, the conjoint

generation of the biological CH4 and H2 can serve as a new

source for this fuel mixture.

Figs. 4 and 5 illustrate the hydrogen production rate (HPR)

and hydrogen yield (HY) across the different phases. In both

reactors, the HPR decreasedwith the addition of vinasse in the

feed. This indicates that the reactors were affected by the

presence of a new substrate.

The HY showed an irregular behavior from phases I to IV.

In R5 (5 g COD L�1), HY was maximum when vinasse was

75% of the organic load, reaching 3.07 mmol H2 g COD�1added

Fig. 4 e HPR and HY throughout the operation phases in R5 (5 g COD L¡1).

i n t e rn a t i o n a l j o u r n a l o f h y d r o g e n en e r g y 4 0 ( 2 0 1 5 ) 8 4 9 8e8 5 0 98502

(14.74 mmol H2 g COD�1removed). It can be seen that glucose

acted in a positive way for the consumption of organic

matter. However, because of the CH4 production, H2 yield

dropped significantly in phase IV (100%, 6H), reaching

0.3 mmol H2 g COD�1added (4.87 mmol H2 g COD�1

removed). In

R10 (10 g COD L�1) there was a peak in stage II (25%, 6H),

1.32 mmol H2 g COD�1added (6.60 mmol H2 g COD�1

removed),

and due to the appearance of CH4 there was a reduction of

approximately 50% in stage IV (100%, 6 h). In stage VII (100%,

1 h), CH4 production ceased and HY increased to approxi-

mately 1.96 mmol H2 g COD�1added. This indicates a good

performance of AFBR for the H2 production from sugarcane

vinasse when CH4 production is inhibited. The best yield

when the reactor is operated solely with vinasse was

1.96 mmol H2 g COD�1added.

HPR in reactor R10 (10 g COD L�1) was gradually decreased

as the HRT was decreased. CH4 production continued even in

low HRT. Similar to R5, HY in R10 showed a downward trend

with decreasing HRT. The minimum value was obtained in

phase VII (100%, 1 h), reaching 1.96 mmol H2 g COD�1added.

The results of the present study are in accordance with the

literature on H2 production from complex wastes. Gahde et al.

[31] reported a yield of 10.95 mmol H2 g COD�1removed from

dairy wastewater. Compared with the present study, during

Fig. 5 e HPR and HY throughout the ope

the best conditions for biogas generation, a yield of

18.37 mmol H2 g COD�1removed in phase VII (100% vinasse, 1 h)

in reactor R5 was obtained.

A comparison between some studies using distillery

wastewaters for H2 production is shown in Table 3.

Intermediate metabolites production

Reactor R5 and R10 presented a similar trend in relation to

metabolites. The distribution values can be found in the

Supplementary material Table S1 and Table S2, respectively.

In phase I (0%, 6 h), when glucose was the only carbon source,

there was a preponderance of ethanol and butyric acid,

together amounting to more than 85% of the metabolites

produced. In Phase II (25%, 6 h), with the addition of vinasse

there was a shift in metabolic pathways. Ethanol production

ceased and a predominance of propionic acid production was

observed. When vinasse became the main substrate in phase

III (75%, 6 h), ethanol was again produced by the reactor and

methanol appeared as the predominant metabolite. From

phase IV to VII, both reactors presented mixed fermentation.

Recent studies [34e36] observed that ethanol can be pro-

duced concurrently with H2 without any loss in the produc-

tivity of the reactor. An alternative route (Equation (1)) was

ration phases in R10 (10 g COD L¡1).

Table 3 e Comparison of H2 production from distillery wastewaters in various studies.

Reactor HY Temperature (�C) Metabolites Reference

ASBRa 7.98 mmol H2 g COD�1d 35 HAc, HPr, HBu, HiBu, EtOH, Acetone [8]

ASBRa 6.77 mmol H2 g COD�1removed

d 37 HAc, HPr, HBu, HVal [9]

ASBRa 4.83 mmol H2 g COD�1removed

d 55 HBu, HVal, HAc, HPr, EtOH [10]

Batch 2.31 mmol H2 g COD�1added 55 HBu, HAc, HPr,EtOH [32]

APBRb 0.7 mmol H2 g COD�1added 55 HAc, HBu, EtOH, HPr [33]

AFBRc 2.23 mmol H2 g COD�1added 55 HLa, HBu, HiBu, HPr [14]

AFBRc 1.96 mmol H2 g COD�1added

18.37 mmol H2 g COD�1removed

22 MeOH, EtOH, HBu, HMal, HVal, HPr This study

a ASBR-anaerobic sequencing batch reactor.b APBR-anaerobic packed bed reactor.c AFBR-anaerobic fluidized bed reactor.d Based on data from article.

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 en e r g y 4 0 ( 2 0 1 5 ) 8 4 9 8e8 5 0 9 8503

launched [37] as a possible deviation followed during the

acidogenic process for producing H2 and ethanol:

C6H12O6 þH2O/C2H5OHþ CH3COOHþ 2H2 þ 2CO2 (1)

Propionic acid is harmful for anaerobic digestion because it

consumes the H2 generated by the system. In fact, during

Phase II (25%, 6 h) there was a decrease in HY. This can also be

attributed to the introduction of vinasse. According to equa-

tion (2), for each mole of propionic acid generated, there are

2 mol of H2 consumed:

C6H12O6 þ 2H2/2CH3CH2COOHþ 2H2O (2)

Yu et al. [38] also found considerable amounts of propionic

acid in an acidogenic system producing H2 from rice wine

vinasse. The authors showed 40% propionic acid among the

metabolites generated under pH 4, which decreased as the pH

was increased to 6 (20%).

Methanol was produced in large proportions when vinasse

was introduced to the system. Methanol can be produced

when CH4 is generated and oxidized by ruminant animals,

stabilization ponds, lakes and landfills. Furthermore, meth-

anol, formate and formaldehyde are intermediates in the

process of CH4 oxidation according to the equation (3) [39].

Thus, methanol is then used for the growth of methylotrophic

microorganisms and is readily converted into formaldehyde

dehydrogenase and the formate. There is little information

regarding methanol production such that both the process of

production and consumption are favorable, and preferably in

liquid media [39].

CH4/CH3OH/HCHO/HCOOH/CO2 (3)

In relation to the metabolites generated, the use of single

substrates such as glucose favored the production of butyrate.

While the use of a complex wastewater such as vinasse

favored the production of organic acids and other solvents

(mixed fermentation).

Composition of the microbial communities

In Fig. 6 a lower similarity coefficient of 55% was found for

biomass of phase I (0%, 6 h) compared with other phases of

operation for the populations of bacteria. At this stage, only

glucose was used as organic source at HRT 6 h. It was found

that between phase II (25%, 6 h) and phase III (75%, 6 h), the

similarity coefficient increased to 83%, probably due to the

presence of a new substrate, vinasse (25%), favoring the se-

lection of better adaptedmicroorganisms. Relative to phase IV

(100%, 6 h), there was a decrease in the similarity between the

populations in comparison with phase III, where a similarity

of 76% was observed. Possibly because of the use of vinasse as

the only carbon source, new microorganisms were favored.

Reactor R10 presented a similar behavior as compared to

reactor R5. The lowest similarity coefficient was obtained in

phase R10 I, where it had a value of 67% over the following

phases. The introduction of vinasse in the system led to an

increase in the coefficient of similarity, reaching 87% between

phases II and III, which were the most similar among the

samples. Between Phase IV and V, the coefficient of similarity

decreased to 84%, but stayed greater than 80%. Between

phases VI and VII the similarity coefficient was 77%. It was

found that increasing vinasse concentration in phases II, III

and IV, as well reducing HRT from 6 to 1 h in phases V, VI and

VII resulted in a decrease in the similarity. This is an indica-

tion that there may have been changes in bacterial pop-

ulations caused bywashing in lowHRT or by adaptation to the

presence of a more complex substrate such as vinasse.

Fig. 7a shows the bands profile of the reactor R5 (5 g COD

L�1) for the Archaea domain. At the beginning of the reactor

operation, the similarity between this phase and the other

was 70% and was the most distinctive of the operation. As the

operating phases were changed, there was an increase in

similarity between phases, indicating that there was a selec-

tion of Archaea as a function of the operating conditions

imposed. The greatest similarity in the latter stages was 95%,

which coincided with the highest CH4 production in the

reactor. The band profile for the reactor R10 (Fig. 7b) shows

that Phase II presents a similarity (55%) lower than the

following operational phases. During this step vinasse was

introduced in the system. Possibly, owing to a shock in this

population, only the most adapted were maintained under

this condition. From Phase IV, there are similarities ranging

from 77 to 97% between phases I, VI, VII and V, which may

indicate the adaptation of the microbial population to a

complex substrate such as vinasse.

Samples from phase VII (100%, 1 h) of reactor R5 were

chosen to be subjected to cloning and sequencing of frag-

ments of 16S rRNA gene, once it reached the best stage for H2

Fig. 6 e Coefficient of similarity by Pearson correlation of DGGE profile bands in relation to the operational steps applied in

reactors R5 (a) and R10 (b) for the domain Bacteria.

i n t e rn a t i o n a l j o u r n a l o f h y d r o g e n en e r g y 4 0 ( 2 0 1 5 ) 8 4 9 8e8 5 0 98504

production with no CH4 production. For Bacteria Domain, 164

clones belonging to the phylum Firmicutes and Bacteroidetes

were obtained. Clones were grouped into five operational

taxonomic units (OTUs), as shown in Fig. 8.

Among all of the sequences obtained and shown in Table 4,

55% belonged to the phylum Bacteroidetes and uncultured

Prevotella, and 28% belonged to the phylum Firmicutes genus

Megasphaera. The presence of 3% of uncultured Clostridia also

belonging to the phylum Firmicutes was observed.

OTUs 1 and 2 were identified as belonging to the phylum

Bacteroidetes. OTU 1were related to uncultured Prevotella. The

Fig. 7 e Coefficient of similarity by Pearson correlation of DGGE

reactors R5 (a) and R10 (b) for the domain Archaea. NOTE: R10 II

populations were below the detection limit of the technique.

microorganisms affiliated to Prevotella spp. are non-

endospore-forming bacteria and strictly anaerobic. In gen-

eral, the Bacteroidia class, towhich Prevotella sp. is affiliated, is

characterized by the ability to grow on a variety of carbohy-

drates, and it can convert cellulose into acetate, propionate

and succinic acid [40]. During sucrose fermentation, succinic

acid is generated. Lactic acid and acetic acid are produced in

minor quantities. Some species can produce H2 at low con-

centrations when the process is directed to the acetate-

pathway [41]. In biogas production from food waste in batch

reactor, Shin et al. [42] also found Prevotella sp. According to

profile bands in relation to the operational steps applied in

I - this sample could not be amplified. Probably the archeal

Fig. 8 e Phylogenetic tree based on the consensus sequences of the microbial consortium clones with primers for Domain

Bacteria. Methanosarcina acetivorans was used as outgroup.

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 en e r g y 4 0 ( 2 0 1 5 ) 8 4 9 8e8 5 0 9 8505

the study, this species can produce acetic acid, succinic acid,

isobutyric acid and isovaleric acid from the degradation of

starch, sucrose and glucose. Li et al. [43] found the presence of

Prevotella spp. in fixed bed reactors used for H2 production

from sucrose at HRT 2 h. Won et al. [44] verified that Prevotella

spp. was present in greatest proportions at a pH of approxi-

mately 4.5, similar to the present study. The authors verified a

satisfactory H2 production (2.18 L H2 L�1) and hydrogen yield

(1.29 mol�1 H2 mol sucrose). Mariakakis et al. [41] observed

that Prevotella sp. wasmore frequent by increasing the organic

load under low HRT when producing H2 from hexose. Similar

to the present study, the best conditions for H2 production

occurred at the phase of the highest organic load, indicating

that these species is able to produce H2 in conditions of high

organic load. The cited studies demonstrated the feasibility of

H2 production by Prevotella sp. concomitant with organic acids

in a suitable environment, i.e., with a pH approximately 4.5

and availability of organic matter, such as observed in this

study with sugarcane vinasse.

In the present study, Megasphaera sp. accounted for 28% of

microorganisms found (OTU 2). According to Marchadin et al.

[45], these microorganisms belong to the phylum Firmicutes

and are able to metabolize carbohydrates producing lactate,

succinic acid, butyric acid, valeric acid and caproic acid. Acetic

acid and propionic acid can also be produced in low amounts.

Furthermore, this specie is sensitive to low pH and high con-

centrations of ethanol.

Castell�o et al. [46] used non-sterilized cheese whey in a

UASB reactor and pointedMegasphaera sp. as a possible source

of microorganisms for H2 production. However, it can also

convert lactic acid into propionic acid. The authors also found

Table 4 e 16S rRNA clone analysis of Bacteria.

OTU Organism affiliation Identity(%)

Phylum GeAcce

1 Uncultured bacterium 97 Bacteroidetes GQ4

Uncultured Prevotella sp. 96 JX57

2 Uncultured Prevotellaceae

bacterium

100 Bacteroidetes JF80

3 Megasphaera sp. 98 Firmicutes HM9

4 Uncultured bacterium 98 Firmicutes JQ07

5 Uncultured Clostridia bacterium 99 Firmicutes EU8

Prevotella sp. Similarly, in the present study, the authors found

out that H2 productionwas reduced due to the CH4 generation.

On the other hand, the H2 remained constant, showing that

thesemicroorganisms are also capable of generating H2. In the

study by Ohnishi et al. [47], Megasphaera sp. could produce H2

in low amounts, besides the organic acids cited. Santos et al.

[14] used vinasse from sugarcane produced in the same dis-

tillery plant of the present study (S~aoMartinho, Prad�opolis, SP,

Brazil) and verified the presence of the following microor-

ganisms: Lactobacillus sp. and Megasphaera sp. in thermophilic

AFBR for H2 production.

OTUs 4 and 5 were also related to the Firmicutes phylum.

OTU 5 is represented as an Uncultured Clostridia Bacterium

and these microorganisms corresponded to 3% of clones

present in this sample. Microorganisms belonging to the

Clostridia class are gram-positive and endospore-forming.

Microorganisms belonging to this class are able to degrade a

variety of organic substrates producing acetate, propionate

and butyrate. In minor proportions, these microorganisms

can also generate formate, lactate, succinate and caproate

[48,49].

The presence of these bacteria related to Clostridium spp. is

common in the process of H2 production from complex waste,

such as the vinasse that was used in the present study, and

are observed in other studies in the literature [13,32,50].

Table 5 illustrates the fermentation products generally

generated by the predominant microorganisms found in the

present study according to literature. Propionic acid and lactic

acid are among these products. In the present work, the

concentration (721 ± 360 mg L�1) of propionic acid from phase

II (25%, 6 h) was the highest among the metabolites.

nBankssion n�

Number of sequence Sequencelength (pb)

Relativeabundance

(%)

77873.1 11 399 7

5984.1

6757.1 91 321 55

90965.2 46 525 28

2158.1 11 576 7

87973.1 5 471 3

Table 5 e Comparison among the fermentation products from microorganisms predominantly found in R5 reactor(5 g COD L¡1).

Organism affiliation Relativeabundance (%)

Fermentationproducts

Substrate Reference

Uncultured Prevotellaceae

bacterium

55 HSu,HAc,HLa,H2a Sucrose [41]

Megasphaera sp. 28 HBu, HVal, HAc, HPr, H2a Food waste [47]

a Low amounts.

i n t e rn a t i o n a l j o u r n a l o f h y d r o g e n en e r g y 4 0 ( 2 0 1 5 ) 8 4 9 8e8 5 0 98506

Megasphaera sp. is able to synthesize propionate from lactic

acid, increasing the quantity of this acid in the system [46]. A

decrease in lactic acid concentration was verified in both re-

actors while the concentration of propionic acid increased in

the system.

In Phase IV (100%, 1 h), the reactor R5 resulted in a wide

distribution of metabolites in terms of soluble microbial

products (SMP) such as propionic acid (14.37% SMP), isobutyric

acid (13.09% SMP) butyric acid (11.95% SMP), propionic acid

(10.71% SMP), malic acid (9.99% SMP), valeric acid (6.37% SMP),

acetic acid (6.34% SMP) and lactic acid (1% SMP). These me-

tabolites are consistent with the major metabolites generated

by the microorganisms in the reactor R5 (Table 5).

Methanol (22% SMP) and ethanol (12.26% SMP) were also

produced in large quantities, but they are not reported as

fermentation products in Table 5. Some species relating to this

class are able to generate methanol under the conditions

shown in present study, for example, Clostridium butyricum in

pectin fermentation [49]. As already mentioned, microorgan-

isms affiliated to the Clostridia class were also observed in the

samples of present study.

Zeikus and Schink [51] conducted an experiment to

demonstrate the degradation of pectin by various microbial

cultures such as C. butyricum. It was found that the major

fermentation product from this strain was methanol. The

control sample with glucose in the absence of methanol led to

generation of other metabolites. In contrast with the present

study, vinasse sample can contain pectin residue in its

composition as a result of bagasse processing in the ethanol

production stage. Then, pectin can be used by certain Clos-

tridium species for methanol formation.

Table 6 shows the 16S rRNA clone analysis of Archaea

found in this sample. For the Domain Archaea, four OTUS

were identified (Fig. 9). OTUs 1, 2 and 3 were related to un-

cultured archaeon affiliated to genus Methanobacterium. These

microorganisms are hydrogenotrophic methanogens, i.e.,

they use CO2 and H2 to form CH4 [52]. As expected, the

presence of this microorganism is responsible for CH4 pro-

duction, and consequently the decrease in H2 production in

Table 6 e 16S rRNA clone analysis of Archaea.

OTU Organism affiliation Identity(%)

Genus

1 Uncultured Methanobacterium sp. 99 Methanobacterium

2 Uncultured archaeon 97 Methanobacterium

3 Uncultured archaeon 98 Methanobacterium

4 Uncultured Methanosphaera sp.

Uncultured archaeon

94 Methanosphaera

the present study. The production of CH4 arises from the

consumption of 4 mol of H2 and 1 mol of CO2 releasing CH4

and H2O.

In a study employing synthetic stillage, Rodriguez et al. [53]

also noted the presence of microorganisms belonging to the

genus Methanobacterium in low sulfate concentrations (higher

ratio of substrate/sulfate). In contrast, when there was a

smaller substrate/sulfate ratio, these organisms disappeared

from the reactor. Rosa et al. [54] found this microbial culture

(Methanobacterium) in a study using cheese whey as organic

substrate for H2 production in a fluidized bed reactor. Simi-

larly, in the present study, the presence of CH4 occurred at low

pH of approximately 4.5.

OTU 4 was related to the uncultured Methanosphaera,

accounting for 1% of the representative sequences. This

archaea is capable of reducing CH4 to methanol using H2. It is

present in animal gut. The presence of such microorganism

explains the appearance of CH4 in the reactor due to the large

presence of methanol in the system. Kim et al. [55] also noted

the presence of Methanosphaera stadtmanae in sludge from the

treatment of swine manure.

Low pH (approximately 4.5) and low HRT (1 h) are harsh

conditions for the CH4 formation by methanogenic archaea.

However, this adverse condition can be offset by other con-

ditions such as the provision of substrate, favoring the

development of this species. The presence of important me-

tabolites can lead to the growth of these microorganisms in

the reactor, beyond the duration of operation of the reactors,

up to 200 days, favoring the adaptation of the methanogenic

archaea in the reactor.

Conclusions

From this study, it was possible to verify the feasibility of

producing biogas from sugarcane vinasse using anaerobic

fluidized-bed reactors. However, the maintenance of a low pH

(4e5) was not sufficient to inactivate the methanogenic

organisms.

GenBankaccession n�

Number ofsequence

Sequencelength (pb)

Relativeabundance (%)

KC533586.1 81 385 94

EU447673 1 540 1

KF848976.1 3 480 3

AB935212.1

KJ676678.1

1 360 1

Fig. 9 e Phylogenetic tree based on the consensus sequences of the microbial consortium clones with primers for Domain

Archaea. Uncultured Prevotella was used as outgroup.

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 en e r g y 4 0 ( 2 0 1 5 ) 8 4 9 8e8 5 0 9 8507

The use of glucose as an auxiliary substrate was effective

as a co-substrate for H2 production. In general, the maximum

H2 production in the concentration of 5 g COD L�1 and under

the hydraulic retention time of 1 h: 0.57 L h�1 L�1 and

1.96 mmol H2 g COD�1removed was obtained when vinasse was

the only organic source.

Mainmetabolites were ethanol, butyric acid andmethanol.

Microbial characterization identified Prevotella sp. and Mega-

sphaera sp. belonging to the domain Bacteria and Meth-

anobacterium sp. and Methanosphaera sp. belonging to the

domain Archaea.

Acknowledgements

The authors gratefully acknowledge the financial support of

CNPq e National Council for Scientific and Technological

Development, CAPES e Coordination for the Improvement of

Higher Education Personnel, and FAPESPe S~ao Paulo Research

Foundation.

Appendix A. Supplementary data

Supplementary data related to this article can be found at

http://dx.doi.org/10.1016/j.ijhydene.2015.04.136.

r e f e r e n c e s

[1] Goldemberg J. The Brazilian biofuels industry. BiotechnolBiofuels 2008;1:1e7.

[2] Siqueira LM, Damiano ESG, Silva EL. Influence of organicloading rate on the anaerobic treatment of sugarcane vinasseand biogas production in fluidized bed reactor. J Environ SciHeal A 2013;48:1707e16.

[3] Pant D, Adholeya A. Biological approaches for treatment ofdistillery wastewater: a review. Bioresour Technol2007;98:2321e43.

[4] Wilkie AC, Riedesel KJ, Owens JM. Stillage characterizationand anaerobic treatment of ethanol stillage fromconventional and cellulosic feedstocks. Biomass Bioenerg2000;19:63e102.

[5] Bauer CG, Forest TW. Effect of hydrogen addition on theperformance of methane-fueled vehicles. Part 1: effect on S.I.engine performance. Int J Hydrogen Energy 2001;26:55e70.

[6] Yang P, Zhang R, McGarvey JA, Benemann JR. Biohydrogenproduction from cheese processing wastewater by anaerobicfermentation using mixed microbial communities. Int JHydrogen Energy 2007;32:4761e71.

[7] Shi XY, Jin DW, Sun QY, Li WW. Optimization of conditionsfor hydrogen production from brewery wastewater byanaerobic sludge using desirability function approach.Renew Energy 2010;35:1493e8.

[8] Buitr�on G, Carvajal C. Biohydrogen production from tequilavinasses in an anaerobic sequencing batch reactor: effect ofinitial substrate concentration, temperature and hydraulicretention time. Bioresour Technol 2010;101:9071e7.

[9] Searmsirimongkol P, Rangsunvigt P, Leethochavalit M,Chadej S. Hydrogen production from alcohol distillerywastewater contanaining high potassium and sulfate usinganaerobic sequencing batch reactor. Int J Hydrogen Energy2011;36:12810e21.

[10] Intanoo P, Rangsunvigit P, Namprohm W,Thamprajamchit B, Chavadej J, Chavadej S. Hydrogenproduction from alcohol wastewater by an anaerobicsequencing batch reactor under thermophilic operation:nitrogen and phosphorous uptakes and transformation. Int JHydrogen Energy 2012;37:104e12.

[11] Camarillo R, Rinc�on J. Effect of inhibitory compounds on thetwo-phase anaerobic digestion performance of dilutedwastewaters from the alimentary industry. Chem Eng J2012;193e194:68e76.

[12] Mohan SV, Mohanakrishna G, Ramanaiah SV, Sarma PN.Simultaneous biohydrogen production and wastewatertreatment in biofilm configured anaerobic periodicdiscontinuous batch reactor using distillery wastewater. Int JHydrogen Energy 2008;35:550e8.

[13] Nasr N, Elbeshbishy E, Hafez H, Nakkla G, Naggar MHE.Biohydrogen production from thin stillage usingconventional and acclimatized anaerobic digester sludge. IntJ Hydrogen Energy 2001;36:2761e9.

[14] Santos SC, Rosa PRF, Sakamoto IS, Varesche MBA, Silva EL.Organic loading rate impact on biohydrogen production anmicrobial communities at anaerobic fluidized thermophilicbed reactors treating sugarcane vinasses. Bioresour Technol2014;159:55e63.

[15] Maintinguer SI, Fernandes BS, Duarte ICS, Saavedra NK,Adorno MAT, Vareche MBA. Fermentative hydrogenproduction with xylose by Clostridium and Klebsiella species inanaerobic batch reactors. Int J Hydrogen Energy2011;36:3508e17.

[16] Kim IS, Hwang MH, Jang N, Hyun SH, Lee ST. Effect of low pHon the activity of hydrogen utilizing methanogen in bio-hydrogen process. Int J Hydrogen Energy 2006;29:133e40.

[17] Leite JAC, Fernandes BS, Pozzi E, Barboza M, Zaiat M.Application of an anaerobic packed-bed bioreactor for the

i n t e rn a t i o n a l j o u r n a l o f h y d r o g e n en e r g y 4 0 ( 2 0 1 5 ) 8 4 9 8e8 5 0 98508

production of hydrogen and organic acids. Int J HydrogenEnergy 2008;33:579e86.

[18] Standard methods for the examination of water andwastewater. 20th ed. Washington, D.C., USA: AmericanPublic Health Association/American Water WorksAssociation/Water Environment Federation; 1998.

[19] Dilallo R, Albertson OE. Volatile acids by direct titration. JWater Pollut Control Fed 1961;33:356e65.

[20] Ripley LE, Boyle WC, Converse LC. Improved alkalimetricmonitoring for anaerobic digestion of high-strength wastes. JWater Pollut Control Fed 1986;58:406e11.

[21] Dubois M, Gilles KA, Hamilton JK, Rebers PA, Smith F.Colorimetric method for determination of sugars and relatedsubstances. Anal Chem 1956;28:350e6.

[22] Griffiths RR, Whiteley AS, O'Donnell AG, Bailey MJ. Rapidmethod for coextraction of DNA and RNA from naturalenvironments for analysis of ribosomal DNA- andrRNAbased microbial community composition. Appl EnvironMicrobiol 2000;66:5488e91.

[23] Nubel U, Engelen B, Felske A, Snaidr J, Wieshuber A,Amann RI, et al. A sequence heterogeneities of genesencoding 16S rRNAs in Paenibacillus polymyxa detected bytemperature gradient gel electrophoresis. J Bacteriol1996;178:5636e43.

[24] Vissers E, Bodelier PLE, Muyzer G, Laanbroek HJ. A nestedPCR approach for improved recovery of archaeal 16SrRNAgene fragments from fresh water samples. FEMS MicrobiolLett 2009;298:193e8.

[25] Lane DJ. 16S/23S rRNA sequencing. In: Stackebrandt E,Goodfellow M, editors. Nucleic acid techniques in bacterialsystematics. New York: John Wiley&Sons; 1991. p. 115e75.

[26] Oh S, Van Ginkel S, Logan BE. The relative effectiveness of pHcontrol and heat treatment for enhancing bio-hydrogen gasproduction. Environ Sci Technol 2003;37:5186e90.

[27] Xia Y, Cai L, Zhang T, Fang HHP. Effects of substrate loadingand co-substrates on thermophilic anaerobic conversion ofmicrocrystalline cellulose and microbial communitiesrevealed using high-throughput sequencing. Int J HydrogenEnergy 2012;37:807e12.

[28] Chaganti SR, Kim DH, Lalman JA. Impact of oleic acid on thefermentation of glucose and xylose mixtures to hydrogenand other byproducts. Renew Energy 2012;42:60e5.

[29] Oh ST, Martin AD. Glucose contents in anaerobic ethanolstillage digestion manipulate thermodynamic driving forcein between hydrogenophilic and acetoclastic methanogens.Chem Eng J 2014;243:526e36.

[30] Espinoza-Escalante FM, Ortiz CP, Corona JN, Garcia YG,Bories A, Pulido HG. Anaerobic digestion of the vinasses fromthe fermentation of agave tequilana weber to tequila: theeffect of pH, temperature and hydraulic retention time onthe production of hydrogen and methane. BiomassBioenergy 2009;33:14e20.

[31] Gahde A, Sonawane S, Varna MN. Kinetic analysis ofbiohydrogen production from complex dairy wastewaterunder optimized condition. Int J Hydrogen Energy2014;39:1306e14.

[32] Lazaro CZ, Perna V, Etchebehere C, Varesche MBA.Sugarcane vinasse as substrate for fermentative hydrogenproduction: the effects of temperature and substrateconcentration. Int J Hydrogen Energy 2014;39:6407e18.

[33] Ferraz Junior ADN,Wenzel J, Etchebehere C, Zaiat M. Effect oforganic loading rate on hydrogen production from sugarcanevinasse in thermophilic acidogenic packed bed reactors. Int JHydrogen Energy 2014;39:16852e62.

[34] Wu KJ, Chang CF, Chang JS. Simultaneous production ofbiohydrogen and bioethanol with fluidized-bed and packed-bed bioreactors containing immobilized anaerobic sludge.Process Biochem 2007;42:1165e71.

[35] Reis CM, Silva EL. Effect of upflow velocity and hydraulicretention time in anaerobic fluidized-bed reactors used forhydrogen production. Chem Eng J 2011;172:28e36.

[36] Barros AR, Silva EL. Hydrogen and ethanol production inanaerobic fluidized bed reactors: performance evaluation forthree support materials under different operatingconditions. Biochem Eng J 2012;61:59e65.

[37] Ren N, Wang B, Huang J. Ethanol type fermentation fromcarbohydrate in high rate acidogenic reactor. BiotechnolBioeng 1997;54:428e33.

[38] Yu H, Zhu Z, Hu W, Zhang H. Hydrogen production from ricewinery wastewater in an upflow anaerobic reactor by usingmixed anaerobic cultures. Int J Hydrogen Energy1994;19:807e12.

[39] Galbally IE, Kirstine W. The production of methanol byflowering plants and the global cycle of methanol. J AtmosChem 2002;43:195e229.

[40] Goud RK, Raghavulu SV, Mohanakrishna G, Naresh K,Mohan SV. Predominance of Bacilli and Clostridia in microbialcommunity of biohydrogen producing biofilm sustainedunder diverse acidogenic operating conditions. Int JHydrogen Energy 2012;37:4068e76.

[41] Mariakakis I, Bischoff P, Krampe J, Meyer C, Steinmetz H.Effect of organic loading rate and solids retention time onmicrobial population during bio-hydrogen production bydark fermentation in large lab-scale. Int J Hydrogen Energy2011;36:10690e700.

[42] Shin HS, Youn JH, Kim SH. Hydrogen production from foodwaste in anaerobic mesophilic and thermophilicacidogenesis. Int J Hydrogen Energy 2004;29:1355e63.

[43] Li C, Zhang T, Fang HHP. Fermentative hydrogen productionin packed-bed and packing-free upflow reactors. Water SciTechnol 2006;54:95e103.

[44] Won SG, Baldwin SA, Lau AK, Rezadehbashi M. Optimaloperational conditions for biohydrogen production fromsugar refinery wastewater in an ASBR. Int J Hydrogen Energy2013;38:13895e906.

[45] Marchadin H, Juvonen R, Haikara A. In: Bergey's manual ofsystematic bacteriology. Volume 3 Firmicutes. 2nd ed., vol.4. Baltimore, Md.: Williams and Wilkins Co.; 2010.p. 1082e90.

[46] Castello E, Garcia y Santos C, Iglesias T, Paolino G, Wenzel J,Borzaconni L, et al. Feasibility of biohydrogen productionfrom cheese whey using a UASB reactor: links betweenmicrobial community and reactor performance. Int JHydrogen Energy 2009;34:5674e82.

[47] Ohnishi A, Bando Y, Fujimoto N, Suzuki M. Development of asimple bio-hydrogen production system through darkfermentation by using unique microflora. Int J HydrogenEnergy 2010;35:8544e53.

[48] Laothanachareona T, Kanchanasutab S, Mhuanthonga W,Phalakornkuleb C, Pisutpaisalb N, Champredaa V. Analysis ofmicrobial community adaptation in mesophilic hydrogenfermentation from food waste by tagged 16S rRNA genepyrosequencing. J Environ Manage 2014;144:143e51.

[49] Ljungdahl LG, Hugenholtz J, Wiegel J. Acetogenic and acid-producing clostridia biotechnology handbooks 3. 1989.p. 145e91.

[50] Kalamaskar LB, Dhakephalkar PK, Meher KK, Ranade DR.High biohydrogen yielding Clostridium sp. DMHC-10 isolatedfrom sludge of distillery waste treatment plant. Int JHydrogen Energy 2009;35:10639e44.

[51] SchinkB,Zeikus JG.Microbialmethanol formation:amajorendproduct of pectin metabolism. Curr Microbiol 1980;4:387e9.

[52] Demirel B, Scherer P. The roles of acetotrophic andhydrogenotrophic methanogens during anaerobicconversion of biomass to methane: a review. Rev Environ SciBiotechnol 2008;7:173e90.

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 en e r g y 4 0 ( 2 0 1 5 ) 8 4 9 8e8 5 0 9 8509

[53] Rodriguez E, Lopes A, Fdz-Polanco M, Stams AJM,Garcia-Encina P. Molecular analysis of the biomass offluidized bed reactor treating synthetic vinasse atanaerobic and micro-aerobic conditions. EnvironBiotechnol 2011;93:2181e91.

[54] Rosa PRF, Santos SC, Sakamoto IK, Varesche MBA, Silva EL.Hydrogen production from cheese whey with ethanol-type

fermentation: effect of hydraulic retention time on themicrobial community composition. Bioresour Technol2014;161:10e9.

[55] Kim W, Cho K, Lee S, Hwang S. Comparison of methanogeniccommunity structure and anaerobic process performancetreating swine wastewater between pilot and optimized labscale bioreactors. Bioresour Technol 2013;145:48e56.