Optimization of Outdoor Cultivation in Flat Panel

Airlift Reactors for Lipid Production by Chlorella

vulgaris
Ronja M unkel,
1
Ulrike Schmid-Staiger,
2
Achim Werner,
1
Thomas Hirth
1,2
1
University of Stuttgart/Institute of Interfacial Process Engineering and Plasma Technology,
Nobelstrabe 12, 70569 Stuttgart, Germany; telephone: 49-711-970-4069;
fax: 49-711-970-4200; e-mail: ronja.muenkel@igvt.uni-stuttgart.de
2
Fraunhofer Institute for Interfacial Engineering and Biotechnology, Stuttgart, Germany
ABSTRACT: Microalgae are discussed as a potential renewable
feedstock for biofuel production. The production of highly
concentrated algae biomass with a high fatty acid content,
accompanied by high productivity with the use of natural
sunlight is therefore of great interest. In the current study an
outdoor pilot plant with ve 30 L Flat Panel Airlift reactors
(FPA) installed southwards were operated in 2011 in
Stuttgart, Germany. The patented FPA reactor works on
the basis of an airlift loop reactor and offers efcient
intermixing for homogeneous light distribution. A lipid
production process with the microalgae Chlorella vulgaris (SAG
211-12), under nitrogen and phosphorous deprivation, was
established and evaluated in regard to the fatty acid content,
fatty acid productivity and light yield. In the rst set of
experiments limitations caused by restricted CO
2
availability
were excluded by enriching the media with NaOH. The
higher alkalinity allows a higher CO
2
content of supplied air
and leads to doubling of fatty acid productivity. The second
set of experiments focused on how the ratio of light intensity
to biomass concentration in the reactor impacts fatty acid
content, productivity and light yield. The specic light
availability was specied as mol photons on the reactor
surface per gram biomass in the reactor. This is the rst
publication based on experimental data showing the
quantitative correlation between specic light availability,
fatty acid content and biomass light yield for a lipid
production process under nutrient deprivation and outdoor
conditions. High specic light availability leads to high fatty
acid contents. Lower specic light availability increases fatty
acid productivity and biomass light yield. An average fatty
acid productivity of 0.39 g L
1
day
1
for a 12 days batch
process with a nal fatty acid content of 44.6% [w/w] was
achieved. Light yield of 0.4 g mol photons
1
was obtained for
the rst 6 days of cultivation.
Biotechnol. Bioeng. 2013;xxx: xxxxxx.
2013 Wiley Periodicals, Inc.
KEYWORDS: Chlorella vulgaris; outdoor cultivation; Flat Panel
Airlift reactor; lipid production; carbon dioxide; light
availability
Introduction
Algal biofuels have the potential to contribute to improving
the sustainability of the transportation sector, but innova-
tions and R&D are needed to realize their full potential.
(Sustainable Development of Algal Biofuels, The National
Academies Press, Washington, US, 2012.) Discussed in the
scientic literature, as well as in the world press, microalgae
are a promising alternative biomass source for biofuel
production. However, there are still challenges that need to be
addressed to ensure stable large-scale production with
positive net energy balance (Chisti, 2007; Mata et al.,
2010). Thereby special attention has to be paid to increasing
the productivity of the production systems, minimizing
energy demand and cost requirements (Grifths et al., 2011),
as well as increasing product quality to ensure an efcient
downstream process (Stephenson et al., 2010). Whether
microalgae production will be feasible for energy purposes
cannot be answered based on data currently available.
Especially robust data fromeld experiments are necessary to
study the potential in detail (Wilhelm and Jakob, 2011).
Although several data sets from outdoor experiments are
available for biomass production (Eriksen, 2008; Lee, 2001)
only very few have been published concerning the lipid
production process under nutrient deprivation (Bondioli
et al., 2012; Rodol et al., 2009). The same applies to
photosynthetic efciency and biomass light yield. These
values are available for biomass production for both
laboratory experiments (Cuaresma et al., 2011; Meiser
et al., 2004) and outdoor cultivations (Acien Fernandez
et al., 2003) as well as for different strains. As mentioned by
Schlagermann et al. (2012) no quantitative measurements
Correspondence to: R. Munkel
Received 7 January 2013; Revision received 28 March 2013; Accepted 19 April 2013
Accepted manuscript online xx Month 2013;
Article rst published online in Wiley Online Library
(wileyonlinelibrary.com).
DOI 10.1002/bit.24948
ARTICLE
2013 Wiley Periodicals, Inc. Biotechnology and Bioengineering, Vol. xxx, No. xxx, 2013 1
have yet been released for photosynthetic efciency during
lipid production.
As shown in laboratory experiments and outdoor
cultivations, deprivation of nutrients like nitrogen and
phosphorus may lead to an increase in lipid content (Illman
et al., 2000; Rodol et al., 2009). This is caused by the
accumulation of triacylglycerides (TAG), which generally
serve as storage molecules in microalgae cells. Along with the
accumulation of TAG, the degree of fatty acid saturation is
shifted towards saturated and monounsaturated fatty acids.
When focusing on biodiesel production such a composition
is closer to the requirements for biodiesel, according to
standards such as EN 14214 (Grifths et al., 2011). Recent
studies show that overall lipid productivity may increase with
limited nutrient supply (Grifths et al., 2011; Lv et al., 2010;
Stephenson et al., 2010). Another parameter inuencing the
lipid content along with the lipid productivity is the specic
light availability or effective irradiance (Su et al., 2011).
Hereafter, specic light availability describes the ratio of light
striking the reactor surface, to biomass concentration in the
reactor. It is specied as mol photons per gram of biomass
and refers to a dened time interval. Under natural light
conditions, light availability varies and depends upon
weather conditions and the diurnal cycle. During the outdoor
production process, specic light availability can be
inuenced solely by either changing the light path of the
reactor or by adjusting the biomass concentration.
The objective of the current investigations was to set up a
lipid production process under outdoor conditions, generat-
ing highly concentrated biomass with a high lipid content
accompanied by a high productivity. The set-up was a lipid
production process with Chlorella vulgaris in an outdoor pilot
plant with 30 L photobioreactors under phosphorous and
nitrogen deprivation. The established batch process was
solely limited by the parameters light and temperature, which
cannot be inuenced. Process limitations caused by restricted
CO
2
or nutrient availability were avoided. In pre-experiments
the CO
2
content of supplied air was determined, warranting a
non-CO
2
-limited lipid production process. Following this, in
a second set of experiments the specic light availability was
varied by setting different initial biomass concentrations. The
investigations resulted in a quantitative correlation between
specic light availability, lipid content and biomass light
yield.
Materials and Methods
Reactor Design and Plant Construction
All experiments were performed in an outdoor pilot plant
located in Stuttgart, Germany (N48

44
0
24 E009

05
0
56 ) in
summer and autumn of 2011. The plant consisted of ve Flat
Panel Airlift reactors facing south with a light path of 3 cm,
total reactor volume of 30 L and a single-edge surface area of
1.3 m
2
(Fig. 1). The Flat Panel Airlift reactor (FPA) (Schmid-
Staiger et al., 2009) was originally developed at the
Fraunhofer IGB and is now commercially produced by
Subitec GmbH (Stuttgart, Germany). The reactor works on
the principle of an airlift reactor and thus allows efcient
intermixing based solely on aeration. Figure 2 shows a cross-
section of the reactor. Static mixers cause a circulating current
in every chamber of the reactor, leading to uniform light
distribution. Every reactor was equipped as shown in
Figure 3. The CO
2
ow rate was continuously adjusted by
a mass owcontroller (Bronkhorst, Kamen, Germany, FIC in
Fig. 3) and added to the air ow. The total gas owrate was set
by an electropneumatic control valve (Bosch Rexroth, Lohr,
Germany, PIC in Fig. 3) and supplied through a perforated
silicone hose at the bottom of the reactor. Temperature
and pH value were measured by a combined electrode
(Endress und Hauser, Stuttgart, Germany, TIC/QI in Fig. 3)
and transmitted to the process control unit (Siemens CPU
1214, Mnchen, Germany). All process parameters were
measured continuously and recorded every minute.
Figure 1. Outdoor pilot plant with ve 30 L Flat Panel Airlift Reactors for a two-stage lipid production process.
2 Biotechnology and Bioengineering, Vol. xxx, No. xxx, 2013
Process Conditions
To evaluate the relevant process correlations of a lipid
production process with C. vulgaris (SAG 211-12) a two-stage
outdoor process was established. In the rst stage two
reactors were operated to produce biomass in a fed-batch
mode with repeated nutrient feeding. In stage two three
reactors were operated in parallel for lipid production, which
was induced by ammonium and phosphate deprivation and
operated as a batch process. Considerations and results
concerning the biomass production stage are not subjects of
this publication.
Modied DSNmedium(Pohl et al., 1987) was used in both
process stages. The medium consists of 3.5 g L
1
sea salt;
1.38 g L
1
MgSO
4
H
2
O; 0.56 g L
1
CaCl
2
; 3.2 mg L
1
Fe(III)
citrate and 40 ml L
1
micronutrient solution (20 mg L
1
MnCl
2
4H
2
O; 5 mg L
1
ZnSo
4
7H
2
O; 5 mg L
1
CoSO
4

7H
2
O; 5 mg L
1
; Na
2
MoO
4
2H
2
O; 0.5 mg L
1
CuSO
4
5H
2
O).
When biomass was transferred from stage one reactors
for biomass production into stage two reactors for lipid
production, the culture was diluted with N/P-free DSN-
medium to achieve the desired biomass concentration. After
dilution, the ammonium concentration in the lipid produc-
tion reactors was <30 mg L
1
. The phosphate concentration
was <20 mg L
1
. Both nutrients were assimilated by 24 h
after start of experiments, at the latest. According to Liu et al.
(2008) lipid productivity may be improved by additional iron
feeding. Hence, during the lipid production stage, 1 mg L
1
Fe was added three times a week by addition of dissolved iron
citrate.
The reactor was kept at a constant volume of 24.5 L.
Evaporation losses were substituted regularly with deionized
water. The aeration rate was set to 500 L h
1
corresponding to
0.34 vvm. The pH value was kept constant at 6.9 by setting
the CO
2
content of the supplied air. The upper reactor
temperature was xed at 29

C by periodical spray cooling.

During all experiments the optical density in all reactors was
determined ve times a week in the morning. Samples from
all reactors were taken to analyze the lipid content of the
biomass three times per week.
Experimental Design
The aim of the rst experiment was to identify the proper
CO
2
content of supplied air to avoid CO
2
-limitation during
the lipid production process. During the lipid production
stage, no buffering nutrients like phosphate are present and a
constant CO
2
content in the supplied air would result in
severe variations of the pH value. To avoid those variations a
PID-controller continuously set the CO
2
content of supplied
air to compensate the pH shift caused by changing CO
2
solubility and a variable CO
2
uptake rate.
In order to evaluate the proper CO
2
content of the supplied
air, the culture medium of the three reactors was enriched
with different amounts of NaOH. Cultivation was conse-
quently possible in all reactors at identical pH values but
different CO
2
contents of supplied air. NaOH neutralized the
different amounts of dissociated carbon dioxide in the culture
medium. During preliminary tests with algae-free medium,
it was shown that a NaOH concentration of 3.9 mM
corresponds to a CO
2
content of supplied air of 2.0% under
equilibrium conditions at a pH value of 6.9 and a reactor
temperature of 28

C. Under equivalent conditions, a

concentration of 7.7 mM NaOH corresponds to 4.0% CO
2
in supplied air. During cultivation the rst reactor was set to a
NaOH concentration of 3.9 mM, while the second reactor
was set to 7.7 mM NaOH. The third reactor was operated
Figure 3. Flow diagram of an outdoor FPA reactor for the lipid production stage.
All three reactors were equipped identically.
Figure 2. A: Cross-section of a Flat Panel Airlift reactor with 3 cm light path.
Arrows show the circular current generated by ascending gas bubbles. B: Schematic
representation of the different light zones in the reactor and the owprole leading to a
homogenous light distribution.
Mnkel et al.: Outdoor Cultivation of C. vulgaris for Lipid Production 3
Biotechnology and Bioengineering
without NaOH addition as a reference. The CO
2
content of
supplied air was constantly regulated to keep the pH value at
6.9. The CO
2
content of supplied air varied according to the
NaOH content in the medium, the photosynthetic activity
and the culture temperature. The reactors were inoculated
simultaneously with an identical initial biomass
concentration.
The aim of the second set of experiments was to evaluate
the inuence of the specic light availability by setting
different initial biomass concentrations. Three reactors were
inoculated simultaneously with biomass concentrations
between 1.3 and 4.1 g L
1
. All reactors were set to a NaOH
concentration of 3.9 mM. The experiment was performed
twice, in order to achieve a sufcient quantity of data (run 1
and 2). The time-dependent courses of biomass concentra-
tion, lipid content and lipid concentration of run 1 are
presented in detail. Time-dependent courses for run 2 are
shown in the Appendix. However, results of run 1 and 2 were
taken into account for the evaluation of the inuence of
specic light availability on biomass light yield and nal lipid
content.
Determination of Irradiance
The radiation was measured by three vertically installed light
sensors aligned to the west, south and east. Additionally, one
sensor measured the diffuse light. The dataset was recorded
every tenth minute in Wm
2
. The position of the reactors
was rotated by 4

on the eastwest axis. Taking into account

their trigonometric relationships, the three sensors aligned
west, south and east were used to calculate the light intensity
on the reactor surface facing the sun. The data from the
diffuse light sensor were taken into account when the
radiation on the reactor surface facing away from the sun was
determined. The total radiation on both reactor surfaces was
calculated as the sum of the radiation on the sun-facing and
the shaded reactor surfaces. The photosynthetic active
radiation (PAR) was estimated as 45% of the total radiation.
By integrating the photosynthetic active radiation for the
time interval of 24 h, the irradiance I
PAR
in mol photons per
day and reactor surface (front and rear) was determined.
Another sensor recorded the horizontal radiation in Wm
2
.
These data were used to calculate the horizontal radiation in
MJ m
2
day
1
.
Determination of Optical Density and Dry Weight
Concentration
The optical density was determined at 750 nm in a
spectrophotometer (Hitachi 150-20, Japan). The following
correlation was used to convert the optical density OD
750
into
biomass concentration:
DW 0:1907 OD
750
; r
2
0:98
The dry weight concentration was determined to conrm
the correlation. Sample with a volume of 10 mL were
centrifuged and washed twice (20 min per step at 5,000 rpm).
The samples were dried for 24 h at 100

C and weighed.
Determination of Fatty Acid Prole and Fatty Acid Content
The fatty acid prole was analyzed by a method based on that
described by Lepage and Roy (1984). The modied method
used in this work is described by Meiser et al. (2004). Prior to
transesterication the samples were centrifuged and washed
twice as described in the section above. The samples were
analyzed in a gas chromatograph (Agilent 7890A, column:
SupelcoSPBTm-PUFA Fused Silicia Capillary Column). A
sum of all fatty acids provided an estimate of the total fatty
acid content.
Calculations
Biomass productivity and fatty acid productivity describe the
change in biomass concentration respective fatty acid
concentration according to a dened time interval. The total
fatty acid content refers to the total biomass and the content
of a single fatty acid refers to the total amount of fatty acids in
the biomass. To evaluate the inuence of irradiance along
with biomass concentration on the lipid production process,
the specic light availability I
PAR,spec
is dened according to
Equation (1):
I
PAR;spec

R
I
PAR
dt
1DWV
1
I
PAR,spec
gives the ratio of irradiance on reactor surface to
biomass in the reactor and refers to a dened time interval.
I
PAR
describes the irradiance in mol photons on the reactor
surface (front and back) and is integrated over the dened
time interval. DWdescribes the mean biomass concentra-
tion during the considered time interval, V is the total culture
volume and t is the cultivation time. The photosynthetic
efciency PE
PAR
describes the ratio of chemically xed energy
to radiation energy impinging the reactor surface and is
calculated according to Equation (2):
PE
PAR

DFAW37; 4kJg
1
DDWDFAW 16; 7 kJg
1
V
217kJ mol photons
1

R
I
PAR
dt
100
2
DFAW corresponds to the change in total fatty acid
concentration in the culture broth. DDW describes the
change in biomass concentration. The mean energy content
of the algae biomass was assumed to be 37.4 kJ g
1
for fatty
acids and 16.7 kJ g
1
for the residual biomass (Atwater and
Benedict, 1902). I
PAR
describes the irradiance in mol photons
on the reactor surface (front and back) and is integrated over
the dened time interval. The mean energy content of the
photosynthetic active radiation is 217 kJ mol photons
1
(Tredici, 2010).
4 Biotechnology and Bioengineering, Vol. xxx, No. xxx, 2013
Equation (3) gives the biomass light yield Y
PAR
. The
biomass light yield describes the amount of biomass
produced per mol photons impinging the reactor surface.
Y
PAR

DDWV
R
I
PAR
dt
3
Results
Inuence of CO
2
Content of Supplied Air on Lipid Production
Enriching the reactors with different amounts of NaOH
permitted us to evaluate the effect of different dissolved CO
2
concentrations without changing the pH value of the system.
The identical pH values ensured the comparability of these
cultivations and kept the ratio of dissolved CO
2
to HCO
3

and CO
3
2
constant.
Figure 4 shows the oscillating course of the CO
2
content of
supplied air during the lipid production stage of all three
reactors. The time-dependent course of the CO
2
content of
supplied air can be described with Equation (4):
dc
CO
2
;l
dt
k
L
ac

CO
2
;l
c
CO
2
;l
m
CO
2
4
The saturation concentration c

corresponds to the CO
2
content of supplied air, where c
CO
2
;l
is the actual concentra-
tion in the culture media and m
CO
2
the CO
2
uptake rate
caused by CO
2
-assimilation. The volumetric mass transfer
coefcient k
L
a is a system-specic variable and identical for
all reactors during the experiment. Assuming a constant
pH value, the change in c
CO
2
;l
is insignicant. During the
night the CO
2
uptake rate m
CO
2
was zero, c

equaled c
CO
2
;l
disregarding CO
2
production caused by respiration. The
system was close to equilibrium. Depending on the NaOH
concentration, CO
2
contents of supplied air of 0%, 1.6%, and
3.5% were sufcient to keep the pH value at 6.9. During
daytime there was CO
2
uptake by the algal cells along with
photosynthetic CO
2
-assimilation and therefore m
CO
2
in-
creased. The gasliquid equilibrium could not be sustained.
Initiated by the PID-controllers that kept the pH value
constant, CO
2
contents of supplied air increased to 0.5%,
3.0%, and 5.5% at noontime. The increase in supplied CO
2
was additionally caused by the higher reactor temperature,
which resulted in lower solubility of CO
2
in the liquid phase.
According to Figure 5A the biomass concentration
increased during cultivation under nitrogen and phosphorus
deprivation from 1.9 to 6.2 g L
1
in 8 days in both reactors
enriched with NaOH. However, the biomass concentration in
the reference reactor remained at 3.0 g L
1
after Day 4. The
nal fatty acid content in all three reactors did not differ
signicantly and reached 3540% of the biomass. The nal
fatty acid concentration was mostly inuenced by the
different biomass concentrations. The lipid concentration
reached 2.5 and 2.6 g L
1
in the two NaOH-enriched reactors
compared with 1.2 g L
1
in the reference reactor. A summary
of biomass and lipid productivities for all three reactors is
shown in the Appendix (Table AI).
Inuence of Biomass Concentration on Lipid Production
In order to quantify the inuence of biomass concentration
on lipid production and hence light availability per cell, the
three reactors were inoculated with different initial biomass
concentrations. The experiment was carried out twice (run 1
und 2). For both runs the three reactors were inoculated
simultaneously and operated in parallel under nitrogen and
phosphorus deprivation. Figure 6 shows the time-dependent
courses of biomass concentration, fatty acid content and fatty
acid concentration of run 1. Time-dependent courses for run
2 are shown in the Appendix (Fig. A1). The initial biomass
concentrations of run 1 were 1.3, 2.5, and 3.8 g L
1
.
According to Figure 6A, the initial biomass concentration
increased more than threefold during cultivation. The highest
increase was observed for the culture with the highest initial
biomass concentration, where the biomass concentration
increased from 3.8 to 12.4 g L
1
in 14 days. This corresponds
to a mean biomass productivity of 0.67 g L
1
day
1
.
The fatty acid content of the biomass for each reactor was
analyzed eight times during the experiment. In all cultivations
the fatty acid content increased continuously and reached the
highest fatty acid content at Day 12. The fatty acid content
increased from an initial gure of 10% to 53.7% for the reactor
with the lowest initial biomass concentration, and up to 44.6%
for the reactor with the highest initial biomass concentration.
The fatty acid concentration was calculated according
to these data (Fig. 6C). This highlighted the inuence of
biomass concentration on the lipid concentration. The reactor
with an initial biomass concentration of 1.3 g L
1
reached a
fatty acid concentration of 2.4 g L
1
at Day 14. The reactor
with the highest biomass concentration showed the highest
fatty acid concentration of 5.2 g L
1
after the same period of
time. A summary of biomass and fatty acid productivities for
all three reactors is shown in the Appendix (Table AII). In a
0 1 2 3 4 5 6 7 8
0
1
2
3
4
5
6
cultivation time [d]
C
O
2

c
o
n
t
e
n
t

i
n

s
u
p
p
l
i
e
d

a
i
r

[
%

v

v

1
]


Figure 4. Time dependent course of the CO
2
content of supplied air during the
lipid production stage in three FPA reactors with increasing NaOH concentrations
added: 0 mM (~), 3.9 mM (&), and 7.7 mM (). A PID controller kept the pH value
constant at 6.9 by varying the CO
2
content of supplied air.
Mnkel et al.: Outdoor Cultivation of C. vulgaris for Lipid Production 5
Biotechnology and Bioengineering
0
1
2
3
4
5
6
7
b
i
o
m
a
s
s

c
o
n
c
e
n
t
r
a
t
i
o
n

[
g

L

1
]


0
10
20
30
40
50
60
f
a
t
t
y

a
c
i
d

c
o
n
t
e
n
t

[
%

w

w

1
]


A
B
0 2 4 6 8 10
0
0.5
1
1.5
2
2.5
3
cultivation time [d]
f
a
t
t
y

a
c
i
d

c
o
n
c
e
t
r
a
t
i
o
n

[
g

L

1
]


C
Figure 5. Parallel outdoor cultivation of C. vulgaris in three parallel 30-L-FPA
reactors under nitrogen and phosphorus deprivation at pH 6.9. Time-dependent
courses of the biomass concentration (A), the fatty acid content (B) and the fatty acid
concentration (C) for different NaOH concentrations in the media: 0 mM (5), 3.9 mM
(&) and 7.7 mM ().
0
2
4
6
8
10
12
14
b
i
o
m
a
s
s

c
o
n
c
e
n
t
r
a
t
i
o
n

[
g

L

1
]


0
10
20
30
40
50
60
f
a
t
t
y

a
c
i
d

c
o
n
t
e
n
t

[
%

w

w

1
]


A
B
0 2 4 6 8 10 12 14 16 18
0
1
2
3
4
5
6
cultivation time [d]
f
a
t
t
y

a
c
i
d

c
o
n
t
e
n
t
r
a
t
i
o
n

[
g

L

1
]


C
Figure 6. Parallel outdoor cultivation of C. vulgaris in three parallel 30-L-FPA
reactors under nitrogen and phosphorus deprivation at pH 6.9. Time-dependent
courses of the biomass concentration (A), the fatty acid content (B) and the fatty acid
concentration (C) for different initial biomass concentrations: 1.3 g L
1
(5), 2.5 g L
1
(&) and 3.8 g L
1
(). Run 1.
6 Biotechnology and Bioengineering, Vol. xxx, No. xxx, 2013
12 days batch process with an initial biomass concentration of
3.8 g L
1
, an average fatty acid productivity of 0.39 g L
1
day
1
with a nal fatty acid content of 44.6% was achieved.
The fatty acid proles of three samples taken from the
reactor with the highest nal fatty acid content of 53.7% at
Days 1, 6, and 12 are shown in Figure 7. In comparison with
Day 1 only the fatty acid C18:1 increased from2.3% to 48.5%
of the total amount of fatty acids at Day 6. Regarding the
content of fatty acids related to the total biomass, the amount
of C16:0 and C18:3 increased as well. C16:0 increased from
2.7% to 7.1% DW, C18:1 from 0.3% to 20.5% DWand C18:3
from 4.7% to 9.5% DWat Day 6 (data not shown in gures).
From Day 6 to Day 12, the fatty acid prole did not change
signicantly but the total fatty acid content further increased
from 42.2% to 53.7%.
Effect of Specic Light Availability on Fatty Acid Content
and Biomass Light Yield
The initial biomass concentrations of all reactors of run 1 and
2 varied between 1.3 and 4.1 g DWL
1
. The horizontal
radiation during the two cultivation periods is shown in
Figure 8. For run 1 the daily mean radiation was nearly
constant at 16 MJ m
2
day
1
. However, during run 2 the daily
mean radiation varied between 5.2 and 20.5 MJ m
2
day
1
.
These two datasets were used to compare the biomass light
yields and nal fatty acid contents achieved for various
biomass concentrations and time intervals of lipid produc-
tion stage. In Figure 9 biomass light yield is plotted over
specic light availability calculated according to Equation (3).
The considered time intervals cover the cultivation time from
Day 0 to 6, fromDay 0 to 8, fromDay 0 to 10 days, fromDay 0
to 11 days and from Day 0 to 12. The mean biomass
concentration of the corresponding time interval was taken
into account for the calculation. Evidently, the biomass light
yield decreases with increasing specic light availability.
Maximum biomass light yield for the lipid production stage
was achieved in the reactor with the highest initial biomass
concentration of 3.8 g L
1
during the rst run, which resulted
in a specic light availability of 1.9 mol photons g
1
. This low
specic light availability resulted in a biomass light yield of
0.4 g mol photons
1
and refers to the time interval from
Day 0 to 6. The fatty acid light yield for the identical time
interval was 0.18 g fatty acids per mol photons and
corresponds to the highest value achieved during the
experiments. Biomass light yield decreased to 0.1 g mol
Figure 7. Fatty acid prole of C. vulgaris cultivated outdoor in a 30-L-FPA reactor
under nitrogen and phosphorus deprivation with an initial biomass concentration of
1.3 g L
1
and a nal fatty acid content of 53.7% of the biomass.
0
10
20
0 2 4 6 8 10 12
0
10
20
cultivation time [d]
h
o
r
i
z
o
n
t
a
l

r
a
d
i
a
t
i
o
n

[
M
J

m

2

d

1
]
run 1
run 2
Figure 8. Daily horizontal radiation during run 1 and run 2 of the experiments
concerning the specic light availability. Test runs took place from September to
October 2011.
Figure 9. Outdoor cultivation of C. vulgaris in 30-L-FPA reactors under nitrogen
and phosphorus deprivation at pH 6.9. Two runs with three reactors each operated in
parallel were set up. The initial biomass concentration varied between 1.3 and
4.1 g L
1
. Biomass light yield (& run 1, & run 2) describes the change in biomass
concentration per irradiance on the reactor surface for different time intervals (Day 0
6, Day 08, Day 010, Day 011, and Day 012) and is plotted over the corresponding
specic light availability. The specic light availability describes the ratio of the sum of
irradiance on the reactor surface to mean biomass in the reactor. Final fatty acid
contents of all reactors are also shown ().
Mnkel et al.: Outdoor Cultivation of C. vulgaris for Lipid Production 7
Biotechnology and Bioengineering
photons
1
under experimental conditions with high specic
light availability of 8.5 mol photons g
1
. The nal fatty acid
content for all six reactors is also shown in Figure 9. Specic
light availabilities lower than 4 mol photons g
1
increases the
fatty acid content from 10% in the biomass production stage
to a nal lipid content of between 35% and 45% within
12 days. For a nal fatty acid content of more than 50% a
higher specic light availability is needed.
The photosynthetic efciency was calculated according to
Equation (2). Maximal PE
PAR
was achieved during the second
run with the highest initial biomass concentration. For a time
interval of 6 days in the beginning of the lipid production
stage, the mean PE
PAR
was 4.9% corresponding to a biomass
light yield of 0.4 g mol photons
1
. However, when aiming to
achieve a nal fatty acid content of more than 50% a mean
PE
PAR
lower than 2% is inevitable.
Discussion
The aim of the rst experiment was to identify the proper
CO
2
content of supplied air to avoid CO
2
-limitation during
the lipid production process. To evaluate the proper CO
2
content of supplied air, the culture medium of the three
reactors was enriched with different amounts of NaOH.
Consequently, cultivation was possible in all reactors at
identical pH values but different CO
2
content of supplied air.
As shown in Figure 5A, the course of biomass concentration
is affected by the amount of NaOH in the media and thus by
the concentration of dissolved CO
2
. These results permit the
assumption that, in the lipid production stage, biomass
productivity was limited due to a low CO
2
concentration
when no NaOH was added and the CO
2
content of supplied
air was around 0.5% during noontime. Neither a minimum
value of NaOH nor a minimum value of CO
2
of supplied air
can be derived from the results. An increase of NaOH from
3.85 to 7.70 mM showed no effect on biomass productivity.
Therefore, running the process with a NaOH concentration
of 3.85 mM stands to reason.
In regard to Figure 5B, the fatty acid content of the biomass
is not affected by different amounts of NaOH in the culture
medium and thus by the concentration of dissolved CO
2
in the liquid phase. However, there is a clear difference
concerning the fatty acid concentration due to the inuence
of biomass productivity (Fig. 5C).
The performed experiments show the importance of
optimizing the production process with regard to the CO
2
supply. This corresponds to the results gained by Chiu et al.
(2009) and Lv et al. (2010). Both reported dependence of the
biomass and lipid productivity on the CO
2
content of supplied
air. Due to the differences in the employed production systems
and algae strains, their quantitative results are not comparable
to the present experiments. To describe the inuence of the
CO
2
content of supplied air on the production process in
detail, all process conditions such as temperature, reactor type,
gassing conditions, algae strain and media composition have
to be taken into account. According to Equation (4), the
dynamics of CO
2
transport from gas phase to liquid phase can
be described by the volumetric mass transfer coefcient k
L
a
and depends on the uid dynamics of the system, as well as the
gassing conditions, including bubble size, their distribution
and retention time. Various inuences on the k
L
a value of
photobioreactors and its identication are already described in
detail in the literature (Acien Fernandez et al., 2003; Talbot
et al., 1990; Zhang et al., 2002).
On the other hand, the saturation concentration of CO
2
in
the culture media inuences the dynamics of CO
2
transport
and depends on the temperature and salinity of the liquid
phase. Further, the CO
2
uptake mechanismof the species may
inuence the optimal CO
2
concentration for a non-limited
production process. The dynamics of CO
2
uptake by algae
cells and the impact on the production system is described in
detail by Nedbal et al. (2010).
Based on the knowledge obtained, CO
2
limitation could be
precluded in a second set of experiments. The initial biomass
concentrations were varied to investigate the inuence of
different specic light availabilities. The specic light
availability was specied as mol photons per gram of
biomass in the reactor and refers to a dened time interval.
According to Figure 6A and B, both the biomass concentra-
tion and the lipid content was inuenced by the specic light
availability. The reactor with the highest specic light
availability showed the highest nal lipid content but
obtained the lowest biomass productivity. Vice versa, the
highest initial biomass concentration accompanied with the
lowest specic light availability led to the highest biomass
productivity along with the lowest nal lipid content. Here,
the biomass concentration increased 3.4-fold in 14 days. Even
if the increase is not only due to a rise in lipid concentration, a
strong increase in biomass concentration under N-depleted
conditions is well documented. Liu et al. (2008) reported a
1.4- to 7.8-fold increase in biomass concentration for
different strains during N-depleted cultivation. Biomass
productivity under N-depleted conditions may be very
different depending on algae strain and cultivation con-
ditions. However, it is a signicant parameter for designing an
economically feasible production process.
In the lipid productivity the inuence of the specic light
availability on the biomass productivity was crucial.
Obviously, a low specic light availability led to the high
lipid productivity. Hsieh and Wu (2009) described a strong
impact of the correlation between biomass concentration and
irradiance on productivity in photobioreactors with different
depths. Furthermore, Su et al. (2011) focused on a lipid
production process under N-decient conditions in labora-
tory with varying initial biomass concentrations. They
reported an increase of inoculum concentration leading to
an increase of volumetric lipid productivity, accompanied by
a decrease of lipid content when operating the process at the
same level of irradiance. Although the work was performed
with Nannochloropsis, their conclusions are consistent with the
current work. Similar experiments were performed in at
panel reactors under outdoor conditions by Feng et al. (2011)
with Chlorella zongiensis. For the inoculation of photo-
bioreactors with different optical densities under nitrogen
8 Biotechnology and Bioengineering, Vol. xxx, No. xxx, 2013
deprivation, an increase in lipid content has been reported
when initial optical density decreases. The increase of lipid
content is accompanied by a decrease in lipid productivity. It
has to be pointed out that the initial biomass concentration in
the current work was around 50 higher compared with the
experiments performed by Feng et al. (2011).
Quantifying the inuence of specic light availability on
the lipid production process, Figure 9 shows the decrease of
biomass light yield with increasing specic light availability.
The correlation shown is applicable for different time
intervals of the experiment, for different levels of radiation
and for varying initial biomass concentrations. The increase
in nal lipid content over the specic light availability is also
plotted. The combination of both correlations allows the
settings for a production process solely based on the targeted
product quality and on the local radiation to be dened.
When explaining the strong decrease in biomass light yield
with increasing specic light availability, various different
facts must be considered. The literature contains evidence
that lowcell densities especially in thin layer reactors can lead
to light inhibition that reduces the photosynthetic efciency
along with biomass light yield (Richmond et al., 2003). This
author suggests the application of the highest cell density
possible. Furthermore, the decrease of biomass light yield is
enforced by the increase in nal lipid content. The higher the
lipid content, the higher is the amount of energy bound per
gram of biomass. According to the energy content of lipids
and residual biomass used in Equation (6), the heating value
of the biomass increases from 18.8 to 27.1 MJ kg
1
when the
lipid content increases from 10% to 50% of the biomass.
Therefore, the light energy demand for the synthesis of lipid-
rich biomass is higher. Concurrently, during lipid synthesis
the photon demand increases compared with the demand for
primary sugar synthesis, due to additional reduction steps
(Wilhelm and Jakob, 2011). Zemke et al. (2010) describe a
lower efciency of the biochemical synthesis of lipids based
on a photon balance.
The photosynthetic efciency depends on specic light
intensity in the same way as biomass light yield. With a
maximum PE
PAR
of 4.9% for the rst 6 days observed in the
reactor with the highest initial biomass concentration and an
overall PE
PAR
lower than 2% for a nal lipid content of more
than 50%, the values were in the lower range for outdoor
cultivation. Tredici (2010) estimated a realistic PE in relation
to the total spectrum of 5.4% for a biomass production
process under outdoor conditions with sufcient nutrient
supply. However, it has also been noted that the PE will be
signicantly lower for a lipid production process under
nitrogen deprivation. Described in Zemke et al. (2010), these
values are far below the theoretical maximum PE for lipid
synthesis where thermodynamics becomes the limiting
factor. Concerning the cell-internal conversion fromphotons
to lipid molecules, an efciency of 26% is theoretically
possible (Zemke et al., 2010).
According to Schlagermann et al. (2012), no quantitative
measurements or assessments of the photosynthetic efcien-
cy for a lipid production process under nitrogen deprivation
can be found in the literature. Prior to this work the PE of
such a process had not been published either for laboratory or
for outdoor experiments.
Comparing the maximum lipid productivity observed in
this work with data fromthe literature, only a small number of
data sets are available. In an acclaimed work by Rodol et al.
(2009) a lipid productivity of 0.204 g L
1
day
1
was reported
for Nannocloropsis cultivated in N- and P-free media in 110-L
photobioreactors under outdoor conditions in Florence
(Italy). The experiments were set up with a daily harvested
culture volume of 40%, which resulted in lower biomass
concentrations compared with the current study. In the work
by Rodol et al. (2009) the specic light availability was higher
and therefore the nal lipid content reached more than 60%.
Bondioli et al. (2012), working in the same group, reported a
lipid productivity of 6.5 g m
2
day
1
for a similar production
system(grams per square meter of illuminated reactor surface
per day) with a nal lipid content of 68.5% for Nannochloropsis
under nitrogen deprivation. In comparison, the highest fatty
acid productivity of 0.39 g L
1
day
1
shown in the current
work can be recalculated on the basis of the illuminated
reactor surface and is hence 7.4 g m
2
day
1
. The area for the
biomass production stage has to be considered as well when
calculating the areal fatty acid productivity. Realistically, at
least one third of the total area is needed for the rst process
stage. Assuming that the oor space equals the illuminated
reactor surface, the areal fatty acid productivity would be
4.9 g m
2
day
1
. Even compared with lipid productivities
obtained in laboratory experiments, the current results are
among the highest for photoautotrophic processes. Only Pal
et al. (2011) reported a lipid productivity of 0.41 g L
1
day
1
for Nannochloropsis under nitrogen deprivation and continuous
illumination in the laboratory.
Aside from lipid productivity, the composition of the fatty
acid prole plays an important role for future applications.
The ratio of chain lengths as well as the degree of saturation
may change, depending on cultivation conditions. In general,
when algae species are able to accumulate storage lipids
during nitrogen-depleted cultivation the amount of TAG
increases. TAGs produced mainly consist of saturated and
monounsaturated fatty acids. Polyunsaturated fatty acids are
typical components of membrane lipids and production is
not enhanced by nitrogen deprivation. High amounts of
saturated and monounsaturated fatty acids are the basis for
an economically feasible production of biofuels from
microalgae. The importance of such acids in accordance
with the biodiesel standard EN 14214 is described in Grifths
et al. (2011). Among others, C. vulgaris is one of the most
promising candidates for the accumulation of triglycerides
and therefore interesting for biodiesel production (Stephen-
son et al., 2010).
In Figure 7 the amount of the four main fatty acids is
shown for Days 1, 6 and 12 of the lipid production stage. From
Day 1 to Day 6 the ratio of saturated and monounsaturated
fatty acids to the total amount of fatty acids increases from an
initial level of 26.2% to 66.7% at Day 6. From Day 6 to 12 the
ratio does not show a further increase. For the design of an
Mnkel et al.: Outdoor Cultivation of C. vulgaris for Lipid Production 9
Biotechnology and Bioengineering
industrial process, the temporal course of fatty acid concen-
tration together with the temporal course of the fatty acid
prole is the basis for determining the optimal harvesting time.
Conclusions
When focusing on a two-stage lipid production process under
outdoor conditions, process parameters such as CO
2
availability and specic light availability must be chosen
carefully. Depending on the reactor type and algae strain, the
target pH value, the CO
2
concentration of supplied air and
the alkalinity of the media have to be considered to avoid CO
2
limitations. The current investigation showed highly im-
proved lipid productivity with enhanced CO
2
concentration
in the media. For a non-limited production process the
product quality can be set by adjusting the specic light
availability. Here, the specic light availability was specied
as mol photons per gram of biomass in the reactor and refers
to a dened time interval. High specic light availability leads
to high fatty acid contents. Lower specic light availability
increases fatty acid productivity and biomass light yield. If all
this is taken into account, a production process under
nitrogen and phosphorus deprivation allows high fatty acid
productivity along with a high nal fatty acid concentration.
Such a process may lead to a nal fatty acid content of around
45% of the harvested biomass, even under the outdoor
conditions common in central Europe.
Nomenclature
DW biomass concentration (g L
1
)
CO
2
content in the supplied air (% vv
1
)
m
CO
2
CO
2
uptake rate (g L
1
day
1
)
c
CO
2
;l
concentration of CO
2
in the culture media (g L
1
)
t cultivation time (day)
FAW fatty acid concentration (g L
1
)
I
hor
horizontal radiation (MJ m
2
)
Y
PAR
light yield (g mol photons
1
)
I
PAR
Irradiance on reactor surface (mol photons day
1
)
OD
750
optical density
PE
PAR
photosynthetic efficiency (%)
A reactor surface (front and rear) (m
2
)
V reactor volume (L)
c

saturation concentration of CO
2
in culture media
(g L
1
)
I
PAR,spec
specific light availability (mol photons g
1
)
k
L
a volumetric mass transfer coefficient (s
1
)
The work was carried out within the project EtaMax: Mehr Biogas
aus lignozellulosearmen Abfall- und Mikroalgenreststoffen durch
kombinierte Bio-/Hydrothermalvergasung (Bundesministerium fr
Bildung und Forschung FKZ-Nr 03SF0350A).
Appendix
0
2
4
6
8
10
b
i
o
m
a
s
s

c
o
n
c
e
n
t
r
a
t
i
o
n

[
g

L

1
]


0
10
20
30
40
50
60
f
a
t
t
y

a
c
i
d

c
o
n
t
e
n
t

[
%

w

w

1
]


A
B
cultivation time [d]
0 2 4 6 8 10 12 14
0
1
2
3
4
5
f
a
t
t
y

a
c
i
d

c
o
n
c
e
n
t
r
a
t
i
o
n

[
g

L

1
]


C
Figure A1. Parallel outdoor cultivation of C. vulgaris in three parallel 30-L-FPA
reactors under nitrogen and phosphorus deprivation at pH 6.9. Time-dependent
courses of the biomass concentration (A), the fatty acid content (B) and the fatty acid
concentration (C) for different initial biomass concentrations: 1.3 g L
1
(5), 2.7 g L
1
(&) and 4.1 g L
1
(). Run 2.
10 Biotechnology and Bioengineering, Vol. xxx, No. xxx, 2013
References
Acien Fernandez FG, Hall DO, Canizares Guerrero E, Krishna Rao K, Molina
Grima E. 2003. Outdoor production of Phaeodactylum tricornutum
biomass in a helical reactor. J Biotechnol 103(2):137152.
Atwater WO, Benedict FG. 1902. Experiments on the metabolism of matter
and energy in the human body. United States. Ofce of experiment
stations. Bulletin No. 109. Government Printing Ofce, Washington,
DC.
Bondioli P, Della Bella L, Rivolta Chini G, Zittelli G, Bassi N, Rodol L, Casini
D, Prussi M, Chiaramonti D, Tredici MR. 2012. Oil production by the
marine microalgae Nannochloropsis sp. F&M-M24 and Tetraselmis suecica
F&M-M33. Bioresour Technol 114:567572.
Chisti Y. 2007. Biodiesel from microalgae. Biotechnol Adv 25(3):294306.
Chiu S-Y, Kao C-Y, Tsai M-T, Ong S-C, Chen C-H, Lin C-S. 2009. Lipid
accumulation and CO
2
utilization of Nannochloropsis oculata in response
to CO
2
aeration. Bioresour Technol 100(2):833838.
Cuaresma M, Janssen M, Vlchez C, Wijffels RH. 2011. Horizontal or vertical
photobioreactors? How to improve microalgae photosynthetic efcien-
cy. Bioresour Technol 102(8):51295137.
Eriksen N. 2008. The technology of microalgal culturing. Biotechnol Lett 30
(9):15251536.
Feng P, Deng Z, Hu Z, Fan L. 2011. Lipid accumulation and growth of
Chlorella zongiensis in at plate photobioreactors outdoors. Bioresour
Technol 102(22):1057710584.
Grifths M, van Hille R, Harrison S. 2011. Lipid productivity, settling
potential and fatty acid prole of 11 microalgal species grown under
nitrogen replete and limited conditions. JAppl Phycol 113.
Hsieh C-H, Wu W-T. 2009. A novel photobioreactor with transparent
rectangular chambers for cultivation of microalgae. Biochem Eng J 46
(3):300305.
Illman AM, Scragg AH, Shales SW. 2000. Increase in Chlorella strains
caloric values when grown in low nitrogen medium. Enzyme Microb
Technol 27(8):631635.
Lee Y-K. 2001. Microalgal mass culture systems and methods: Their
limitation and potential. J Appl Phycol 13(4):307315.
Lepage G, Roy CC. 1984. Improved recovery of fatty acid through direct
transesterication without prior extraction or purication. J Lipid Res
25(12):13911396.
Liu ZY, Wang GC, Zhou BC. 2008. Effect of iron on growth and lipid
accumulation in Chlorella vulgaris. Bioresour Technol 99(11):4717
4722.
Lv J-M, Cheng L-H, Xu X-H, Zhang L, Chen H-L. 2010. Enhanced lipid
production of Chlorella vulgaris by adjustment of cultivation conditions.
Bioresour Technol 101(17):67976804.
Mata TM, Martins AA, Caetano NS. 2010. Microalgae for biodiesel
production and other applications: A review. Renew Sust Energ Rev
14(1):217232.
Meiser A, Schmid-Staiger U, Trsch W. 2004. Optimization of eicosapen-
taenoic acid production by Phaeodactylum tricornutumin the at panel
airlift (FPA) reactor. J Appl Phycol 16(3):215225.
Nedbal L, Cerven J, Keren N, Kaplan A. 2010. Experimental validation of a
non-equilibrium model of CO
2
uxes between gas, liquid medium,
and algae in a at-panel photobioreactor. J Ind Microbiol 37(12):
13191326.
Pal D, Khozin-Goldberg I, Cohen Z, Boussiba S. 2011. The effect of light,
salinity, and nitrogen availability on lipid production by Nannochloropsis
sp. Appl Microbiol Biotechnol 90(4):14291441.
Pohl P, Ohlhase MK, Rautwurst SK, Laus-Kinnerk Baasch K. 1987. An
inexpensive inorganic medium for the mass cultivation of freshwater
microalgae. Phytochemistry 26(6):16571659.
Richmond A, Cheng-Wu Z, Zarmi Y. 2003. Efcient use of strong light for
high photosynthetic productivity: Interrelationships between the optical
path, the optimal population density and cell-growth inhibition. Biomol
Eng 20(46):229236.
Rodol L, Chini Zittelli G, Bassi N, Padovani G, Biondi N, Bonini G, Tredici
MR. 2009. Microalgae for oil: Strain selection, induction of lipid
synthesis and outdoor mass cultivation in a low-cost photobioreactor.
Biotechnol Bioeng 102(1):100112.
Schlagermann P, Gttlicher G, Dillschneider R, Rosello-Sastre R, Posten C.
2012. Composition of algal oil and its potential as biofuel. J Combust
2012:114.
Table AII. Parallel outdoor cultivation of C. vulgaris in three parallel 30-L-FPA reactors under nitrogen and phosphorus deprivation at pH 6.9 with
different initial biomass concentrations.
Day 07 Day 014
Initial biomass
concentration (g L
1
)
Biomass
productivity
(g L
1
day
1
)
Fatty acid
productivity
(g L
1
day
1
)
Fatty acid
content at
Day 7 (% DW)
Biomass
productivity
(g L
1
day
1
)
Fatty acid
productivity
(g L
1
day
1
)
Fatty acid
content at
Day 14 (% DW)
Run 1 1.3 0.43 0.24 42.6 0.28 0.18 49.5
2.5 0.68 0.42 43.8 0.49 0.33 51.7
3.8 0.95 0.44 32.1 0.67 0.38 41.8
Run 2 1.4 0.41 0.29 50.1 0.28 0.18 49.6
2.7 0.55 0.29 35.0 0.41 0.26 45.7
4.1 0.58 0.27 26.7 0.41 0.27 40.1
Table AI. Parallel outdoor cultivation of C. vulgaris in three parallel 30-L-FPA reactors under nitrogen and phosphorus deprivation at pH 6.9 with
different NaOH concentrations in the media.
NaOH
concentration
(mM)
CO
2
content in supplied
air at noon (%)
Days 08
Biomass productivity
(g L
1
day
1
)
Fatty acid productivity
(g L
1
day
1
)
Fatty acid content at
Day 8 (% DW)
0 0.5 0.16 0.13 36.8
3.9 3.0 0.55 0.29 39.7
7.7 5.5 0.54 0.31 41.5
Mnkel et al.: Outdoor Cultivation of C. vulgaris for Lipid Production 11
Biotechnology and Bioengineering
Schmid-Staiger U, Preisner R, Trsch W, Marek P. 2009. Kultivierung von
Mikroalgen im Photobioreaktor zur stofichen und energetischen
Nutzung. Chemie Ingenieur Technik 81(11):17831789.
Stephenson AL, Dennis JS, Howe CJ, Scott SA, Smith AG. 2010. Inuence of
nitrogen-limitation regime on the production by Chlorella vulgaris of
lipids for biodiesel feedstocks. Biofuels 1(1):4758.
Su C-H, Chien L-J, Gomes J, Lin Y-S, Yu Y-K, Liou J-S, Syu R-J. 2011. Factors
affecting lipid accumulation by Nannochloropsis oculata in a two-stage
cultivation process. J Appl Phycol 23(5):903908.
Talbot P, Lencki R, Noe J. 1990. Carbon dioxide absorption characterization
of a bioreactor for biomass production of Phormidium bohneri:
Comparative study of three types of diffuser. J Appl Phycol 2(4):341350.
Tredici MR. 2010. Photobiology of microalgae mass cultures: Understanding
the tools for the next green revolution. Biofuels 1(1):143162.
WilhelmC, Jakob T. 2011. Fromphotons to biomass and biofuels: Evaluation
of different strategies for the improvement of algal biotechnology based
on comparative energy balances. Appl Microbiol Biotechnol 92(5):909
919.
Zemke PE, Wood BD, Dye DJ. 2010. Considerations for the maximum
production rates of triacylglycerol from microalgae. Biomass Bioenerg
34(1):145151.
Zhang K, Kurano N, Miyachi S. 2002. Optimized aeration by carbon dioxide
gas for microalgal production and mass transfer characterization in a
vertical at-plate photobioreactor. Bioprocess Biosyst Eng 25(2):97101.
12 Biotechnology and Bioengineering, Vol. xxx, No. xxx, 2013