Factors affecting the longterm stability of biomass and hydrogen productivity in outdoor photofermentation

ww.sciencedirect.com

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 3 7 ( 2 0 1 2 ) 1 6 4 2 1e1 6 4 2 9

Available online at w

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

Amelioration of photofermentative hydrogen production frommolasses dark fermenter effluent by zeolite-based removalof ammonium ion

Dominic Deo Androga a,*, Ebru Ozgur b, Inci Eroglu b, Ufuk Gunduz c, Meral Yucel c

aDepartment of Biotechnology, Middle East Technical University, Ankara 06800, TurkeybDepartment of Chemical Engineering, Middle East Technical University, Ankara 06800, TurkeycDepartment of Biological Sciences, Middle East Technical University, Ankara 06800, Turkey

a r t i c l e i n f o

Article history:

Received 11 November 2011

Received in revised form

24 February 2012

Accepted 28 February 2012

Available online 28 March 2012

Keywords:

Dark fermentation

Photofermentation

Ammonium

Clinoptilolite

Rhodobacter capsulatus

* Corresponding author. Tel.: þ90 312 210 26E-mail address: [email protected]

0360-3199/$ e see front matter Copyright ªdoi:10.1016/j.ijhydene.2012.02.177

a b s t r a c t

One of the challenges in the development of integrated dark and photofermentative bio-

logical hydrogen production systems is the presence of ammonium ions in dark fermen-

tation effluent (DFE). Ammonium strongly inhibits the sequential photofermentation

process, and so its removal is required for successful process integration. In this study, the

removal of ammonium ions from molasses DFE using a natural zeolite (clinoptilolite) was

investigated. The samples were treated with batch suspensions of Na-form clinoptilolite.

The ammonium ion concentration could be reduced from 7.60 mM to 1.60 mM and from

12.30 mM to 2.40 mM for two different samples. Photofermentative hydrogen production

on treated and untreated molasses DFE samples were investigated in batch photo-

bioreactors by an uptake hydrogenase deleted (hup�) mutant strain of Rhodobacter capsu-

latus. Maximum hydrogen productivities of 1.11 mmol H2/Lc$h and 1.16 mmol H2/Lc$h and

molar yields of 79% and 90% were attained in the treated DFE samples, while the untreated

samples resulted in no hydrogen production. The results showed that ammonium ions in

molasses DFE could be effectively removed using clinoptilolite by applying a cost-effective,

simple batch process.

Copyright ª 2012, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights

reserved.

1. Introduction biological hydrogen production processes can be improved by

Biological hydrogen production processes, namely bio-

photolysis, dark fermentation and photofermentation, offer

the prospect of producing hydrogen from renewable sources

derived from waste streams and agricultural residues. Low

hydrogen yields and production rates, however, are the major

barriers in developing these technologies [1]. Nevertheless,

studies have demonstrated that the performance of these

96; fax: þ90 312 210 2600.(D.D. Androga).2012, Hydrogen Energy P

integrating them in 2- or 3-stage processes [2e5].

One widely investigated integrated system is the sequen-

tial dark and photofermentation system [6]. In the dark

fermentation stage, anaerobic bacteria like the Clostridium

species break down carbohydrate-rich substrates (sugars) to

produce H2, CO2 and short-chain organic acids such as acetic

acid and butyric acid Eq. (1). The organic acids produced in the

dark fermentation stage can be used in a sequential

ublications, LLC. Published by Elsevier Ltd. All rights reserved.

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 3 7 ( 2 0 1 2 ) 1 6 4 2 1e1 6 4 2 916422

photofermentation stage where photoheterotrophic bacteria

convert them to hydrogen and carbon dioxide Eq. (2) [1,6]. The

photofermentation process occurs under anaerobic and

nitrogen limited conditions and utilizes light energy to supply

ATP that is needed in the complete conversion of the organic

acids to hydrogen and carbon dioxide [1].

C6H12O6þ2H2O/2CH3COOHþ4H2þ2CO2 DG : �206 kJ (1)

2CH3COOHþ4H2O/8H2þ4CO2 DG: 104:6 kJ�2¼209:2 kJ (2)

Theoretically, presuming acetic acid to be the major by-

product of the dark fermentation process, 12 mol of H2

(4 mol of H2 from dark fermentation þ 8 mol of H2 from pho-

tofermentation) can be produced per mol of glucose in the

integrated processes. However, actual yields are lower

because part of the glucose is used for microbial growth

and maintenance [2,6]. Approximately 9 mol of H2/mol of

glucose (75% conversion efficiency) can be attained in an

integrated thermophilic fermentation and photofermentation

process [2].

One of the major challenges in these systems is the pres-

ence of fixed nitrogen in the dark fermentation feed, which

leads to the formation of ammonium ions that are inhibitory

to the photofermentation process [7]. Ammonium ions are

known to repress the nitrogenase enzyme which is respon-

sible for catalyzing hydrogen production in the photosyn-

thetic bacteria [8]. High ammonium concentrations inhibit the

nitrogenase enzyme either directly via feedback inhibition

(reversible) or at the genetic level by the repression of nif genes

that encode subunits of the nitrogenase enzyme [9]. An

ammonium ion concentration not exceeding 2mM in the dark

fermenter effluent has been reported to be suitable for pho-

tofermentative hydrogen production [3,4,8].

Several alternatives have been suggested to address the

issue of inhibitory ammonium concentrations in the dark

fermenter effluent. The commonly applied method is dilution

of the effluent to reduce ammonium concentration

[3,4,10e13]. This technique is effective in reducing ammo-

nium ion inhibition, but it also reduces the concentration of

other essential components in the feed and thus may cut

down hydrogen production efficiency. Also, dilution incurs

extra costs, which are particularly unfavorable for large scale

systems.

Another suggestion is the use of ammonium ion tolerant

photosynthetic bacteria (wildtype or mutant). Various

researchers have applied genetic modification to develop

mutant bacterial strains that can grow and produce hydrogen

at high ammonium concentrations [9,14]. However, great

metabolic and genetic diversity exists among the purple non

sulfur bacteria (PNS) such that further studies are required to

isolate or develop strains that can grow and produce hydrogen

at high rates and yields under high ammonium ion concen-

trations. Moreover, the use of mutant strains in industrial

settings can be problematic and expensive [7].

Another method of circumventing the inhibitory effects of

ammonium ions on photofermentative hydrogen production

is the use of wastewater treatment methods. A variety of

biological and physicochemical methods and technologies

such as air stripping, breakpoint chlorination, nitrification-

denitrification and ion exchange are commonly applied to

remove ammonium from wastewater [15]. Among these

techniques, ion exchange using natural zeolites is more

attractive because of its low cost and simplicity of application

and operation. Moreover, the zeolites can be regenerated for

reuse as fertilizer or animal feed supplement [16]. Previous

studies by Eroglu et al. [17] have demonstrated the successful

application of olive mill wastewater treatment using clay and

zeolite for photofermentative hydrogen production.

Clinoptilolite is a zeolite mineral with an aluminosilicate

tetrahedral crystalline structure that contains exchangeable

cations such as Kþ, Naþ, Mg2þ and Ca2þ. It has a Si/Al > 4 and

its structure is: (Na,K,Mg)6(Al6Si30O7)$24H2O [18]. Owing to its

sorptive properties, clinoptilolite has been applied in the

removal of heavy metals [19], dye contaminants [20] and

ammonium [15,18,21] from wastewater. It was reported to

selectively adsorb cations in the order: Pb2þ > NHþ4 >

Ba2þ > Cu2þ > Zn2þ > Cd2þ > Co2þ and have a maximum

exchange capacity of 2.6 mEq/g [22].

This study aimed to investigate the possibility of using

natural clinoptilolite zeolite to reduce the ammonium ion

content in molasses dark fermenter effluent that has high

ammonium concentrations. The molasses dark fermenter

effluent was treated with zeolite in batch processes, and

parameters such as ammonium ion concentration, pH levels,

color change, and duration of treatment were investigated.

The resulting effluents were analyzed for both metal cation

content and organic acid concentrations and tested for pho-

tofermentative hydrogen production by Rhodobacter capsulatus

YO3 (hup�).

2. Materials and methods

2.1. Clinoptilolite zeolite pre-treatment

Clinoptilolite samples used in this study were taken from

Manisa-Gordes, Western Anatolia region of Turkey. The

zeolite samples were ground and sieved between mesh 30/60

(0.25 to 0.60 mm) particle size and pre-treated with brine

solution. Pre-treatment aimed to replace the exchangeable

cations in the zeolite (Kþ, Mg2þ, Ca2þ) with the more easily

removable ones such as Naþ ions, therefore creating a near

homoionic form. This improved the zeolite’s effective

exchange capacity, subsequently providing better perfor-

mance in ion exchange application [22]. The clinoptilolite

samples were batch pre-treated using the procedure applied

by Bayraktaroglu [23]. Mixtures of 50 g of clinoptilolite and

500 ml of 1 M NaCl solution were stirred using a wrist-action

shaker (Fisher-Kendall mixer) at ambient temperature for 13

days. The NaCl exchange solution was replaced with freshly

prepared 1 M NaCl at 24 h intervals. The pre-treatment

procedure was stopped when the cation concentration (Kþ,Ca2þ and Mg2þ) in the exchange solution effluent reached

saturation. The mixture was then decanted and the zeolite

washed using deionizedwater to get rid of NaCl traces. During

washing, presence of Naþ was checked using 1% (v/v) AgNO3

solution, then it was dried in an oven set at 70 �C for 24 h Table

1 lists the chemical composition of the zeolite before and after

pre-treatment. The analyses were carried out using lithium

metaborate fusion dissolution method [24].

Table 1 e The chemical composition of the Gordesclinoptilolite zeolite.

Component Original (% w/w) Na-form (% w/w)

SiO2 68.12 68.06

Al2O3 11.15 11.10

CaO 1.70 0.40

MgO 1.50 0.90

Fe2O3 3.47 2.46

Na2O 2.32 5.54

K2O 1.98 0.99

Othersa 9.76 10.55

LOIb 10.06 10.14

a Mixture of MnO, TiO2, P2O5, SO3 and H2O.

b Loss if ignition.

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 3 7 ( 2 0 1 2 ) 1 6 4 2 1e1 6 4 2 9 16423

2.2. Ammonium ion reduction in the molasses DFE

Molasses DFEs were delivered by DLO-FBR (Wageningen UR

Food and Biobased Research, The Netherlands). The samples

M1 and M2 were the effluents of two separate thermophilic

dark fermentation processes [3] which were carried out with

different ammonium chloride concentrations. Composition of

the effluents is shown in Table 2. The ammonium concen-

trations of M1 and M2 were 7.60 mM and 12.30 mM,

respectively.

The molasses DFE samples (200 ml) were treated with Na-

form clinoptilolite zeolite (10 g) in an Erlenmeyer flask placed

in a water bath shaker (Clifton NE25) at 25 �C and 70 rpm.

During treatment, ammonium concentration, pH and color of

the liquid samples were measured and recorded at 30 min

intervals.

2.3. Bacteria culture and media

R. capsulatus YO3 (hup�), an ‘uptake hydrogenase’ deleted

strain of R. capsulatus MT1131 [25] was used in this study. The

inoculums were prepared in modified Biebl and Pfennig (BP)

media [26]. Compositions of treated and untreated molasses

DFEs are given in Table 2. The hydrogen production

Table 2 e Comparisons of the change in the organic acidconcentrations, cation concentrations, total nitrogen(TN), total organic carbon (TOC) and the chemical oxygendemand (COD) of the two molasses dark fermentereffluent samples before (M1 and M2) and after treatment(M1T and M2T) with the Na-form clinoptilolite zeolite.

Molasses DFE sample M1 M1T M2 M2T

Ammonium (mM) 7.60 1.60 12.30 2.30

Mg2þ(mg/L) 54.72 70.25 70.75 61.44

Fe3þ(mg/L) 1.87 0.75 1.98 0.52

Lactic acid (mM) 4.40 4.30 1.10 0.90

Formic acid (mM) 0.50 0.40 4.00 4.00

Acetic acid (mM) 85.40 82.60 98.00 95.00

Butyric acid (mM) 1.40 1.30 13.00 13.00

TN (M) 0.026 0.020 0.027 0.016

TOC (M) 0.078 0.078 0.079 0.078

C/N molar ratio 3.00 3.90 2.93 4.88

COD (mg/L) 3380 3340 3180 3120

experiments were carried out using two times diluted

molasses DFE samples. Samples were centrifuged and steril-

ized by autoclaving to remove contaminants and any colloidal

materials that may interfere with light penetration into the

photobioreactors. Untreated molasses DFE that contained

high ammonium concentrations were used as controls. The

DFE sampleswere supplementedwith iron (Fe citrate, 0.1mM)

andmolybdenum (Na2MoO4$2H2O, 0.16 mm) and 5mM sodium

carbonate (Na2CO3) buffer. Hydrogen production experiments

were carried out in duplicate glass bottle photobioreactors

(55 ml) inoculated with freshly grown bacterial cultures (10%).

The photobioreactors were maintained at 30e32 �C in an

incubator. Continuous illumination was provided by 60 W

tungsten lamps adjusted to provide a uniform light intensity

of 170W/m2 at the surface of the reactors. The initial pH in the

photobioreactors was 6.6. Hydrogen production was followed

by water displacement method, using calibrated glass

columns.

The performance of hydrogen production was analyzed

based on the maximum hydrogen production rate and

maximum molar yield. The hydrogen production rate (mmol/

Lc/h) was determined as the number of moles of hydrogen

produced in a given time period by the bioreactor culture

volume. The percent molar yield (%) was calculated as the

ratio of the moles of hydrogen produced from the experiment

to the moles of theoretical hydrogen that would have been

produced if all the acetic acid utilized had been converted to

hydrogen [27].

2.4. Analytical methods

During the clinoptilolite pre-treatment procedure, the

concentrations of Kþ, Mg2þ and Ca2þ were determined using

atomic absorption spectroscopy (Philips, PU9200X) and Naþ

ions using a flame photometer (Jenway Model PFP7). The

ammonium ion concentration, total nitrogen (TN), total

organic carbon (TOC) and chemical oxygen demand (COD)

were determined spectrophotometrically using their respec-

tive Hach-Lange kits and DR/2400 Hach-Lange Spectropho-

tometer, Germany. The bacterial cell concentration, organic

acid analyses, evolved gas analyses and pH were measured as

previously described [27]. The color of the molasses DFE was

analyzed using a DR/2400 Hach-Lange spectrophotometer.

3. Results and discussion

3.1. Pre-treatment of the clinoptilolite zeolite

The sodium content of the clinoptilolite zeolite was success-

fully increased from 2.32% to 5.54% after pre-treatment (Table

1). The exchangeable cations in the clinoptilolite zeolite (Kþ,

Mg2þ and Ca2þ) decreased as theywere replacedwith Naþ ions

from the exchange solution. No further ion exchange was

observed after 10 days of pre-treatment (Fig. 1). Bayraktaroglu

[23] also reported that 10 days were adequate to attain Na-

form clinoptilolite applying the same batch pre-treatment

procedure.

Fig. 1 e The change in the concentration of exchangeable

cations in the clinoptilolite zeolite during pre-treatment

with 1 N brine solution. (A) KD, (-) Mg2D and (:) Ca2D.

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 3 7 ( 2 0 1 2 ) 1 6 4 2 1e1 6 4 2 916424

3.2. Treatment of molasses dark fermenter effluent

Table 2 summarizes ammonium ion, organic acid, cation

concentrations, total nitrogen (TN), total organic carbon

(TOC), carbon to nitrogenmolar ratio and the chemical oxygen

demand (COD) of the molasses dark fermenter effluent

samples (M1 and M2) and the treated samples with the Na-

form clinoptilolite zeolite (M1T and M2T).

Ammonium ion concentration decreased from 7.60 mM to

1.60 mM in sample M1 and from 12.30 to 2.30 mM in sample

Fig. 2 e The variation of ammonium ion concentration, pH and

with different initial NHD4 concentrations using Na-form clinop

([NHD4 ]Initial [ 12.30 mM). (-) Ammonium, (,) pH and (D) color

M2 (Fig. 2A and B). The duration of treatment varied depend-

ing on the initial NHþ4 concentration; sampleM1 took a shorter

time (120 min) compared to sample M2 (180 min). Treatment

was stopped when the NHþ4 concentration decreased to

around 2 mM, which was the threshold concentration for

photofermentative hydrogen production using R. capsulatus

YO3 (hup�) [3]. High ammonium ion removal efficiency

(around 80%) was achieved for both samples. Successful NHþ4

removal obtained in this study confirmed the high selectivity

and suitability of using clinoptilolite for ammonium ion

removal as previously reported in literature [15,18,21]. The

presence of organic compounds in dark fermenter effluent

might enhance the ammonium ion removal by reducing the

surface tension of the aqueous phase to the point of raising

access of the aqueous phase to the macropores of the zeolite

[28].

The real exchange capacity (REC) of the Na-form clinopti-

lolite was determined as 2 mg NHþ4 =g and 3 mg NHþ

4 =g, for

samples M1 and M2, respectively. These values are compa-

rable to the adsorption capacity (2e30 mg NHþ4 =g) of clinopti-

lolite reported for ammonium in literature [29]. However, they

were lower than the calculated theoretical exchange capacity

(TEC) of the pre-treated zeolite which was 45.54 mg NHþ4 =g.

The difference between the REC and TEC could be attributed to

incomplete removal of cations from the zeolite during treat-

ment. That might occur either due to strong bonding of

cations on zeolite structure or interference of competing ions

color during treatment of molasses dark fermenter effluent

tilolite. (A) M1 ([NHD4 ]Initial [ 7.60 mM) and (B) M2

.

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 3 7 ( 2 0 1 2 ) 1 6 4 2 1e1 6 4 2 9 16425

existing in the molasses DFE samples. Also, parameters such

as the cation concentration, pH, temperature and agitation

speed may affect ion exchange [18,21,22].

During the treatment process, pH ranged between 7.0 and

7.6 in sample M1 and between 8.8 and 9.0 in sample M2. In

literature, pH 6e7 was suggested for optimal ammonium

removal [30], however pH was not significant for the removal

of ammonium ion from molasses DFE.

Negligible color change was observed during the treatment

of the DFEs. Color ranged between 2090 and 2120 PtCo APHA

and 2560 and 2780 PtCo APHA in the M1 and M2 samples,

respectively (Fig. 2A and B).

Elemental analysis of themolasses DFE samples before and

after treatment revealed that some of the cations (Kþ, Mg2þ,Ca2þ, Fe3þ) were released from the zeolite while others were

taken up from the molasses DFE. After treatment the

concentration of Mg2þ increased in sample M1, while it

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

6.0

6.4

6.8

7.2

7.6

8.0

8.4

8.8

9.2

9.6

10.0

0 1 2 3 4 5 60.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

4.5

5.0

Bio

mas

s (g

DC

W/L

c)pH

Cum

ulat

ive

volu

me

of H

2 pro

duce

d (L

/Lc)

Tim

Fig. 3 e Biomass growth (A), pH change (B) and cumulative hyd

during photofermentation on treated and untreated molasses d

times. The initial ammonium concentrations of the treated sam

untreated samples (DM1 (B) and DM2 (D)) were 3.8 mM and 6.2

decreased in sample M2 (Table 2). The variation in the uptake

and release of the cations by the zeolite could be attributed to

the interactions between the exchangeable cations and the

differences in pH of the samples. It was reported that pH

affected the surface charges on the zeolite, therefore

impacting cation exchange [31]. Under high pH, OH� ions

could form complexeswith heavymetals, while under low pH,

Hþ ions tend to competewith cations for exchange sites on the

zeolite, thereby decreasing its efficiency [32]. Generally, pH

range between 3 and 11 is applied in metal cation removal

processes using clinoptilolite [33].

It was also observed that the amount of Fe (III) in the

molasses DFE decreased after treatment with clinoptilolite e

circa 62% and 76% in the M1 and M2 samples, respectively

(Table 2). Iron is an important co-factor of the nitrogenase

enzyme that is responsible for hydrogen production in the

photosynthetic bacteria. Supplementation of molasses DFE

7 8 9 10 11 12 13

e (days)

B

A

C

rogen production (C) by Rhodobacter capsulatus YO3 (hupL)

ark fermenter effluents. The samples were diluted two

ples (DM1T(,) and DM2T(>)) were 0.2 mM and 1.2 mM,

mM, respectively.

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 3 7 ( 2 0 1 2 ) 1 6 4 2 1e1 6 4 2 916426

with Fe was reported to improve hydrogen yield and produc-

tivity [3,4]. In practice, although the addition of Fe to the

molasses DFE is to be carried out after the removal of

ammonium using clinoptilolite (before photofermentative

hydrogen production), further investigations are needed to

determine treatment conditions that would prevent Fe uptake

during the ammonium removal process. This would reduce

the amount of Fe to be supplemented to the DFE and save on

raw materials costs in an industrial setting.

The ammoniumremoval procedure using clinoptilolite had

negligible effect on the organic acid concentrations, TOC and

CODof themolassesDFEs (Table 2). Clinoptilolitewas reported

to have a high affinity for cations like ammonium and heavy

metals like lead and cadmium [16,31e33]. This is a desirable

trait for the treatment of the DFEs. The inhibitory effects of

0

10

20

30

40

50

0

10

20

30

40

50

Con

cent

rati

on o

f or

gani

c ac

id (

mM

)

0

10

20

30

40

50

0 1 2 3 4 5 6

0

10

20

30

40

50

Time (d

Fig. 4 e Organic acid and ammonium ion utilization by Rhodoba

molasses dark fermenter effluents. (A) DM1T ([NHD4 ]Initial [ 0.2

([NHD4 ]Initial [ 3.8 mM); (D) DM2 ([NHD

4 ]Initial [ 6.2 mM). (C) Acetic

ammonium ion.

ammonium ions and heavy metals on photofermentation can

be reduced, therefore improving hydrogen production rates

and yield. TN in themolasses DFE decreased as a consequence

of the NHþ4 removal during treatment (Table 2).

3.3. Photofermentation

The hydrogen production experiments were carried out using

two times dilutedmolasses DFE samples. Dilution was carried

out to adjust the concentration of acetic acid in the DFE

samples to around 40 mM, which is reported to be optimum

for photofermentative hydrogen production [34e36]. Higher

biomass concentrations were obtained in the batch experi-

ments using the treated molasses DFE samples (DM1T and

DM2T) compared to the untreated ones (DM1 and DM2). On

0

1

2

3

4

5

6

7 8 9 10 11 12 13

ays)

012345678

Concentration of am

monium

B

A

C

D

cter capsulatus YO3 (hupL) on treated and untreated

mM); (B) DM2T ([NHD4 ]Initial [ 1.2 mM); (C) DM1

acid, (>) formic acid, (6) butyric acid ( ) lactic acid and (:)

Fig. 5 e The average daily variation in total gas

composition during photofermentation by Rhodobacter

capsulatus YO3 (hupL) on treated molasses dark fermenter

effluents DM1T and DM2T with initial ammonium

concentrations 0.2 mM and 1.2 mM, respectively. H2 (-)

and CO2 (,).

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 3 7 ( 2 0 1 2 ) 1 6 4 2 1e1 6 4 2 9 16427

average, maximum biomass concentrations of 0.7 gdcw/Lcand 0.4 gdcw/Lc were observed in the treated and untreated

samples, respectively (Fig. 3A). A rapid increase in pH was

observed as the bacteria grew. In the experiments using

samples DM1T and DM2T, pH increased from 6.6 to around 7.8

on the second day of the experiment and then fluctuated

between 7.6 and 8.1 during the rest of the experiments

(Fig. 3B). However, in the experiments using samples DM1 and

DM2, pH sharply increased from 6.6 to around 9 on the second

day of the experiment. Then, it ranged between 8.4 and 9.6

(Fig. 3B). The high pH (>9) could have led to the low bacterial

growth observed in the untreated DFE samples. During pho-

tofermentation, bacterial cells maintain their membrane

potential by the efflux of negatively charged OH� ions from

the cell to counteract the effect of taking in negatively charged

organic acids [3]. Therefore, high pH in the feed could prevent

the cell from maintaining its membrane potential, therefore

affecting cellular metabolism and eventually hindering cell

growth. PNS bacteria such as R. capsulatus are reported to

optimally grow between pH 6e9 [37].

The findings in this study also showed that in the presence

of high ammonium concentrations (as in samples DM1 and

DM2), 5 mM Na2CO3 buffer used in the photofermentation

experiments was inadequate to maintain pH at tolerable

levels (pH 7e7.5). Further increase in buffer concentration

using 10, 15 and 20 mMNa2CO3 or 22 mM KH2PO4 gave similar

results (data not shown). Hence, higher buffering capacity is

required to maintain pH at the desirable levels.

Maximum hydrogen productivities of 1.11 mmol H2/Lc/h

and 1.16 mmol H2/Lc/h and molar yields of 79% and 90% were

obtained in the experiments using samples DM1T and DM2T,

respectively. No hydrogen production was observed in the

DM1 and DM2 samples (Fig. 3C). This could be because of the

inhibition of the nitrogenase enzyme by the initially high

ammonium concentrations (3.8 mM and 6.2 mM NHþ4 ). In

experiments using batch cultures of Rhodobacter sphaeroides

and growth media containing different NH4Cl concentrations

(1 to 10 mM), Akkose et al. [8] observed no hydrogen produc-

tion in NH4Cl concentration above 2 mM. They attributed this

to the inhibition of nitrogenase by ammonium. Similarly,

Ozgur et al. [3] observed no hydrogen production in batch

cultures of R. capsulatus hup� fed with sugar beet molasses

DFE containing ammonium >2 mM. However, they obtained

hydrogen productivity of 1.37mmol H2/Lc/h andmolar yield of

58% inmolasses DFE samples lacking ammonium. The results

obtained in this study compare to that in literature

[1,3,4,6,10e12].

Organic acid analysis results showed that acetic acid was

well utilized for growth and hydrogen production by the

bacteria in the DM1T and DM2T containing photobioreactors

compared to their DM1 and DM2 counterparts (Fig. 4). In the

DM1T and DM2T experiments, hydrogen production stopped

after the depletion of acetic acid on the 6th day of the exper-

iment (Fig. 4A and B). In contrast, no hydrogen productionwas

observed and around 15 to 20 mM of acetic acid remained at

the end of the DM1 and DM2 experiments (Fig. 4C and D).

Lactic acid and propionic acid were consumed in all the pho-

tobioreactors while butyric acid was only depleted in low

ammonium containing photobioreactors. Accumulation of

formic acid, ranging between 15 and 20mMwas also observed

in the low NHþ4 photobioreactors (Fig. 4C and D). It could have

either been produced by NAD-dependent formate dehydro-

genase which reversibly catalyzes formic acid synthesis from

carbon dioxide and NADH, Hþ or as a fermentation end

product through pyruvate by the action of pyruvate formate

lyase. Avcioglu et al. [11] also observed formic acid production

in fed-batch cultures of R. capsulatus in experiments using

molasses DFE lacking NHþ4 ions.

The analyses of the evolved gas showed that at the

beginning of the experiment (during exponential growth) the

gas composition was 98% hydrogenwith the rest being carbon

dioxide. The gas quality decreased as the bacteria approached

stationary phase (86% hydrogen) and ranged between 71 and

76% hydrogen for the rest of the experiment (Fig. 5).

4. Conclusions

Ammonium ions in molasses dark fermenter effluent can be

effectively removed using natural clinoptilolite zeolite by

applying a cost-effective batch process. The treatment was

successfully carried outwith pH levels ranging from 7 to 9. The

reduction of ammonium ions from the dark fermenter

effluent samples led to the attainment of a high hydrogen

productivity and molar yield in the photofermentation

experiments. This is a significant improvement in integrating

dark and photofermentation processes for biohydrogen

production, which is a sustainable way of producing hydrogen

from renewable resources such as biomass.

Acknowledgments

This research study was supported by the EU 6th Framework

Integrated Project 019825 (HYVOLUTION). The authors would

like to thank Dr Truus de Vrije from DLO FBR (Wageningen UR

Food and Biobased Research, The Netherlands), for providing

the molasses dark fermenter effluent samples, Dr. Yavuz

Ozturk from GMBE TUBITAK-MAM, Gebze, Turkey for

providing the R. capsulatus YO3 (hup�) mutant strain used in

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 3 7 ( 2 0 1 2 ) 1 6 4 2 1e1 6 4 2 916428

this study and Kerime Guney from the Chemical Engineering

Department at the Middle East Technical University, for

carrying out the elemental analysis of the molasses dark

fermenter effluent samples and the component analysis of the

natural clinoptilolite zeolite. Dominic Deo Androga acknowl-

edges the Scientific and Technological Research Council of

Turkey (TUBITAK-BIDEB) for providing financial support

through the PhD Fellowships for Foreign Citizens (Code 2215)

program.

Acronyms

BP Biebl and Pfennig

COD Chemical oxygen demand, mg/L

DFE Dark fermentation effluent

DM1 Two times diluted molasses dark fermenter effluent

sample containing 3.80 mM ammonium ion

DM2 Two times diluted molasses dark fermenter effluent

sample containing 6.20 mM ammonium ion

DM1T Treated (ammonium ion reduced) and two times

diluted molasses dark fermenter effluent sample

containing 0.80 mM ammonium ion

DM2T Treated and two times diluted molasses dark

fermenter effluent sample containing 1.20 mM

ammonium ion

gdcw Gram dry cell weight

hup� Membrane bound uptake hydrogenase deficient,

(mutant)

Lc Liter culture

LOI Loss of ignition

M1 Molasses dark fermenter effluent sample containing

7.60 mM ammonium

M2 Molasses dark fermenter effluent sample containing

12.30 mM ammonium

M1T Treated molasses dark fermenter effluent sample

containing 1.60 mM ammonium ion

M2T Treated molasses dark fermenter effluent sample

containing 2.30 mM ammonium ion

NADH Reduced nicotinamide adenine dinucleotide

NHþ4 Ammonium ion

PNS Purple non sulfur

PtCoAPHA Color unit based on the standards of American

Public Health Association. One color unit is equal

to 1 mg/l platinum as chloroplatinate ion

REC Real exchange capacity, mg NHþ4 =g

TEC Theoretical exchange capacity, mg NHþ4 =g

TOC Total organic carbon, M

TN Total nitrogen, M

r e f e r e n c e s

[1] Argun H, Kargi F. Bio-hydrogen production by differentoperational modes of dark and photofermentation: anoverview. Int J Hydrogen Energy 2011;36:7443e59.

[2] Claassen PAM, de Vrije T, Koukios E, van Niel E, Eroglu I,Modigell M, et al. Non-thermal production of pure hydrogenfrom biomass: HYVOLUTION. J Clean Prod 2010;18:S4e8.

[3] Ozgur E, Mars AE, Peksel B, Louwerse A, Yucel M, Gunduz U,et al. Biohydrogen production from beet molasses bysequential dark and photofermentation. Int J HydrogenEnergy 2010;35:511e7.

[4] Afsar N, Ozgur E, Gurgan M, Akkose S, Yucel M, Gunduz U,et al. Hydrogen productivity of photosynthetic bacteria ondark fermenter effluent of potato steam peels hydrolysate.Int J Hydrogen Energy 2011;36:432e8.

[5] Lo YC, Chen CY, Lee CM, Chang JS. Sequential dark-photofermentation and autotrophic microalgal growth for high-yield and CO2-free biohydrogen production. Int J HydrogenEnergy 2010;35:10944e53.

[6] Keskin T, Abo-Hashesh M, Hallenbeck PC. Photofermentativehydrogen production from wastes. Bioresour Technol 2011;102:8557e68.

[7] Guwy AJ, Dinsdale RM, Kim JR, Massanet-Nicolau J,Premier G. Fermentative biohydrogen production systemsintegration. Bioresour Technol 2011;102:8534e42.

[8] Akkose S, Gunduz U, Yucel M, Eroglu I. Effects of ammoniumion, acetate and aerobic conditions on hydrogen productionand expression levels of nitrogenase genes in Rhodobactersphaeroides O.U.001. Int J Hydrogen Energy 2009;34:8818e27.

[9] Pekgoz G, Gunduz U, Eroglu I, Yucel M, Kovacs K, Rakhely G.Effect of inactivation of genes involved in ammoniumregulation on the biohydrogen production of Rhodobactercapsulatus. Int J Hydrogen Energy 2011;36:13536e46.

[10] Ozgur E, AfsarN, deVrije T, YucelM,GunduzU, Claassen PAM,et al. Potential use of thermophilic dark fermentationeffluents in photofermentative hydrogen production byRhodobacter capsulatus. J Clean Prod 2010;18:S23e8.

[11] Avcioglu SG, Ozgur E, Eroglu I, Yucel M, Gunduz U.Biohydrogen production in an outdoor panel photobioreactoron dark fermentation effluent of molasses. Int J HydrogenEnergy 2011;36:11360e8.

[12] Ozkan E, Uyar B, Ozgur E, Yucel M, Eroglu I, Gunduz U.Photofermentative hydrogen production using darkfermentation effluent of sugar beet thick juice in outdoorconditions. Int J Hydrogen Energy 2012;37:2044e9.

[13] Azbar N, Dokgoz FTC. The effect of dilution and l-malic acidaddition on bio-hydrogen production with Rhodopseudomonaspalustris from effluent of an acidogenic anaerobic reactor. IntJ Hydrogen Energy 2010;35:5028e33.

[14] Li X, Liu T, Wu Y, Zhao G, Zhou Z. Derepressive effect of NH4þ

on hydrogen production by deleting the glnA1 gene inRhodobacter sphaeroides. Biotechnol Bioeng 2010;106:564e72.

[15] Demir A, Gunay A, Debik E. Ammonium removal fromaqueous solution by ion exchange using packed bed naturalzeolite. Water SA 2002;28:329e36.

[16] Leung S, Barrington S, Wan Y, Zhao X, El-Husseini B. Zeolite(clinoptilolite) as feed additive to reduce manure mineralcontent. Bioresour Technol 2007;98:3309e16.

[17] Eroglu E, Eroglu I, Gunduz U, Yucel M. Treatment of olive millwastewater by different physicochemical methods andutilization of their liquid effluents for biological hydrogenproduction. Biomass Bioenergy 2009;33:701e5.

[18] Karadag D, Koc Y, Turan M, Armagan B. Removal ofammonium ion from aqueous solution using natural Turkishclinoptilolite. J Hazard Mater 2006;B136:604e9.

[19] Tao YF, Qiu Y, Fang SY, Liu ZY, Wang Y, Zhu JH. Trapping thelead ion in multi-component aqueous solution by naturalclinoptilolite. J Hazard Mater 2010;180:282e8.

[20] Armagan B, Ozdemir O, Turan M, Celik MS. The removal ofreactive azo dyes by natural and modified zeolites. J ChemTechnol Biotechnol 2003;78:725e32.

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 3 7 ( 2 0 1 2 ) 1 6 4 2 1e1 6 4 2 9 16429

[21] Zorpas AA, Inglezakis VJ, Stylianou M, Irene V. Sustainabletreatment method of a high concentrated NH3 wastewater byusing natural zeolite in closed-loop fixed bed systems. OpenEnv Sci 2010;4:1e7.

[22] Inglezakis VJ, Hadjiandreou KJ, Loizidou MD,Grigoropoulou HP. Pretreatment of natural clinoptilolite ina laboratory-scale ion exchange packed bed. Wat Res 2001;35:2161e6.

[23] Bayraktaroglu K. Multicomponent ion exchange onclinoptilolite. Turkey: Chemical Engineering Department,MiddleEastTechnicalUniversity; 2006,MasterofScience thesis.

[24] Feldman C. Behavior of trace refractory minerals in thelithium metaborate fusion-acid dissolution procedure. AnalChem 1983;55:2451e3.

[25] Ozturk Y, Yucel M, Daldal F, Mandacı S, Gunduz U, Turker L,et al. Hydrogen production by using Rhodobacter capsulatusmutants with genetically modified electron transfer chains.Int J Hydrogen Energy 2006;31:1545e52.

[26] Biebl H, Pfennig N. Isolation of members of the familyRhodosprillaceae. In: Starr MP, Stolp H, Truper HG, Balows A,Schlegel HG, editors. The prokaryotes, vol. 1. New York:Springer; 1981. p. 267e73.

[27] Androga DD, Ozgur E, Gunduz U, Yucel M, Eroglu I. Factorsaffecting the longterm stability of biomass and hydrogenproductivity in outdoor photofermentation. Int J HydrogenEnergy 2011;36:11369e78.

[28] Jorgensen TC, Weatherley LR. Ammonia removal fromwastewater by ion exchange in the presence of organiccontaminants. Water Res 2003;37:1723e8.

[29] Wang S, Peng Y. Natural zeolites as effective adsorbents inwater andwastewater treatment. Chem Eng J 2010;156:11e24.

[30] Ji ZY, Yuan JS, Li XG. Removal of ammonium fromwastewater using calcium form clinoptilolite. J Hazard Mat2007;141:483e8.

[31] Martinez RC, Miranda VM, Rios MS, Sosa IG. Evaluation ofnatural and surfactant- modified zeolites in the removal ofcadmium from aqueous solutions. Sep Sci Technol 2004;39:2711e30.

[32] Sprynskyy M, Buszewski B, Terzyk AP, Namiesnik J. Study ofthe selection mechanism of heavy metal (Pb2þ, Cu2þ, Ni2þ

and Cd2þ) adsorption on clinoptilolite. J Colloid Interface Sci2006;304:21e8.

[33] Payne KB, Abdel-Fattah TM. Adsorption of divalent lead ionsby zeolites and activated carbon: effects of pH, temperatureand ionic strength. J Environ Sci Health Part A Toxic/HazardSubst Environ Eng 2004;39:2275e91.

[34] Ozgur E, Uyar B, Ozturk Y, Yucel M, Gunduz U, Eroglu I.Biohydrogen production by Rhodobacter capsulatus on acetateat fluctuating temperatures. Resour Conserv Recycl 2010;54:310e4.

[35] Boran E, Ozgur E, van der Burg J, Yucel M, Gunduz U,Eroglu I. Biological hydrogen production by Rhodobactercapsulatus in solar tubular photobioreactor. J Cleaner Prod2010;18:S29e35.

[36] Androga DD, Ozgur E, Gunduz U, Yucel M, Eroglu I.Significance of carbon to nitrogen ratio on the long-termstability of continuous photofermentative hydrogenproduction. Int J Hydrogen Energy 2011;36:15583e94.

[37] Sasikala K, Ramana CV, Rao PR, Kovacs KL. Anoxygenicphototropic bacteria: physiology and advances inhydrogen production technology. Adv Appl Microbiol1993;38:211e95.