Renewable and Sustainable Energy Reviews 40 (2014) 806813

Contents lists available at ScienceDirect

Renewable and Sustainable Energy Reviews

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

Review on the start-up experiences of continuous fermentative

hydrogen producing bioreactors
P. Bakonyi n, N. Nemestthy, V. Simon, K. Bla-Bak
Research Institute on Bioengineering, Membrane Technology and Energetics, University of Pannonia, Egyetem ut 10, 8200 Veszprm, Hungary

art ic l e i nf o

a b s t r a c t

Article history:
Received 23 April 2014
Received in revised form
16 July 2014
Accepted 5 August 2014
Available online 23 August 2014

The start-up of continuous biohydrogen fermentations is a complex procedure and a key to acceptable
hydrogen production performance and successful long-term operation. In this review article, the
experiences gained and lessons learned from relevant literature studies dealing with various aspects
of H2 producing bioreactor start-up are comprehensively surveyed. Firstly, the importance of H2-forming
biosystem start-up including its main steps is outlined. Afterwards, the role of main inuencing factors
and methods (e.g. strain selection, seed pretreatment and inocula stimulation, switch-over time,
bioreactor design, operating conditions) in avoiding the deterioration of starting a reactor is analyzed
and presented in detail. Finally, the so far suggested applicable start-up strategies and the corresponding
ndings are critically discussed pointing out the advantages and disadvantages of each strategy.
& 2014 Elsevier Ltd. All rights reserved.

Keywords:
Biohydrogen
Production
Continuous system
Bioreactor
Start-up

Contents
1.
2.
3.

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
The role of start-up in the efcacy of continuous hydrogen production . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Factors affecting the initiation of continuous H2 fermenters. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.1.
H2 producing strains . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.2.
Pretreatment and stimulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.3.
Switch-over time . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.4.
Bioreactor conguration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4. Experiences and lessons of continuous H2 producing bioreactor start-up . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1. Introduction
Hydrogen is an emerging candidate among the various alternative energy carriers. H2 is believed to help the transition of
current fossil-based economy to a renewable-based one [1],
however, only if it is derived by sustainable processes. Though
H2 can be prepared by many conventional and mature methods
(e.g. steam reformation of hydrocarbons), environmental-friendly
methods such as with biological routes are required and still
subjects to extensive research [2].

Corresponding author. Tel.: 36 88 624385; fax: 36 88 624292.

E-mail address: [email protected] (P. Bakonyi).

http://dx.doi.org/10.1016/j.rser.2014.08.014
1364-0321/& 2014 Elsevier Ltd. All rights reserved.

806
807
807
807
807
808
809
810
811
811
811

Nowadays, microbiologically produced hydrogen is recognized as

an emerging way ahead, especially when formed via dark fermentation because of its inherent advantages such as relatively low energy
demand (attributed to the gentle reaction conditions), the usability of
wide range of feedstocks e.g. derivates of biomass, waste streams and
agricultural residues [35], and the possibility to integrate with other
e.g. membrane-based processes in order to accomplish the sufcient
reuse of hydrogen producing cells [6] or to upgrade bioH2 [79] so that
it could be a viable feedstock in energy efcient fuel cells.
Nevertheless, additional efforts are still essential to make
biohydrogen generation more attractive. From practical aspect,
the two major criteria to be considered are H2 production yields
and rates. As a result of the investigations in the past decades,
several factors were identied that signicantly affect the main

P. Bakonyi et al. / Renewable and Sustainable Energy Reviews 40 (2014) 806813

technological indicators mentioned above. Among them, bioreactor conguration and operation are apparent ones [10]. Regardless
the type of the fermenter, it can be concluded that feasible
biohydrogen fermentation should be conducted in continuous
rather than batch systems [11], e.g. due to higher expectable
process efciencies.
The establishment of continuous ow bioreactors usually starts
in batch mode, and it is to note that successful transition and
reliable, long-term operation is highly inuenced by the start-up
strategy applied [1214]. However, up to the authors' best knowledge, there is no recent review paper comprehensively surveying
start-up experiences in continuous hydrogen producing bioreactors. Hence, in this work, the experiences gained and lessons
learned from batch to continuous shifts are reviewed and suggestions are given to achieve proper continuous operation.

2. The role of start-up in the efcacy of continuous hydrogen

production
Process instability is a frequently observed drawback in
fermentative H2 production [15] that could be attributed to
multiple reasons as specied later in this paper. In fact, beyond
steady-state operational parameters and medium composition,
stable and continuous bioreactor operation to obtain acceptable
hydrogen production performance is strongly dependent on the
start-up phase [16]. It could involve the following steps:
 Selection of the hydrogen producing biocatalysts.
 Enhancement and acclimatization of H2-forming strains to
fermentation circumstances.
 System transition until steady-state is reached.
These steps in a line require great attention and comprehensive
control in terms of environmental and operational circumstances
to develop robustful H2 fermenting culture [17]. Otherwise, starting a reactor may easily be deteriorated e.g. due to the insufcient
growth and H2 production capacity of microorganisms. Such
bottlenecks can be avoided or at least mitigated by properly
designed start-up strategy.
In the next sections of the paper, the aforementioned parts of
the continuous dark fermentative bioreactor establishment are
outlined and discussed in details.

3. Factors affecting the initiation of continuous H2 fermenters

3.1. H2 producing strains
Fermentative biohydrogen generation can be realized either by
pure cultures [18] such as Escherichia coli [19,20] or mixed
bacterial consortia [21,22] and both have their own benets. For
example, cultures of pure isolates may be easier to control but
need constantly sterile environment to prevent contamination
that is difcult and costly to maintain out of laboratories. Considering their application in a non-sterile environment, pure
cultures may be used in the bioaugmentation of diverse H2
producing population to attain better gas turnouts [23]. The
restrictions of sterility criteria are the main reasons why mixed
bacterial communities are preferred to their pure counterparts in
real-case, scaled-up applications.
3.2. Pretreatment and stimulation
Conceptually, anaerobic, mixed H2-producing consortia (e.g. in
sewage sludge, biogas plants, etc.) are built up by co-existing and

807

synergic species [24]. However, in most cases, they naturally occur

together with H2-consumer microorganism such as methanogenic
archaea, homoacetogenic (producing acetate from CO2 and H2)-,
lactic- and propionic acid bacteria which must certainly be
suppressed or more preferentially totally eliminated [25]. As a
consequence, regardless the source of mixed inocula, it should
undergo initial pretreatment in order to select the desired
whole cell biocatalysts. For such purposes, a lot of tools have been
developed based on heat shock, addition of chemicals, swinging
the oxidationreduction potential (ORP) e.g. by aeration, high
energy irradiation, alteration of pH, freezing and thawing
[2628]. These pretreatment techniques exploit the distinct sensitivity of strains present in the mixture and in general could
provide a satisfactory starter culture to be used as seed inocula
for subsequent biohydrogen fermentation. In other words, these
procedures aim to eliminate hydrogen-consuming vegetative cells
and on the other hand, are devoted to enhance acidogenic- and
often sporulative H2-forming cells [29].
Although culture pretreatments can effectively suppress undesired microbiological activity, they may also reduce the number of
indigenous H2-former bacteria, especially the ones with low stress
tolerance. For these reasons, as a next step after culture pretreatment, treated inocula should be submitted to stimulative environment (e.g. to a batch reactor) to let the microbes proliferate so that
a reasonable amount of active cellmass can be accumulated,
harvested and further applied. Also, batch cultivation can play a
role to help biolm development on carrier materials (e.g. powdered- and granulated activated carbon) if an immobilized, continuous H2 production system is to be implemented [30,31].
According to literature reports, pretreated inocula are more
often than not dominated by spore-forming and robust H2producer species such as Clostridium sp. [13]; however some
organisms of no utility (e.g. propionic acid and homoacetogenic
bacteria) may also survive and reclaim their niche over time
[21,25]. Changes in the microbial background can be revealed by
the modern technical apparatus of molecular biology [32,33].
Furthermore, it is presumable that the age of the seed source
most commonly sewage or biogas (anerobic fermenter) sludge as
suggested by Tables 1 and 2 may also be a factor to take into
account. It is assumable that the microbial community structure of
anaerobic mixed cultures varies constantly during storage due to
changes (e.g. concentration differences) within microenvironments. Consequently, aging of an anaerobic seed culture
over time can result in the variation of the obtainable bacterial
populations and their activity. Thus, it might lead to alterations in
the attainable biohydrogen performances even though standardized, identical pretreatment conditions are ensured time after
time to prepare H2 producing inocula.
Moreover, beyond the goal of activating H2-producer organisms
[16,34,35], preliminary cultivation mostly in batch may also
serve as a tool to acclimatize the microora to certain substrates
and their loadings e.g. to overcome inhibitory effect [36], which
will induce a dynamic competition between the various groups of
bacteria. Although batch-continuous start-up strategy was proposed by various authors to follow (for examples, please refer to
Tables 1 and 2), some researchers reported adequate start-up
directly in continuous operation [3741].
The advantage of this strategy lacking initial discontinuous cell
growth might be that in batch operation the nutrient concentrations as the time passes, especially at the end hours of the cycle,
are insufciently low and consequently a shift in the dominant
strains could occur, depressing H2 production [12]. During careful
continuous adaptation, broth is constantly supplemented and such
disadvantageous phenomenon may be avoided. Moreover, continuous (hydraulic detention time inuenced) acclimatization strategy encompasses the so-called biokinetic control which causes the

808

P. Bakonyi et al. / Renewable and Sustainable Energy Reviews 40 (2014) 806813

Table 1
Start-up experiences during H2 fermentation in CSTR.
Reactor
type

Inoculum

Inoculum
Pretreatment

Substrate

pH

T
Start-up experiences
(1C)

CSTR

5 Different
thermophilic
sludge
Indigenous
microora of
substrate
Anaerobic digester
sludge
E. coli
Anaerobic granular
sludge
Digester sludge
Waste activated
sludge
Anaerobic digester
sludge
Anaerobic digester
sludge
Anaerobic digester
sludge
Anaerobic digester
sludge
Wastewater
treating sludge

Starch

N.C.

Sweet sorghum
extract

3.56.5

[70]
Continuous feeding was started after obtaining exponential growth
phase in batch operation. Stable hydrogen production was attained in
less than 30 days of start-up
35C 24 h in batch mode to activate the indigenous microora contained in [74]
substrate

Heat shock (85 1C,

45 min)

Heat shock (boiling

for 40 min)
N.M.
Heat shock (70 1C,
30 min)
N.M.

Cheese whey
wastewater
Na-formate
Cheese whey

5.5

55

6.5
7.5

Cellulose
Glucose

N.C.
5.56.5

CSTR

CSTR
CSTR
CSTR
CSTR
CSTR
CSTR
CSTR
CSTR
CSTR
CSTR

CSTR
CSTR

CSTR
CSTR

CSTR

Heat shock
20 min)
Heat shock
10 min)
Heat shock
15 min)
Heat shock
45 min)

55

[75]

37
37

2 days in batch mode, conversion to continuous operation when

hydrogen production reached its peak value
Batch operation until exponential growth phase took place
Batch operation for 12 h

70
37

90 days until steady-state

15 h in batch mode, 10 days to reach steady-state

[72]
[39]

32

Continuous operation was commenced once signicant hydrogen

production occurred
When the yield reached 60 mL H2/g dcw, the operation was shifted to
continuous mode
When cumulative H2 production was equivalent to 0.5 mol H2/mol
hexose, the reactors were put into continuous mode
20 days long start-up

[76]

Continuous feeding started after 24 h of batch operation. During startup, decreased initial organic loading rate could enhance hydrogen
production efciency
35 The fermenter was rst operated in a batch mode for two days and
then switched to a continuous operation
35 For start-up, the reactor was operated in batch mode for 24 h to
activate the indigenous microora contained in the seed before
initiation of continuous operation
35 Continuous bioreactors were operated as a batch for the rst 40 h
38
37 1 day in batch mode before continuous operation, Steady gas
production was observed after 19 days

[48]

(90 1C,

Sugarbeet water 5.2

extract
L. japonica
5.58

35

(90 1C,

Food waste

5.3

35

(90 1C,

Sucrose

5.3

35

(100 1C,

Starch

5.5

35

Sucrose

Cheese whey

5.2

Whey permeate

45

Glucose

5.5

Terephthalic
acid processing
wastewater

Wastewater sludge Heat shock (100 1C,

45 min)

Indigenous
microora of
substrate
Anaerobic digester N.M.
sludge
Anaerobic digester Heat and acid
sludge
treatment (98 1C,
2 h; pH 2, 24 h)
Sewage sludge

Reference

35

Stabilized gas production was achieved after 25 days

[20]
[49]

[78]
[15]
[47]

[63]
[34]

[79]
[66]

[46]

N.M.: not mentioned; N.C.: not controlled.

wash out of existing microbes possessing inadequately low growth

rates or adaption capabilities [42]. In other words, feeding regime
affects the culture diversity and the relative abundance of the
bacterial species.
Additionally, pH, temperature and ORP adjustment are also of
crucial importance, since their values change the generation time,
growth rate and metabolic pathways of microorganisms present in
a mixed culture [12,4345]. It was also demonstrated that sustained continuous hydrogen formation could be achieved with a
start-up strategy completely lacking preliminary inocula pretreatment and batch propagation. For example, bioreactor inoculated
with untreated consortia achieved the suppression of H2 consuming microorganisms through the simultaneous enrichment of
biohydrogen producers, taking place because of the insistent
acidophilic microenvironment maintained from the beginning of
operation [46]. For more studies skipping inocula pretreatment,
the reader is referred to Tables 1 and 2.
Besides the adequate substrate composition and loading, temperature, pH and OPR there are other parameters such as hydrogen
partial pressure in the bioreactor that may need a control since it
is a potential threat that hydrogenotrophic consortial activity may
be provoked under high H2 concentrations [25]. From another
point of view, reduced H2 partial pressure was proven to increase
hydrogenase activity and making H2 formation thermodynamically favorable [47].

3.3. Switch-over time

Since the establishment of continuous hydrogen fermentation
implicates an initiative batch cycle for most cases, another issue to
be discussed is its duration.
The literature is not consistent about this question, or in other
words, it is not fully obvious when to convert to continuous
hydrogen fermentation. However, as listed in Tables 1 and 2, the
following strategies could be identied as the most popular ones:
 switch-over when signicant biohydrogen production
commences,
 switch-over when reaching the exponential H2 production
phase, and
 switch-over after a few days of batch cultivation.
Regardless the hydrogen fermentation system employed, dilution
rate, substrate loading intensity, pH and temperature applied during
transition-state (caused by the switch between batch and continuous
operation) reactor run will result in the enrichment of certain bacterial
populations and moreover, these factors inherently direct their metabolic pathways. After a period of time when the functional consortia
got used to the environmental conditions and consequently stabilized,
steady-state can take place which is mostly considered to reach when
variations in H2 gas production, pH and efuent (spent media) quality

P. Bakonyi et al. / Renewable and Sustainable Energy Reviews 40 (2014) 806813

809

Table 2
Start-up experiences during H2 fermentation in reactors other than CSTR.
Reactor
type

Inoculum

Inoculum
Pretreatment

Substrate

pH

Start-up experiences

Reference

UASBR

Mixture of
precultured and
granulated sludge
Sewage sludge

Heat shock (100 1C,

2 h)

Starch

55

After conrming the exponential production of the biogas, the

operation was turned into continuous mode

[73]

Sucrose

6.7

35

Wastewater
treating sludge
Anaerobic digester
sludge

Heat shock (100 1C,

45 min)
Heat shock (90 1C,
15 min)
Heat shock (90 1C,
20 min)

N.M.

30

5.5

35

UASBR

Anaerobic digester
sludge

Heat shock (90 1C,

20 min)

5.5

35

When the yield of produced H2 reached 0.5 mol H2/mol hexose,

continuous operation started

[59]

ASBR

Anaerobic digester
sludge

Heat shock (90 1C,

10 min)

Glucose, cheese
whey
Coffee drink
manufacturing
wastewater
Coffee drink
manufacturing
wastewater
Food waste

Time-consuming start-up, almost 40 days were taken to reach steady- [52]

state
Initially, the reactors were operated as batch for 76 h prior to switching [29]
to continuous mode
Continuous operation was delayed until a yield of 0.5 mol H2/mol
[77]
hexose achieved in the batch, start-up took 10 days in continuous mode

5.3

35

[57]

AFBR

Wastewater
treating sludge
Anaerobic digester
sludge

Heat shock (90 1C,

10 min)
Heat shock (105 1C,
2 h)

Glucose

N.C.

30

Tapioca
wastewater

6.5a;
9b

32

ASBR

Anaerobic digester
sludge

Heat shock (100 1C,

30 min)

37

ASBR

Anaerobic digester
sludge

Heat shock (boiled,

30 min)

Enriched
facultative
anaerobic culture
Wastewater
treating sludge

Enriched culture
with Clostridium
pasteurianum
N.M.

5a;
4.4
5.6b
7

37

UASBR

Liquid swine
manure mixed
with glucose
Liquid swine
manure mixed
with glucose
Citric acid
wastewater

When cumulative H2 production of 0.5 mol H2/mol hexose was

observed, the reactors were put into continuous operation. Steady-state
was reached in 1030 days depending on the HRT and inoculation
conditions
The bioreactor was initially run as a batch for 2 days to stimulate the
hydrogen-producing biomass
Multistep batch operation and gradual acclimatization of mixed
consortia to substrate. First 3 days in batch operation. 37 days were
required to reach steady-state H2 production
Firstly, the bioreactor was operated in a batch mode for 24 h until the
established biogas production took place

Sucrose

6.1
9.5

UASBR
AFBR
UASBR,
CSTR

ABR

UASBR

Firstly, the bioreactor was operated in a batch mode for 24 h until

established biogas production took place

[16]
[64]

[56]

[58]

35 More than a month long acclimatization before starting continuous

38 mode. UASB start-up took 20 days, excellent system stabilization

[55]

39

[53]

The start-up of the UASB reactor lasted for 300 days to enrich H2producing microbes and establish a stable gas generation. Afterwards,
a successful operation was achieved with the formation of the H2producing granules

ABR: anaerobic bafed reactor; AFBR: anaerobic uidized bed reactor; N.M.: not mentioned; N.C.: not controlled.
a
b

During start-up.
After start-up.

e.g. in terms of soluble metabolic product (SMP) distribution and

related concentrations are below 10% on a daily average base [48].
Therefore, appropriate threshold levels of the parameters
mentioned can develop an attractive hydrogen-generating biocommunity and govern the whole bioreactor towards better
performances e.g. volumetric production rates and yields [49].

3.4. Bioreactor conguration

The conguration of the bioreactor set-up is also a concern to
keep in mind since different kinds of reactors can be characterized by
distinct start-up stage features, for example in terms of its duration
[6,14]. Nowadays, the suspended-cell, completely stirred tank reactor
(CSTR) is the most routinely applied one, however, up-ow anaerobic
sludge blanket (UASBR) reactors, anaerobic membrane bioreactors
and immobilized (e.g. uidized bed) bioreactors [16,29] became
popular due to their improved H2 producing potentials.
Generally, the CSTR is featured by a relatively short induction
phase [50] as compared to other applications (e.g. UASBR) due to
better mass transfer, but it needs rigorous supervision due to the
disposition of cells to wash out at inadequate operating bioreactor
regimen. Troubleshooting the risk of wash out can be performed
by biomass-rejective systems such as the membrane bioreactors
[6] or immobilized e.g. xed bed systems [51,70]. These

alternatives may demonstrate more robust operation and

enhanced hydrogen production efciency even in smaller reactor
volumes that is of economical importance.
Another option is the usage of UASBR. Basically, this construction
is described by extended start-up phase [52,53] since the occulation
of bacterial communities in the sludge-bed demands longer time.
However, start-up period of granular systems for biohydrogen
generation could be considerably shortened through the transformation of methanogenic granules (obtained e.g. from already existing
and well-established anaerobic, methane forming UASB reactors)
into hydrogen producing ones, as recently reported [54].
An important trait of UASBR is the fact that it does not apply
mechanical mixing and therefore pH gradients can occur which is not
easy to control. For pH regulation purposes, the buffer capacity of the
fermentation media may be adjusted to withstand progressive pH
depression caused by the formation of acidic by-products that is
always expectable in parallel to H2 bioproduction. However, in return
to laborious start-up, granulated reactors reect remarkably improved
operational stability [55]. This is because granulation reactors enhance
the active biomass concentration and thereby able to sustain under
increased substrate dosage and greater dilution rates so that higher H2
production intensity can be accomplished [56].
Last but not least, the anaerobic sequencing batch reactor is also
among the available design options for continuous H2 fermentation
[5658].

810

P. Bakonyi et al. / Renewable and Sustainable Energy Reviews 40 (2014) 806813

4. Experiences and lessons of continuous H2 producing

bioreactor start-up
Alternatively to the strategy described in [54], it has been
reported that the start-up time of UASBR could also be signicantly decreased by thriving H2-producing cells in CSTR arrangement prior to transferring them to the up-ow anaerobic sludge
blanket reactor as seeding source [59]. As it was found, despite the
high shearing forces in CSTR, self-occulation of hydrogengenerating bacteria was notably faster than in UASBR and
explained by the more intense mass transfer capacity of the
former reactor type.
Besides, organic loading rate uctuation was reported to accelerate the start-up process both in suspended- and immobilized
cell applications [31]. This strategy was shown as an efcient way
to rapidly and effectively establish a good biohydrogen evolver
culture and less than 3 weeks were required to obtain stabilized
operation, which is considerably shorter in comparison with other
similar systems [60].
In relation to start-up duration, anaerobic sequencing batch
reactor (ASBR) was demonstrated as a feasible concept to provide
quick steady-state. Studies indicated that start-up time requirements in the range of 1214 days were far below the values
revealed for UASBR and CSTR [56,61].
As communicated [62], operational conditions employed in
UASBR start-up could have signicant impact on the microbial
ngerprint of mixed H2 producing sludge, depending on the seed
inocula structure.
Likewise, pH adjustment was noticed to express marked
inuence during H2-formation bioreactor start-up [46,58]. It has
appeared that pH values out of optimal range may induce
population- and metabolic shifts (i.e. solventgenesis). Moreover,
the hydrogenase enzyme activity and growth rate of microorganism responsible for H2 fermentation can also be hindered under
inappropriate pH conditions. Hence, non-optimal pH probably
causes an unambiguous delay in attaining steady-state conditions
with the desired H2 production rate and yield.
Moreover, pre-culturing seed inocula and initial batch strategy
were observed to be efcient start-up concept that supported
biomass growth and consistent H2 production well in subsequent continuous operation [63]. The adaption of H2 producing
microora to a given substrate can be conducted in complex
and multistep batch operation, where bacterial consortia are
periodically supplied with gradually increasing feed concentrations before putting into continuous mode [64]. Nevertheless, as
mentioned above, there are studies that skipped preliminary
batch operation and adapted the H2 generating cultures during
continuous operation. Regarding these start-up strategies, it is to
conclude that the hydraulic retention time (HRT) is usually
stepwise rened from long to short time intervals to allow the
acclimatization of microorganisms to new environments and
prevent washing out the bacteria of interest [65]. As a result
of shifting HRT, the microbial population dynamically changes
leading to the disappearance of certain species while others show
up [66].
Additionally to HRT, altering organic loading rate (OLR) is also a
stress source to the strains that are forced to get accustomed to
new surroundings [52]. Varying OLR might cause sporulation in
hydrogen producing cultures and contribute to the observable
uctuations in the H2 production efcacy during start-up stage
[48,67].
It is an ongoing progress in biohydrogen research that various
waste substances are utilized to improve process viability. However, depending on the nature of these problematic raw materials
(e.g. cheese whey), it is presumptive that they contain indigenous
microora which of course, from the beginning, might affect

bioreactor performance in a negative manner [75]. In this regard,

it was evaluated [15] that attention should be paid to the
pretreatment (e.g. by means of alkali) and sufcient storage
circumstances (preferentially at cold temperature) of such
streams. This is attributed to the fact that indigenous strains, for
example non-H2-producing acidogens such as lactic- and propionic bacteria present in non-aseptic substrate could cause strong
contamination and even outcompete the H2-producing microbes
[68,69].
Even though the biosystems may resist the perturbations
caused by contamination and might be able to express the
performance required, dominance of disadvantageous nonhydrogen producing bacteria potentially leads to problematic
start-up, unsuccessful continuous operation in long terms and
consequently may force to reinoculate and restart the H2 fermenting bioreactor.
Recently, it was experimentally demonstrated that nonhydrogen producing cells, as a result of long batch cultivation,
were promoted alongside their useful H2 evolver counterparts and
could take over, causing tough start-up failure. In order to avoid
such undesired microbiological activity, an early switch-over time
was recommended that highly increased the chance of successful
and long-term continuous H2 fermentation [69]. It is also extractable from the literature that inocula source can be a determining
factor of the time necessary for starting-up a continuous H2
producing fermenter [70].
In addition to contamination related issues, operational failures
caused by unforeseen technical difculties (e.g. broken pumps,
leaking tubes) may also challenge the adequate start-up since they
can cause insufcient or shocking organic loading rates, altered
hydraulic detention times and thus, disturb the developing microbial consortia and affect its survivability.
To aid bioreactor start-up, monitoring the soluble metabolites
such as volatile fatty acids is apparently benecial. The ratios of
acetic-, butyric-, and propionic acids produced by the various
groups of microorganisms can be useful feedbacks about the state
of the hydrogen producing consortia. When propionic acid concentration gradually increases in fermentation broth, it assumes
the occurrence of microbes with no utility for H2 production and
gives the sign that troubleshooting of the reactor is required to
avoid strong deterioration of its performance.
In cases when an unusual decline in hydrogen production
efciency occurs (e.g. as a consequence of population shift or
appearance of methanogenic activity) during start-up that threats
achieving steady-state, a temperature shift strategy may be carried
out which includes heating the bioreactor to higher temperature
ranges (e.g. to 7080 1C) for a short time (e.g. 1 h) to reclaim
hydrogen producing bacteria and reactor performance [71]. However, it may be ineffective for granulated systems, since it was
proven that granules serve as protective structure. In such cases,
disintegration of granules prior to temperature shift or combined
methods (e.g. temperature- and pH shift together) might work
[62]. On the contrary, it was demonstrated that simple washing
and subsequent boiling of granular sludge could be a feasible
approach to inactivate hydrogen consuming microorganisms [49].
It is to mention that process temperature even in the same
bioreactor design could inuence the reactor behavior during
start-up. As a matter of fact, comparison of meso- and (hyper)
thermophilic H2 production in CSTRs indicates that start-up of the
latter group could last as long as 90 days to achieve stabilized H2
fermentation [72], while mesophilic H2 production in similar
system congurations was reported to reach steady-state circumstances in shorter times [39,47,67].
As a summary, the ow chart depicted in Fig. 1 presents the
connection network of the various steps involved in continuous
hydrogen producing bioreactor start-up.

P. Bakonyi et al. / Renewable and Sustainable Energy Reviews 40 (2014) 806813

811

Fig. 1. Flow chart illustrating the various steps of start-up process.

5. Conclusions
In this review, the experiences of continuous hydrogen
fermentation start-up were scoped and analyzed. The lessons of
relevant literature papers about the routes leading to continuous
and efcient, steady-state hydrogen production imply that startup is of high concern to avoid signicant performance losses. As a
general guideline, the establishment of reliable, continuous hydrogen producing bioreactors should start with proper inocula
selection and its pretreatment (if necessary), followed by an
acclimatization period conducted mainly as a batch to adopt
the living biocatalysts to the intended substrate which may also
require preliminary treatment to eliminate native and undesirable
microora present in it. Subsequently, switch-over strategy
assigned to ensure viable and smooth batch to continuous shift
must be designed e.g. timed properly in order to preserve and
ensure microorganism with as high hydrogen producing capacity
as possible. Besides timing, transition from batch to continuous
mode hydrogen fermentation should take also into account the
suitable adjustment of major environmental (physiological) factors
such as pH, temperature, etc. and the operating conditions (e.g.
hydraulic retention time, organic loading rate, etc.) applied with
respect to bioreactor conguration (CSTR, UASB, AnMBR, etc.).
Currently, CSTRs are the most widely used reactors for continuous
hydrogen production due to their relatively rapid start-up phase.
Nevertheless, as a result of the efforts made to cut start-up time
demand of other devices, more wide-spread employment of
granular- and immobilized systems and that of reactors integrated
with downstream (membrane bioreactors) is presumable in the
future, which is also attributed to their potential benets (e.g.
higher long-term performance and enhanced stress tolerance)
over the conventional set-ups.

Acknowledgments
Pter Bakonyi thanks the TMOP 4.2.4.A/2-11-1-2012-0001
National Excellence Program supported by the European Union
and the State of Hungary, co-nanced by the European Social

Fund. Nndor Nemestthy acknowledges the Jnos Bolyai Research

Scholarship of the Hungarian Academy of Sciences. The research
infrastructure was supported by TMOP 4.2.2/A-11/1/KONV-20120071 nanced by the European Union and the European Social
Fund.

References
[1] Azwar MY, Hussain MA, Abdul-Wahab AK. Development of biohydrogen
production by photobiological, fermentation and electrochemical processes:
a review. Renew Sustain Energy Rev 2014;31:15873.
[2] Ren NQ, Guo WQ, Liu BF, Cao GL, Ding J. Biological hydrogen production by
dark fermentation: challenges and prospects towards scaled-up production.
Curr Opin Biotechnol 2011;22:36570.
[3] Cheng CL, Lo YC, Lee KS, Lee DJ, Lin CY, Chang JS. Biohydrogen production from
lignocellulosic feedstock. Bioresour Technol 2011;102:851423.
[4] Guo XM, Trably E, Latrille E, Carrre H, Jean-Philippe S. Hydrogen production
from agricultural waste by dark fermentation: a review. Int J Hydrog Energy
2010;34:1066073.
[5] Lin CY, Lay CH, Sen B, Chu CY, Kumar G, Chen CC, et al. Fermentative hydrogen
production from wastewaters: a review and prognosis. Int J Hydrog Energy
2012;37:1563242.
[6] Bakonyi P, Nemestthy N, Simon V, Bla-Bak K. Fermentative hydrogen
production in anaerobic membrane bioreactors: a review. Bioresour Technol
2014;156:35763.
[7] Bakonyi P, Nemestthy N, Bla-Bak K. Biohydrogen purication by membranes: an overview on the operational conditions affecting the performance
of non-porous, polymeric and ionic liquid based gas separation membranes.
Int J Hydrog Energy 2013;38:967387.
[8] Bakonyi P, Kumar G, Nemestthy N, Lin CY, Bla-Bak K. Biohydrogen
purication using a commercial polyimide membrane module: studying the
effects of some process variables. Int J Hydrog Energy 2013;38:150929.
[9] Ramrez-Morales JE, Tapia-Venegas E, Nemestthy N, Bakonyi P, Bla-Bak K,
Ruiz-Filippi G. Evaluation of two gas membrane modules for fermentative
hydrogen separation. Int J Hydrog Energy 2013;38:1404252.
[10] Hallenbeck PC, Ghosh D. Advances in fermentative biohydrogen production:
the way forward? Trends Biotechnol 2009;27:28797.
[11] Wang J, Wan W. Factors inuencing fermentative hydrogen production:
a review. Int J Hydrog Energy 2009;34:799811.
[12] Hawkes FR, Dinsdale R, Hawkes DL, Hussy I. Sustainable fermentative hydrogen production: challenges for process optimisation. Int J Hydrog Energy
2002;27:133947.
[13] Hawkes FR, Hussy I, Kyazze G, Dinsdale R, Hawkes DL. Continuous dark
fermentative hydrogen production by mesophilic microora: principles and
progress. Int J Hydrog Energy 2007;32:17284.
[14] Jung KW, Kim DH, Kim SH, Shin HS. Bioreactor design for continuous dark
fermentative hydrogen production. Bioresour Technol 2011;102:861220.

812

P. Bakonyi et al. / Renewable and Sustainable Energy Reviews 40 (2014) 806813

[15] Kim SH, Shin HS. Effects of base-pretreatment on continuous enriched culture
for hydrogen production from food waste. Int J Hydrog Energy 2008;33:
526674.
[16] Shida GM, Barros AR, Reis DCM, Amorim ELCD, Damianovic MHRZ, Silva ED.
Long-term stability of hydrogen and organic acids production in an anaerobic
uidized-bed reactor using heat treated anaerobic sludge inoculum. Int J
Hydrog Energy 2009;34:367988.
[17] Hernndez-Mendoza CE, Moreno-Andrade I, Buitrn G. Comparison of
hydrogen-producing bacterial communities adapted in continuous and discontinuous reactors. Int J Hydrog Energy 2014;39:142349. http://dx.doi.org/
10.1016/j.ijhydene.2014.01.014.
[18] Lee DJ, Show KY, Su A. Dark fermentation on biohydrogen production: pure
culture. Bioresour Technol 2011;102:8393402.
[19] Bakonyi P, Nemestthy N, Lvitusz , Bla-Bak K. Application of Plackett
Burman experimental design to optimize biohydrogen fermentation by E. coli
(XL1-BLUE). Int J Hydrog Energy 2011;36:1394954.
[20] Bakonyi P, Nemestthy N, Ramirez J, Ruiz-Filippi G, Bla-Bak K. Escherichia
coli (XL1-BLUE) for continuous fermentation of bioH2 and its separation by
polyimide membrane. Int J Hydrog Energy 2012;37:562330.
[21] Valdez-Vazquez I, Poggi-Varaldo HM. Hydrogen production by fermentative
consortia. Renew Sustain Energy Rev 2009;13:100013.
[22] Wong YM, Wu TY, Juan JC. A review of sustainable hydrogen production using
seed sludge via dark fermentation. Renew Sustain Energy Rev 2014;34:
47182.
[23] Pattra S, Lay CH, Lin CY, O-Thong S, Reungsang A. Performance and population
analysis of hydrogen production from sugarcane juice by non-sterile continuous stirred tank reactor augmented with Clostridium butyricum. Int J Hydrog
Energy 2011;36:8697703.
[24] Hung CH, Chang YT, Chang YJ. Roles of microorganisms other than Clostridium
and Enterobacter in anaerobic fermentative biohydrogen production systems
a review. Bioresour Technol 2011;102:843744.
[25] Saady NMC. Homoacetogenesis during hydrogen production by mixed cultures dark fermentation: unresolved challenge. Int J Hydrog Energy 2013;38:
1317291.
[26] Bakonyi P, Borza B, Orlovits K, Simon V, Nemestthy N, Bla-Bak K.
Fermentative hydrogen production by conventionally and unconventionally
heat pretreated seed cultures: a comparative assessment. Int J Hydrog Energy
2014;39:558996.
[27] Ren NQ, Guo WQ, Wang XJ, Xiang WS, Liu BF, Wang XZ, et al. Effects of
different pretreatment methods on fermentation types and dominant bacteria
for hydrogen production. Int J Hydrog Energy 2008;33:431824.
[28] Wang J, Wan W. Comparison of different pretreatment methods for enriching
hydrogen-producing bacteria from digested sludge. Int J Hydrog Energy
2008;33:293441.
[29] Ferreira Rosa PR, Santos SC, Silva EL. Different ratios of carbon sources in the
fermentation of cheese whey and glucose as substrates for hydrogen and
ethanol production in continuous reactors. Int J Hydrog Energy 2014;39:
128896.
[30] Lee KS, Lin PJ, Chang JS. Temperature effects on biohydrogen production in a
granular sludge bed induced by activated carbon carriers. Int J Hydrog Energy
2006;31:46572.
[31] Ren NQ, Tang J, Liu BF, Guo WQ. Biological hydrogen production in continuous
stirred tank reactor systems with suspended and attached microbial growth.
Int J Hydrog Energy 2010;35:280713.
[32] Li RY, Zhang T, H.H.P. Fang. Application of molecular techniques on heterotrophic hydrogen production research. Bioresour Technol 2011;102:844556.
[33] Tolvanen KES, Karp MT. Molecular methods for characterizing mixed microbial communities in hydrogen-fermenting systems. Int J Hydrog Energy
2011:3652808.
[34] Venetsaneas N, Antonopoulou G, Stamatelatou K, Kornaros M, Lyberatos G.
Using cheese whey for hydrogen and methane generation in a two-stage
continuous process with alternative pH controlling approaches. Bioresour
Technol 2009;100:37137.
[35] Wu X, Lin H, Zhu J. Optimization of continuous hydrogen production from cofermenting molasses with liquid swine manure in an anaerobic sequencing
batch reactor. Bioresour Technol 2013;136:3519.
[36] Badiei M, Jahim JM, Anuar N, Abdullah SRS. Effect of hydraulic retention time
on biohydrogen production from palm oil mill efuent in anaerobic sequencing batch reactor. Int J Hydrog Energy 2011;36:59129.
[37] Chen CC, Lin CY, Chang JS. Kinetics of hydrogen production with continuous
anaerobic cultures utilizing sucrose as the limiting substrate. Appl Microbiol
Biotechnol 2001;57:5664.
[38] Cheng CH, Hung CH, Lee KS, Liau PY, Liang CM, Yang LH, et al. Microbial
community structure of a starch-feeding fermentative hydrogen production
reactor operated under different incubation conditions. Int J Hydrog Energy
2008;33:52429.
[39] Hafez H, Baghchehsaraee B, Nakhla G, Karamanev D, Margaritis A, Naggar HE.
Comparative assessment of decoupling of biomass and hydraulic retention
times in hydrogen production bioreactors. Int J Hydrog Energy 2009;34:
760311.
[40] Li J, Zheng G, He J, Chang S, Qin Z. Hydrogen-producing capability of anaerobic
activated sludge in three types of fermentations in a continuous stirred-tank
reactor. Biotechnol Adv 2009;27:5737.
[41] Lin CY, Wu CC, Hung CH. Temperature effects on fermentative hydrogen
production from xylose using mixed anaerobic cultures. Int J Hydrog Energy
2008;33:4350.

[42] Tapia-Venegas E, Ramirez JE, Donoso-Bravo A, Jorquera L, Steyer JP, RuizFilippi G. Bio-hyrogen production during acidogenic fermentation in a multistage stirred tank reactor. Int J Hydrog Energy 2013;38:218590.
[43] Ren NQ, Zhao D, Chen X, Li J. Mechanism and controlling strategy of the
production and accumulation of propionic acid for anaerobic wastewater
treatment. Sci China Ser B 2002;45:31927.
[44] Ren NQ, Chua H, Chan SY, Tsang YF, Wang YJ, Sin N. Assessing optimal
fermentation type for bio-hydrogen production in continuous-ow acidogenic
reactors. Bioresour Technol 2007;98:177480.
[45] Wang L, Zhou Q, Li FT. Avoiding propionic acid accumulation in the anaerobic
process for biohydrogen production. Biomass Bioenergy 2006;30:17782.
[46] Biohydrogen production from puried terephthalic acid (PTA) processing
wastewater by anaerobic fermentation using mixed microbial communities.
In: Zhu GF, Wu P, Wei QS, Lin JY, Gao YL, Liu HN, editors. Int J Hydrog Energy,
35; 2010. p. 83506.
[47] Kim DH, Han SK, Kim SH, Shin HS. Effect of gas sparging on continuous
fermentative hydrogen production. Int J Hydrog Energy 2006;31:215869.
[48] Lin CY, Chang CC, Hung CH. Fermentative hydrogen production from starch
using natural mixed cultures. Int J Hydrog Energy 2008;33:244553.
[49] Davila-Vazquez G, Cota-Navarro CB, Rosales-Colunga LM, Len-Rodrguez A,
Razo-Flores E. Continuous biohydrogen production using cheese whey:
improving the hydrogen production rate. Int J Hydrog Energy 2009;34:
4296304.
[50] Yuan Z, Yang H, Zhi X, Shen J. Increased performance of continuous stirred
tank reactor with calcium supplementation. Int J Hydrog Energy 2010;35:
26226.
[51] Keskin T, Giusti L, Azbar N. Continuous biohydrogen production in immobilized biolm system versus suspended cell culture. Int J Hydrog Energy
2012;37:141824.
[52] Chang FY, Lin CY. Biohydrogen production using an up-ow anaerobic sludge
blanket reactor. Int J Hydrog Energy 2004;29:339.
[53] Zhao QB, Yu HQ. Fermentative H2 production in an upow anaerobic sludge
blanket reactor at various pH values. Biroresour Technol 2008;99:13538.
[54] Hernndez-Mendoza CE, Buitrn G. Suppression of methanogenic activity in
anaerobic granular biomass for hydrogen production. J Chem Technol Biotechnol 2014;89:1439.
[55] Yang H, Shao P, Lu T, Shen J, Wang D, Xu Z, et al. Continuous bio-hydrogen
production from citric acid wastewater via facultative anaerobic bacteria. Int J
Hydrog Energy 2006;31:130613.
[56] Wu X, Zhu J, Dong C, Miller C, Li Y, Wang L, et al. Continuous biohydrogen
production from liquid swine manure supplemented with glucose using an
anaerobic sequencing batch reactor. Int J Hydrog Energy 2009;34:663645.
[57] Kim SH, Han SK, Shin HS. Optimization of continuous hydrogen fermentation
of food waste as a function of solids retention time independent of hydraulic
retention time. Process Biochem 2008;43:2138.
[58] Wu X, Yao W, Zhu J. Effect of pH on continuous biohydrogen production from
liquid swine manure with glucose supplement using an anaerobic sequencing
batch reactor. Int J Hydrog Energy 2010;35:65929.
[59] Jung KW, Kim DH, Kim SH, Shin HS. A simple method to reduce the start-up
period in a H2-producing UASB reactor. Int J Hydrog Energy 2011;36:146673.
[60] Wu SY, Hung CH, Lin CY, Lin PJ, Lee KS, Lin CN, et al. HRT-dependent hydrogen
production and bacterial community structure of mixed anaerobic microora
in suspended, granular and immobilized sludge systems using glucose as the
carbon substrate. Int J Hydrog Energy 2008;33:15429.
[61] Kim SH, Han SK, Shin HS. Performance comparison of a continuous-ow
stirred-tank reactor and an anaerobic sequencing batch reactor for fermentative hydrogen production depending on substrate concentration. Water Sci
Technol 2005;52:239.
[62] Carrillo-Reyes J, Celis LB, Alatriste-Mondragn F, Razo-Flores E. Different startup strategies to enhance biohydrogen production from cheese whey in UASB
reactors. Int J Hydrog Energy 2012;37:5591601.
[63] Lin CY, Wu SY, Lin PJ, Chang JS, Hung CH, Lee KS, et al. Pilot-scale hydrogen
fermentation system start-up performance. Int J Hydrog Energy 2010;35:
134527.
[64] Thanwised P, Wirojanagud W, Reungsang A. Effect of hydraulic retention time
on hydrogen production and chemical oxygen demand removal from tapioca
wastewater using anaerobic mixed cultures in anaerobic bafed reactor (ABR).
Int J Hydrog Energy 2012;37:1550310.
[65] Chen CC, Lin CY. Start-up of anaerobic hydrogen producing reactors seeded
with sewage sludge. Acta Biotechnol 2001;21:3719.
[66] Zhang ZP, Show KY, Tay JH, Liang DT, Lee DJ, Jiang WJ. Effect of hydraulic
retention time on biohydrogen production and anaerobic microbial community. Process Biochem 2006;41:211823.
[67] Lin CY, Cheng CH. Fermentative hydrogen production from xylose using
anaerobic mixed microora. Int J Hydrog Energy 2006;31:83240.
[68] Jo JH, Jeon CO, Lee DS, Park JM. Process stability and microbial community
structure in anaerobic hydrogen-producing microora from food waste
containing kimchi. J Biotechnol 2007;131:3008.
[69] Kim DH, Kim SH, Ko IB, Lee CH, Shin HS. Start-up strategy for continuous
fermentative hydrogen production: early switchover from batch to continuous
operation. Int J Hydrog Energy 2008:33153241.
[70] Akutsu Y, Li YY, Tandukar M, Kubota K, Harada H. Effects of seed sludge on
fermentative characteristics and microbial community structures in thermophilic hydrogen fermentation of starch. Int J Hydrog Energy 2008;33:65418.
[71] Chang JS, Lee KS, Lin PJ. Biohydrogen production with xed-bed bioreactors.
Int J Hydrog Energy 2002;27:116774.

P. Bakonyi et al. / Renewable and Sustainable Energy Reviews 40 (2014) 806813

[72] Gadow SI, Jiang H, Hojo T, Li YY. Cellulosic hydrogen production and microbial
community characterization in hyper-thermophilic continuous bioreactor. Int
J Hydrog Energy 2013;38:725967.
[73] Akutsu Y, Lee DY, Chi YZ, Li YY, Harada H, Yu HQ. Thermophilic fermentative
hydrogen production from starch-wastewater with bio-granules. Int J Hydrog
Energy 2009;34:506171.
[74] Antonopoulou G, Gavala HN, Skiadas IV, Lyberatos G. Inuence of pH on
fermentative hydrogen production from sweet sorghum extract. Int J Hydrog
Energy 2010;35:19218.
[75] Azbar N, Dokgz FTC, Keskin T, Korkmaz KS, Syed HM. Continuous fermentative hydrogen production from cheese whey wastewater under thermophilic
anaerobic conditions. Int J Hydrog Energy 2009;34:74417.
[76] Hussy I, Hawkes FR, Dinsdale R, Hawkes DL. Continuous fermentative hydrogen production from sucrose and sugarbeet. Int J Hydrog Energy 2005;30:
47183.

813

[77] Jung KW, Kim DH, Kim SH, Shin HS. Continuous fermentative hydrogen
production from coffee drink manufacturing wastewater by applying UASB
reactor. Int J Hydrog Energy 2010;35:133708.
[78] Jung KW, Kim DH, Kim SH, Shin HS. Continuous fermentative hydrogen and
methane production from Laminaria japonica using a two-stage fermentation
system with recycling of methane fermented efuent. Int J Hydrog Energy
2012;37:1564857.
[79] Yang P, Zhang R, McGarvey JA, Benemann JR. Biohydrogen production from
cheese processing wastewater by anaerobic fermentation using mixed microbial communities. Int J Hydrog Energy 2007;32:476171.