understanding
bioprocesses
Report No. 3
Introduction
Understanding
Bioprocesses.
BlueSens Report No. 3, January 2014
© 2014 by BlueSens gas sensors GmbH, Herten, Germany, www.BlueSens.com
Page layout, artwork and cover design:
Marcus Riepe, ebene1 Kommunikation GmbH, Düsseldorf, www.ebene1-kommunikation.de
We are pleased to present the third BlueSens Report. Nearly two years after Report No.2, we can
introduce to you the new report with interesting application
reports about our gas sensors for bioprocesses. We
always receive very good feedback for the BlueSens Report
from our customers and prospects, and have now
brought together many meaningful contributions for the
third time. Many universities have used the BlueSens
analysis technology for their research and present their
results in the individual project reports. Like its predecessor report, in the third edition the contributions also
come from our global science competition, the BlueCompetition. A special thank you goes to the independent
jury which assessed the individual contributions. In
2013, we were also able to again win Prof. Dr. Gesine
Cornelissen from the HAW Hamburg, Prof. Dr. Lars
Blank, RWTH Aachen, and Reps. Prof. Dr. Frank Eiden,
W-HS Gelsenkirchen, for the evaluation team. We also
thank all participants in the BlueCompetition 2013, who
again demonstrated pioneering and novel applications
for the BlueSens measurement solutions with their contributions. We would especially like to draw attention to
the innovative winning entry from Clemson University.
The group comprised of Matthew Pepper, Li Wang, Ajay
Padmakumar, Assoc. Prof. Timothy C. Burg, Assoc. Prof.
Sarah W. Harcum and Assoc. Prof. Richard E. Groff from
the Institute of Electrical and Computer Engineering, as
well as the Institute of Bioengineering, developed a realtime bioprocess control for aerobic metabolic processes.
The runners-up from West Chester University were also
BlueSens.com
Dr. Holger Mueller
persuasive with their contribution and took a close second.
The scientists employed photobioreactors to produce
biogas sustainably using algae. In conclusion it can be
stated that the BlueCompetition was a huge success in
2013 and that we gained great many new friends in our
work with the science groups.
Since Report No.2, BlueSens has achieved a great deal.
In 2012, our company moved to larger premises, and
our products and measurement solutions are also constantly evolving. With BlueVis we introduced bioprocessing software which automatically carries out many important calculations for the user, such as OUR, CER, and
RQ calculations. The software shows the direction in
which BlueSens is to develop in the coming years: we
will not only continue to offer our highly successful and
reliable sensors for gas analysis but also provide complete solutions for measurement tasks in fermentations,
biofuels production, and the life sciences, in order to
help you to fully understand your bioprocess.
BlueSens Report No. 3
3
Contents
understanding
bioprocesses
3 BlueSens. Introduction - The BlueSens Report No. 3
Application Reports
6 A Real-time Adaptive Oxygen Transfer Rate Estimator for Metabolism Tracking in Escherichia coli cultures
By Matthew Pepper, Li Wang, Ajay Padmakumar, Timothy C. Burg, Sarah W. Harcum, Richard E. Groff
12 Assessing biodegradation of oil with marine bacteria by monitoring O and CO concentrations
2
2
online in a closed loop
Fr A.G. Valladares Juárez, N. M. Enwena, M. Schedler, R. Müller Institute of Technical Biocatalysis,
Hamburg University of Technology, Germany
17 Bio-hydrogen production by fermentative bacteria using crude glycerol as substrate in experimental testsystems
Hydrogen and Gas Sensors Laboratory Institute of Solid State Physics, University of Latvia
20 Application of BlueSens® H sensor in a gas separation study
2
P. Bakonyi, N. Nemestóthy, K. Bélafi-Bakó Research Institute on Bioengineering, Membrane Technology
and Energetics University of Pannonia, Egyetem street 10, Veszprém, Hungary
24 Enhanced waste to fuel conversion with a bioelectrochemically controlled autotrophic bioreactor
John M. Pisciotta, Joe Mossman, Zehra Zaybak and William Schultz. Department of Biology,
West Chester University.
Information
28 BlueVis: Start QbD with BlueVis.
BlueInOne-Series: The combined gas analyzers for CO2 and O2: The BlueInOne series
29 BCP-Series: In-situ gas analysis for p­ rocess controlled fermentations
30 BlueSens’ sensors overview
32 Analyzer systems
BCpreFerm and YieldMaster
BlueSens.com
BlueSens Report No. 3
5
Application Report
A Real-time Adaptive Oxygen Transfer
Rate Estimator for Metabolism Tracking
in Escherichia coli cultures
Application Report
, (3)
By Matthew Pepper, Li Wang, Ajay Padmakumar, Timothy C. Burg, Sarah W. Harcum, Richard E. Groff,
Department of Electrical and Computer Engineering/Department of Bioengineering, Clemson University.
where Mf is the mass flow rate. The concentration b2 is
sampled by the BlueInOne sensor, and the sensor output, b3, is modeled as a first order linear system with
time constant 2,
Estimator Mathematics
The adaptive estimator is based on a structure from [6].
The mathematics for the estimator will be briefly summarized. The estimator is based on the sensor and
transport dynamics, Eqns (3) to (5), rewritten as a state
space model in observable canonical form,
. (4)
Introduction
Oxygen and carbon dioxide off-gas measurement enables the application of advanced estimation and control methods for aerobic bioprocesses. BlueSens, GmbH
(Herten, Germany) offers an exhaust gas sensor capable
of measuring mole ratios of both oxygen and carbon dioxide with accuracy comparable to mass spectrometry,
the benchmark standard, but at a fraction of the cost
[1]. The reduced cost makes it feasible to dedicate a
sensor to each bioreactor to enable more sophisticated
real-time estimation and control of bioprocesses. Direct
computation of the oxygen transfer rate (OTR) from any
type of off-gas measurement is problematic for use in
control, since transport dynamics and sensor dynamics
cause the computed OTR to be an attenuated and delayed version of the true OTR signal. The primary objective of this report is to describe a state estimator that
combines exhaust gas, stir speed, and dissolved oxygen
measurements to predict the true OTR in real time without attenuation or delay. The predicted OTR can be used
for a variety of controls and estimation purposes, such
as determining when the culture is in oxidative or overflow metabolism.
Background: OTR Dynamics
In a stirred-tank bioreactor, the input gas passes through
two stages before being measured by the exhaust gas
sensor. Figure 1 shows a bioreactor and attached offgas sensor. The input gas contains a known concentration of oxygen, b0. The gas enters from the bottom of the
vessel and a rotating impeller breaks up the bubbles,
facilitating the absorption of oxygen into the culture. The
difference in oxygen concentration between the entering
gas, b0, and the exiting gas, b1, represents the amount of
6
BlueSens Report No. 3
(6)
Clearly, computing OTR directly from the concentration
b3 would be significantly different from the true value of
OTR computed from concentration b1.
Figure 1 – Oxygen concentration in a stirred-tank bioreactor, the input gas
concentration is b0, the concentration exiting the liquid, b1, the concentration
of the head-space b2, and b3 the concentration reported by the sensor.
oxygen transferred into the liquid V1. Assuming the mass
inflow is equal to the outflow, this difference can be
used to calculate the OTR in the liquid,
(1)
where Mf is the mass flow, 04 is the density of oxygen at
25ºC, and V1 is the volume of the liquid. OTR can also
be modeled using the stir speed and dissolved oxygen
concentration,
(2)
where kLa is the oxygen transfer coefficient, C* is the
maximum oxygen carrying capacity of the liquid, C is the
current dissolved oxygen concentration, and N is stir
speed. The dependence of kLa on stir speed N is modeled as linear about stir speed N0.
The gas leaving V1 mixes with the gas in the headspace
volume V2, resulting in the oxygen concentration b1. This
process is a standard mixing model,
BlueSens.com
Estimator Design
A state estimator predicts the unmeasurable state variables of a system by using the measureable outputs and
inputs in combination with the known dynamics of the
system. When an adaptive state observer is used, the
system dynamics are known except for some unknown
or time-varying parameters.
In the bioreactor system presented in the previous section, the oxygen concentration dynamics are represented
by the state variables b2 and b3, which follow the dynamics given by Eqns (3) and (4). The input to the system is
concentration b1, which cannot be measured directly, but
can be computed from stir speed using Eqns (1) and (2),
The adaptive estimator simultaneously estimates the
state variables b2 and b3 and the parameters a0 and a1.
The estimator is driven by known variables b3, C, N, V1,
and V2. The adaptive estimator is given by
. (7)
. (5)
All parameters in the model may be reliably characterized
in advance, with the exception of the parameters a0 and
a1 which relate stir speed to kLa. These parameters are
notoriously hard to find and tend to vary slowly over the
length of the culture [2,4]. The estimator described in this
report uses an adaptive law based on measurements
from the BlueSens sensor to identify the possibly-timevarying parameters a0 and a1 in real time. With good estimates of these parameters, OTR can be accurately predicted from stir speed and dissolved oxygen concentration
using Eqn (2).
BlueSens.com
where
v0 and v1 are auxiliary variables. The
adaptive law for the two unknown parameters are dewhere 0 and 1
signed as
are observer gains. Other variables are designed as
Note that the argument s denotes the differential operator d/dt .
BlueSens Report No. 3
7
Application Report
Table 1 contains all the parameters for the oxygen
dynamics model. kLa is known to slowly change over the
course of a fermentation and is influenced by multiple
factors such as anti-foam addition and liquid viscosity
changes; therefore, the values for the linear model seen
in Eqn. (2) are time-varying and can only be defined
approximately.
Parameter
Name
Value
Mass Flow rate
3 L/min
Culture volume
1.67 L
Headspace volume
5.03 L
BlueSens time constant
55 s
Oxygen density
1.331 g/L
Initial value
0.0035 1/s
Initial value
0.00022
1/(s RPM)
Input Oxygen Volume
20.97%
Maximum Oxygen carrying
capacity
6.71 mg/L
Initial Stir speed
200 RPM
Experiments
Bioreactor System
The bioreactor system is a Sartorius Biostat B 5-L glass
vessel and digital control unit (DCU). The DCU connects
to probes for pH (Hamilton Company, Reno, NV),
temperature (Sartorius AG), and dissolved oxygen
(Hamilton). The data is sampled every 15 seconds using
the OPC protocol in a Simulink model running on Matlab
2012a (Mathworks Inc. Natick, MA). The model also
samples the relative humidity, temperature, pressure,
and oxygen and carbon dioxide volumetric percentages
coming from the BlueSens BlueInOne exhaust gas
analyzer every 10 seconds. The mass flow of the sparged
gas is measured with a mass flow controller (Omega
Engineering Inc, Stamford, CT) and sampled every 5 seconds. Lastly, two balances (Ohaus Corp, Parsippany, NJ)
keep track of the amount of glucose and base dispensed
and report weight every 5 seconds.
Table 1: Adaptive Estimator Parameters
Application Report
Sensor Characterization
In order to implement the adaptive OTR estimator on the
system, the behavior of the BlueInOne sensor was
analyzed in several characterization experiments.
The first experiment was a study of the delay and response time of the sensor, 4, Previous publications
have numbers for response time and measurement
delay, but for a different model and at a much smaller
flow rate (0.040 vs. 2 L/min (lpm)). The input gas, bo,
was connected directly to the input of the BlueInOne
sensor, bypassing the stirred-tank vessel; this allowed
for measurement of only the sensor dynamics. In this
experiment, the composition and flow of the input gas
was varied. The input gas was switched from nitrogen to
air in 5 minute increments. The gas would cycle twice,
and then the flow rate would increase by 1 lpm, from 2
to 8 lpm. The delay (15 seconds) represents the time it
took the sensor to respond once the gas composition
had changed. The response time (55 seconds) is the
time constant for the sensor, i.e., the time for the sensor
to reach 63% of its final value. The data showed that the
response time did not change with increasing gas flow,
indicating the measuring chamber in the BlueInOne was
being flushed well above maximum requirements.
The second experiment explored the accuracy and
stability of the sensor measurements to changes in
pressure, ranging from 1.02 to 1.67 bar. This experiment
determined the usability of the sensor in experiments
requiring pressurizing of the headspace. The Biostat-B
vessel has a maximum pressure limit of 1.63 bar. The
manufacturer’s certified, accurate operational pressure
range for the BlueInOne sensor is between 0.8 and 1.3
bar, with a maximum of 2 bar. The experiment began by
performing a one-point calibration on the BlueInOne;
the input gas of air had a constant assumed oxygen
volume of 20.97%. As in the previous experiment, the
input gas was connected directly to the input port of the
BlueInOne sensor. The mass flow rate was varied between
1 and 9 lpm in steps of 1 lpm. The steps in mass flow
caused the internal pressure of the sensor to rise from 1
to 1.67 bar (Table 2).
Mass
Flow
(lpm)
Pressure
(bar)
Pressure
Step (bar)
O2
Volume
(%)
O2
Spike
(%)
O2 Spike
Duration
(s)
1
1.02
2
1.06
0.045
20.97
1.35
35
3
1.12
0.055
20.97
1.85
45
20.97
4
1.18
0.064
20.96
2.41
75
5
1.27
0.081
20.97
3.17
55
6
1.35
0.085
20.98
3.53
85
7
1.45
0.095
21.01
3.78
55
8
1.56
0.105
21.04
3.57
75
9
1.67
0.110
21.10
3.49
85
Table 2: Measurement Stability Tests
Figure 3 – Measurement Stability Results, the pressure compensation
algorithm in the BlueInOne quickly corrected any error caused by the step
changes in pressure. The final corrected value changed by less than 1%
over the tested pressure range.
After each step of the flow rate and pressure, the reading would spike due to the increased amount of oxygen
in the chamber. The compensation software acted quickly to correct the reported measurement, and on average
only 64 (±19) seconds of data was affected. The
reported oxygen volume drifted from 20.97% to 21.1%
over this range of pressure, representing <1% error. As
expected, the base reading began to drift above 1.3 bar.
The pressure compensation software produced accurate
measurements even outside its certified range, indicating
the sensor should yield accurate stable readings under
conditions of constant or slowly varying pressure.
Figure 2 – The Biostat-B bioreactor system, shown with 5 L vessel and motor, DCU, BlueSens BlueInOne sensor, and Balances.
8
BlueSens Report No. 3
BlueSens.com
BlueSens.com
BlueSens Report No. 3
9
Application Report
Bioreactor Conditions
Two experiments were performed to test the accuracy of
the OTR estimator. In the first experiment, the OTR
estimator was implemented on a bioreactor simulator.
The delays on input and output data characteristics
mirror that of the Sartorius Biostat-B, the BlueInOne
sensor, and the two balances. The simulated behavior of
the Escherichia coli (E. coli) is based on the work of
1999 Xu [8]. In the second experiment, E. coli MG1655
pTV1GFP [5] was cultured for 12 hours in minimal
media. The bioreactor was inoculated at an optical
density (measured by spectrophotometer) of 0.5 OD. In
both experiments, the growth rate was kept at 0.25 h-1.
The volume of the culture was 1.67 L. The mass flow of
the input gas was 3 liters per minutes.
Validation of the OTR Estimator
A feed rate disturbance in the form of a pulse was used
to test the capability of the OTR estimator to track E. coli
metabolism. The pulse increased the feed rate to four
times the current feed rate. This pulse was designed to
increase the glucose concentration such that the E. coli
enter overflow metabolism. While overflow metabolism is
not desirable for E. coli, the feed rate pulse is short and
any negative byproduct is quickly consumed. Tracking the
E. coli metabolism transition from oxidative to overflow
will help determine a feed rate to keep the culture on the
verge of overflow, maximizing oxidative metabolism.
In Figure 4, the experimental results (A, B) show the OTR
and OUR rise sharply as the E. coli metabolism increased
in response to the glucose pulse. This behavior duplicates
the behavior seen in simulation, Figure 4 D and E, indicating that experimental results are reasonable. Notice the
response of the OTR formed from the BlueInOne measurements, OTRsens. The filtering effects are seen in the attenuation of the maximum height of OTRsens versus the OTRest
and the delay between both their respective peaks.
The response of the OTRest is almost immediate while the
peak in OTRsens does not occur until after the 5 minute
pulse is over. The plateau effect seen in OTRest indicates
that the culture entered overflow metabolism. In overflow
metabolism, the oxidative metabolism becomes saturated with glucose and any excess glucose absorbed by the
cells is anaerobically converted to acetate, which can
10
BlueSens Report No. 3
Application Report
values. The kLa for the actual fermentation was plotted in
Figure 5. The distribution of the data points indicates the
kLa varied linearly with stir speed, validating our linear
model. A linear model is appropriate since the short culture length did not allow the cell density to affect the
viscosity of the media.
Figure 5 – Experimental Oxygen Transfer Coefficient, the estimates for the
kLa calculated by the OTR estimator. The kLa and stir speed maintained an
approximately linear relationship during the 12 hour fermentation. This
result validates the kLa model.
Figure 4 – Glucose Pulse Experiments, a pulse (C) in the feed rate was
implemented to test tracking of the OTR estimator on actual (A, B)
and simulated (D, E) bioreactor systems. In A, the OTR estimator was
able to accurately track the OTR and provided a better indication of E.
coli metabolism than the OTR using the BlueInOne measurements. The
attenuation and delay effects of the system are clearly seen (A, B).
inhibit growth at high concentrations (2 g/L). The profile
of OTRest indicates that the oxidative metabolism could
process glucose at a higher rate, thus the feed rate could
be increased significantly. A much longer pulse would be
required to obtain the same data from OTRsens, resulting
in significantly more acetate production.
The behavior of the E. coli in actual and simulated experiments was very similar, validating the tuning of the
OTR estimator gains (Figure 4 A, B, D, and E). In the
simulated experiments (Figure 4 D, E), the OTRest tracks
the actual OTR very well. This indicates that the estimate
of the kLa parameters a0 and a1 converged to their true
BlueSens.com
Conclusions
The OTR estimator provided a more accurate OTR profile
by removing the influences of the headspace and sensor
dynamics on the exhaust gas oxygen volume
measurement. The BlueInOne measurements provided
both accurate and stable measurements, allowing the
adaptive algorithm to estimate both the unknown kLa
and OTR values. The OTR estimator accurately tracked
the E. coli metabolism and will be an integral part in the
development of maximizing controllers for oxidative
metabolism and improving biomass yield.
Researchers have been using exhaust gas sensors to
develop advanced estimator and control algorithms for
over 20 years. The implementation of these advanced
estimator and control algorithms in an industrial context
has been very slow, much to the detriment of the industry
[7]. Ranging from neural networks to model predictive
control, the BlueInOne sensor enables academic and
industrial researchers to explore and implement the next
level of bioprocess control techniques.
BlueSens.com
References
[1] M
. Aehle, R. Simutis, and A. Lubbert, “Comparison of viable cell
concentration estimation methods for a mammalian cell cultivation
process,” Cytotechnology, vol. 62, no. 5, pp. 413–422, Oct 2010.
[2] M
. Akesson, P. Hagander, and J. P. Axelsson, “A pulse technique for
control of fed-batch fermentations,” in Proceedings of the 1997 IEEE
International Conference on Control Applications, pp. 139–144.
[3] V. Babaeipour, S. A. Shojaosadati, S. M. Robatjazi, R. Khalilzadeh,
and N. Maghsoudi, “Over-production of human interferon by hcdc of
recombinant Escherichia coli,” Process Biochemistry, vol. 42, no. 1,
pp. 112–117, Jan 2007.
[4] G
. Bastin and D. Dochain, On-line Estimation and Adaptive Control of
Bioreactors. Amsterdam, Netherlands: Elsevier Science, 1990.
[5] E . Garcia-Fruitos et al., “Aggregation as bacterial inclusion bodies does
not imply inactivation of enzymes and fluorescent proteins,” Miocrob.
Cell Fact. vol. 4, no. 27.
[6] K
. S. Narendra, A. M. Annaswamy, Stable Adaptive Systems, Chp 4,
Prentice Hall, Englewood Cliffs, NJ, 1989.
[7] U
. S. Food and D. A. (USFDA), “Guidance for industry process analytical
technology (PAT) - a framework for innovative pharmaceutical
development, manufacturing, and quality assurance,” Sep 2004.
[8] B
. Xu, M. Jahic, and S. O. Enfors, “Modeling of overflow metabolismin
batch and fed-batch cultures of Escherichia coli,” Biotechnology
progress, vol. 15, no. 1, pp. 81–90, Jan-Feb 1999.
i
The Cell-Culture and Optimization Lab is part of the
Department of Bioengineering (BioE) at Clemson
University under the direction of Sarah Harcum (harcum@
clemson.edu). The work presented is in collaboration the
Bio-Inspired Systems Lab in the Electrical and Computer
Engineering (ECE) Department at Clemson University
under the direction of Richard E. Groff (regroff@clemson.
edu). The goal of this project is to develop advanced
estimation and control strategies for the fermentation
and culture of bacteria and mammalian cells. The
research is supported by National Institute of Health
(NIH) Centers of Biomedical Research Excellence
(COBRE) under grant # P20GM103444.
Project Team
Matthew E. Pepper (right) Li Wang (center) Ajay Padmakumar (left)
Dr. Timothy C. Burg Dr. Sarah W. Harcum Dr. Richard E. Groff PhD Student, ECE
MS Student, ECE
MS Student, ECE
Associate Professor, ECE
Associate Professor, BioE
Associate Professor, ECE
BlueSens Report No. 3
11
Application Report
Assessing biodegradation of oil with
marine bacteria by monitoring O2 and CO2
concentrations online in a closed loop
Because of the hydrophobic
nature of the oil, at least­
two immiscible phases appear when mixed with seawater and bacteria, making it
impossible to sample the system representatively. Moreover the low molecular weight
components of the oil are
volatized during sampling and
that affects the experiment.
Currently, we have been quantifying the degradation of oil
indirectly by offline analysis
of bacterial concentrations or
12
BlueSens Report No. 3
surement of oxygen or carbon dioxide changes is not
possible. Therefore, a closed loop reactor was designed
in order to monitor cumulative oxygen consumption and
carbon dioxide production.
by sacrificing one reactor for gas chromatography analysis per time point, which is labour intensive. In order to
quantify the extent of oil biodegradation in real time, it
is necessary to measure one or more of these parameters: bacterial, oxygen, carbon dioxide or oil concentration online. A continuous measurement of oxygen and
carbon dioxide can be correlated to the disappearance
of the oil and to the growth of the bacteria.
Setup description
One of the low-pressure reactors was adapted with
airtight connections and was connected to the BlueInOne
Cell analyzer via flexible tubing (Figure 1). To transport
the gas in closed-loop modus through the analyzer, a
peristaltic pump recirculated the air at 3 rpm. The reactor
was stirred with a magnetic stirrer. In the same way an
airtight 3-port-Schott flask was connected to the
analyser. Two ports were used for the gas recirculation;
the third port was used for taking samples for determining
the concentration of bacteria (Figure 2). A general
scheme of the setup is shown in Figure 3.
Research objective
The objective of this research was to monitor the rate of
oil biodegradation by newly isolated marine bacteria
using the oxygen and carbon dioxide BlueInOne analyzer
(BlueSens gas sensor GmbH, Herten, Germany). Since
the biodegradation is a rather slow process a direct mea-
Materials and methods
The oil-degrading bacterial strains used in this study
were kindly provided by Prof. Joe Lepo and Prof. Wade
Jeffrey (University of West Florida). They were isolated
from sea samples taken in expeditions to the Gulf of
Mexico before (reference strains) and after (test strains)
Fr A.G. Valladares Juárez, N. M. Enwena, M. Schedler, R. Müller
Institute of Technical Biocatalysis, Hamburg University of Technology, Germany
Introduction
In April 2010, the Deep-Water-Horizon (DWH) well
explosion at the Gulf of Mexico caused the worst oil spill
in the deep sea to date. About 5 million oil barrels
flowed during 86 d into the sea1. The effects on the
marine habitat and the fate of the oil are not well
understood. Bacteria play a major role in the degradation
of petroleum in marine spills. The high pressure, high
salinity and low temperatures in that environment make
it difficult to study the oil-degradation mechanisms in
situ. It is therefore important to study the mechanisms of
the biodegradation of oil in the laboratory. For that
purpose, we have built high-pressure reactors (400 bar)
and low-pressure reactors (1 bar) where it is possible to
compare the biodegradation rates of mineral oil at room
condition (1 bar, 20°C) vs.
those in the deep sea around
the DWH well (150 bar, 4°C).
Application Report
Figure 1: Closed-loop setup used for monitoring of the CO2 and O2 concentration during the incubation of marine
bacteria with oil
BlueSens.com
Figure 3: Diagram of the reactor setup with the BlueInOne analyzer and
pump connected in a closed recirculation loop in the gas phase. A liquid
sampling port as used in the setup of Figure 2 is depicted
the DWH explosion. The strains were used as inocula for
the experiments. The fermentations were carried out in
batch mode.
In order to overcome oxygen limitation, the reactor was
partially filled up to 1/4th of its volume with minimal
medium and the rest with air. Sweet Louisiana crude oil
or some of its main compounds: naphthalene, xylene,
toluene and hexadecane were used as the only carbon
source in 0.03% to 0.1%
(v/v) concentrations. Finally
10% of bacterial inoculum
was added. The O2 and CO2
values were recorded with the
BlueVis software. Samples
were collected at the start and
end points or in case of the
setup with the sampling port,
throughout the experiment in
order to determine the number of bacteria. The BlueInOne
analyzer was calibrated with
air, when the analyzer requested it and the calibration values were automatically saved.
Data storage was reliable and
Figure 2: Closed-loop setup used for monitoring of the CO2 and O2 and the bacterial concentration during the
user-friendly.
incubation of marine bacteria with oil
BlueSens.com
BlueSens Report No. 3
13
Application Report
14
Application Report
Figure 4: CO2 ( ), O2 ( ) and bacterial ( ) concentration of the
i­ncubation of strain DWHO6A in minimal medium with 0.1% (v/v) oil
Figure 6: CO2 ( ), O2 ( ) and bacterial ( ) concentration of the
­incubation of strain DF8 in minimal medium with 1 mM hexadecane
Figure 8: CO2 ( ), O2 ( ) and bacterial ( ) concentration of the
i­ncubation of strain DWHO6B in minimal medium with 1 mM toluene
Figure 5: CO2 ( ), O2 ( ) and bacterial ( ) concentration of the
i­ncubation of strain GM2 in minimal medium with 1 mM hexadecane
Figure 7: CO2 ( ), O2 ( ) and bacterial ( ) concentration of the
­incubation of strain DS4P5 in minimal medium with 0.1% (v/v) Louisiana oil
Figure 9: CO2 ( ), O2 ( ) concentration of the incubation of strain SY1
in minimal medium with 1.8 mM naphthalene
Results
Growth of oil-degrading bacteria
in oil and oil-compounds
The strain DWHO6A could degrade Louisiana oil, hexadecane, xylene and toluene on agar. In liquid medium
the strain DWHO6A was able to biodegrade Louisiana oil
in mineral medium as seen from the bacterial cell number of Figure 4. The system was leaking through the silicon hoses, leading to diffusing of air into the reactor.
This can be observed by the increase in the oxygen concentration to its initial value of 21% after 35 h. For the
next experiments the setup was air-tight, the hoses were
changed to PVC material and with this, the leak was
sealed.
The strain GM2 is milky coloured in appearance when
grown on agar plates. GM2 showed good growth in minimal medium with hexadecane (Figure 5), oil and moder-
ate growth with naphthalene. GM2 reached the highest
bacterial cell numbers during this investigation.
The strain DF8 grew well on selective mineral medium
agar supplemented with Louisiana oil and hexadecane.
The results of the incubation of strain DF8 in liquid
medium containing 1 mM hexadecane are shown in
Figure 6. The strain DS4P5 could not degrade any of the
oil components and showed residual growth on agar
plates. Using Louisiana oil as carbon source this strain
did not grow and the oxygen and carbon dioxide values
remained constant (Figure 7). The BlueInOne analyzer
helped us to corroborate the information from the bacterial density.
The strain DWHO6B is orange when grown on naphthalene agar plates and can degrade Louisiana oil, naphthalene, toluene and xylene. This strain was the only one
that could degrade naphthalene in a medium supple-
mented with 3% NaCl, which is the concentration of salt
in the open ocean. This strain can degrade toluene in
liquid mineral medium as shown in Figure 8 and it can
also grow well at 145 bar. Both pressure regimes gave
similar bacterial cell values, but the O2 and CO2 cannot
be analysed at the high pressure. It would be interesting
to know if the rates of degradation of toluene are similar.
The fermentation shown in Figure 8 was not recorded
completely due to a communication error between the
analyzer and the computer, but this strain could grow to
1x108 with toluene as only carbon source. The degradation capabilities of this interesting strain should be further studied.
The strain SY1 (Figure 9) was able to degrade
naphthalene without any lag phase, but consumes less
oxygen than strains GM2, DF8 and DWHO6b do with
hexadecane and toluene.
BlueSens Report No. 3
BlueSens.com
BlueSens.com
BlueSens Report No. 3
15
Application Report
Conclusion
Spilled oil can be degraded by marine bacteria. Oxygen
is a key substrate for aerobic microbial growth and
biodegradation. Carbon dioxide and biomass are the
major products of the biodegradation and can be used
to estimate the amount of oil that has been degraded.
The degradation capabilities of some deep-sea and
surface-isolated bacteria could be tested and some
promising strains were found. Sampling in this
multiphase system was impossible and this was
circumvented by the measurements with the BlueInOne
analyzer. This accelerated the screening of the new
strains. The use of the BlueInOne analyzer provided
valuable, online and real-time information of the
biodegradation of oil.
Moreover we proved that the BlueInOne analyzer can be
used to measure batch fermentations with rather low
metabolic rates by recirculation of the exhaust flow into
the fermenter. This new application is attractive for
laboratory research because the analyzer can be
connected to any type of reactor or fermenter via a
couple of hoses. Its application is high versatile, exact
and a cost effective solution for monitoring oxygen and
carbon dioxide.
Outlook
Further experiments, changing various parameters like
temperature and type of strains can be conducted using
the BlueSens system efficiently. This BlueInOne analyzer
would also be highly useful for conducting long-term
biodegradation studies.
A further useful application of the BlueInOne analyzer
would be to take samples from the gas in the pressurized
reactors, decompress it, and supply it to the analyzer.
These data would be extremely useful for comparing the
metabolism of the bacteria between atmospheric and
deep-sea pressures.
References
1. Marcia K. McNutt, Rich Camilli, Timothy J. Crone, George D. Guthrie,
Paul A. Hsieh, B. Ryerson, Omer Savas and Frank Shaffer.
Review of flow rate estimates of the Deepwater Horizon oil spill.
PNAS 2012 109 (50) 20260-20267
Application Report
Bio-hydrogen production by fermentative
bacteria using crude glycerol as substrate
in experimental testsystems
Hydrogen and Gas Sensors Laboratory
Institute of Solid State Physics, University of Latvia
Introduction
Our scientist group in the Institute of Solid State Physics
and Microbiology department of Faculty of Biology is
working on bio-hydrogen and biogas production, harvesting, storage and usage technology research. Our
goal is to explore possibilities for bio-hydrogen and
biogas production using alternative local resources –
industrial and agricultural wastes as well as byproducts
of food industry. Industrial and agricultural organic waste
used as feedstock for bacteria is a perspective way for
alternative energy production and it noticeably
decreases the raw material cost. During the conversion
of organic wastes, in anaerobic environment, hydrogen
or methane gas is produced as by-products. One of the
substrates that can be effectively used for microbial
hydrogen and methane production is glycerol, which is a
by-product from the process of biodiesel production.
Because of large quantities available of crude glycerol
and the highly reduced nature of carbon in glycerol per
se, microbial conversion of it seems to be economically
Experimental set-up
i
The main focus of the group environmental biotechnology led by
Prof. Müller of the Institute of Technical biocatalysis lies in the
elucidation of novel pathways in the biological degradation of
environmental pollutants. For this purpose new bacteria with the
ability to degrade problem-causing substances are isolated from
environmental samples and the intermediates in the degradation pathways are determined. For interesting new reaction steps
the corresponding enzymes are purified, cloned, sequenced and
characterized.
Biodegradation is the key to understand the environmental fate
of chemicals released into the environment either intentionally
(e.g. pesticides) or by accidents (e.g. oil spills). For many chemicals the biological degradation pathways are not known yet. For
others the degradation is only known under certain conditions.
Therefore, research is conducted to close these gaps in our
knowledge, and to find safe solutions for environmental problems.
Figure 1
16
BlueSens Report No. 3
BlueSens.com
BlueSens.com
BlueSens Report No. 3
17
Application Report
and environmentally viable possibility, especially
because, over the last several years, the demand and
production of biodiesel has remarkably increased [1].
Methods ans materials
Growth media, cultivation and experimental set up
Different anaerobic and facultatively anaerobic bacteria
were used for hydrogen production measurements from
Microbial Strain Collection of Latvia (MSCL) - Clostridium
sporogenes MSCL 764, Enterobacter asburiae MSCL
839, Enterobacter cloacae MSCL 778, Enterobacter
aerogenes 758 MSCL. Experiments were continued with
the best hydrogen producer using crude glycerol as
substrate - Enterobacter aerogenes 758 MSCL. Bacterial
cultures were inoculated in 200 ml flasks containing AB
medium (2,5 g/L yeast extract, 1g/L tryptone, 12,5g/L
glycerol), adapted from Tolvanen et al. (2011) [2].
Strains were cultivated overnight aerobically in shaken
flask at 37°C for 12 hours at 150 rpm using a multishaker PSU-20 (BioSan, Latvia). Optical density (OD)
calibration curve was used to find out number of cells in
1mL of culture [3]. The overnight culture in AB liquid
medium was put in a measurement flask sterilized for
measurements. The measurement flask was kept in a
termostat (Precisterm 2-110, 2L), in order to maintain
temperature around 37±2 ˚C. Analytical glycerol (AG)
(97%) and crude glycerol (CG) (40% wt/wt, determined
with HPLC analysis) from biodisel fuel production was
used as substrate, final concentration of glycerol used
was 135 mM. Glycerol was autoclaved for 60 min at
121 oC. Argon gas (99.99 % purity) bubbling through
the media was used to sustain anaerobic environment.
BlueSens gas sensors were used for the exhaust gas
analyses the bacterial test-system (H2, CO2 and CH4
sensors). Sensors were calibrated in AB medium using
pure hydrogen, carbon dioxide and methane and air and
argon gas for zero measurements. The evolved gas was
also subsequently injected in the mass-spectrometer
RGAPro-100 (HyEnergy, Setaram, France) for hydrogen
gas measurement. All cultivation experiments were
carried out in three independent repeats.
H2 production from crude glycerol
H2 concentration, mmol/l
4.50
4.00
3.50
3.00
E. aerogenes I
2.50
Cl. sporogenes I
2.00
1.50
E. cloacea I
1.00
E. asburiae I
0.50
0.00
0
20
40
60
80
100
120
Time, h
Application Report
Substrate
Bacteria
max,
CG
AG
Enterobacter
aerogenes
CG
Clostridium
sporogenes
AG
CG
AG
CG
AG
Enterobacter
cloacea
Enterobacter
asburiae
BlueSens Report No. 3
BlueSens.com
mmol/L/h
H2
production
rate average,
mmol/l/h
1.700
1.063
0.831
0.736
0.019
0.014
1.084
0.546
0.938
0.260
1.223
0.620
0.009
0.006
0.013
0.011
References
1 – Adhikari, S., Fernando, S.D., Haryanto, A. 2009. Hydrogen production
from glycerol: An update. Energy Conversion and Management, 50,
2600–2604.
2 - Tolvanen, K.S., Mangayil, R.K. Karp, M.T., Santala, V.P. 2011. Simple
Enrichment System for Hydrogen Producers. - Appl Environ Microbiol,
77(12): 4246–4248
3 – Widdel, F. 2007. Theory and measurement of Bacterial growth,
Grundpraktikum Mikrobiologie, 4. Sem. (B.Sc.). Bremen: Universität
Bremen.
4 - Sarma, S.J., Brar, S.K., Sydeny, E.B., Bihan, Y., Buelna, G., Soccol, C.R.
2012. Microbial hydrogen production by bioconversion of crude glycerol:
A review. – International Journal of hydrogen energy, 37: 6473-6490
Table 1
Results
Representative results from gas production experiments
of each type are shown in Table 1 and Figure2. There
were no CH4 gas production in the samples and CO2 gas
was measured only for determination of fermentation
process and results are not displayed in the report.
When comparing analytical glycerol and crude glycerol
as hydrogen fermentation substrates in liquid phase
with different anaerobic and facultatively anaerobic
bacteria, substantial differences were observed. The
highest rates were gained using crude glucose as substrate (H2 production rate max - 1.700 mmol/L/h) with
Enterobacter aerogenes.
Increased rate using crude glycerol (concentration 40%, comparably analytic glycerol concentration was
97%) must be regarded to different impurities in crude
glycerol as biodisel by-product that increase rude glycerol’s energy value- 19 MJ/kg (analytical glycerol), 25,3
MJ/kg (crude glycerol) [4].
Acknowledgements
We acknowledge BlueSens Company for providing the
gas sensing equipment.
Figure 2
18
H2 production
rate
BlueSens.com
i
by Dr. phys. Janis Kleperis,
MSc. Ilze Dimanta,
BSc. Arturs Gruduls,
MSc. Laimonis Jekabsons,
Hydrogen and Gas Sensors Laboratory, Institute of Solid
State Physics, University of Latvia
Hydrogen and Gas Sensors Laboratory:
Research and development of new materials and devices
for hydrogen production via electrolysis, using direct current
and short voltage pulses;
Development of technology for bio-hydrogen production in
anaerobic fermentation process; elaboration and research
of hydrogen separation membranes;
Development of new materials for hydrogen storage in
metal hydrides and composites;
Development and research of proton exchange membranes and electrodes for fuel cells, rechargeable MH/Ni
and Lithium batteries and their applications in electric –
hydrogen vehicles;
Research on gas sensors and sensor arrays with application for specific tasks (searching the rapid diagnostic
methods to detect lung diseases for patients and wheat
diseases).
BlueSens Report No. 3
19
Application Report
Application of BlueSens® H2 sensor
in a gas separation study
P. Bakonyi, N. Nemestóthy, K. Bélafi-Bakó
Research Institute on Bioengineering, Membrane Technology and Energetics
University of Pannonia, Egyetem street 10, Veszprém, Hungary
Introduction
Moreover, evaluating the influences of notable process
Biohydrogen is an emerging alternative energy carrier
variables, namely separation temperature and recovery
with high potential for sustainable development. It can
factor was also aimed.
be utilized in fuel cells or internal combustion engines.
Materials and Methods
However, purified hydrogen should be provided for these
Hydrogen purification experiments using BlueSens®
end-use applications thus its enrichment – through
H2 sensor
getting rid of gaseous side-products especially CO2
generated in the fermentative bioreaction – is very
In the current experiments a membrane module
important.
containing polyimide hollow-fibers (UBE Industries Ltd.,
A wide range of absorptive, adsorptive, cryogenic
NM-B01A) was investigated. This polymer is considered
techniques and alternatives relying on membranes may
as a highly promising material for membrane gas
be employed to concentrate biologically produced
separation including H2 purification, as well.
The measurements aiming the concentration of hydrogen
hydrogen. In recent decades as a result of advancements
were carried out in a laboratory-scale device
made in the field of membrane technology it is now
schemiatically presented in in Fig. 1.
accounted among the efficient, energy-saving and
ecological-friendly opportunities.
Membranes for gas separation are mature, commercially
available nowadays and mostly fabricated from high-performance polymers such as polyimides. Membranes for H2 purification, however,
are featured with challenges. On one
hand, membranes possess sufficiently high permeability and selectivity should be developed. On the
other hand, revealing the membranes
real potential under practical separation conditions e.g. with mixed gases
is still of interest but such studies are
not frequently conducted despite
their importance.
Therefore, in this work we wanted to
test a lab-scale polyimide memFig. 1 – Experimental membrane testing set-up
brane module for fermentative hy- 1; gas cylinder containing binary H2/CO2 mixture, 2; pressure controller, 3; feed valve, 4; feed pressure
drogen recovery applying binary H2/ indicator, 5; module thermostat, 6; membrane module, 7; differential pressure indicator, 8; permeate
pressure indicator, 9,10; permeate valves, 11,15; digital gasflow meter, 12,16; BlueSens® H2 sensor,
CO2 gaseous mixtures.
13; permeate fraction, 14; retentate valve, 17; retentate fraction
20
BlueSens Report No. 3
BlueSens.com
Application Report
Results and discussion
Statistical analysis (ANOVA) was performed to evaluate
the experimental design (Table 1). Based on ANOVA
results, feed H2 concentration, temperature and recovery
factor were all identified as significant process
parameters (p<0.05).
During the experiments, the hydrogen concentrations of
the feed, permeate and retentate streams were
continuously followed by using on-line BlueSens® H2
analyzers (BlueSens Gas Sensor GmbH). The equipment
is able to measure hydrogen in the range of 0-100 vol%
with a quick response time of 20 seconds. The additional
characteristics can be accessed by visiting the
manufacturer’s homepage (www.bluesens.com).
The feed gas pressures during the experiments were
adjusted as low as approximately 3 bar since considering
practicality, a high compression degree of biologically
produced gases comprising hydrogen would consume a
noticeable portion of its energy content and hence
making the process economically unattractive. The
permeate side of the module was kept under ambient
conditions.
The influences of three process variables (temperature,
feed hydrogen content and retentate/feed flow ratio
referred to as recovery factor) were determined by
experimental design approach using binary H2/CO2
mixtures.
To estimate the mixed gas selectivities (ai/j) according to
Eq. 1, only the data recorded at steady-state – when no
change in the retentate and permeate flows and in their
compositions occured – were considered:
ai/j = (xiP/xjP)/(yiF/yjF)
The response plots displaying the impacts of the
independent variables mentioned can be seen in Fig.
2-Fig. 4.
Impact of gas composition
The composition of the feed gas is apparently a factor
that determines the performance of the membrane
module. However, the separation efficiencies of the
membranes are more often than not sought only in
single gas experiments. This ideal value can be notably
different to those obtained under practical (mixed gas)
separation circumstances. The theoretical and real
selectivities would expectedly be similar only when the
constituents of a gas mixture do not strongly interact
with the membrane material. In the case of CO2 and H2
separation, however, the carbon-dioxide is easily
dissolved in membrane materials such as polyimide.
This phenomenon can remarkably affect the permeation
of the other gases e.g. hydrogen and thus the efficiency
of the overall separation process. From the tentative
results, we found that higher hydrogen concentration
could be attained in the permeate fraction by feeding
H2/CO2 mixtures with higher initial hydrogen
concentration. Based on the experiments it was
(1)
where x iP and x jP are the volumetric fractions of compound
„i” and „j” in the permeate, meanwhile yiF and yjF are the
their volumetric fractions in the feed mixture.
ANOVA – Dependent Variable: Mixed gas selectivity; R2: 0.98
Factor
SS
df
MS
F
p
Separation Temperature
0.007200
1
0.007200
11.2941
0.02828
Feed H2 content
0.068450
1
0.068450
107.3725
0.00049
Recovery factor
0.051200
1
0.051200
80.3137
0.00086
Error
0.002550
4
0.000637
Total SS
0.129400
7
SS: Sum of squares; df: Degree of freedom; MS: Main square; F: F-value, p: Significance value
Table 1 – ANOVA table showing the significance of the variables studied
BlueSens.com
BlueSens Report No. 3
21
Application Report
concluded that the more amount of carbon-dioxide in
the feed mixture represented a limiting factor to achieve
better selectivities. This might be explained by the fact
that carbon dioxide likely influences the migration of
hydrogen across the membrane capillaries. As CO2
concentration has decreased, improved separation
efficiency could be observed.
Fig. 2 – Influence of separation temperature and feed composition on mixed
gas selectivity
Influence of recovery factor
Membranes, in general, can enrich the slowly permeating
compounds in the concentrate stream, however, only
when low recoveries are applied. In our investigation,
almost pure (>99 vol%) CO2 could be obtained in the
retentate at certain, extremely low recoveries when the
vast majority of the gas introduced to the membrane
module was taken at the permeate side (Table 2).
CO2 Permeate
(vol%)
CO2 Retentae
(vol%)
Recovery
value
Mixed Gas
Selectivity
67.17
99.61
0.05
1.12
65.42
83.8
0.17
1.21
63.82
76.13
0.38
1.3
62.8
71.74
0.69
1.36
Table 2 – Results of separating H2/CO2 gas mixture
(Separation temperature: 37 oC; initial H2 content: 30 vol%)
Fig. 3 – Influence of feed composition and recovery factor on mixed gas
selectivity
It was also observed that the separation of the faster
permeating H2 is determined by the membrane’s selectivity, since H2 concentration in the permeate was only
slightly increased when recovery factor was increased.
Hence, if hydrogen is to be concentrated in the permeate fraction, membranes reflecting higher selectivities
need to be developed or multi-step, cascade applications should be employed. Furthermore, the change in
the recovery value also changes the amount of permeate which together with the total feed flow and the surface velocities should not be neglected to avoid concentration polarization. Basically, maintaining a high
recovery value is favorable so as to get a more sufficient
membrane performance. This might be the reason for
the increment of selectivity with increased recoveries.
Application Report
Separation temperature effect
It is quite obvious from the research outcomes that the
mixed gas selectivity has increased at higher
temperatures. It is understandable if one keeps in mind
that increasing temperature has contradictory impacts
on the solubility and diffusivity of the permeating gases.
In general, the former decreases, whilst the latter
increases with raising temperatures. Furthermore,
hydrogen and carbon dioxide have distinct characteristics
from diffusion and solubility points of views in rigid,
glassy polymers such as polyimide. CO2 usually
expresses a relatively higher solubility in comparison to
hydrogen which is in turn more rapid and possesses
faster diffusion coefficient. These properties of the gases
are in correlation with molecular size, their affinity to
polymers and the developing interactions with the
membrane material during permeation. Polyimide, as
mentioned, is a glassy polymer which basically achieves
selectivity mainly on diffusivity-difference basis. Since
CO2 is a larger molecule than hydrogen, the increment in
its diffusivity is less pronounced than that of hydrogen at
elevated temperatures. This was assumed to be the
reason for the increasing separation efficiency of H2/
CO2 with increasing temperature.
Conclusions
In our research work we established that the polyimide
membrane module had a potential to enrich hydrogen
from gaseous mixtures. The results demonstrated that
process variables such as gas composition, temperature
and the recovery factor had remarkable influence on
hydrogen separation and could affect the achievable
separation efficiency. Thus, these parameters are to be
considered when designing a suitable enrichment
process. It has been proven that higher H2/CO2 mixed
gas selectivity could be accomplished by increasing the
feed hydrogen concentration as well as the temperature
and the recovery value within the studied design
boundaries.
Fig. 4 – Influence of separation temperature and recovery factor on mixed
gas selectivity
22
BlueSens Report No. 3
BlueSens.com
BlueSens.com
i
DipI. Eng Péter Bakonyi, University of Pannonia
Research Institute on Bioengineering , Membrane Technology and Energetics
Ph.D. Nándor Nemestóthy, University of Pannonia
Research Institute on Bioengineering , Membrane Technology and Energetics
Prof. KataIin Bélafi-Bakó, University of Pannonia
Research Institute on Bioengineering , Membrane Technology and Energetics
The main tasks of the Research Institute on Bioengineering , Membrane Technology and Energetics are:
>> to teach and train students to get BSc In bioengineering
>> to introduce them to a wide range of biochemical and
biotechnological processes
>> to introduce them to the world of membranes and their
applications
Application possibilities of various membrane processes,
improvements of bioprocesses and manufacture of renewable , “green” energy sources are studied at the institute
involving the PhD students of the Doctoral School of
Chemical Engineering and Material Sciences (University of
Pannonia).
BlueSens Report No. 3
23
Application Report
Enhanced waste to fuel conversion
with a bioelectrochemically controlled
autotrophic bioreactor
John M. Pisciotta, Joe Mossman, Zehra Zaybak and William Schultz. Department of Biology,
West Chester University.
1. Introduction
Photobioreactors (PBRs) are increasingly being used to grow algae for production of biofuel and other products.
Light and CO2 are two major limiting
resources that can restrict algal growth
in PBRs. Atmospheric CO2 is only
0.04% and to overcome this limitation
researchers have started to couple biogas released from the anaerobic digestion (AD) of organic waste with PBRs as
biogas consists of mostly CO2 and CH4.
The CO2 component of biogas is normally considered a nuisance and its
abiotic removal can be expensive. Unfortunately, the conventional mesophilic AD is a slow process. Recently, bioelectrochemical systems (BES) have
Fig 1.: Simplified schematic overview of BES biogas fed PBR fitted with in-line BlueSens O2, CO2 and
been described that accelerate the an- CH4 gas sensors.
aerobic breakdown of organic waste
erated biogas via potentially enhanced growth of photointo biogas (Booth, 2009, Cheng et al., 2009). BESs
autotrophs. This gas sensor integrated arrangement
accomplish this via anaerobic cultures of red pigmentshould provide real time characterization of the overall
ed, exoeletrogenic Geobacter species grown on the anmetabolic activity inside a hybrid BES-PBR linked sysode with poorly understood electroactive methanogens
tem. Two PBR systems were built and tested head-tothat grow in association with biocathodes (Cheng et al,
head to study Chlorella vulgaris growth in BG-11 media
2009, Pisciotta et al, 2012, Zaybak et al., 2013).
on standard AD derived biogas versus BES derived
2. Aim of Investigation
In this report, BlueSens CO2, CH4 and O2 Gas Sensors
were integrated into a new type of BES-fed PBR built to
test if bioelectrochemically produced biogas can be
used to grow algae. If successful, this could provide a
method for more rapidly treating organic waste streams
while biologically processing bioelectrochemically-gen-
24
BlueSens Report No. 3
Application Report
0.5 cm holes were drilled into both ends of the glass
covers using diamond tipped drill bits. Gas incurrent and
out-current silicone tubing lines were passed through
these holes and sealed to prevent gas leaks. The input
gas line was connected to an airstone affixed to the
bottom of each tank. This provided for even gas
dispersion and bubble mediated mixing of the media.
Gas bubbling up into the headspace then exited PBR via
the out-current gas line connected to an air pump sealed
inside a hollow plastic reservoir. The pump set up
provides for constant gas flow through the closed loop
PBR systems. Prior to reentering the PBRs gas was
pumped through in-line O2, CO2 and CH4 sensors via gas
line tubing. For each PBR, a biogas delivery line tied into
the circulating gas just upstream of the pump and was
fitted with a one way valve. The one way valves prevent
oxygen from entering the BES or AD chambers and inhibiting anaerobes therein. Log grown C. vulgaris was inoculated into 5 L of illuminated (12hr light/12hr dark) BG11 per PBR at an initial density of OD 0.025 (Fig 2).
biogas (Fig 1).
3. Materials and Methods
3.1 Photobioreactors
Two rectangular 18.9 L glass aquaria were converted
into biogas-fed, closed-loop photobioreactors. Each
aquarium was sealed with a custom cut sheet of glass
epoxyed into place over the open top of each tank. Next,
BlueSens.com
Fig 2.: Side-view photograph detailing Chlorella containing PBRs fed by either BES (left) or AD biogas (right). Components of the system including pumps and sensor indicated by numbered arrows on side bar legend.
BlueSens.com
BlueSens Report No. 3
25
Application Report
3.2. Anaerobic Digesters and Bioelectrochemical Cells.
Triplicate gas-linked anaerobic digesters were
constructed from 100 ml serum vials containing 75 mls
of synthetic wastewater; ie 50 mM Phosphate Buffered
Saline (PBS) amended to contain 1 g/l sodium acetate
plus Wolfe’s vitamins and minerals. All vials were
degassed with nitrogen and crimp sealed prior to
injection of 0.5 mls of starter culture from a 5 week old,
acetate fed, biogas producing digester. Headspace of
the three anaerobic digesters was connected by needles
passing through the rubbers stoppers and connected by
tubing. Passive pressurization of biogas formation forced
the biogas through a one way valve just prior to biogas
entry into the recirculated PBR gas stream. Triplicate
BESs were similarly produced by inserting two 10 cm
electrodes into each BES vial and a potential of 0.7 V
was applied using a Biologic MPG-2 that also monitored
current over time; an indicator of bioelectrochemical
activity and acetate consumption. Cell density was
periodically measured at OD 600nm via
spectrophotometer.
4. Results and Conclusions
In-line BlueSens CH4 Sensors revealed a considerably
more rapid accumulation of methane in the BES-PBR
headspace compared to the conventional control AD
reactor (Fig 3A). Potentiostat measured current peaked
within four days for all three BESs indicating rapid
catabolism of acetate (Fig 3B). The optical cell density
of both C. vulgaris cultures increased from the initial
inoculum baseline level (0.025) during the course of the
incubation indicating algal growth in BG-11 media in
both PBRs (Fig 4A). The control AD linked PBR increased
more rapidly but the BES-PBR system reached a higher
final cell density. Oxygen sensor also indicated higher
algal photosynthesis in the late stage BES-PBR (Fig 4B).
For both reactors CO2 remained near baseline levels of
detection; likely a function of the relatively high
headspace volume and photosynthetic demand of the
cultures. Increasing the volume of the BESs and ADs
relative to the PBR culture volume could further improve
productivity. These results suggest integration of
26
BlueSens Report No. 3
BlueSens Sensors with photobioreactors can enable
optimized biofuel production from renewable resources
like sunlight and bioelectochemically treated organic
waste.
Application Report
References
Booth B. 2009. Electromethanogenesis: the direct bioconversion of current
to methane. Environ Sci Technol. 43(13):4619.
Cheng S, Xing D, Call DF, Logan BE. 2009. Direct biological conversion of
electrical current into methane by electromethanogenesis. Environ Sci
Technol. 43(10):3953-8.
Pisciotta JM, Zaybak Z, Call DF, Nam JY, Logan BE. 2012. Enrichment of
microbial electrolysis cell (MEC) biocathodes from sediment microbial fuel
cell (sMFC) bioanodes. Appl Environ Microbiol. 2012 Aug;78(15):5212-9.
Zaybak Z, Pisciotta JM, Tokash JC, Logan BE. 2013. Enhanced start-up of
anaerobic facultatively autotrophic biocathodes in bioelectrochemical
systems. J Biotechnol. In press. doi:pii: S0168-1656(13)00425-2.
Acknowledgments
We would like to thank BlueSens for use of the sensors
and the West Chester University College of Arts and
Sciences for providing funding to build the
photobioreactors.
i
Fig 3.: Methane accumulation and CO2 consumption in PBRs (A). Current
generation in triplicate BESs attached to the BES-PBR (B).
The Pisciotta Laboratory at West Chester University develops
­microbial systems for conversion of wastes into biofuels and
value-added products. Current research focus is directed at
electrical enhancement of CO2 fixation by microorganism using
biocathodes and novel systems like the MEC-linked photobio­
reactor described. Laboratory instruments and methods used
include potentiostats, bioelectrochemical systems, optical and
light microscopy, HPLC, GC, mass spectrometry, spectrophoto­
metry, molecular biology and experimental culture techniques.
Other projects include research and development of microfluidic
diagnostics to combat infectious diseases like malaria and low
cost water sanitation tools for developing nations.
Project team pictured from left: Students William Schultz, Zehra Zaybak, Joseph Mossman and Assistant Professor Dr. John M.
Pisciotta. Dana Charitonchick carried out stock culture maintenance and logistical support for this project.
Fig 4.: C. vulgaris optical cell density in PBRs (A) and sustained O2 detected in PBRs (B).
BlueSens.com
BlueSens.com
BlueSens Report No. 3
27
Information
Information
BCP- Series In-situ gas
Blue
analysis for p
­ rocess
controlled fermentations
Start QbD with BlueVis.
The BlueVis software is the most convenient solution to operate the BlueInOne gas analyzers. It logs
and visualizes the data and is a maintenance-tool for the monthly calibration as well. Now you can
easily connect up to 12 analyzers to one interface of your PC. This Bioprocessing software calculates
the oxygen uptake rate (OUR), carbon dioxide emission rate (CER) and respiratory quotient (RQ)
automatically by the collected data of the BlueInOne analyzer and the given process
parameters. No additional corrections of the data or manual calculations will be
needed. The BlueVis software is easy to use and to install. The BlueVis software comes
with an integrated OPC server and can easily communicate with process lead systems.
BlueVis will make QdB as easy and as convenient as possible.
Data
24V DC
ID=1
Gas in
Data Bus RS 485
Off gas
24V DC
ID=2
Gas in
Process 1
USB-adaptor
Off gas
Data
Off gas
Data
Data
Terminating
resistor
24V DC
ID=3 etc.
Gas in
Process 2
Process 3 etc.
The combined gas analyzers for CO2 and O2: The BlueInOne series
The state-of-the-art BlueInOne gas analyzers combine the parallel measurement of CO2 and O2 in one
space-saving analyzer. The BlueInOne Ferm is the modern tool for monitoring fermentations and the
BlueInOne Cell is the perfect device for the measurement of cell growth processes. The BlueInOne
analyzers will help you to monitor and understand your bioprocess in real-time.
Advantages
>> Parallel measurement of O2 and CO2
>> Compact stainless steel housing (IP65)
>> QbD and PAT
>> PAT conform in-situ-measurement
>> Auto compensated humidity and pressure
>> No gas cooler, pumps or valves needed
>> Connectable to any hose/tube or pipe
>> RS485 Modbus, RS232, USB, 2x4-20mA Output,
Modbus OPC server available
BlueSens Report No. 3
Data
Your advantages:
>> Run up to 12 parallel processes
>> Gain all the process data
>> Visualization of the measurements
>> Automatic CER, OUR und RQ calculation
>> Easy to use
>> Overview of all current measuring results
>> Log all data
>> OPC server included
>> Real RQ calculation
28
nection
New con
The BCP sensors are reliable devices for the laboratory and the industrial use. You
can choose between a robust version with an aluminum housing (IP 65) and a compact version with a PA-plastics housing. Whatever you decide for – BCP series optimizes bioprocessing easily, efficiently and precisely. So far we offer sensors of the
BCP series for ethanol, CH4, O2, CO2, H2 and CO. The sensors are already used in large
numbers for the Biotechnologies, Life Science, the production of Biofuels like Biogas
and Bioethanol, the pharmaceutical industry and numerous other areas of application. You can order various optional accessories like a ball valve (e.g. for the industrial scale Biofuels production). There a lot of optional features like an external
pressure sensor and a display and buttons for the on-site operation. The field of
application is large – whether in the bioreactor or in the fermenter, in the laboratory
or on the industrial large scale. The sensors are reasonably priced, long-term stable
and already often used in parallel bioreactors. BCP series works independently of
the gas flow quantity with an established dual wavelength infrared measurement
technology. The maintenance is more than easy and the operation costs are very low.
Due to its low weight of 150g, the compact sensor in a PA housing can be screwed
sterile directly at the shake flask. BlueSens sensors are always measuring on the
spot, where the process takes place. Thanks to its universal connections, BCP sensors can be easily integrated in existing gas lines. Wherever your bioprocess takes
place – in a shake flask, the fermenter or of course also for the use of disposals – the
BCP series is always the right choice. Sampling, gas cooling or additional complicated gas lines are no longer needed. Thanks to the standard interfaces the data can
be readout in real time by any electronic process control system.
Get the sensors in you process:
Easy installation:
> > Direct installation in the exhaust gas line
> > No sampling, no danger of contamination
>> No exhaust gas treatment (exhaust gas cooler) necessary
> > Minimum maintenance without test gases
> > Suitable for every fermenter and easy to integrate
Advantages:
> > Easiest process control
> > Increase of reliability and repeatability
> > Log all data
> > Metabolic flux analysis and mass-balance calculations
BlueSens.com
BlueSens.com
Gas tight and sterilizable
installation on shake flask
Optional display and button
Tri-Camp-installation
Biofuels-Production in the lab
Biogas-application with ball valve
and external pressure sensor
BlueSens Report No. 3
29
Information
Everything at a glance: product overview
Information
All the BlueSens’ gas analyzers can be easily integrated
directly into the gas lines independent of the gas flow.
Additional gas coolers, pumps and valves are not needed
a lot of maintenance are the result. With the aid of
standard interfaces, the sensors can be connected to
any process control system or computer.
to make the measurements.
The sensors measure at the point where things are
happening. Fast and reliable measurement data without
Sensor
CO2
CH4
CO
EtOH
Measuring range
0 … 10 Vol. %
0 … 100 Vol. %
0 … 30 Vol. %
0.3 … 25 Vol. %
0 … 25 Vol. %
6
0 … 100 Vol. %
BlueInOne Ferm
BlueInOne Cell
O2
O2ec
H2
Sensor
0 … 10 Vol. % CO2,
0.1 … 25 Vol. % O2
0 … 10 Vol. % CO2,
0 … 100 Vol. % O2
0.1 … 25 Vol. %
0 … 100 Vol. %
0 … 100 Vol. %
0 … 25 Vol. % CO2,
0.1 … 25 Vol. % O2
0 … 25 Vol. % CO2,
0 … 100 Vol. % O2
1 … 50 Vol. %
Infrared, dual wavelengths
(CO2) / Galvanic cell (O2)
ZrO2
Galvanic cell
Thermal conductivity
7
Measuring range
0 … 10 Vol. % CO2,
1 … 50 Vol. % O2
0 … 50 Vol. %1
0 … 25 Vol. % CO2,
1 … 50 Vol. % O2
Infrared, dual wavelengths
Measuring principle
Accuracy
Long-term stability
Infrared, dual wavelengths
(CO2) / ZrO2 (O2)
< ± 0.2 % FS* ± 3% reading
< ± 0.2 % FS* ± 3% reading
< ± 2% reading / year
< ± 2% reading / year
2
> 3 years
Lifetime sensor element
Accuracy
Long-term stability2
Approx. 3 years (CO2),
15,000 operating hours (O2)
Approx. 3 years (CO2), approx. 15,000 operating hours
900 000 Vol. % h operating
hours at 1 bar (14.5 psi) (O2)
Approx. 900.000 Vol.% h
operating hours at 1 bar
(14.5 psi)
> 3 years
Lifetime sensor element
Housing Aluminum, IP 65
Dimension (WxDxH) mm
Dimension (WxDxH) inch
Weight
100 x 131 x 118
3.94 x 5.16 x 4.64
900 g (1.98 lb)
100 x 131 x 118
3.94 x 5.16 x 4.64
900 g (1.98 lb)
100 x 131 x 118
3.94 x 5.16 x 4.64
900 g (1.98 lb)
100 x 131 x 118
3.94 x 5.16 x 4.64
3000 g (6.61 lb)
170 x 150 x 1205
6.69 x 5.91 x 4.725
4000 g (8.82 lb)
170 x 150 x 1205
6.69 x 5.91 x 4.725
4000 g (8.82 lb)
100 x 131 x 118
3.94 x 5.16 x 4.64
900 g (1.98 lb)
100 x 131 x 118
3.94 x 5.16 x 4.64
900 g (1.98 lb)
100 x 131 x 118
3.94 x 5.16 x 4.64
900 g (1.98 lb)
Housing Aluminum, IP 65
Dimension (WxDxH) mm
Dimension (WxDxH) inch
Weight
Housing PA6
Dimension (DxH) mm
Dimension (DxH) inch
Weight
80 x 130
3.15 x 5.12
350 g (0.77 lb)
80 x 130
3.15 x 5.12
350 g (0.77 lb)
80 x 130
3.15 x 5.12
350 g (0.77 lb)
80 x 130
3.15 x 5.12
350 g (0.77 lb)
Not available
Not available
80 x 130
3.15 x 5.12
350 g (0.77 lb)
80 x 130
3.15 x 5.12
350 g (0.77 lb)
80 x 130
3.15 x 5.12
350 g (0.77 lb)
Housing PA6
Dimension (DxH) mm
Dimension (DxH) inch
Weight
Steel 1.4404 / Sapphire / Viton / PTFE
Stainless steel / Viton /
Sapphire / PTFE /
Polymer H.L. / Nitrile
Stainless steel / Viton /
Sapphire / PTFE /
Polymer H.L. / Nitrile
G 1¼”, GL 45, Tri-Clamp, hose connection 4-12mm etc.
¼“ – 1 ¼“6, hose
connection 4-12mm etc.
¼“ – 1 ¼“6, hose
connection 4-12mm etc.
Material in contact
with gas
Connection6
Steel 1.4404 / Viton / PTFE
Steel 1.4404 / Viton
G 1¼”, GL 45, Tri-Clamp, hose connection 4-12mm etc.
General
Max. -25 – 55 °C / -13 – 131 °F6
Operating temperature
Pressure range (absolute):
15 - 40 °C / 59 - 104 °F
0.8 – 1.3 bar / 11.6 – 18.85 psi absolute pressure
0.8 – 1.3 bar / 11.6 – 18.85 psi absolute pressure
75% RH noncondensing
0 ... 100% RH
Operating humidity
3
12 or 24 VDC, 1-2 A
Power supply (max.)
RS232, 4-20mA, USB, Ethernet
Output
EN61326-1:1997 +A2:1998
accuracy < ± .0.5 % FS* ± 5% reading
BlueSens Report No. 3
2
with monthly 1-point calibration
3
not compensated
4
Storage temperature
Pressure range (absolute):
> 75% RH noncondensing
5...100% RH noncondensing > 75% RH noncondensing
5...100% RH noncondensing 75% RH noncondensing
Storage humidity
0…100% RH
5...100% RH noncondensing 0 ... 100% RH
5...100% RH noncondensing 0 ... 100% RH
Operating humidity
Integrated humidity
compensation
Integrated humidity
compensation
24 VDC, 1 A
24 VDC, 1 A
3
12 or 24 VDC, 1-2 A
24 VDC, 1-2 A
RS232, RS485 Modbus, 2x
4-20mA, USB, Modbus OPC Server
RS232, 4-20mA, USB, Ethernet
One point calibration with
ambient air once a month
(other conditions on request)
One point calibration with
ambient air once a month
(other conditions on request)
1-point calibration with ambient air or nitrogen
(other conditions on request)
EN61326-1:2006 / FCC
15:2009 Subpart 107/109,
ICES - 001:2006
EN61326-1:2006 / FCC
15:2009 Subpart 107/109,
ICES - 001:2006
Optional factory calibration with certified gases
Maintenance yearly
CE
Operating temperature
6
RS232, RS485 Modbus, 2x
4-20mA, USB, Modbus OPC Server
4
1-point calibration with ambient air or nitrogen
(other conditions on request)
Maintenance once a month
Max. -25 – 55 °C / -13 – 131 °F6
0 – 60 °C / 32 – 140 °F
6
Storage humidity
1
15 - 40 °C / 59 - 104 °F
0 – 60 °C / 32 – 140 °F
Storage temperature
*full scale
Material in contact
with gas
Connection6
General
30
Measuring principle
4
Optional factory calibration with certified gases
Ethernet with BACCom
5
BlueSens.com
BlueSens.com
6
others on request
Power supply (max.)
Output
Maintenance once a month
Maintenance yearly
EN61326-1:1997 +A2:1998
stainless steel housing, dimensions depends on flow adapter
3
7
CE
binary mixture
BlueSens Report No. 3
31
Information
Notes
Analyzer systems
Yield Master
Measure the gas yield and quality in every anaerobic process
The unique structure of the CH4 devices from BlueSens facilitate measuring methane concentrations
in processes that diversify in gas production. The use of sample taking is impossible there, so conventional systems fail.
The CH4 sensors are easily screwed onto the fermentation container and obtain the methane content
directly over the sample in the gas phase. Even at 55 °C (131 °F) in water-saturated atmospheres. The accruing
volumes are precisely registered via a precision volumenometer (Milligascounter®*).
The data is registered online with the corresponding software and visualized on the computer. Optionally, BlueSens
can provide everything as accessories; from the stirrer to the incubator. Additional sensors: To cover as many
measurement parameters as possible, BlueSens also offers sensors
for ethanol (C2H6O), hydrogen (H2)
and carbon monoxide (CO).
* Registered trademark. The MilliGascounter was
developed at the University of Applied Science
Hamburg under the leadership of Prof. Dr. Paul
Scherer.
BC preFerm
BlueSens Report No. 3
ab
tig
s
32
Ga
The same sensors are also integrated in the BCpreFerm system, which
is designed for process optimization (scale up) from flasks up to large-scale
fermenters. The system comprises up to 12 sensors that are linked to a computer via an electronic multiplexer. The related software visualizes the results
and calculates vital parameters such as the oxygen
uptake-rate (OUR), the carbon dioxide emission rate
(CER) and the respiration quotients (RQ) both on
fermenters as well as on flasks.
>> Visualization of the process
>> Increase of reliability and repeatability
>> Dedicated process optimization without limitations (e.g. oxygen, nutrients etc.)
>> Predictions for your scale up
le
w connection
e
N
Simple tool for process optimization
h t & s t e r il i z
BlueSens.com
BlueSens.com
BlueSens Report No. 3
33
Notes
34
BlueSens Report No. 3
BlueSens.com
Notes
BlueSens.com
BlueSens Report No. 3
35
Questions?
Please ask directly!
Phone
+49 2366 4995 500
Or visit our homepage:
www.BlueSens.com
BlueSens gas sensor GmbH
Snirgelskamp 25 • D-45699 Herten (Germany)
Phone +49 2366 4995-500 • Fax +49 2366 4995-599
e-mail: info@BlueSens.de
BlueSens.com