Energy balance of dark anaerobic fermentation as a tool for sustainability analysis
-
Upload
independent -
Category
Documents
-
view
2 -
download
0
Transcript of Energy balance of dark anaerobic fermentation as a tool for sustainability analysis
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 5 ( 2 0 1 0 ) 1 0 2 0 2e1 0 2 1 1
Avai lab le at www.sc iencedi rect .com
journa l homepage : www.e lsev ie r . com/ loca te /he
Energy balance of dark anaerobic fermentation as a toolfor sustainability analysis
Bernardo Ruggeri*, Tonia Tommasi, Guido Sassi
Dept. of Material Science and Chemical Engineering, Politecnico di Torino Corso Duca Degli Abruzzi 24, 10129 Turin, Italy
a r t i c l e i n f o
Article history:
Received 8 April 2010
Received in revised form
4 August 2010
Accepted 4 August 2010
Keywords:
BioH2
Dark fermentation
Sustainable energy balance
Methane
* Corresponding author. Tel.: þ39 011 090 46E-mail addresses: bernardo.ruggeri@polit
0360-3199/$ e see front matter ª 2010 Profedoi:10.1016/j.ijhydene.2010.08.014
a b s t r a c t
A process aimed at producing energy needs to produce more energy than the energy
necessary to run the process itself in order to be energetically sustainable. In this paper, an
energy balance of a batch anaerobic bioreactor has been defined and calculated, both for
different operative conditions and for different reactor scales, in order to analyze the
sustainability of hydrogen production through dark anaerobic fermentation. Energy
production in the form of hydrogen and methane, energy to warm up the fermentation
broth, energy loss during fermentation and energy for mixing and pumping have been
considered in the energy balance. Experimental data and literature data for mesophilic
microorganism consortia have been used to calculate the energy balance. The energy
production of a mesophilic microorganism consortium in a batch reactor has been studied
in the 16e50 �C temperature range. The hydrogen batch dark fermentation resulted to only
have a positive net production of energy over a minimal reactor dimension in summer
conditions with an energy recovery strategy. The best working temperature resulted to be
20 �C with 20% of available energy. Hydrogen batch dark fermentation may be coupled with
other processes to obtain a positive net energy by recovering energy from the end products
of hydrogen dark fermentation. As an example, methane fermentation has been consid-
ered to energetically valorize the end products of hydrogen fermentation. The combined
process resulted in a positive net energy over the whole range of tested reactor dimension
with 45e90% of available energy.
ª 2010 Professor T. Nejat Veziroglu. Published by Elsevier Ltd. All rights reserved.
1. Introduction processing and water electrolysis [3] potentiates the biological
The search for renewable energy sources is at the present
a necessity [1] as a consequence of the shortage of hydro-
carbon and of global warming due to CO2 emission. In this
context, the use of hydrogen, produced from non-fossil
sources, can be considered as a clean and sustainable source
of energy. Hydrogen can be produced in several ways, e.g.,
through the electrolysis of water, the steam reforming of
no-renewable hydrocarbons, and the gasification of biomass
and biological processes [2]. The high cost of hydrocarbon
47; fax: þ39 011 090 4699.o.it (B. Ruggeri), tonia.tomssor T. Nejat Veziroglu. P
production of hydrogen, especially when organic wastes are
used, as a component of the sustainable energy market [4].
The biological processes are classified as photo fermentation
and dark fermentation processes. The photosynthesis of H2
by bacteria or algae offers an opportunity for direct trans-
formation of solar energy. In spite of the theoretical oppor-
tunity, the real application is currently difficult [5] mainly due
to the bioreactor scale up [6] and to the low efficiency in the
use of solar light [7,8]. The dark fermentation approach is
based on the well-known technology of the anaerobic
[email protected] (T. Tommasi).ublished by Elsevier Ltd. All rights reserved.
Nomenclature
cp specific heat of the broth, [MJ kg�1 K�1]
D diameter of the bioreactor, [m]
DB diameter of the bench scale bioreactor, [m]
di diameter of the impeller, [m]
EH2 energy produced as H2 per unit volume of the
reactor, [MJ m�3]
ECH4 energy produced as CH4 per unit volume of the
reactor, [MJ m�3]
Ew energy for the initial warming per unit of the
reactor, [MJ m�3]
El energy loss per unit of the reactor, [MJ m�3]
Ee energy formixing and pumping per unit volume of
the reactor, [MJ m�3]
En net energy production per unit volume of the
reactor, [MJ m�3]
F filling fraction of the bioreactor volume, [e]
g acceleration of gravity, [m s�2]
he external convective heat transfer coefficient,
[kJ h�1 m�2 K�1]
hi internal convective heat transfer coefficient,
[kJ h�1 m�2 K�1]
HCH4 lower heating value of methane, 36.18 MJ/Nm3,
[MJ Nm�3]
HH2 lower heating value of hydrogen, 10.8 MJ/Nm3,
[MJ Nm�3]
L height of the bioreactor, [m]
ksteel thermal conductivity of steel, [MJ h�1 m�1 K�1]
kfoam thermal conductivity of the expanded polystyrene
foam, [MJ h�1 m�1 K�1]
N impeller rotation velocity, [rpm]
NB impeller rotation velocity of the bench scale
bioreactor, [rpm]
PH2 ðTwÞ H2 production vs. temperature per unit volume of
broth, [Nm3 m�3]
PCH4 ðTwÞ CH4 production vs. temperature per unit volume
of broth, [Nm3 m�3]
Pn Power number, [e]
Pw electrical power, [MJ h�1]
Re rotational Reynolds number Re a ND2, [e]
ROP RedOx Potential, [mV]
ssteel Thickness of the reactor wall, [m]
sfoam thickness of the reactor insulator, [m]
Ta outdoor ambient temperature, [�C]Tw working temperature in the bioreactor, [�C]U global heat transfer coefficient, [kJ h�1m�2 K�1]
VFA volatile fatty acids, [mg L�1]
VSS volatile suspended solid, [mg L�1]
VG produced gas volume per broth volume, [L L�1]
Greek
r biomass density, [mg L�1]
h global thermal efficiency of the heating system,
[e]
hcomb combustion efficiency, [e]
he electrical/mechanical conversion efficiency, [e]
hheat exc efficiency of the heat exchanger, [e]
Dt total duration of the batch fermentation, [h]
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 5 ( 2 0 1 0 ) 1 0 2 0 2e1 0 2 1 1 10203
digestion of organic wastes for biogas production. Anaerobic
digestion has been reported to operate in continuous and in
batchmodes, depending on the carbon source availability and
seasonality, the reacting time, the mixing approach and the
technological approach [9].
Dark fermentation is a promisingway of using inexpensive
feedstock, from several organic waste streams, as a substrate
source for hydrogen production [10]. Reviews on recent
studies on anaerobic digestion to produce hydrogen are
available in Refs. [11,12]. Nevertheless, several non-linear
effects, which influence the fermentative hydrogen produc-
tion, still remain to be clarified. The main factors under study
are: microorganism inoculum procedures, carbon sources,
reactor types, nitrogen phosphate and metal concentrations
(especially Fe2þ), fermentation temperature, agitation and
fermentation pH [13,14]. Studies on the macro effects of these
factors on fermentative hydrogen production have recently
been reviewed [15]. Fermentation temperature and pH have
been demonstrated to be the most relevant parameters that
affect the process performances [15e17], because of their
influence on the activity of hydrogen producing bacteria.
Essential enzymes for fermentative hydrogen production,
e.g., Hydrogenase and Ferrodoxin, require adequate pH and
temperature ranges [18,19]. Basification is necessary to
maintain the pH and the redox potential in the broth at the
most appropriate range to enhance the activity of the Ferro-
doxin enzyme pool [20].
Temperature affects the general cell viability and the
catabolic pathways. A low temperature depletes cell activity
and the hydrogen production rate, while high temperature
leads to more oxidized compounds to be available as electron
acceptors. This low temperature depletes hydrogen produc-
tion [21,22]. The temperature range depends on the type of
microorganism consortium that is used. The lower boundary
of the temperature range is generally similar, i.e., around
12e16 �C, while the optimal temperature may vary in the
mesophilic range, i.e., around 37 �C, and the thermophilic
range, i.e., around 55 �C, [23]; however 22 �C [24] and 60 �C[25,26] have also been reported. From an energetic point of
view, the fermentation temperature also determines the
energetic level to which all the materials must be warmed to
and the driving force for heat dispersion to the external
environment. The best working temperature can be different
from the optimal temperature for the microbial consortium
activity. The total energy balance is the key equation for
fullscale reactor design and the fermentation temperature is
the most relevant parameter [27,28]. Nevertheless very few
papers have dealt with the total energy balance in hydrogen
production [28]. The relevance of the energetic reasoning is
evident for the production of hydrogen as an energetic vector;
the value of the fermentation temperature determines the net
energy produced. Hydrogen production through dark
fermentation involves a partial oxidation of the carbon source
to fatty acids [13]. The outlet stream from the reactor still has
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 5 ( 2 0 1 0 ) 1 0 2 0 2e1 0 2 1 110204
a chemical energetic content that can be used for further
energy production, e.g., through anaerobic digestion with
a resulting methane production. The overall process, e.g.,
hydrogen production and subsequent methane production,
should have a higher specific net energy production due to the
thermal integration, which leads to an energy saving. In the
present paper, the performances of a mesophilic consortium
that produces hydrogen in a batch reactor are studied as
a function of the fermentation temperature. The total energy
balance is applied to investigate the energetic performances of
the process as a function of the basic technological parame-
ters of a fullscale bioreactor, i.e., reactor dimension and
reactor insulation. The net energy produced by the hydrogen
process alone and by the hydrogen process coupled with
a subsequent anaerobic digestion, i.e., methane production
from the outlet stream of the hydrogen reactor, has been
calculated. The net energy is reported as a function of the
reactor dimension and fermentation temperature in order to
identify the acceptable conditions for both summer and
winter operations. The importance of energy recovery strat-
egies has been verified.
2. Materials and methods
2.1. Seed microflora for H2 and CH4 production
The seed microorganism consortium for H2 production was
prepared according to Chen et al. [29] and Mu et al. [30]. The
sludge of an anaerobic digester working at 38 �C fed by
municipal wastewater (SMAT S.p.A., Turin) was pretreated by
leaving for 24 h at pH 3 in anaerobic conditions. The pH value
was reached by adding a 1 N HCl solution to the sludge. The
initial conditions of the sludgewere: pH 7.2; density 890 kg/m3,
volatile suspended solid (VSS) 11,560 mg/L and total solid
concentration (TSS) 16,500 mg/L. The seed microorganism
consortium for CH4 production was prepared from the
untreated sludge of the same anaerobic digester (SMAT S.p.A.,
Turin). The sludge was withdrawn once from the anaerobic
digester and partially treated as previously described. The
treated and untreated sludge portions were divided into 50 ml
amounts and frozen. Each amount was unfrozen at ambient
temperature under anaerobic conditions before being inocu-
lated into the reactor.
2.2. Reactor and medium composition
A 2 L working volume stirred-batch reactor (Minifors HT,
Switzerland) was filled whit a glucose synthetic medium [31]
(NaHCO3 1.25 g/l, NH4Cl 2.5 g/l; KH2PO4 0.25 g/l; K2HPO4
0.250 g/l; CaCl2 0.5 g/l; NiSO4 32 mg/L; MgSO4$7H2O 320 mg/l;
FeCl3 20 mg/l; Na2BO4$H2O 7.2 mg/l; Na2MoO4$2H2O 14.4 mg/l;
CoCl2$6H20 21 mg/l; MnCl2$4H2O 30 mg/l; Yeast extract
50 mg/l). 10% v/v seed was inoculated into the medium, N2
was sprinkled for 10 min to reach anaerobic conditions and
the broth was warmed to the chosen temperature. The C/N
ratio was set at 30 and the glucose concentration at 60 g/L in
order to maximize the hydrogen production rate, according to
previously kinetic studies [19].
2.3. Test runs for H2 and CH4 production
The batch test runs were conducted at 16, 20, 35, 40 and 50 �C.The initial pH was set at 7.2 by 1 N HCl solution. After the
initial free decrease till 5.2, the pH value was maintained by
pumping in a 2 N NaOH solution [20]. The test runs were shut
down when the biogas production stopped.
The broth of one of the test runs, conducted at 35 �C, was
inoculated with 10% untreated sludge after the end of the run
when the pH reached 7 by free evolution of the broth. The
test run was shut down when biogas production stopped
again.
2.4. Analytical tests
The pH (HA405-DXK-S8, Mettler Toledo), RedOx Potential
(ROP, Pt4805-DXK-S8, Mettler Toledo), temperature, NaOH
flow rate (Iris, Infors HT) and total gas produced volume
(Milligas Counter, Ritter) were monitored and registered every
5 min. Liquid samples were withdrawn twice a day during the
test runs, the glucose concentration was measured by an
enzymatic method (Biopharm-Roche) and the volatile fatty
acids and ethanol concentrations were measured by a gas-
chromatographic method (Model 6580, Agilent). The gas was
collected in bags whichwere changed on the basis of time and
the gas quantity; average concentrations of H2, O2, CH4, CO2,
CO, N2 were measured in each bag by means of off-line gas-
chromatographic analysis (Varian, CP 4900).
2.5. Uncertainty evaluation
The uncertainty of net energy estimated values was evalu-
ated in accord to the rules reported in Ref. [40]. At the end
to evaluate the uncertainty, considering that Guide to
Uncertainty Measurement (GUM) defines uncertainty as
a quantifiable parameter associated with the results of
a measurement procedure, the suggested approach has been
utilized either to evaluate the uncertainty or to estimate the
most affecting parameters. This latest was obtained by using
the expression reported in Ref. [41] known as the law of the
propagation of uncertainty based on the evaluation of the
partial derivatives of the parameters on the estimation of
the net energy, called sensitivity coefficients which describe
how the output estimate, varies with changes in the value of
the input estimates.
3. Energy balance methodology
The net energy produced (En) in a bioreactor by dark fermen-
tation is the difference between the energy contained in the
generated gas, i.e., contained in the hydrogen (EH2 ) and/or the
methane (ECH4 ) produced, and the energy spent to obtain and
maintain the reaction conditions. The last term involves the
warm up of the inlet materials to start fermentation (Ew), the
heat loss to the environment (El) during fermentation and the
electrical energy for mixing and pumping (Ee).
The net energy production may be expressed as:
En ¼ EH2þ ECH4
� ðEw þ El þ EeÞ (1)
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 5 ( 2 0 1 0 ) 1 0 2 0 2e1 0 2 1 1 10205
The energy balance considers a reference volume and
a reference time period onwhich each termmust be evaluated
and calculated. The system under study is a batch reactor into
which raw materials are introduced, agitated, warmed to the
working temperature, inoculated by seed and left to evolve,
controlling the pH by adding an alkaline solution till the gas
production stopped. In this case, the reference volume is the
reactor volume and the reference time period is the duration
of a single batch run. Intensive terms may be calculated,
referring to the unit working volume of the reactor. A batch
process was considered because hydrogen produced as
a function of the working temperature has mainly been
studied in detail and recognized in Ref. [15] for batch
processes. Table 1 summarizes the data from several litera-
ture cases of batch hydrogen production at the optimal
working fermentation temperature. In order to investigate the
temperature and scale-up effects, the net energy per unit
volume of the reactor was calculated at different working
fermentation temperatures and different dimensions of the
reactor. The reactor diameter D was chosen as the reactor
dimension parameter, and the reactor was considered to be
cylindrical with a height-to-diameter ratio equal to 4. Two
different outdoor ambient conditions were considered in
order to evaluate the effect of seasonal variations of the
ambient temperature. Each term of Equation (1) is discussed
hereafter.
3.1. Energy production
The energy produced per unit volume of the reactor is the total
energy embedded in the gas produced by a single batch run
referring to the reactor volume, i.e., the energy contained in
the amount of hydrogen and methane retrieved from a single
batch run referring to the reactor volume.
For hydrogen and methane, this can be calculated as:
EH2¼ F$PH2
ðTwÞ$HH2; ECH4 ¼ F$PCH4 ðTwÞ$HCH4 (2)
where PH2ðTwÞ and PCH4 ðTwÞ are the specific production of H2
and CH4 respectively, i.e., the amount of H2 or CH4 produced in
a single batch run per unit volume of broth at the working
temperature. These are expressed as Nm3 of H2 or CH4 per m3
of fermenting broth, and their value, which was determined
experimentally, depends on the working temperature. In
Table 1 e Highest specific hydrogen production and duration fmicroorganism consortium producing hydrogen.
References Substrate Microorganisms In
[15] Glucose Mixed culture
[38] Sucrose Mixed culture
[19] Glucose Mixed culture
[17] Glucose Mixed culture
[Present study] Glucose Mixed culture
[26] Sucrose Mixed culture
[26] Sucrose Mixed culture
[18] Cattle wastewater Mixed culture
[39] Glucose Geothermal hot
spring sediments
addition, HH2is the lower heating value of hydrogen
(10.8 MJ/Nm3) and HCH4 that of methane (36.18 MJ/Nm3). F is
the liquid hold up in the reactor, i.e., the fraction of reactor
volume filled by liquid.
3.2. Warming energy
The energy required to warm the fermenting broth mainly
depends on its specific heat (cp), the difference between the
working temperature (Tw) and the outdoor ambient tempera-
ture (Ta) and the efficiency of the heating system (h). The
heating energy per unit volume of the bioreactor may be
calculated as:
Ew ¼�r$cp$ðTw � TaÞ$F
�h
(3)
Where the r and cp of water were used for the fermenting
broth. The warming device was considered to be composed
of a combustion boiler (hcomb z 0.8) and a heat exchanger
(hheat exc z 0.6). The global efficiency of the warming system
was calculated as the product (hz 0.48). The outdoor ambient
temperaturewas considered for different seasonal conditions,
i.e., summer and winter conditions. Ta was calculated on the
basis of historical data from northern Italy; mean night and
day values over the season were considered in order to avoid
an increase in the computational complexity: Ta ¼ 5 �C for the
winter and Ta ¼ 15 �C for the summer. A heat recovery of 50%
was considered, i.e., at the end of the batch run 50% of the heat
of the broth is considered to have been recovered. The
warming of the reactor wall and insulator and that of the
NaOH solution was neglected in order to consider the most
optimistic situation in which specific strategies are performed
to enhance energy saving.
3.3. Heat loss
The difference between the working temperature of broth Tw
and the ambient temperature Ta outside the reactor is
responsible for the heat loss from the fermenting broth. The
energy lost must be supplied from the heating system of the
temperature control system and it depends on the insulation
of the fermenting broth from the external environment, the
surface area exposed to the environment and the duration of
or batch operative conditions with mesophilic
itialpH
pHControl
Tw (�C) Dt (h) Produced H2
(mmol/L)
7.0 no 40 25 123
8.0 no 35 25 207
at 5.5 41 24 91
7.5 no 25e26 50 60
7.2 at 5.2 35 340 442
8.5 no 22 466 120
8.5 no 37 90 50
at 5.5 45 30 8
6.5 no 51.7 16 625
Table 2 e Experimental H2 production results: H2 produced, run duration and relative mass abundance of the measuredVOC.
Tw
(�C)H2 produced(mmol/L)
Dt(days)
Acetate(%)w
Butyrate(%)w
Propionate(%)w
Formate(%)w
Lactate(%)w
Ethanol(%)w
16 15.4 6.6 1.2 60.2 0.4 1.5 18.0 18.4
20 215.3 23.9 2.9 81.1 0 2.8 0.04 13.3
35 442.0 13.8 7.9 90.8 0.1 0.6 0 0.6
40 96.0 8.4 4.9 15.3 0.1 10.0 68.2 1.5
50 1.1 12.5 1.7 4.0 0.01 3.1 90.4 0.9
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 5 ( 2 0 1 0 ) 1 0 2 0 2e1 0 2 1 110206
the batch run (Dt). The energy loss per unit volume of reactor
may be calculated as:
El ¼
�4:5D
$kfoam
sfoam$DtðTwÞ$ðTw � TaÞ
�
h(4)
in which the bottom and the top of the bioreactor walls were
considered insulated with the same thickness as the vertical
wall. The bioreactor walls were considered to be constituted
by a steel 2.5 mm thick wall, as the structural material and
300 mm thick expanded polystyrene foam as the insulating
material. The total duration of fermentation Dt depends on
the working temperature, which was experimentally deter-
mined as reported in Tables 1e3.
The total resistance, i.e., the reciprocal of the total heat
transfer coefficient U, accounts for the total insulation of the
broth from the outside environment, and it can be calculated
as the sum of the single resistances, i.e., internal broth, steel,
foam and external air resistances:
U�1 ¼ h�1i þ ssteel
ksteelþ sfoamkfoam
þ h�1e (5)
where h is the convective heat transfer coefficient for the
internal and external fluid, s the thickness of the steel and
foamand k the thermal conductivity of the steel and foam [32].
Very thick polystyrene foam makes the foam resistance the
only relevant contribution to the total resistance, as graphi-
cally reported in Fig. 1, therefore both the convective internal
and external fluid resistance, as well as the steel resistance,
may be neglected to simplify the calculation. The resistance to
heat transport is here only considered in the insulating
material (sfoam/kfoam). This assumption leads to over-
estimating the insulator thickness at the same energy loss.
The term 4.5/D accounts for the exchange surface per unit
volume of the reactor, and it depends on geometrical
assumptions, i.e., the height-to-diameter ratio.
Table 3 e Numerical values to calculate the net energyproduction for integrated H2 and CH4 production.
Tw
(�C)H2 produced(mmol H2/L)
CH4 produced(mmolCH4/L)
Total Energy(kJ/L)
Test RunDurationDt (days)
16 15.4 391 317.5 22.6
20 220 452 409.0 39.9
35 442 530 519.2 29.8
3.4. Electrical energy for mixing and pumping
The electrical energy consumed to run the bioreactor is for
mixing, filling up and empting the bioreactor using a pump. In
batch fermentation, the raw material and the inoculum are
pumped in at the beginning of a run and the broth is pumped
out at the end of the run. The reactor ismixed all along the run.
The energy for pumping may roughly be calculated as the
energy necessary to lift the broth to the top of the reactor, i.e.,
E ¼ rgL; it accounts for the efficiency due to the pressure drop
and the partial recovery from empting the reactor, where L is
the height of the reactor.
The power number and the rotational Reynolds number
were considered to evaluate the mixing performances of the
bioreactor [33]. As the turbulence scale-up criteria, the rota-
tional Reynolds number was considered independent of the
reactor diameter in order to evaluate the energy necessary to
mix the fermenting broth at different diameters, i.e., the
power number is independent of the reactor diameter [33,34].
Geometrical similitude was assumed for the vessel and
impeller scale up, i.e., an equal impeller-to-reactor diameter
ratio. The impeller-to-reactor diameter ratio in the bench
scale bioreactor was set at 0.5. The following equation allows
us to estimate the electrical power necessary tomix the broth:
Pw ¼ Pn$r
8gp$N3
B$D6B
D4(6)
The procedure reported by Bailey and Ollis [34] was used to
calculate the power number (Pn) for the bench scale reactor. In
Fig. 1 e (a) Geometrical similarity used in the scale-up
procedure D/di [ 2 (b) Assumption used for the evaluation
of the heat loss through the bioreactor wall.
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 5 ( 2 0 1 0 ) 1 0 2 0 2e1 0 2 1 1 10207
the Equation (6) where subscript B refers to the bench scale
bioreactor, Pw is the power per volumeunit volume required to
reach the target mixing performances, i.e., the value of the
Reynolds number equal to that of the bench reactor. The total
electrical energy is the sum of the pumping and the mixing
terms, and the mixing power must be multiplied by the
duration of the run:
Ee ¼ Pw$DtðTwÞhel
þ 2rgL ¼ Pn$r$N3B$D
6B
8gp$hel
$DtðTwÞD4
þ 8rgD (7)
An efficiency factor of electrical energy conversion into
mixing energy of 0.75 was considered. The pumping term
resulted to be around 100 J/L per meter of reactor diameter. In
the considered diameter range this term resulted to be negli-
gible with respect to the mixing term, which means that it
is not interesting to calculate the pumping term more
accurately.
4. Results
4.1. Experimental results
4.1.1. Hydrogen productionThe hydrogen evolution was measured in the bench scale
bioreactor at each investigated temperature and is reported in
Fig. 2, while Table 2 reports, for each temperature, the main
experimental results, i.e., the specific hydrogen productivity,
the batch test run duration and the relative abundance of the
main metabolites at the end of the fermentation. The meso-
philic microorganism consortium, obtained from a waste-
water treatment sludge operating at 38 �C and treated at
acid conditions, has resulted to have a maximal hydrogen
production at 35 �C; this result agrees with the literature data
reported in Table 1. The mesophilic microorganism consor-
tium stops hydrogen production at 50 �C and reduces the
production of hydrogen to a very low level under 16 �C. Themetabolites reported in Table 2 are those that were recognized
[19] and measured in the test runs at the end of fermentation.
The data are reported as the mass percentage of the total
measured metabolites at the end of the test run. Acetate,
propionate and butyrate are known to be the compounds that
are stoichiometrically related to hydrogen production [13,19]
Fig. 2 e Hydrogen production vs. time in the bench scale
bioreactor (2 L, pH 5.2; 60 g/l glucose, N [ 100 rpm).
their relative amount depends on the temperature, which is
in agreement with literature findings [22,35]. At a lower
temperature than 35 �C, the main product is butyrate; its
relative abundance diminishes as the temperature decreases
while the ethanol and lactate abundance increases. Over
35 �C, lactate becomes the main product and little ethanol is
observed. It is therefore evident that temperature variations
shift the metabolic pathways. Acetate and butyrate abun-
dances have been confirmed to be closely related to hydrogen
production [39].
4.1.2. Methane productionAcetate and butyrate constitute 84% and 99% of the total
volatile organic compounds at 20 and 35 �C respectively. Thus
the broth at the end of fermentation contains the main
residual chemical energy in the form of acetate and butyrate;
they constitute the most ideal substrate for the methanogen
consortium [9]. The methanogen consortium can be inocu-
lated at the end of hydrogen production to produce methane
from fatty acids in order to utilize the energy contained in the
initial carbon source asmuch as possible. The gas evolution of
a 35 �C test run with hydrogen and the subsequent methane
production are reported in Fig. 3a. The first part of the test run
involved the dark fermentation to produce hydrogen and
volatile fatty acids after an inoculum with sludge treated at
acid conditions. The second part was the secondary fermen-
tation of the fatty acids to producemethane after an inoculum
with untreated sludge. As an average, 80% v/v ofmethanewas
measured in the biogas produced during the second part of
Fig. 3 e Time curve evolution of glucose fermentation
(initial concentration 60 g/l) (a) cumulative gas production
H2 and CH4 (b) pH and ROP evolution.
Fig. 5 e Net energy production vs. diameter of a bioreactor
for H2 production with 50% warming up energy recovery:
(a) winter and (b) summer conditions.
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 5 ( 2 0 1 0 ) 1 0 2 0 2e1 0 2 1 110208
fermentation. Fig. 3b reports the pH and ROP evolution during
the 35 �C test run; the pH of the broth evolved naturally till
reaching neutrality after the end of the hydrogen production.
At neutral pH, the reactor was inoculated by the untreated
sludge. The ROP remains at a reductive range at around
�500mV over the duration of the whole run. The correction of
pH at the end of hydrogen production did not lead to
a subsequent methane production. The natural evolution of
pH seems to allow microorganism to modify their environ-
ment and activity to reach the right condition for secondary
inoculum. The methanogenic consortium lag phase was
around 10 days at pH 7.2. As the gas production restarted, the
pH increased till 9, and the methane concentration in the gas
reached 80e85%. In order to evaluate the methane production
in the range less than 35 �C, a methanogenic smoothing
function, retrieved from Ref. [27], was used to extrapolate data
from themethane production data experimentally obtained at
35 �C. The experimental data at 35 �C and the data extrapo-
lated at 16 and 20 �C using a smoothing function [36] are
reported in Table 3. The total energy production, as well as the
hydrogen and methane contribution, is reported in Fig. 4 as
a function of the working temperature; the methane contri-
bution to energy production was calculated to be greater than
80% over the whole temperature range.
4.2. Net energy evaluation
The net energy was calculated for a single batch run using
Equations (1)e(7) for each working temperature for the
summer and winter reference ambient temperatures for
a bioreactor diameter ranging from 0.5 to 10 m. An expanded
polystyrene foam thickness of 30 cm was considered to
insulate the bioreactor.
4.2.1. Hydrogen productionThe net energy production for a batch run producing
hydrogen is reported in Fig. 5 as a function of the reactor
diameter. Fig. 5a and b refers to winter and summer condi-
tions, respectively. As a first result, the net energy for a 20 �Cworking temperature is always greater than that for a 35 �C in
spite of the lower hydrogen production at 20 �C, in which the
energy production is about 50% of the maximum at 35 �C(See Table 2 and Fig. 4). The net energy production increases
with diameter, due to the diminution of the external surface
area per unit volume which enhances the insulation and the
Fig. 4 e Specific energy production vs. temperature: H2 and
CH4 productions.
reduction of energy loss per unit volume. The energy loss
becomes more relevant under a critical reactor dimension,
around 3m in diameter, for the geometrical assumptions here
considered, i.e., height-to-diameter ratio and foam thickness.
Over the critical reactor dimension, the relevance of the
energy loss is almost independent of the reactor dimension.
At the winter condition, the net energy production is negative
over the whole diameter range, while, at the summer condi-
tion with 50% heat recovery, the net energy becomes positive
over a 1 m diameter at a 20 �C working temperature and over
a 3 m diameter at 35 �C. At 20 �C, the net energy is around 20%
of the total energy produced, while at 35 �C, the net energy is
around 5% of the total energy produced. The main energy
consumption is due to the initial warming of the broth, even
though an energy recovery was considered and the warming
of the reactor structure was neglected. If the energy recovery
is not considered, the net energy production is also negative
for summer conditions. Similar results were obtained using
the specific production data of other researchers, reported in
Table 1. The production of hydrogen for energy purposes by
batch runs has proved to be critical, and particular care has to
be taken in the raw material treatment, the energy recovery
and the warming strategies in order to design a low energy
requirement pretreatment and warming process, e.g., the use
of solar warming or of thermal wastes to warm the broth
and more favorable climatic conditions. In general, from
a sustainable energy point of view, a secondary process
should be coupled to hydrogen production. An energetic
analysis of the whole process should be performed. On the
other hand the Volatile Fatty Acids (VFA) and the other ener-
getic metabolites available in the broth at the end of hydrogen
fermentation may be energetically valorized, and several
approaches are available [9,37,38].
Fig. 7 e Net energy production vs. diameter for integrated
H2 and CH4 production with 50% warming energy recovery:
(a) winter and (b) summer conditions.
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 5 ( 2 0 1 0 ) 1 0 2 0 2e1 0 2 1 1 10209
4.2.2. Coupled hydrogen and methane productionAnaerobic digestion to produce biogas at a high methane
concentration is a possible way of energetically valorizing
the VFAs and the other compounds available in the broth at
the end of acidogenic fermentation. The methanation of the
residue seems to be one of the easiest processes to perform,
because the same reactor can be used without any further
addition or modification. In a first case, the net energy was
calculatedwithout any heat recovery; the results are shown in
Fig. 6 for the winter (6a) and summer (6b) conditions at the
lowest working temperature, i.e., 16, 20 and 35 �C. A higher
working temperature leads to a lower net energy production,
due to the lower energy production and the higher energy
requirement. The net energy production is positive over the
reactor dimension whole range both for the winter and the
summer average conditions. A 3m diameter was confirmed to
be the critical reactor dimension under which energy loss
plays amore important role. The 20 �Cworking temperature is
the most convenient from an energetic point of view over the
whole diameter range. Nevertheless, over a 2 m reactor
diameter, the net energy production resulted to be at about
the same level at 20 and 35 �C, i.e., the greater energy
production at 35 �C is balanced by the greater energy neces-
sary to warm and maintain the working temperature. This
means that, over a 2 m reactor diameter, the energetic criteria
are not relevant in order to choose the working temperature.
The net energy resulted to be 45e88% of the produced energy
as H2 and CH4 forms, depending on the working temperature
and the average seasonal conditions, the maximal percentage
is at 20 �C for the summer.
In the second case, net energy was calculated considering
the recovery of 50% of the warming energy, the results are
shown in Fig. 7 for winter (7a) and summer (7b) conditions.
The 35 �C working temperature is the most convenient, from
Fig. 6 e Net energy production vs. diameter for integrated
H2 and CH4 production without energy recovery: (a) winter
and (b) summer conditions.
an energetic point of view, over almost the whole reactor
dimension range, because of the greater relevance of the
warming energy in the balance. Nevertheless for a very small
reactor dimension the energy loss becomes very relevant and
20 �C could be considered as an energetically convenient
choice. The net energy resulted to be 70e80% of the produced
energy, depending on the working temperature and the
average seasonal conditions, while themaximal percentage is
at 35 �C for the summer; in Fig. 8 is reported a Sankey diagram
for the best situation. Once again the importance of the
running strategy is stressed; the recovery of energy makes it
energetically convenient to work at higher temperatures.
Working at higher temperatures reduces the run time and
enhances the hydrogen production, and it also leads to
Fig. 8 e Sankey diagram for the best situation: summer
time, working temperature 35 �C, 50% heat recovery,
reactor diameter > 4 m.
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 5 ( 2 0 1 0 ) 1 0 2 0 2e1 0 2 1 110210
economic advantages. The optimal reactor dimension
depends on many factors such as economic considerations,
robustness, reliability and flexibility. The present paper has
offered a rough evaluation for energetic sustainability. The
same procedure may be considered for a design criterion to
calculate the optimal insulator thickness.
4.2.3. UncertaintyAs concerns the main affecting parameters on the net energy
estimation, the warming energy compared to other energy
terms has the highest predominance. It reaches almost the
same numerical value both in winter time and in summer
time: 85% and 70% respectively. In regard to themain affecting
parameter, by applying the procedure recalled in Section 2.5
the variability either of ambient temperature Ta or that of
the global thermal efficiency h are the main sources of the
uncertainty. With a Ta variability around 5 �C and a relative
variability of the efficiency around 10% on the used values for
the calculation, the uncertainty of the net energy is around
50 kJ/L; the evaluation of the uncertainty without any vari-
ability of Ta and h, i.e., considering only the uncertainty due to
the experimental data results on the range (5e15) kJ/L, this
confirms that the Ta is the main affecting parameter in
anaerobic digestion. According to the value of the net energy
estimation reported in Figs. 6 and 7, we can consider the
suggested energy balance sufficiently acceptable in the eval-
uation of the sustainability of biohydrogen and biogas
productions. In any case for a specific design of a detailed
plant the uncertainty could be reduced.
5. Conclusions and comments
This paper has considered the energy balance to evaluate the
net energy produced, by batch runs of a bioreactor producing
hydrogen, as a function of the working temperature and
reactor scale. The procedure proposed in this paper allows an
evaluation of the net energy in order to define themain criteria
to choose the scale up and design strategies. For detailed
design purposes, it is necessary to define a more refined
calculation procedure. The evaluation of the net energy allows
maximizing the energy produced by the plant differently
to maximize the yield of hydrogen, as reported in Fig. 6.
The most relevant conclusion is that particular attention
must be devoted to the process energy saving strategy to reach
a positive net energy production. The main part of the energy
is required for the warming of the broth; an adequate strategy
needs to be designed in order to reduce this energy cost, e.g.,
the use of solar warming or thermal wastes and heat recovery
and thermal integration if it is possible. However, the effects
of seasonal differences on the net energy are relevant; by
uncertainty analysis the ambient temperature and its varia-
tion play the utmost important role. In the case of the
production of H2 alone, the best calculated available energy
was around 20%.
The broth at the end of dark fermentation contains
substances which embed a part of the energy originally in the
substrate. These substances can be valorized by integrating
a process with the hydrogen production. The anaerobic
fermentation of the residual metabolites to methane can
easily be thermally integrated with hydrogen production
through a two-step hydrogenemethane production. Energetic
sustainability is reached, for the integrated process, over the
whole range of operative conditions considered in the anal-
ysis. The net energy was calculated to be higher than 50% and
to even reach 80% in the best case. However, it is possible to
work at higher temperatures by enhancing the process strat-
egies to save energy, this have a consequent positive
economic impact. The reactor scale plays an important role
under a critical dimension due to the higher weight of the
energy loss. All the calculations have been performed at the
typical climatic conditions of northern Italy; warmer climatic
conditions would lead to an increase in the net energy
production.
Finally, it has been observed that the shift of the pH
towards neutral values and the addition of methanogens
inoculum are able to start a methanogenic fermentation to
produce methane using the VFAs and the other products at
the end acidogenic fermentation as the substrate. However,
the way the shift of the pH towards neutral is performed, may
influence the lag phase of themethanogenic consortium. This
aspect needs to be evaluated in detail both from a biological
and from an energetic point of view.
Lastly, the present evaluation of the net energy does not
take into account the quantities of energy spent to construct
the plant as well the energy spent to produce the chemicals,
(mainly NaOH) to control the pH during the fermentation. This
would be accomplished by a more comprehensive analysis
using Life Cycle Approach (LCA) in order to evaluate the
amount of energy delivered to society. By LCA approach the
net energy produced needs to be compared with the total
energy required to produce thematerial, to assemble the plant
and process, deliver and upgrade otherwise the sources, the
organic refuse in the present contest, in a useful form to give
an energy service. In the present paper only the net energy
was evaluated. The evaluation of “useful” energy delivered
into the society reserves several advantages: it assesses the
change in the physical scarcity of energy resources, it is
a measure of the potential of such technology to do useful
work in sustainable way and finally it is possible to rank
alternative energy supply technologies according to their
capacity to do useful work. Researchers in this area are in due
course; reliable results are welcomed in order to acquire an
evaluation of energetic of biohydrogen and biogas
productions.
r e f e r e n c e s
[1] Balat M. Potential importance of hydrogen as a futuresolution environmental and transportation problem. Int JHydrogen Energy 2008;33:4013e29.
[2] Das D, Veziroglu TN. Advances in biological hydrogenproduction processes. Int J Hydrogen Energy 2008;33:6046e57.
[3] Turner JA. Sustainable hydrogen production. Science 2004;305:972e4.
[4] Westermann P, Jorgensen B, Lange L, Ahring BK,Christensen CH. Review: maximizing renewable hydrogenproduction from biomass in a bio/catalytic refinery. Int JHydrogen Energy 2007;32:4135e41.
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 5 ( 2 0 1 0 ) 1 0 2 0 2e1 0 2 1 1 10211
[5] Roger CP, Kheshgi HS. The photobiological production ofhydrogen: potential efficiency and effectiveness asa renewable fuel. Critic Rev Microbiol 2005;31(1):19e31.
[6] Akkerman I, Janssen M, Rocha J, Wijffeld RH. Photobiologicalhydrogen production: photochemical efficiency andbioreactor design. Int J Hydrogen Energy 2002;27:1195e208.
[7] Anastasios M. Green alga hydrogen production: progress,challenges and prospects. Int J Hydrogen Energy 2002;27:1217e28.
[8] Miyake J, Kawamura S. Efficiency of light energy conversionto hydrogen by the photosynthetic bacterium Rhodospirillumsphaeraides. Int J Hydrogen Energy 1987;12(3):147e9.
[9] Reith JH, Wijffeels RH, Barten H. Status and perspectives ofbiological methane and hydrogen production. Nederland:Dutch Biological Foundation; 2003.
[10] Kapdan IK, Kargi F. Review e biohydrogen production fromwaste material. Enzym Microb Technol 2006;38:569e82.
[11] Nishio N, Nakashimada Y. High rate production of hydrogen/methane from various substrate and wastes. Adv BiochemEng Biotechnol 2004;90:63e87.
[12] Hawkes FR, Hussy I, Kyazze G, Dinsdale R, Hawkes DL.Continuous dark fermentative hydrogen production bymesophilic microflora: principles and progress. Int JHydrogen Energy 2007;32:172e84.
[13] Das D, Veziroglu TN. Hydrogen production by biologicalprocesses: a survey of literature. Int J Hydrogen Energy 2001;26:13e28.
[14] Sen U, Shakdwipee M, Banerjee R. Status of biologicalhydrogen production. J Sci Ind Res 2008;76(11):980e93.
[15] Wang JL, Wan W. Factor influencing fermentative hydrogenproduction: a review. Int J Hydrogen Energy 2009;34:799e811.
[16] Kaual SK, Chen WH, Li L, Sung S. Biological hydrogenproduction: effect of pH and intermediate product. Int JHydrogen Energy 2004;29:1123e31.
[17] Wang JL, Wan W. Effect of temperature on fermentativehydrogen production by mixed cultures. Int J HydrogenEnergy 2008;33:5392e7.
[18] Batie CJ, Kamin H. The relation of pH andoxidationereduction potential to the association state of theferrodoxin.ferrodoxin:NADPþ reductase complex. J BiolChem 1981;256(15):7756e63.
[19] Ruggeri B, Tommasi T, Sassi G. Experimental kinetics anddynamics of hydrogen production on glucose by hydrogenforming bacteria (HFB) culture. Int J Hydrogen Energy 2009;34:753e63.
[20] Adams MWW, Mortenson LE, Chen TS. Hydrogenase.Biochem Biophys Acta 1981;594:105e16.
[21] Tang GL, Huang J, Sun ZJ, Tang QQ, Yan CH, Liu GQ.Biohydrogen production from cattle wastewater by enrichedanaerobic mixed consortia: influence of fermentationtemperature and pH. J Biosci Bioeng 2008;106(1):80e7.
[22] Mu Y, Zhung XJ, Yu HQ, Zhu RF. Biological hydrogenproduction by anaerobic sludge at various temperatures.Int J Hydrogen Energy 2006;31:780e5.
[23] Li CL, Fang HHP. Fermentative hydrogen production fromwastewater and solid waste by mixed culture. Crit RevEnviron Sci Technol 2007;31:1e39.
[24] Gadhamshetty V, Johnson Dc DC, Nirmalakhandan N,Smith GB, Deng S. Feasibility of biohydrogen production atlow temperatures in unbuffered reactor. Int J HydrogenEnergy 2009;34:1233e43.
[25] Yokoyama H, Waki M, Moriya N, Yasuda T, Tanaka Y,Haga K. Effect of fermentation temperature on hydrogenproduction from cow waste slurry by using anaerobicmicroflora within the slurry. Appl Microbiol Biotechnol 2007;74(2):474e83.
[26] O-Thong S, Presertsan P, Karakashev D, Angelidaki I.Thermophilic fermentative hydrogen production by thenewly isolated Thermoanaerobacterium thermosaccharolyticumPSU-2. Int J Hydrogen Energy 2008;33(4):1204e14.
[27] Ruggeri B. Thermal analysis of anaerobic digesters. Chem Ind1984;66(7e8):477e83.
[28] Bonallagui H, Haonari O, Tauhami Y, Ben Cheikh R,Maronani L, Hamdi M. Effect of temperature on theperformance of an anaerobic tubular reactor treatingfruit and vegetable waste. Process Biochem 2004;39:2143e8.
[29] Chen CC, Lin CY, Liu MC. Acidebase enrichment enhancesanaerobic hydrogen production process. Appl MicrobiolBiotechnol 2002;58:224e8.
[30] Mu Y, Yu HQ, Wang G. Evaluation of three methods forenriching H2-producing cultures from anaerobic sludge.Enzyme Microb Technol 2006;40(4):947e53.
[31] Fang HP, Li C, Zhang M. Acidophilic biohydrogen productionfrom rice slurry. Int J Hydrogen Energy 2006;31:683e92.
[32] Rohsenaw WM, Hartnett JP. Handbook of heat transfer.New York: McGraw Hill; 1973.
[33] Nagata S. Mixing: principles and applications. New York:John Wiley & Sons; 1975.
[34] Bailey JE, Ollis DF. Biochemical engineering fundamentals.2nd ed. Singapore: McGraw Hill Education InternationalEditions; 1986.
[35] Lee KS, Lin PJ, Chang JS. Temperature effects onbiohydrogen production in a granular sludge bed inducedby activated carbon carriers. Int J Hydrogen Energy 2006;31:465e72.
[36] Ruggeri B. Thermal and kinetic aspects of biogas production.Agric Wastes 1986;16:183e91.
[37] Pham TH, Rabaey K, Aelterman P, Clauwaert P, DeShampelaire L, Boon N, Verstraete W. Microbial fuel cells inrelation to conventional anaerobic digestion technology. EngLife Sci 2006;6(3):285e92.
[38] Kapdan IK, Kargi F, Oztekin R, Argun H. Bio-hydrogenproduction from acid hydrolyzed wheat starch by photo-fermentation using different Rhodobacter sp. Int J HydrogenEnergy 2009;34:2201e7.
[39] Lee HS, Rittmann BE. Evaluation of metabolism usingstoichiometry in fermentative biohydrogen. BiotechnolBioeng 2009;102:749e58.
[40] ISOTAG4. Guide to the expression of the uncertainty inmeasurements (GUM). Geneva: ISO; 1994.
[41] Ruggeri B. Chemicals exposure: scoring procedure anduncertainty propagation in scenario selection for riskanalysis. Chemosphere 2009;77:330e8.