bluesens report no.1

Report No. 1

BlueSens Report No. 1, October 2010

© 2010 by BlueSens gas sensors GmbH, Herten, Germany, www.BlueSens.com

This report was worked out in co-operation with APZ Ruhr-Lippe, www.apz-rl.de

Page layout and cover design: Marcus Riepe, Krefeld, Germany, [email protected]

Printing press: Offset Company, Wuppertal, Germany, www.offset-company.de

Introduction

BlueSens.com 3BlueSens Report No. 1

You may wonder why we have published a

report with detailed product information and user data

only recently after almost 10 successful years and several

thousand sensors already sold?

The answer is quite simple – time.

It is only recently we have increased the human resourc-

es in our PR-department and in addition we have

received the support of the APZ (center for applications

Biotechnik Ruhr-Lippe) which has made it possible for

us to publish now.

Almost every new customer wanted this type of information

however until recently we referred them to our “reference

customers” (at this point we want to thank them sin-

cerely).

Nevertheless, we knew that sooner or later we had to

produce a report with case studies and information

regarding our customer experience with our sensors.

Consequently we contacted our customers and asked

them for a case study where they describe how they use

our sensors.

Obviously, until recently many pharmaceutical companies

were unable to cooperate due to the confidential nature

of their work. However, as we do not require confidential

data about the microorganisms or the specified micro-

bial strains, we only require the official statement: “Yes,

we use BlueSens sensors and they operate as speci-

fied.”

Stunning are also statements as follows: “BlueSens sen-

sors? You don’t have to explain them to me, we solely

use them and no other sensor.” (stated by an anony-

mous customer during a call).

Where has this customer bought our sensors?

The answer is absolutely clear – of course we couldn’t

have achieved the global supply of our products without

our sales partners or OEM-distributors, like DASGIP AG,

Sartorius Stedim Biotech GmbH, Infors, Applikon, Bioen-

gineering or many other plant developers. We also want

to thank them for the longtime and good cooperation.

Longtime double-digit growth rates give us the motivation

to continue to achieve these results in future. In order to

do this, we listen to our customers and respond quickly

to their needs. So our sensors are specified for each

customer’s application: temperature, respective gas flow

or different pressure ranges – we have a solution for their

requirements. Thanks to the PAT-initiative of the FDA,

which deals with the analysis of the process and not

only with the end product, our online-sensors are well-

accepted by our customers. Our sensors can be inte-

grated with little effort directly in the process and so are

made for current requirements.

We also want to take a look to the future. For aerobic

fermentations, up to now, you had to connect one sensor

for the measurement of CO2 and another one for the

measurement of O2. Although that is not difficult, it would

be more convenient to receive all required measurement

data with one device.

We have listened to our customers and have reacted to

them:

BlueInOne. The most compact gas analyzer on the market

for the measurement of up to 4 gases with an automated

pressure and humidity compensation.

In this spirit I wish you to enjoy reading our report and

want to thank all of our customers and sales partners for

their confidence in us.

The first BlueSens Report

Dr. Holger Mueller

Contents


6 BlueSens. Advanced information

8 Application of a self constructed off gas analyser in the education of bioengineers

Dr. Michael Maurer, FH Campus Wien

10 Continuous bio-ethanol production by means of yeast

Dr.-Ing. Eva Maria del Amor Villa, Technical University Dortmund

12 Model based optimization of biogas plants

Dr. F. Uhlenhut, Prof. Dr. S. Steinigeweg, Prof. Dr. A. Borchert, Prof. Dr. Reiner Lohmüller,

University of Applied Sciences Emden/Leer

14 Online observation of oxygen uptake and carbon dioxide production and characterisation of oxygen

transfer capacity

by Prof. Dr.-Ing. Reiner Luttmann et al., Hamburg University of Applied Sciences

16 The precultivation in shake flasks for the execution of bioreactor cultivations


18 Automated Design of Experiments (DoE) in a multi-bioreactor system BIOSTAT® Qplus 6


20 Monitoring of baker’s yeast fermentations

PD Dr.-Ing. Lars M. Blank, Technical University Dortmund

22 Application of BlueSens® Gas Analyzers in a Cell Culture Process

Mathias Aehle, Martin-Luther-University Halle-Wittenberg

Application Reports

28 Connections for every application

30 BlueSens’ sensors overview

32 We help you understand, control and optimize your process!

BCpreFerm and YieldMaster

33 The freedom of software choice

FermVis and BACVis

34 Parallel systems · Measuring according to PAT

Information

For controlling biotechnological processes you primarily

depend on continuous information. BlueSens has

made it to its business to provide this information for

every customer by means of gas analysis directly in

the process. Reliable measurement engineering

makes the results available in highest measurement

density and in real time. So biotechnological processes

can be analyzed better and, as a result, of course

also optimized.

With this report we also want to give you advanced

information. Numerous examples of applications

show exactly how the products of BlueSens are used

under real conditions. With this booklet you can also

learn concretely how our sensors are connected and

readout.

Dr. Holger Mueller

(Sales and Marketing)

Dr. Udo Schmale

(R&D and Production)

Furthermore, you can inform yourself about the

accurate specifications of the particular sensors with

the help of a clearly arranged spreadsheet.

Bluesens: advanced information for your process,

advanced information about the products.

Nearly a decade after the foundation the dynamic

company is well-known in the world of biotechnology.

BlueSens stands for reasonably priced quality sen-

sors – made in Germany. The strength of the company

is the personal contact to every single customer.

The Managers

6

During the calibration process each sensor is tested

and set up particularly. This process can take up to

one week. It involves a lot of time, but it‘s worth it. In

the detailed test procedure BlueSens solely uses certified

test gases. Depending on the gas component to be

measured, 10 to 18 different test gases are used. So

it is assured that the sensors provide best results for

each application the customer requires. Each sensor

so becomes a handmade piece and is individually

tested by BlueSens.

BlueSens stands for sensors which are as uncompli-

cated as possible and therefore as competitive as

possible. Based on the ever latest developments

BlueSens would like to pass on its competitive edge

to its customers. With the measuring systems of

BlueSens the corresponding process parameters can

already be determined before the actual process

takes place. In the production range of active compo-

nents, fermentation and also biogas generation, the

productivity of the raw material can thus be optimized

in preliminarily tests based on the gas measurement.

In research and development BlueSens sensors mean

that results are achieved faster and products can be

positioned in the market quicker. The use of BlueSens

sensors also means that production online can be

BlueSens is exclusively producing the sensors in Ger-

many. We have the highest requirements regarding

the utilised components. So the company guarantees

long-lasting quality and reliability of the products.

Production goes hand in hand with research. By

short ways the results of our Research & Develop-

ment department can be integrated quickly into the

production.

optimized when controlling industrial processes on

the spot, directly where the process takes place. This

saves both personnel and production capacities and

maximizes the outcome. The investment costs amor-

tize very quickly (Return on investment).

Already installed bioreactors can also be upgraded

with the sensors of BlueSens with minimum effort.

Therefore older installations can be modernized cost-

effectively.

Many customers confirm ever gain:

BlueSens: “We cannot afford not to have it!”

Every sensor a unique piece

Keeping an eye on costs

7

Application Report

BlueSens Report No. 1 BlueSens.com8

Our University of Applied Sciences, FH

Campus Wien, offers a degree program in ‘Bioengineering’.

In the course of this study a fermentation laboratory has

to be attended. The aim of this course is the design,

operation and analysis of a bioprocess experiment. The

students have to use their biological, mathematical and

technical skills to solve this exercise.

One of the experiments involved cultivation of the meth-

ylotrophic yeast Pichia pastoris (X33); a well known host

for recombinant protein expression (Cregg et al. 2000),

as well as for applications in white biotechnology (e.g.

riboflavin (Marx et al. 2008)). An overnight shake culture

was used to inoculate a defined 2 l batch medium (as

described in Maurer et al. 2006) with 40 g glucose L-1

as sole carbon source, to a starting optical density

(OD600) of 1.0.

The cultivation was carried out in a 5.0 l bioreactor

(Minifors, Infors, Bottmingen-Basel, Switzerland; figure 1

B) with a tailored off gas analyser. This off gas analyser

consists of a BCP-CO2, a BCP-O2 probe (BlueSens,

Herten, Germany) and a mass flow controller (Vögtlin,

Aesch, Switzerland) with a power supply in a separate

control box (figure 1 A). The analogue signals were

directly led to an I/O input of the bioreactor and mea-

sured as control parameters in the monitoring software

(IRIS, Infors).

The fermentation temperature was controlled at 25°C,

pH was controlled at 5.0 with addition of 25% ammo-

nium hydroxide and the dissolved oxygen concentration

was maintained above 20% saturation by controlling the

Application of a self constructed off gas analyser in the education of bioengineers by DI Dr. Michael Maurer, FH Campus Wien – University of Applied Sciences, Bioengineering degree programme

Figure 1: A) self assembled off gas analyser

Figure 1: B) bioreactor with off gas analyser

Application Report


stirrer speed between 250 and 1200 rpm and the air

flow between 2.0 and 5.0 l min-1.

Samples were taken frequently over the whole process

and analysed as described below. Three aliquots of

10 ml of culture broth were centrifuged and the super-

natant saved for HPLC analysis. The pellets were washed

in distilled water and recentrifuged, transferred into

weighed beakers and dried at 105°C until a constant

weight was attained. The biomass concentration was

also monitored with an on-line probe (Fogale nanotech,

Nimes, France), which had previously been calibrated

with dry cell mass data (CDW).

Glucose and ethanol were analysed by HPLC (Shimadzu,

DI Dr. Michael Maurer, FH Campus Wien – University of Ap-plied Sciences, Bioengineering degree programme. The University of Applied Sciences, FH Campus Wien, is an ed-ucational institution which offers a rich variety of academ-ic studies. The bioengineering degree programme edu-cates students for their work in the field of biotechnological industry. www.fh-campuswien.ac.at

i

Figure 2: A) Trends of measured cultivation parameters glucose - (squares), ethanol – (triangles) and bio mass concentration (crosses), as well as the carbon balance (circles).

Figure 2: B) RQ trend read out of the P. pastoris batch cultivation.

Application Report


Japan) using an ion exchange column Aminex HPX-87H

(Bio Rad). The mobile phase was 15 mM sulphuric

acid.

The aim of the exercise was the calculation of typical

fermentation parameters such as biomass concentration,

substrate uptake rate, specific growth rate, and so on, as

well as the respiratory quotient (RQ) and the over all

carbon balance (OCB). Using the universal gas equation

and the recorded oxygen and carbon dioxide concentration

[%] and the air flow data. The students were able to

calculate the oxygen uptake rate (OUR), the carbon dioxide

evolution rate (CER) and hence the required RQ and

OCB.

Figure 2 A shows the diauxic behaviour of this yeast

strain, first using up glucose as preferred substrate (spe-

cific glucose uptake rate qGlucose= 0.44 g g-1 h-1) and

forming ethanol with a rate of qP ethanol= 0.08 g g-1

h-1 as by product. After a first stationary phase the ethanol

was utilised with a rate of qethanol = 0.04 g g-1 h-1.

The online measurement of the oxygen and carbon dioxide

concentrations enabled the simultaneous determination

of the shift based on the calculated RQ, which changed

from 1.2 during the aerobic glucose consumption to 0.5

during the ethanol utilization. The carbon utilisation was

therefore balanced with a tolerance of 93-105%. These

online measurements therefore serve as teaching vehicles

enabling the students to grasp application and value of

off-gas analysis.

LiteratureCregg, J., J. Cereghino, J. Shi & D. Higgins (2000) Recombinant protein expression in Pichia pastoris. Mol Biotechnol, 16, 23-52. | Marx, H., D. Mattanovich & M. Sauer (2008) Overexpression of the riboflavin biosynthetic pathway in Pichia pastoris. Microb Cell Fact, 7, 23. | Maurer, M., M. Kuehleitner, B. Gasser & D. Mattanovich (2006) Versatile modeling and optimization of fed batch processes for the production of secreted heterologous proteins with Pichia pastoris. MICROBIAL CELL FACTORIES, 5, -.

One example for applying the BlueSens

technology at the Biochemical Engineering Laboratory is

the gas online-monitoring for the continuous bio-ethanol

production in the field of the biotechnological produc-

tion of alternative fuels (so-called biofuels). Yeast is able

to metabolize under anaerobic conditions several car-

bon sources (particularly sucrose and glucose) into car-

bon dioxide and ethanol, conventionally in a batch or

fed batch mode. However, if the ethanol concentration

exceeds the concentration threshold – ca. 115 g/l, de-

pending on the strain – an inhibition of the metabolism

is initiated: ethanol becomes a toxic substance and the

maximum product concentration achieves a biological

limit. Keeping the product content under the tolerance

limit of the cells will allow increasing the bio-ethanol-yield

to its maximum.

Continuous bio-ethanol production by means of yeastby Dr.-Ing. Eva Maria del Amor Villa, Biochemical Engineering Laboratory, Biochemical and Chemical Department, Technical University Dortmund

Dr.-Ing. Eva Maria del Amor Villa, Biochemical Engineering Laboratory, Biochemical and Chemical Department, Tech-nical University Dortmund. The Biochemical Engineering Laboratory deals with research and teaching in the areas of fermentation and sterilization technology, downstream processing as well as biocatalysis (in aqueous and organ-ic media). Pilot equipment for process scale up is avail-able up to a fermentation capacity of 300 l for interfacing with academic and industrial partners. www.bvt.bci.tu-dortmund.de

i

Application Report


The continuous bio-ethanol production by means of in

sodium alginate entrapped Saccharomyces cerevisiae

(ATCC 7752) was successfully carried out at 40°C in a

stirred bioreactor with an operating volume of 600 ml by

continuous substrate feed over a period of five days.

The sensors were connected gastight, allowing quantita-

tive online records on gases (carbon dioxide, ethanol

and oxygen) present in the headspace of the bioreactor

(see figure 1). By using a suitable calibrated ethanol

sensor a direct calculation of the ethanol content in the

liquid phase could be made based on the ethanol con-

tent in the gaseous phase; those results were validated

by comparative analysis using high performance liquid

chromatography.

The measurement of the unavoidable metabolite CO2 in

the bioreactor and the oxygen content in the flue gas

stream provided the expected results (see figure 2): the

CO2 concentration increased up to 90 Vol.-% and stabi-

lized at that value as no ambient air could enter the

bioreactor. The oxygen content stagnated after reaching

its minimum (approx. 0 Vol.-%), as only CO2 and ethanol

were discharged from the system. The ethanol concen-

tration remained almost constant after the first 60 oper-

ating hours. However, the tolerance limit for yeast with

respect to ethanol was by no means reached, as it was

solely intended to show that such a system could be

operated over a longer period of time.

The proposed measurement method offers the advan-

tage that the analysis is not influenced by further media

components and metabolites (e.g. organic acids). Strik-

ingly, this demonstrates the potential that the arrange-

ment used to determine online ethanol concentrations

can be applied to limit the ethanol content in the me-

dium due to an adequate adjustment.

Actual works dealing with the continuous

production process of bio-butanol

(under anaerobic conditions) and bio-

tensides (rhamnolipids) extent the

field of application of the BlueSens

technology for the gas online-monitor-

ing in biotechnological processes.

Figure 1: Stirred unit reactor with connected CO2, O

2 and ethanol sensors

Figure 2: Gas online-monitoring of the bio-ethanol production process by continuous feed of 40 g glucose/l

Application Report


MotivationIncreasing amount of energy derived from biogas plants

will only be available if a wide variety of different sub-

strates can be used. The feed to a biogas reactor will

change according to the fluctuating supply demand sce-

nario for various substrates. The plant has to deliver

maximum gas yield and hence energy yield for various

substrates. This can only be achieved if the process

parameters are optimized continuously. The model

should be able to predict optimized process parameters

as well as energy yield for a given substrate mix. There-

fore the model has to take biological processes into

consideration which takes place during anaerobic digestion.

The aim of our research at the Emder Institut für Umwelt-

technik (EUTEC) is to develop a sophisticated process

model which is capable of predicting the behavior of an

industrial sized biogas plant. The model should

include:

Simulation of biogas production for different >

substrate mixtures.

Adaptation of appropriate modeling approaches for >

the simulation-based evaluation of complex substrates.

Design of a control concept for biogas plants. >

ExperimentsFollowing experimental facilities have been used:

Batch experiments in 1 liter flasks at 37°C for 2-3

weeks. Aim was to evaluate gas generation rate for vari-

ous substrates continuous reactor in 20 liter scale.

Equipped with screw pumps and BlueSens analytics

system to count gas quantity and gas composition

(methane and carbon dioxide) in a continuous mode.

SimulationSimulation studies have been performed using ADM1

model incorporated into Matlab/Simulink. Parameters

of ADM1 kinetic model have been regressed to experi-

mental data.

ResultsFigure 1 shows experimental results in comparison with

calculated results for the continous recator in semi-

industrial scale. A very good agreement between both

data can be observed indicating that the model is

capable of describing the complex biological processes.

As input parameters only readily available data for the

substrates have been used.

In order to evaluate the capabilities of the model data

from the biogas plant in Wittmund (Germany) have been

Model based optimization of biogas plants by Dr. F. Uhlenhut, Prof. Dr. S. Steinigeweg, Prof. Dr. A. Borchert, Prof. Dr. Reiner Lohmüller, University of Applied Sciences Emden/Leer, EUTEC Institute

Figure 1 Comparison of experimental (black line) and simulated data (red line) for manure (left diagram) and fat mud (right diagram).

Application Report


compared to results predicted by the model (figure 2).

Again just readily available parameters describing the

substrate and the biogas plant have been incorporated

into the process model.

As can be seen a very good agreement between experi-

mental data and data from the biogas plant have been

achieved.

Further research will focus on incorporating a wide

variety of different substrates, to account for substrate

pre-treatment and for biogas purification.

Dr. F. Uhlenhut, Prof. Dr. S. Steinigeweg, Prof. Dr. A. Borchert, Prof. Dr. Reiner Lohmüller, University of Applied Sciences Emden/Leer, EUTEC Institute. Research and development in the following areas:

Optimization of industrial processes with respect to >high level of sustainabilityTechnologies to reduce pollutants in soil water and air >Bioenergy >Renewable resources as new raw materials >

www.technik-emden.de

i

Figure 2 Calculated (red line) and experimental data (black line) from industrial sized biogas plant in Wittmund (Germany).

Application Report


A 5 l stainless steel bioreactor

BIOSTAT® ED5 was used for the production of

the chemokine 1-8 del MCP-1, 1-3 del I-TAC,

vMIP-II as well as for potential Malaria vac-

cines with the yeast Pichia pastoris in HCDC.

The high instrumented reactor is equipped

with BlueSens sensors for the measurement

of oxygen and carbon dioxide (BCP-O2 and

BCP-CO2). The sensors are placed in the off-

gas line of the fermenter, behind the off-gas

filter.

The signal for the molar fraction of oxygen xO2

and carbon dioxide xCO2 are recorded and

stored in the data acquisition system MFC-

Swin. Different gas balance values are

calculated with the control system and stored

online.

The fermentation process starts with a batch

phase with unlimited growth on the substrate

glycerol. In the following glycerol fed batch

phase, limited cell growth is preparing the

cells for the production phase on controlled

methanol concentration.

In figure 1 the off-gas molar fractions xO2 and

xCO2 are shown. With an air aeration the in-

coming molar fractions are known (xOGin =

xOAIR = 0.2094, xCGin = xCAIR = 0.0003).

So the oxygen supply rate QO2, the carbon dioxide pro-

duction rate QCO2, the respiratory quotient RQ and the

oxygen transfer capacity OTC can be calculated online.

The dissolved oxygen tension pO2 is controlled via pO2/

agitation control at a setpoint of 25%. The regulation

starts at t = 12 h, when the pO2 drops below the set-

point.

During the fed batch phase QO2 and QCO2 are increasing

exponentially proportional to the volumetric cell growth rate.

The RQ converges to a stationary endpoint of 0.9 at batch

end. With reduced cell growth both rates drop down at

the beginning of the fed batch phase, but increase expo-

nentially again afterwards. In the production phase the

cell activity is reduced again. This can be observed in a

decreased QO2 and QCO2. The oxygen transfer capacity

OTC is a valuable parameter for the characterization of a

bioreactor plant and a capable scale up criteria.

Online observation of oxygen uptake and carbon dioxide production and characterisation of oxygen transfer capacityby Prof. Dr.-Ing. Reiner Luttmann et al., Research and Transfer Center of Bioprocess Engineering and Analytical Techniques Hamburg University of Applied Sciences

Bioreactor for recombinant protein production research

Application Report


Figure 1: Course of off-gas measurement and gas balance values

Figure 2: Course of O2-transfer rates during cultivation

Application Report


In figure 2 the online estimation of the OTC and the

volumetric O2-transfer coefficient kLa are shown together

with the influencing variables FnG (aeration rate) and

NSt (agitation speed).

Although FnG and NSt are constant in the beginning,

kLa and OTC are slightly decreasing. Due to the

exponential cell growth during pO2-control (since t=12 h)

the oxygen uptake is increasing exponential too. There-

fore the OTC has to be increased. The pO2-controller rises

FnG and later on NSt to keep the kLa on track and there-

with the OTC.

iProf. Dr.-Ing. Reiner Luttmann, Prof. Dr. Gesine Cornelissen, Dipl.-Ing. Ulrich Scheffler, Dipl.-Ing. Hans-Peter Bertelsen. Research and Transfer Center of Bioprocess Engineering and Analytical Techniques Hamburg University of Applied Sciences. The institute is engaged in advanced bioprocess engineering in fields such as production of potential malaria vaccines, optimi-zation of recombinant protein production (DoE), Process Analytical Technology (PAT) and modeling and simulation of bioprocesses.

The precultivation in shake flasks for the execution of bioreactor cultivationsby Prof. Dr.-Ing. Reiner Luttmann et al., Research and Transfer Center of Bioprocess Engineering and Analytical Techniques Hamburg University of Applied Sciences

For the execution of bioreactor

cultivations the precultivation in shake flasks is

from great interest. The cells should be in good

condition to avoid a long adaption phase in the

beginning. For assuring vital cells in the precul-

ture no substrate and no oxygen limitation should

occur during cultivation and cells should be in

exponential growth.

Shaking flask experiments have been carried out

for the optimization of preculture conditions.

Therefore a 1 l glass Erlenmeyer flask was

equipped with the BluSens Sensors BCP-O2 and

BCP-CO2 for the measurement of oxygen and car-

bon dioxide in the gas phase. For comparison an

optical oxygen microsensor was also used.

A recombinant Escherichia coli strain was cultivated. The

experiments were conducted in a shaking flask cabinet

at 200 rpm and 37 °C.

In figure 1 the course of the percentaged molar fraction

of oxygen xO2 and carbon dioxide xCO2 is shown. The

signal from the BlueSens

O2-sensor (BS) is corresponding very well to the signal

of the optical sensor. The BlueSens signal however is

much noiseless comparing to the other.

In the beginning xO2 starts at a value around 21 % which

is equal to the oxygen fraction of air (20.94 %). With

increasing cell growth the oxygen demand is increasing

proportional, so that the xO2 is decreasing to a value

around 15.7 % at t = 6.5 h. The signal of xCO2 is contrary

proportional to xO2.

Figure 1: Course xO2

and xCO2

signals of shaking flask experiment

Application Report


Another experiment was conducted with

additional measurement of the dissolved oxy-

gen tension pO2 in the liquid phase (figure 2).

This gives the opportunity for a better identifi-

cation of oxygen limitation and verification of

the data from the gas phase.

The signals of the xO2 signals are corresponding

still very well in this experiment. The pO2 is

decreasing exponentially with increasing cell

growth. After 4.2 hours oxygen limitation

occurs. This can be seen also in the xO2 signal

in a decreasing slope of the curve. At t = 6.5

h the substrate is exhausted and substrate

limitation begins. The xO2 graph is at the lowest

point at this time.

As mentioned in the beginning, the cells

should be in exponential growth and limita-

tions should be avoided. Therefore the dura-

tion of the preculture should not exceed 3.5

hours. With an optimized preculture consistent

initial conditions for bioreactor cultivations

can be realized. Thus a better reproducibility

and robust cultivation conditions can be

achieved.Shaking flask experiments with BlueSens Sensors

Figure 2: oxygen measurement in gas and liquid phase of shaking flask

Application Report


The bioreactor system BIO-

STAT® Qplus (Sartorius Stedim Biotech

GmbH, Göttingen) was established in the

Laboratory of Bioprocess Automation at

Hamburg University of Applied Sciences.

This multi-reactor system enables the exe-

cution of parallel experiments with inde-

pendent measuring and control of process

parameters. Therefore it is a very powerful

solution for the execution of optimization

experiments following DoE.

The system consists of two supply towers,

a digital control unit DCU 4 and six auto-

clavable 1 l culture vessels. Each vessel is

equipped with probes for measurement of

pO2, pH and foam. Two external gasmix sta-

tions with mass flow controllers are used for aeration up

to 2 vvm. A pump station enables different substrate

limited fed-batch operations with reduced cell specific

growth rates.

With the BlueSens sensors BCP-O2 for oxygen and the

Automated Design of Experiments (DoE) in a multi-bioreactor system BIOSTAT® Qplus 6 by Prof. Dr.-Ing. Reiner Luttmann et al., Research and Transfer Center of Bioprocess Engineering and Analytical Techniques Hamburg University of Applied Sciences

Figure 1: O2- and CO

2-signals from one experiment showing all six vessels

with batch phase followed by a fed batch phase

Multi-bioreactor system BIOSTAT® Qplus 6 for the execution of DoE optimization experiments

Application Report


BCP-CO2 sensors for carbon dioxide the measurement of

these gases in the off-gas of each vessel is possible. The

multiplexer unit BACCom transferring the off-gas values

to the process control system MFCSwin, where data are

recorded and further online calculations are carried

out.

Experiments for the optimization of the space-time-yield

of a recombinant fusion protein expressed in Escherichia

coli are conducted. The process starts with a glucose

batch, followed by a fed batch phase and the IPTG

induced production phase.

Figure 1 shows the course of the off-gas measurement

of all six vessels from the multi-reactor system. The ini-

tial conditions in every single reactor are the same. Also

for the batch part all parameters are identical. This can

be seen in an almost identical course of the six curves

in the batch phase and the very small variation of the

batch end time. In the fed batch phase the cell specific

growth rate µ and the liquid phase temperature JL are

changed to different values (see figure 1). Also the

incoming oxygen mole fraction xO2 was increased step-

wise from 20.94 % (AIR) to 45 % (AIR/O2) to avoid oxy-

gen limited cell growth.

The production phase of two different DoE experiments

is plotted in figure 2. For a better comparison of the two

experiments the timeline of the chart is standardized

onto the point of induction at the beginning of the pro-

duction phase. The plot shows the observable cell spe-

cific growth rate , estimated online from the off-gas sig-

nals xO2 and xCO2, the fluorescence signal S48/53_sol

of the soluble fusion protein measured in relative fluo-

rescence units (RFU) and the cell density cXL deter-

mined from cell dry mass. The setpoint of the cell spe-

cific growth rate µw, realized with an open loop controlled

glucose fed batch, was set to 0.18 h-1 in experiment

1 and 0.21 h-1 in experiment 2. After induction the

growth rate is decreasing due to the change in metabo-

lism and a reduced liquid phase temperature in the pro-

duction phase, but it is increasing afterwards and shows

an almost constant course.

The chosen parameters in experiment 1 yield in a much

higher target protein concentration compared to experi-

ment 2.

Figure 2: -estimation with off-gas measurement and O2-balancing

Application Report


The open question we addressed with the

new setup from BlueSens (CO2 and ethanol sensor)

originated from our previous finding (Blank and Sauer,

2004) that under aerobe and glucose excess conditions

ethanol production and the rate of TCA cycle operation

were dependent on the glucose uptake rate. As ethanol

generally cannot be quantified in shake flasks, the find-

ing relied only on indirect observations from 13C-tracer

metabolic flux analyses. Here we aimed to directly quan-

tify the TCA cycle flux by closing the carbon balance

using the BluesSens sensors for quantification of the

volatile fermentation products ethanol and CO2.

As can be seen in figure 1, the new setup delivered fer-

mentation data of very high quality (lines represent a

simultaneous fit of the experimental data using an expo-

nential growth model). As contribution to the scientific

discussion, a strong negative correlation between glu-

cose uptake rate and the rate of TCA cycle operation

could be communicated (Heyland et al., 2009). The

BlueSens setup was invaluable for the here presented

quantitative physiology project with baker’s yeast. Since

then, numerous co-workers used the setup and

experienced a tremendous increase in data amount and

more importantly in quality.

Monitoring of baker’s yeast fermentationsby PD Dr.-Ing. Lars M. Blank, Laboratory of Chemical Biotechnology, Technical University Dortmund PD Dr.-Ing. Lars M. Blank, Laboratory of Chemical Biotech-

nology, Technical University Dortmund. The group Systems Biotechnology characterizes, designs and constructs met-abolic networks.

www.bci.tu-dortmund.de/bt

i

Shake flasks equipped with CO2, O2 and ethanol sensor in a waterbath shaker

Application Report


Figure 1. Fermentation course of S. cerevisiae during respiro-fermentative growth. (a) CO2 and gaseous ethanol concentrations were monitored in the gas

phase using infrared sensors. (b) Biomass plotted vs. CO2 and gaseous ethanol concentrations. (c) Concentrations of glucose, ethanol, glycerol, and acetate

were quantified by UV-RI-HPLC. (d) Biomass plotted vs. concentrations of glucose, ethanol, glycerol and acetate. Lines represent a best fit of all experimental data using an exponential growth model or by linear fit implemented in the Sigma Plot statistic module during exponential growth until 10 h. Linear fitting for gaseous CO

2 and Ethanol was only conducted until 9 h.

LiteratureBlank, L. M. and U. Sauer, TCA cycle activity in Saccharomyces cerevisiae is a function of the environmentally determined specific growth and glucose uptake rate, Microbiol. 2004 150: 1085-1093Heyland J., J. Fu, and L. M. Blank, Correlation between TCA cycle flux and glucose uptake rate during respiro-fermentative growth of Saccharomyces cerevisiae, Microbiology, 2009, 155: 3827-3837

0 1 2 3 4 5 6 7 8 9 10 110

5

10

15

20

25

30

35

40

0

2

4

6

8

10

12

0 1 2 3 4 5 6 7 8 9 10 110

20

40

60

80

100

120

140

0

2

4

6

8

10

12

14

0 2 4 6 8 10 120

20

40

60

80

100

120

140

0

2

4

6

8

10

12

14

0 2 4 6 8 10 120

5

10

15

20

25

30

0.0

0.4

0.8

1.2

1.6

2.0

Eth

anol

[Vol

-%]

CO

2 [Vol

-%]

CO2

OD

600 [-

], E

than

ol [V

ol-%

]

CO

2 [Vol

-%]

t [h]

Biomass Ethanol

(a)

Glucose Ethanol

Gly

cero

l and

ace

tate

[mM

]

Glu

cose

and

eth

anol

[mM

]

t [h]

(c)

Glycerol Acetate

(b)

Glucose Ethanol

Gly

cero

l and

ace

tate

[mM

]

Glu

cose

and

eth

anol

[mM

] Glycerol Acetate

(d)

Biomass [OD600]

CO2

Biomass [OD600]

Ethanol

Fig. 1. Fermentation course of S. cerevisiae during respiro-fermentative growth. (a) CO2 and gaseous ethanol concentrations were monitored in the gas phase using infrared sensors. (b) Biomass plotted vs. CO2 and gaseous ethanol concentrations. (c) Concentrations of glucose, ethanol, glycerol, and acetate were quantified by UV-RI-HPLC. (d) Biomass plotted vs. concentrations of glucose, ethanol, glycerol and acetate. Lines represent a best fit of all experimental data using an exponential growth model or by linear fit implemented in the Sigma Plot statistic module during exponential growth until 10 h. Linear fitting for gaseous CO2 and Ethanol was only conducted until 9 h.

LiteratureBlank, L. M. and U. Sauer, TCA cycle activity in Saccharomyces cerevisiae is a function of the environmentally determined specific growth and glucose uptake rate, Microbiol. 2004 150: 1085-1093Heyland J., J. Fu, and L. M. Blank, Correlation between TCA cycle flux and glucose uptake rate during respiro-fermentative growth of Saccharomyces cerevisiae, Microbiology, 2009, 155: 3827-3837

Application Report


1. Machinery assemblyThe cultivation system consisted of a fully equipped 2 l

Biostat B (Sartorius, Göttingen) bioreactor installed on a

balance. The BlueSens sensors were installed in series,

where the first one was directly connected to the

exhaust gas filter with a silicon tube. The gassing rate

through the measuring volumes was 3.9 l/h. The adjust-

ment of the sensors was performed under process con-

ditions, so that the initial volume fractions were 20.957

Vol. % O2 and 0.04 Vol. % CO2. Both sensors contained

a pre-installed internal noise filter to prevent high noise

levels. The sensors have not been disconnected from the

Application of BlueSens® Gas Analyzers in a Cell Culture Processby Mathias Aehle, Center for Bioprocess Engineering, Martin-Luther-University Halle-Wittenberg, Halle (Saale)

1 4 5 632

1: Quadrupole mass spectrometer, 2: Needle valve to MS, 3: 2-way valve, 4: Inlet gas line, 5: Exhaust gas line, 6: O2 and CO

2 BlueSens sensors

Application Report


current supply during the entire study. A brief overview of

the features for the used BlueSens sensors is depicted

in table 1.

The already installed quadrupole mass spectrometer

(QMA 200, Balzers, Lichtenstein) was calibrated with

test gas (3 Vol.% CO2, 97 Vol.% N2). The gas flow to the

mass spectrometer in all experiments was adjusted to

2.1 l/h by means of a needle valve. In order to increase

accuracy, the volume fractions were additionally

measured in the gas supply line. For that purpose a

2-way valve was installed to periodically multiplex

between input and output and gauge gas measurements.

The volume fractions of the gases from both measure-

ment devices (BlueSens & mass spectrometer) were

recorded simultaneously in a Siemens SIMATIC PCS7

system and used to calculate the oxygen uptake rate

(OUR) and carbon dioxide production rate (CPR).

2. ExperimentalExperiments determining the response times at the gas

flow rate used in fermentation (3.9 l/h) were performed

with the above mentioned test gas and normal air. The

change of the volume fractions was recorded equidis-

tant (1 s) to determine characteristic time constants

(Td, T95). The experiments were performed separately

for each measurement device.

For fermentation, a serum-free suspension-CHO-cell-line

was used as the host cell system. The process was oper-

ated as a glutamine-limited fed-batch with a starting

volume of 0.8 l and exponential feeding. Further details

of the process conditions can be found in Aehle et al.

(2010) . The cultivations S687 and S693 were inocu-

lated with 4.5·105 cells/ml whereas S691 and S695

were inoculated with 5.4·105cells/ml, respectively. Dur-

ing exponential growth with a specific growth rate of

0.02h-1 the final viable cell concentrations at t = 132h

reached 4.48±0.23·106 cells/ml. Stimulus-response

experiments in real fermentation were performed by

manipulating the glutamine feed rate.

OUR and CPR were calculated as follows:

3. Results and Conclusions 3.1 Stimulus-response to changing gas

compositionsThe responses of both volume fractions recorded during

the measurements are shown in figure 1.

Sensor ID 14031 14026

Gas O2 CO2

Measuring principle

Zirconium oxide

Infrared: Two wavelengths

Concentration range

0.1-25 Vol.% 0-10 Vol.%

Resolution 0.01 Vol.% 0.01 Vol.%

Accuracy < 0.2% MR ±3% Display < 0.2% MR ±3% Display

Measurement chamber volume

35ml 35ml

2

4.48±0.23·106cells/mL. Stimulus-response experiments in real fermentation were performed

by manipulating the glutamine feed rate.


fac100

)CC(

WTRVMp

hkgmgOUR

outO

ininO 22O2

fac100

)CC(

WTRVMp

hkgmgCPR

outininCO 2CO2CO2

,

with %volC ,

hLV , W = [kg], p = [bar],

molg0.32M

2O ,

molg0.44M

2CO ,

molL4.22Vm ,

KmolLbar08314.0

KmolJ314.8R ,

gmg1000fac and

outin VV .

3 Results and Conclusions

3.1 Stimulus-response to changing gas compositions The responses of both volume fractions recorded during the measurements are shown in Fig.

1.

0 200 400 600 800 1000 1200 1400 16000

4

8

12

16

20

24

28

Time [s]

O2 [v

ol%

]

BlueSensMS

0 200 400 600 800 1000 1200 1400 16000

0.5

1

1.5

2

2.5

3

3.5

4

Time [s]

CO

2 [vol

%]

BlueSensMS

Figure 1: Volume fractions of O2 and CO2 to changing gas composition measured by MS and BlueSens (Gas 1: 3 vol% CO2, 97 vol% N2; Gas 2: air, Gas flow rate through BlueSens sensors: 3.9 L/min)

As can be clearly seen, the mass spectrometer employed here reacts faster. The response times

of the MS signals depend on the gas flow rates to the MS-inlet adjusted by the needle valve.

The higher this rate the lower the response times and vice versa. In animal cell bioreactors,

however, the gas flow rates through the reactor is rather low, often lower than it would be

2




fac100

)CC(

WTRVMp

hkgmgOUR

outO

ininO 22O2

fac100

)CC(

WTRVMp

hkgmgCPR

outininCO 2CO2CO2

,

with %volC ,

hLV , W = [kg], p = [bar],

molg0.32M

2O ,

molg0.44M

2CO ,

molL4.22Vm ,

KmolLbar08314.0

KmolJ314.8R ,

gmg1000fac and

outin VV .



1.

0 200 400 600 800 1000 1200 1400 16000

4

8

12

16

20

24

28

Time [s]

O2 [v

ol%

]

BlueSensMS

0 200 400 600 800 1000 1200 1400 16000

0.5

1

1.5

2

2.5

3

3.5

4

Time [s]

CO

2 [vol

%]

BlueSensMS






2




fac100

)CC(

WTRVMp

hkgmgOUR

outO

ininO 22O2

fac100

)CC(

WTRVMp

hkgmgCPR

outininCO 2CO2CO2

,

with %volC ,

hLV , W = [kg], p = [bar],

molg0.32M

2O ,

molg0.44M

2CO ,

molL4.22Vm ,

KmolLbar08314.0

KmolJ314.8R ,

gmg1000fac and

outin VV .



1.

0 200 400 600 800 1000 1200 1400 16000

4

8

12

16

20

24

28

Time [s]

O2 [v

ol%

]

BlueSensMS

0 200 400 600 800 1000 1200 1400 16000

0.5

1

1.5

2

2.5

3

3.5

4

Time [s]

CO

2 [vol

%]

BlueSensMS






2




fac100

)CC(

WTRVMp

hkgmgOUR

outO

ininO 22O2

fac100

)CC(

WTRVMp

hkgmgCPR

outininCO 2CO2CO2

,

with %volC ,

hLV , W = [kg], p = [bar],

molg0.32M

2O ,

molg0.44M

2CO ,

molL4.22Vm ,

KmolLbar08314.0

KmolJ314.8R ,

gmg1000fac and

outin VV .



1.

0 200 400 600 800 1000 1200 1400 16000

4

8

12

16

20

24

28

Time [s]

O2 [v

ol%

]

BlueSensMS

0 200 400 600 800 1000 1200 1400 16000

0.5

1

1.5

2

2.5

3

3.5

4

Time [s]

CO

2 [vol

%]

BlueSensMS






Table 1: Abstract from the data sheets for the O2 und CO

2 BlueSens sensors

Figure 1: Volume fractions of O2 and CO

2 to changing gas composition mea-

sured by MS and BlueSens (Gas 1: 3 vol% CO2, 97 vol% N

2; Gas 2: air, Gas

flow rate through BlueSens sensors: 3.9 l/min)

Application Report


As can be clearly seen, the mass spectrometer employed

here reacts faster. The response times of the MS signals

depend on the gas flow rates to the MS-inlet adjusted by

the needle valve.

The higher this rate the lower the response times and

vice versa. In animal cell bioreactors, however, the gas

flow rates through the reactor is rather low, often lower

than it would be desirable for a small time constant of

the off-gas measurement devices. The gas flow through

the measurement devices must always be lower than the

aeration rate itself. Only then, an overpressure, necessary

for sterility purposes, can be maintained in the reactor.

In the small-scale experiments reported the flow rate

into the mass spectrometer was fixed to 2.1 l/h. For

systems that are operated at higher aeration rates any-

way, such as microbial cultures, this problem does not

play any role.

The BlueSens sensors react much slower on the same

gas composition change. This is due to the slow gas

throughputs through their relatively large measuring

chambers of ca. 35ml. The smaller chambers of ca.

10 ml offered by BlueSens are recommended to shorten

the reaction time. The results of the time constants are

depicted in the following table.

Independent of the direction of the gas change there

was no significant variation of the delay times for the

respective method. The BlueSens sensors have a 4.5

times higher delay time compared to the MS.

For both methods, the change of the O2 signal is faster

than the CO2 signal. At a first glance, the T95 time con-

stants (i.e. the time needed to reach 95 % of the end

value) of both gas components measured with BlueSens

seemed to be critical. Despite of the use of a 3 Vol.%

CO2 test gas resulting in a rather high concentration dif-

ferences, these T95 values are assumed to be valid at

smaller concentration differences that are present dur-

ing cultivation as well. The maximum concentration dif-

ferences for O2 and CO2 at the end of the cultivation

performed in this study are 0.75 Vol.% and 0.6 Vol.%,

respectively. At a second glance, such extreme concen-

tration differences will not appear in real fermentations,

as the gas sensors will be simply located in the exhaust

line. Additionally, the gas flow rate to the sensors will not

be multiplexed in standard applications. Hence, those

T95 time constants will have no influence in practical

fermentation.

Corresponding stimulus-response experiments with the

highest dynamic change of respiration rates expected

will be demonstrated in chapter 3.3.

3.2 Fermentation3.2.1 Oxygen uptake rate (OUR)Typical OUR-profiles simultaneously measured with the

mass spectrometer and the BlueSens sensors are shown

in figure 2. What is easily to notice is that there is a

constant offset between the trajectories.

Device Mass spectrometer BlueSens

Analyte O2 CO2 O2 CO2

Gas change Test gas-Air

Td = 9s T95 = 17s

Td = 8s T95 = 56s

Td = 40 T95 = 225s

Td = 40s T95 = 340s

Gas change Air-Test gas

Td = 10s T95 = 10s

Td = 7s T95 = 52s

Td = 35s T95 = 249s

Td = 40s T95 = 320s

Table 2: Time constants (Td, T95) of the reaction characteristics for O2 and

CO2 measured by mass spectrometry und BlueSens analyzers

Figure 2: OUR of S687 und S691 without offset correction

3

Corresponding stimulus-response experiments with the highest dynamic change of respiration

rates expected will be demonstrated in chapter 3.3.

3.2 Fermentation

3.2.1 Oxygen uptake rate (OUR)

Typical OUR-profiles simultaneously measured with the mass spectrometer and the BlueSens

sensors are shown in Fig. 2. What is easily to notice is that there is a constant offset between

the trajectories.

0 20 40 60 80 100 120 140-10

0

10

20

30

40

50

60

Process time [h]

OU

R [m

g/L/

h]

Mass spectrometerBlueSens

S687

0 20 40 60 80 100 120 140-10

0

10

20

30

40

50

60

70

Process time [h]

OU

R [m

g/L/

h]


S691


To explain this offset one must refer to the equation for the OUR calculation. There, the inlet

(O2in) and outlet (O2

out) oxygen concentrations are incorporated. As already mentioned for the

BlueSens sensors in chapter 1, this O2in value is adjusted before inoculation to 20.957 vol.%.

Thus, this value is fixed in the OUR equation during cultivation. This could be a drawback as

the O2 volume fraction was found to untypical move away from this initial value within the

first 10 h. The courses of the volume fractions within the first hours and the initially adjusted

value (dashed black line) are depicted in Fig.3.

3

Corresponding stimulus-response experiments with the highest dynamic change of respiration

rates expected will be demonstrated in chapter 3.3.

3.2 Fermentation

3.2.1 Oxygen uptake rate (OUR)

Typical OUR-profiles simultaneously measured with the mass spectrometer and the BlueSens

sensors are shown in Fig. 2. What is easily to notice is that there is a constant offset between

the trajectories.

0 20 40 60 80 100 120 140-10

0

10

20

30

40

50

60

Process time [h]

OU

R [m

g/L/

h]


S687

0 20 40 60 80 100 120 140-10

0

10

20

30

40

50

60

70

Process time [h]

OU

R [m

g/L/

h]


S691


To explain this offset one must refer to the equation for the OUR calculation. There, the inlet

(O2in) and outlet (O2

out) oxygen concentrations are incorporated. As already mentioned for the

BlueSens sensors in chapter 1, this O2in value is adjusted before inoculation to 20.957 vol.%.

Thus, this value is fixed in the OUR equation during cultivation. This could be a drawback as

the O2 volume fraction was found to untypical move away from this initial value within the

first 10 h. The courses of the volume fractions within the first hours and the initially adjusted

value (dashed black line) are depicted in Fig.3.

Application Report


To explain this offset one must refer to the equation for

the OUR calculation. There, the inlet (O2in) and outlet

(O2out) oxygen concentrations are incorporated. As

already mentioned for the BlueSens sensors in chapter

1, this O2in value is adjusted before inoculation to

20.957 Vol.%. Thus, this value is fixed in the OUR equa-

tion during cultivation. This could be a drawback as the

O2 volume fraction was found to untypical move away

from this initial value within the first 10 h. The courses of

the volume fractions within the first hours and the ini-

tially adjusted value (dashed black line) are depicted in

figure3.

From figure 3 it turns out that the measured O2 volume

fractions move immediately to higher or lower values

from the adjusted O2 concentration after inoculation. The

consequence is a positive or negative offset with respect

to the MS-OUR which appears randomly. We are cur-

rently investigating the reasons.

To remove the offset between the BlueSens-OUR and the

MS-OUR, the difference in the OUR values was deter-

mined once ca. 10h after inoculation. This difference

was then removed by manually adjusting the O2 in value

in the BlueSens-OUR equation.

If this manual adjustment is made too early, an offset

between the OUR-values will remain.

Very good results were obtained for BlueSens-OUR com-

pared with the MS after appropriate correction. The OUR

4

0 5 10 15 20 2520.92

20.93

20.94

20.95

20.96

20.97

20.98

20.99

21

O2 [v

ol%

]

S687

0 5 10 15 20 2520.92

20.93

20.94

20.95

20.96

20.97

20.98

20.99

21

S689

0 5 10 15 20 2520.94

20.95

20.96

20.97

20.98

20.99

21

21.01

21.02

21.03

S691

0 5 10 15 20 2520.89

20.9

20.91

20.92

20.93

20.94

20.95

20.96

20.97

20.98

S693

0 5 10 15 20 2520.86

20.88

20.9

20.92

20.94

20.96

20.98

Process time [h]

S695

Figure 3: BlueSens O2 signals during the first 20h from 5 CHO fed-batch fermentations. The dashed line depicts the adjusted O2 value prior to inoculation

From Fig. 3 it turns out that the measured O2 volume fractions move immediately to higher or

lower values from the adjusted O2 concentration after inoculation. The consequence is a

positive or negative offset with respect to the MS-OUR which appears randomly. We are

currently investigating the reasons.

To remove the offset between the BlueSens-OUR and the MS-OUR, the difference in the

OUR values was determined once ca. 10h after inoculation. This difference was then removed

by manually adjusting the O2in value in the BlueSens-OUR equation.

If this manual adjustment is made too early, an offset between the OUR-values will remain.

Very good results were obtained for BlueSens-OUR compared with the MS after appropriate

correction. The OUR trajectories determined by BlueSens O2 volume fractions stayed within

the MS-OUR noise and are thus nearly identical (Fig.4). The BlueSens-OUR reflected the

process dynamic accurately.

Figure 3: BlueSens O2 signals during the first

20h from 5 CHO fed-batch fermentations. The dashed line depicts the adjusted O

2 value prior

to inoculation

Dipl. Ing. Mathias Aehle, Martin-Luther-University Halle-Wittenberg, Institute of Biochemistry/Biotechnology Cen-ter for Bioprocess EngineeringCentral objective of the workgroup is the teaching and re-search in the area of biochemical engineering. In research the emphasis is put on bioprocess engineering. The design and optimization of the production processes for recombi-nant proteins, which are predominantly used for therapy or diagnostic applications, are in the focus of the group. De-velopment of improved process control strategies for in-dustrial production processes development of new meth-ods for online characterization of fermentation processes for application in process controlinvestigation of transfer processes in bioreactors in pilot and production scale.

i

Application Report


trajectories determined by BlueSens O2 volume fractions

stayed within the MS-OUR noise and are thus nearly

identical (figure4). The BlueSens-OUR reflected the pro-

cess dynamic accurately.

3.2.2 Carbon dioxide production rate (CPR)Problems with a fixed CO2in-value for calculating the

CPR from BlueSens data, as already discussed in chap-

ter 3.2.1, were not identified.

The cell growth in form of its respiration rate CPR is

exactly determined by the BlueSens CO2 volume

fractions. Even several pH- recalibrations (see S693,

t=20 – 60 h) which resulted in very fast changes of the

dissolved and consequently gaseous CO2 volume

fractions were recognized without significant delay.

Figure 5 shows the CPR profiles of MS and BlueSens.

3.2.3 Stimulus-response experimentsAs indicated in chapter 3.1 stimulus-response experi-

ments by changing the glutamine feed rate were per-

formed to investigate the significance of response times.

Clear responses in the respiration rates are expected

when changing feed rates under tight glutamine limita-

tion. Figure 6 shows the glutamine feed rate along with

the corresponding reaction of the OUR and CPR. The

expected reactions to higher and lower feed pulses were

obtained. A delay time under these conditions compared

to the MS-signal was not identified. As this experiment

described the fastest dynamic changes possible during

cultivation the rather high T95 time constants from chapter

3.1 can be seen as not relevant any more.

3.3 ConclusionsThe BlueSens system is easy plug-and-play measure-

ment system analyzing the exhaust gas composition on-

line. The implementation to the already installed reactor

configuration and process control system was done

without serious problems.

The sensors were used in several CHO fed-batch fermen-

tations. A well established mass spectrometer was used

for direct comparison and evaluation of the signals. Due

to lower aeration rates typically found for mammalian

cell culture processes, the independence of the

5

0 20 40 60 80 100 120 1400

10

20

30

40

50

60

70

OU

R [m

g/L/

h]


0 20 40 60 80 100 120 1400

10

20

30

40

50

60

70


0 20 40 60 80 100 120 1400

10

20

30

40

50

60

70

Process time [h]


0 20 40 60 80 100 120 1400

10

20

30

40

50

60

70


S687 S691

S693 S695

Figure 4: Comparison of OUR from 4 fermentations with offset correction

3.2.2 Carbon dioxide production rate (CPR)

Problems with a fixed CO2in-value for calculating the CPR from BlueSens data, as already

discussed in chapter 3.2.1, were not identified.

The cell growth in form of its respiration rate CPR is exactly determined by the BlueSens CO2

volume fractions. Even several pH- recalibrations (see S693, t=20 – 60 h) which resulted in

very fast changes of the dissolved and consequently gaseous CO2 volume fractions were

recognized without significant delay. Fig. 4 shows the CPR profiles of MS and BlueSens.

Figure 4: Comparison of OUR from 4 fermentations with offset correction

Application Report


BlueSens sensors to the volumetric gas flow rate is quite

advantageous. Initial concerns about the behavior at low

aeration rates and thus too long delay times were rela-

tivized after the first fermentation.

Despite of a low aeration rate (3.9 l/h) used in fermen-

tation experiments, a rather low resolution of 0.01 Vol. %

and the factory-made calibration in a wide concentra-

tion rate (0-25 Vol.% for O2, 0-10 vol % for CO2) the

BlueSens sensors performed surprisingly good, respec-

tively. The maximal volume fraction differences at the

end of the process were 0.75 Vol. % for O2 and 0.6 Vol.

% for CO2. Hence, it should be explicitly stressed that the

CO2 measurement was always performed at the lowest

measurement range. The signals, nevertheless, mirrored

the process dynamics without any problems.

In first experiments, a constant offset between MS and

BlueSens derived OUR data was obtained. A first analysis

revealed the influence of fixing the O2in value which has

6

0 20 40 60 80 100 120 1400

20

40

60

80

100

CPR

[mg/

L/h]


0 20 40 60 80 100 120 1400

20

40

60

80

100


0 20 40 60 80 100 120 1400

20

40

60

80

100


0 20 40 60 80 100 120 1400

20

40

60

80

100

Process time [h]


S687 S691

S693 S695

Figure 5: Comparison of CPR from 4 fermentations.

3.2.3 Stimulus-response experiments

As indicated in chapter 3.1 stimulus-response experiments by changing the glutamine feed

rate were performed to investigate the significance of response times. Clear responses in the

respiration rates are expected when changing feed rates under tight glutamine limitation. Fig.

6 shows the glutamine feed rate along with the corresponding reaction of the OUR and CPR.

The expected reactions to higher and lower feed pulses were obtained. A delay time under

these conditions compared to the MS-signal was not identified. As this experiment described

the fastest dynamic changes possible during cultivation the rather high T95 time constants

from chapter 3.1 can be seen as not relevant any more.

40 60 80 100 120 1400

10

20

30

40

50

60

70

80

OUR

[mg/

L/h]

Process Time [h]

40 60 80 100 120 1400

1

2

3

4

5

6

7

8

Glu

tam

ine

Feed

Rat

e [g

/h]


40 60 80 100 120 1400

10

20

30

40

50

60

70

80

CPR

[mg/

L/h]

Process Time [h]

40 60 80 100 120 1400

1

2

3

4

5

6

7

8

Glu

tam

ine

Feed

Rat

e [g

/h]


Figure 6: Comparison of OUR and CPR reactions to glutamine feed rate pulses

6

0 20 40 60 80 100 120 1400

20

40

60

80

100

CPR

[mg/

L/h]


0 20 40 60 80 100 120 1400

20

40

60

80

100


0 20 40 60 80 100 120 1400

20

40

60

80

100


0 20 40 60 80 100 120 1400

20

40

60

80

100

Process time [h]


S687 S691

S693 S695


3.2.3 Stimulus-response experiments

As indicated in chapter 3.1 stimulus-response experiments by changing the glutamine feed

rate were performed to investigate the significance of response times. Clear responses in the

respiration rates are expected when changing feed rates under tight glutamine limitation. Fig.

6 shows the glutamine feed rate along with the corresponding reaction of the OUR and CPR.

The expected reactions to higher and lower feed pulses were obtained. A delay time under

these conditions compared to the MS-signal was not identified. As this experiment described

the fastest dynamic changes possible during cultivation the rather high T95 time constants

from chapter 3.1 can be seen as not relevant any more.

40 60 80 100 120 1400

10

20

30

40

50

60

70

80

OUR

[mg/

L/h]

Process Time [h]

40 60 80 100 120 1400

1

2

3

4

5

6

7

8

Glu

tam

ine

Feed

Rat

e [g

/h]


40 60 80 100 120 1400

10

20

30

40

50

60

70

80

CPR

[mg/

L/h]

Process Time [h]

40 60 80 100 120 1400

1

2

3

4

5

6

7

8

Glu

tam

ine

Feed

Rat

e [g

/h]





Application Report


to be adjusted prior to inoculation for calculating

BlueSens-OUR. As the O2 volume fraction arbitrary

moved away from the fixed O2in value short after inocu-

lation a constant positive or negative offset occurs

resulting a time shift of the OUR profile. This effect of the

O2 values at start of the fermentation is not yet fully

understood. Further investigations should be made to

get consistent and reproducible OUR data. The usage of

additional BlueSens sensors at the inlet gas line or the

insertion of a multiplexing device to measure the incom-

ing volume fractions as well would be helpful to over-

come this problem. Nevertheless, after manual adaption

of the O2 in value excellent conformity to the MS-OUR

was obtained.

CPR data from CO2 volume fraction measurements

showed very good results as well.

Neither time delays nor significant loss of information

occurred during stimulus-response experiments in real

fed-batch fermentation.

To sum it up, the BlueSens system is suitable for exhaust

gas analysis under cell culture conditions. It is a cost

effective alternative to established mass spectrometers

often used for cell culture off-gas monitoring. The offset

problems could be more discussed when further appli-

cation reports of cell culture processes are available.

The sensors of BlueSens dispose of universal possibilities

of installation. By its multifunctional connections each

sensor can be integrated in nearly every existent system.

So the measuring instrument can be installed easily and

cost-effectively. Existing installations can be upgraded

with the products of BlueSens without any problem. In

general you have the choice between the use of flow

adapters or to use already existing screwed connections.

Then the connection can be realized by the following

accesses:

any hose connector from 4-12 mm >

GL45 screw thread >

1 ¼“ screw thread >

Tri-Clamp >

For the use of flow adapters you can make your choice

between the reasonably priced and robust POM-adapters

or the high-quality stainless steel adapters. Then the gas

flow to/in flow adapters is simply achieved via hose

connections.

Information

Connections for every application

Information


Sensors in PA housing with GL45 screwed connection on shake flask Flow adapter POM with GL45 and plug connections for hoses

Aluminum housing with flow adapter stainless steel

Tube with screwed connection 1 ¼“

Aluminum housing with flow adapter stainless steel

Tri-Clamp

Information


Sensor CO2 CH4 CO EtOH EtOH O2 O2ec H2 Sensor

Measuring range 0 … 10 Vol. % 0 … 100 Vol. % 0 … 30 Vol. % 0.2 … 25 Vol. % 0 … 1 Vol. % 0.1 … 25 Vol. % 0 … 100 Vol. % 0 … 100 Vol. %3 Measuring range

0 … 25 Vol. % 0 … 100 Vol. % 1 … 50 Vol. %

0 … 50 Vol. %1

Measuring Principle Infrared, dual wavelength Infrared, dual wavelength ZrO2 Galvanic cell Thermal conductivity Measuring Principle

Accuracy < ± 0.2 % FS* ± 3% reading < ± 0.2 % FS* ± 3% reading Accuracy

Long-term stability2 < ± 2% reading / year < ± 2% reading / year Long-term stability2

Lifetime sensor element > 3 years > 3 years 15,000 hours Approximately 900 000 Vol. h operating hours

> 3 years Lifetime sensor element

Housing Aluminum, IP 65Dimension (WxDxH) inch Weight lb

3.94 x 3.94 x 4.06 1.65 3.94 x 3.94 x 4.061.65

3.94 x 3.94 x 4.061.65

3.94 x 3.94 x 4.061.65

11,42 x 3,94 x 2,366.61

3.94 x 3.94 x 4.061.65

3.94 x 3.94 x 5.441.70

3.94 x 3.94 x 5.441.70


Housing PA6Dimension (DxH) inch Weight lb

3.15 x 3.940.66

3.15 x 3.940.66

3.15 x 3.940.66

3.15 x 3.940.66

Not available 3.15 x 3.940.66

3.15 x 5.320.70

Not availableHousing PA6Dimension (DxH) inch Weight lb

Disconnectable Measuring cap

possible possible possible possible No possible possible No Disconnectable Measuring cap

Connecting tolerance < ± 0.2 % FS* ± 3% reading – – – – Connecting tolerance

Material in contact with gas

Steel 1.4571 / Sapphire / Viton / PTFE Steel 1.4571 / Sapphire / Viton / PTFE

Steel 1.4571 / Viton / PTFE Stainless steel, Si, SiOxNy, gold,epoxyAcrylnitril-butadien-rubber, Viton


Connection** G 1¼”, GL 45, Tri-Clamp, hose connection 4-12mm etc. Connector for 6mm hose and 8mm tube

G 1¼”, GL 45, Tri-Clamp, hose connection 4-12mm etc. Connection**

General General

Operating temperature max -25 – 55 °C / -13 – 131 °F ** max -25 – 55 °C / -13 – 131 °F ** Operating temperature

Storage temperature 0 – 60 °C / 32 – 140 °F / 75% RH non-condensing 0 – 60 °C / 32 – 140 °F / 75% RH non-condensing Storage temperature

Pressure range (absolute):

0,8 – 1,3 bar / 11.6 – 18.85 psi** 0,8 – 1,3 bar / 11.6 – 18.85 psi** Pressure range (absolute):

Pressure dependence compensated: < ± 3 % reading (range) compensated: < ± 3 % reading (range) Pressure dependence

Operating humidity 0 ... 100% RF 0 ... 100% RF Operating humidity

Power supply (max.) 12 or 24 VDC, 1 A 12 or 24 VDC, 1 A 24 VDC, 1 A Power supply (max.)

Output RS 232, RS 485, 4 – 20 mA, Ethernet RS 232, RS 485, 4 – 20 mA, Ethernet Output

Maintenance once a month

1-point calibration with ambient air or nitrogen 1-point calibration with ambient air or nitrogen Maintenance once a month

Maintenance yearly optional factory calibration with certified gases optional factory calibration with certified gases Maintenance yearly

CE EN61326-1:1997 +A2:1998 EN61326-1:1997 +A2:1998 CE 1 accuracy < ± .0.5 % FS* ± 5% reading 2 with monthly 1-point calibration *full scale ** others on request

Sensors overviewThe BCP series’ exceedingly robust and reasonably

priced sensors can be easily integrated directly into the

gas lines independent of the gas flow. Additional gas

coolers,pumps and valves are not needed to make the

measurements.

Information


Sensor CO2 CH4 CO EtOH EtOH O2 O2ec H2 Sensor

Measuring range 0 … 10 Vol. % 0 … 100 Vol. % 0 … 30 Vol. % 0.2 … 25 Vol. % 0 … 1 Vol. % 0.1 … 25 Vol. % 0 … 100 Vol. % 0 … 100 Vol. %3 Measuring range

0 … 25 Vol. % 0 … 100 Vol. % 1 … 50 Vol. %

0 … 50 Vol. %1

Measuring Principle Infrared, dual wavelength Infrared, dual wavelength ZrO2 Galvanic cell Thermal conductivity Measuring Principle

Accuracy < ± 0.2 % FS* ± 3% reading < ± 0.2 % FS* ± 3% reading Accuracy

Long-term stability2 < ± 2% reading / year < ± 2% reading / year Long-term stability2

Lifetime sensor element > 3 years > 3 years 15,000 hours Approximately 900 000 Vol. h operating hours

> 3 years Lifetime sensor element


3.94 x 3.94 x 4.06 1.65 3.94 x 3.94 x 4.061.65

3.94 x 3.94 x 4.061.65

3.94 x 3.94 x 4.061.65

11,42 x 3,94 x 2,366.61

3.94 x 3.94 x 4.061.65

3.94 x 3.94 x 5.441.70

3.94 x 3.94 x 5.441.70


Housing PA6Dimension (DxH) inch Weight lb

3.15 x 3.940.66

3.15 x 3.940.66

3.15 x 3.940.66

3.15 x 3.940.66

Not available 3.15 x 3.940.66

3.15 x 5.320.70

Not availableHousing PA6Dimension (DxH) inch Weight lb

Disconnectable Measuring cap

possible possible possible possible No possible possible No Disconnectable Measuring cap

Connecting tolerance < ± 0.2 % FS* ± 3% reading – – – – Connecting tolerance


Steel 1.4571 / Sapphire / Viton / PTFE Steel 1.4571 / Sapphire / Viton / PTFE

Steel 1.4571 / Viton / PTFE Stainless steel, Si, SiOxNy, gold,epoxyAcrylnitril-butadien-rubber, Viton


Connection** G 1¼”, GL 45, Tri-Clamp, hose connection 4-12mm etc. Connector for 6mm hose and 8mm tube

G 1¼”, GL 45, Tri-Clamp, hose connection 4-12mm etc. Connection**

General General

Operating temperature max -25 – 55 °C / -13 – 131 °F ** max -25 – 55 °C / -13 – 131 °F ** Operating temperature

Storage temperature 0 – 60 °C / 32 – 140 °F / 75% RH non-condensing 0 – 60 °C / 32 – 140 °F / 75% RH non-condensing Storage temperature

Pressure range (absolute):

0,8 – 1,3 bar / 11.6 – 18.85 psi** 0,8 – 1,3 bar / 11.6 – 18.85 psi** Pressure range (absolute):

Pressure dependence compensated: < ± 3 % reading (range) compensated: < ± 3 % reading (range) Pressure dependence

Operating humidity 0 ... 100% RF 0 ... 100% RF Operating humidity

Power supply (max.) 12 or 24 VDC, 1 A 12 or 24 VDC, 1 A 24 VDC, 1 A Power supply (max.)

Output RS 232, RS 485, 4 – 20 mA, Ethernet RS 232, RS 485, 4 – 20 mA, Ethernet Output

Maintenance once a month

1-point calibration with ambient air or nitrogen 1-point calibration with ambient air or nitrogen Maintenance once a month

Maintenance yearly optional factory calibration with certified gases optional factory calibration with certified gases Maintenance yearly

CE EN61326-1:1997 +A2:1998 EN61326-1:1997 +A2:1998 CE 1 accuracy < ± .0.5 % FS* ± 5% reading 2 with monthly 1-point calibration 3 binary mixture *full scale ** others on request

The sensors measure at the point where things are

happening. Fast and reliable measurement data without

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.

Information


We help you understand, control and optimize your process!

Simple tool for process optimization

The same sensors are also used in the BCpreFerm

system, which is used for process optimization (scale

up) for flasks up to large-scale fermenters. The system

comprises up to 12 sensors that are linked to a com-

puter via an electronic multiplexer. The related software

visualizes the results and can calculate

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 the scale up >

Measure the gas yield and quality in every anaerobic process

The unique structure of the CH4 sensors from BlueSens

facilitate measuring methane concentrations in processes

that sometimes produce much, sometimes little gas. The

use of sample taking is impossible there, so conven-

tional systems fail.

The CH4 sensors are simply screwed onto the fermenta-

tion container and measure the methane content directly

over the sample. Even at 55 °C (131° F) in water-satu-

rated atmospheres. The accruing volumes are precisely

registered via a precision volumenometer (Milligas-

counter®*).

The data are registered online with the corresponding

software and visualized on the computer. Optionally,

BlueSens can provide everything as accessories; from

the stirrer through the incubator.

Additional sensorsTo cover as many measurement parameters as possible,

BlueSens also offers sensors for Ethanol (C2H6O), Hydro-

gen (H2) and Carbon monoxide (CO).

preFermBC

MasterYield

* Registered trademark. The MilliGascounter was developed at the University of Applied Science Hamburg under the leadership of Prof. Dr. Paul Scherer.

Information


The use of the conductible FermVis software is obvious

for the parallel measurement of CO2 and O2. Oxygen or

substrate limitations can be detected along with meta-

bolic transpositions.

Furthermore, a time specific analysis of the respective

products is made possible. For improved comparability,

the BCpreFerm measurement system can be used for

shake flasks and fermenters. FermVis calculates the

oxygen uptake rate (OUR), the carbon dioxide emission

rate (CER) and the respiratory quotient (RC) for fermenters

as well as for shake flasks.

The software BacVis was made for data

recording of different sensors and gas flow

meters (Milligascounter®*). The sensors are

recognized automatically by means of their

identification number. Due to the easy

handling, BacVis is self-explanatory. As the

obtained data are recorded in the ASCI-

format, you can process them without any

problems.

For sure you have also the option to use your own

software for your process control. We are pleased to

support you in finding the best solution for your plants.

BlueSens sensors can be used nearly

everywhere. Both screwed and clamped connections

and the standardized data transfer allow the integration

in nearly every biotechnical plant. You are also free in

the software choice for the process control.

* Registered trademark. The MilliGascounter was developed at the University of Applied Science Hamburg under the leadership of Prof. Dr. Paul Scherer.

The freedom of software choice

FermVis

BACVis

Information


The modern in-situ measurement on parallel

bioreactors offers various advantages compared to the

conventional method with just one central gas analyzer.

The parallel measurement of gas concentration directly

in every single fermenter saves the installation of com-

plicated gas lines to a central analyzer and also the

complicated processing of the gases can be left out. The

identical test preparation in several fermenters reduces

the danger to work with incorrect results.

You rely not just on one analyzer, but on many, indepen-

dently working sensors. Furthermore, contamination

between the particular bioreactors can nearly be excluded.

Acc. to PAT, every single fermenter disposes of an own

sensor which transfers continuous real time data to

control the process. The decisive process parameters

can be recognized and influenced in time. This is a real

advantage in bioprocessing.

Such a continuous data stream can‘t be produced by

means of the conventional measuring method.

The central analyzers are mostly extremely cost-intensive

to purchase and maintain. Often the entire production

process is on hold, if a component has to be changed or

maintained.

With the use of many, decentral sensors this problem

does mostly not come up. If a fermenter is turned off

due to maintenance, the remaining bioreactors can con-

tinue production without any problems.

With the use of parallel systems you mostly achieve

much faster results in research. Under identical terms of

cultivation, alternatives can be tested well-aimed in the

particular bioreactors and therefore the decisive factors

can be determined much faster (DOE).

Parallel systemsMeasuring according to PAT

Questions?Please ask directly!

Phone

+49 2366 305 301Or visit our homepage:

www.BlueSens.com

Konrad-Adenauer-Str. 9-13 • D-45699 Herten (Germany)Phone +49 - (0)2366 - 305-301 • Fax +49 - (0)2366 - 305-300e-mail: [email protected]

Internet:www.BlueSens.dewww.BlueSens.com

bluesens report no.1

Documents