Standard Bioreactor Sensors 13221 Introduction

Temperature, pH and pO2 sensors (or 'probes') are essentially always present on bioreactors for monitoring and control purposes. They are installed in situ and are steam-sterilizable in place. They are covered with a housing that provides pressure balance during sterilization and other pressurization operations, as well as protection from contamination. Additionally, pilot- and production-scale bioreactors are commonly equipped with on-line devices to measure the liquid level in, or weight of, the bioreactor and the pressure in the headspace. The use of an in situ probe for biomass is also becoming common, at least as a complement to off-line analysis (see Section 13.2.3). Figure 13.2 shows a simplified instrumentation diagram of a pilot- or large-scale bioreac-tor operated in batch mode, with the standard sensors. Their characteristics are discussed below. Sensors to measure the flow rate of supplied media and gases are not used as independent monitoring devices but are embedded in a combined monitoring and controlling device; they are discussed at the process-control level (Section 13.3.2).

Bioreactor Sensors

Figure 13.2 Simplified instrumentation diagram of a pilot- or large-scale bioreactor operated in batch mode, with the standard sensors used during cell culture; additional probes (e.g. temperature) and valves used during operations such as cleaning- and sterilization-in-place are not represented. The main control loops (dotted lines) are also shown, in a schematic way (see Section 13.3.2 for details). C: controller; F: flow rate; I: indicator; M: motor; P: pressure; T: temperature; X: biomass; W: liquid weight; '1', '2': indicate probes in duplicate (1 main probe and 1 back-up; see Sections 13.2.2.2-13.2.2.4).

Figure 13.2 Simplified instrumentation diagram of a pilot- or large-scale bioreactor operated in batch mode, with the standard sensors used during cell culture; additional probes (e.g. temperature) and valves used during operations such as cleaning- and sterilization-in-place are not represented. The main control loops (dotted lines) are also shown, in a schematic way (see Section 13.3.2 for details). C: controller; F: flow rate; I: indicator; M: motor; P: pressure; T: temperature; X: biomass; W: liquid weight; '1', '2': indicate probes in duplicate (1 main probe and 1 back-up; see Sections 13.2.2.2-13.2.2.4).

13.2.2.2 Temperature

Animal cells are very sensitive to temperature; this parameter must therefore be monitored and controlled accurately, i.e. within 0.5 °C or less around the setpoint. Temperature probes should be steam-sterilizable in place and stable over several weeks, so that they can be used for in situ measurement in fed-batch and perfusion cultures. In pilot- or large-scale bioreac-tors, duplicate probes are commonly used. Additional temperature probes are also mounted at various points of the tank and associated pipework to monitor sterilization (see Chapter 14, Section 14.4). There are two main types of temperature probe used for bioreactors (Hartnett 1994):

• Resistance temperature devices (RTDs): these devices are characterized by a high accuracy and stability. They have a typical time constant of 5 seconds, which is adequate for animal cell cultures. RTDs are based on the measurement of the electrical resistance of a metal, which is known to change with temperature. The most commonly used metal for RTDs is platinum; a classical device is the Pt-100 sensor, referring to a resistance of 100 Q at 0 °C. Other common metals are nickel, nickel alloys and copper. The non-linear relationship between electrical resistance and temperature requires a careful calibration of the RTD.

• Thermocouples: these are considered to be excellent low-cost alternatives to RTDs; their accuracy can actually be equivalent to the latter. While RTDs are preferred for critical measurements in a bioreactor during a culture, thermocouples are typically used for utilities and sterilization. Thermocouples consist of a closed electrical circuit made of two different metals; if there is a temperature difference between the two junctions, an electrical current is generated. One junction is placed in the sensing region and the other, considered as the reference, in an environment where the temperature is either carefully controlled (e.g. in an ice bath) or measured. The temperature at the location of the sensing element is then determined on the basis of the measured current and of the temperature of the reference junction. Here again, the non-linear relationship between current and temperature requires careful calibration of the thermocouple.

Both probe types are equipped with mechanical protection devices called thermowells, which separate the sensor from the process environment. Several standards from various organizations such as the American Society for Testing and Materials (ASTM), the Bureau International des Poids et Mesures (BIPM), the Instrument Society of America (ISA) and the US National Bureau of Standards (NBS) have been defined for both probe types (see, for instance, Kennedy 1983; Mangum 1989).

Other temperature-measuring devices that can be used in upstream processing but are less accurate or less suitable for control purposes, include:

• thermistors, consisting of a thermally sensitive resistor made of a metal oxide;

• gas- or liquid-filled thermometers, where the fluid expands with increasing temperature, thus producing a mechanical movement proportional to temperature;

• bimetal thermometers, made of two metal strips with different coefficients of thermal expansion.

pH is routinely measured and controlled in bioreactors, since cells are very sensitive to pH changes. The most common type for bioreactors is a potentiometric steam-sterilizable glass probe filled with liquid or gel electrolyte, combining, in one unit, a reference electrode and a pH-sensing electrode (Hartnett 1994). The electrical signal is temperature-dependent and thus requires a temperature probe for correction or calibration at the working temperature. A three-point calibration is typically performed: one standard for the adjustment of the transmitter output, typically at pH 7, a second one for the slope adjustment, and a third one for a check on linearity. A typical problem with pH probes is the drift of the signal due to fouling of the reference electrode diaphragm with components of the cell culture fluid. Recalibration is thus often required during the culture. This can be performed by adjusting the reading on the basis of an off-line measurement (one-point calibration). An additional common safety measure at pilot- and large-scale production is to use probes in duplicate (Figure 13.2). If the first (main) probe fails, i.e. cannot be recalibrated, for instance because of excessive fouling, it is replaced by the second one to provide the input to the controller. An alternative is to use an interchangeable probe device (for instance the InTrac system of Mettler Toledo), so that the probe can be removed during a culture and either properly recalibrated off line with a two-point method, or serviced. At the end of the culture, chemical and enzymatic cleaning of the diaphragm, and replacement of the electrolyte, is recommended. Sensors based on pH-sensitive dyes and field-effect transistors have recently been developed, but the technology is not yet mature enough for routine use in bioprocessing (Sonnleitner 1999).

13.2.2.4 Dissolved oxygen (pO2)

The measurement and control of dissolved oxygen during cell cultivation is essential, particularly at high cell density. This measurement is also widely used during the design and scale-up of bio-reactors, to determine the efficiency of the gassing system.

The most common sensors are in situ sterilizable electrodes, of either galvanic (potentiometric) or, more often, polarographic (amperometric or Clark) type (Bailey & Ollis 1986). An oxygen-permeable membrane separates the electrode from the cell culture fluid. For both types, the reaction at the cathode (typically platinum) is the reduction of oxygen according to the following reaction:

At the anode of the galvanic type, the reaction is:

The resulting current provides a voltage, which is measured.

In the polarographic probe, a constant voltage is applied between the cathode and anode; the reaction at the anode is:

The resulting current is measured.

In both electrode types, the measured electrical signal is, at steady state, proportional to the oxygen flux to the cathode, which, in turn, is proportional to the oxygen partial pressure in the liquid phase. A two-point calibration is usually performed in situ after sterilization but before inoculation, in nitrogen- and then air-saturated culture medium. The main drawback of these electrodes is the drift of the signal, due to the accumulation of hydroxyl or metal ions, the depletion of chloride ions and the external fouling of the membrane surface. Frequent replacement of the membrane and of the internal electrolyte, and cleaning of the electrodes are thus required. As for pH probes, pO2 probes are typically installed in duplicate in pilot- and large-scale bioreactors. When processes are scaled up, the effect of hydrostatic pressure on the reading of the probe should be taken into account (Marks 2003).

Alternatively, oxygen can be measured with fibre-optic sensors, where an oxygen-sensitive dye is placed at the tip of a fluorescent probe (Marose et al. 1999). The very small size and high sensitivity of these probes make them particularly suitable for measuring oxygen at low concentration, in small-scale culture systems, and in biosensors (see Section 13.2.4.2). This technology is, however, still rarely used at pilot and large scale.

13.2.2.5 Dissolved carbon dioxide (dCO2)

Dissolved carbon dioxide has been shown to affect cell metabolism and protein glycosylation (for instance, see deZengotita et al. 1998; Kimura & Miller 1997); the accumulation of CO2 can become particularly severe in large vessels, due to a relatively low removal rate via the gas phase. Several probes have been developed to measure this variable, although initially not specifically for animal cell cultures (Sonnleitner 1999). A fibre-optic sensor using a fluorescent dye has been tested recently and was found suitable to monitor and control, via N2 sparging, dissolved CO2 in a perfusion culture over several weeks (Pattison et al. 2000). At present, pilot- and large-scale bioreactors are rarely equipped with such a probe.

13.2.2.6 Pressure

Pilot-and large-scale bioreactors are commonly equipped with a pressure-sensing device in the tank headspace for various applications. After sterilization of the bioreactor, whether during a culture or a period when the tank stays idle before use, a positive pressure is usually applied to minimize the risk of contamination from the environment; additionally, pressure is commonly monitored during sterilization, for a measurement of the degree of steam saturation and for safety reasons. It is also usual to perform a pressure-hold test before sterilization, to detect any leaks in the tank (see Chapter 14, Section 14.4).

Pressure in bioprocessing is normally measured either as a gauge pressure, i.e. relative to atmospheric conditions, or as a differential pressure, i.e. as a difference between two pressure levels (Hartnett 1994). A measuring device is typically made of a mechanical sensor (orpressure gauge) and a converter, which transforms the mechanical signal into an electrical or a different mechanical one. The whole device is often referred to as a pressure transducer or transmitter.

The most common mechanical sensors are made of a C-shaped or a helical tube (both called Bourdon tubes), a diaphragm, a capsule or a set of bellows. In each case, a displacement or change in shape is created in response to the applied pressure. When no digital monitoring or automatic control is required, the sensor itself can be used for local readout via a pointing device; alternatively, simple pneumatic pressure converters for a local transmission of the signal can be used. The most common mechanical-to-electrical converter is the strain gauge, which is based on the principle that metallic- and semi-conductors subjected to mechanical deformation exhibit a change in their electrical resistance. When the strain gauge is bonded onto the mechanical sensor, the measured electrical resistance of the strain gauge can be correlated with the pressure applied to the sensor. Alternatively capacitance, potentiometric and resonant-wire sensors can be used; in the last of these, the pressure affects the tension of a wire, which in turn changes its resonant frequency; this directly generates a digital signal.

The main requirements of pressure transducers for bioreactors are that they can withstand sterilization temperatures over several cycles and that they are able to provide a reliable measurement at both culture and sterilization temperatures. To maintain sterility, the pressure gauge is normally separated from the bioreactor by a diaphragm.

13.2.2.7 Liquid level and weight

The liquid level in, or weight of, a bioreactor is an essential variable for monitoring and controlling the amount of medium and other nutrient solutions transferred to a bioreactor, both in fed-batch and perfusion cultures. Tanks for the preparation and storage of media and solutions are also often equipped with these devices. Mobile tanks can be placed on a floor scale. For fixed tanks, three types of device are commonly used (Hartnett 1994; Krahe 2002):

• Differential pressure cells are probably the most simple and economical devices; they measure the difference between the pressure at the bottom of the liquid and in the tank headspace, and convert the signal into an equivalent height of liquid. The device is made of two pressure sensors. The main characteristic of this measurement is that it is not influenced by the presence of foam and the amount of gas hold-up in the liquid (which may be a drawback, if it is desired to detect excessive foaming during the culture); it can however be affected by pressure fluctuations due to turbulence from agitation and aeration.

• Conductance and capacitance probes are both electronic sensors, based upon a chemically resistant probe located inside the tank. In a conductance probe, when the liquid level reaches the tip of the probe, an electrical circuit is closed which sends a signal to the reading device. In a capacitance probe, a capacitor is formed between the probe and the vessel;

the measured capacitance is then linearly related to the liquid level. The measurement provided by these probes may be influenced by the presence of foam and the amount of gas hold-up.

With these two level-measuring devices, the determination of the corresponding liquid weight requires knowledge of the tank cross-sectional area as a function of height.

• Load cells are external sensors that are placed at the base of the tank and measure the weight directly. The most common ones are strain-gauge cells, where the load acting on the cells is converted into an electrical signal (see Section 13.2.2.6). In most cases, four gauges are used to obtain maximum sensitivity and temperature compensation. The main problem is avoiding the interference of other external forces on the tank. For this reason, all portions of transfer lines directly connected to the tank should be made of flexible material (e.g. Teflon or silicone) or bent so that they exert only a stable force. Influence of temperature should also be taken into account. Load cells are typically more accurate than the two aforementioned devices, but also more expensive.

13.2.3 Biomass

13.2.3.1 Introduction

Biomass concentration is a key variable in animal cell cultures, and is routinely measured at all stages of process development and manufacturing. It is typically expressed as a mass, volume or number of cells per unit of the bioreactor working volume. The most widely used off-line technique consists in counting the cells in a haemocytometer under a microscope, using a dye such as Trypan blue, in order to distinguish viable from non-viable cells (Freshney 2000) (see also Chapter 32, Table 32.2). Several automated off-line instruments, using the same principle, are now commercially available.

On-line monitoring techniques for biomass can be classified into two main methods: the direct method, where a signal, directly related to some physical property of the biomass, is measured, and the indirect method, where biomass is derived from the measurement and calculation of other variables, such as metabolic rates or metabolite concentrations. The main techniques of both groups are reviewed below and summarized in Table 13.1 with some of the key references from literature. Additional discussions on these techniques can be found in reviews by Junker et al. (1994), Konstantinov et al. (1994a), Marose et al. (1999), Olsson and Nielsen (1997), and Sonnleitner (1999).

13.2.3.2 Direct method

Probably the most simple and popular technique for the direct method is based on the measurement, in the cell culture fluid, of light absorption (turbidity), light scattering (nephelometry) or of a combination of both. This is performed via an in situ optical probe, usually using visible or near-infrared (NIR) light (750 - 1100 nm). The signal is related to the concentration of solid particles in suspension and can thus be used to measure the total cell concentration, typically over a range of 0.5-20 X 106 cell/ml; usually, the correlation is linear only up to 2-5 X 106 cell/ml; at higher cell densities, a polynomial regression must be applied. Several probes are commercially available (Table 13.1). The main advantages of this technique are the simplicity, low cost and robustness of the probe. The main drawbacks are as follows. First, the signal is not specific to cells and can be affected by interference that changes with time, such as non-cellular solid particles and bubbles in the cell culture fluid, as well as biomass and protein build-up on the surface of the probe. Second, the signal also depends on the cell size and morphology, so that the measured cell concentration may become erroneous if these parameters change during the culture. Finally, no discrimination between viable and non-viable cells is possible. This technique is therefore unsuitable for detecting a decrease in viability.

Table 13.1 List of the main techniques for on-line monitoring of biomass.

Direct method

Range of biomass

Measurement principle

Measured variable and correlation with biomass

Application(s)

concentration^1

References

Light absorbance and/

Absorbance of NIR/visible light in the cell culture

Cells in suspension

0.5-

-20 X 106 cell/mL

Akhnoukh et al. 1996; Junker et al. 1994;

or scattering

fluid, correlated with the total cell concentration

Wu et al. 1995

Scattering of NIR/visible light in the cell culture

Cells in suspension

0.5-

-20 X 106 cell/mL

Junker et al. 1994; Wu et al. 1995

fluid, correlated with the total cell concentration

Combined absorbance and scattering of NIR/

Cells in suspension

0.5-

-20 X 106 cell/mL

Junker et al. 1994; Wu et al. 1995; Zhou

visible in the cell culture fluid, correlated with

and Hu 1994

the total cell concentration

Dielectric spectroscopy

Dielectric permittivity of the cell culture fluid, proportional to the concentration of cells with intact plasma membranes, the cell radius to the the fourth power and the cell capacitance per unit of membrane area

Cells in suspension

0.1

-3 X 106 cell/mL

Cannizzaro et al. 2003; Cerckel et al. 1993; Davey et al. 1997; Guan et al. 1998; Noll and Biselli 1998; Siano, 1997; Zeiser et al. 1999

Cells on microcarriers

0.5-

-5.5 X 106 cell/mL

Ducommun et al. 2002; Noll and Biselli 1998

Cells in a packed bed

1-

-2 X 10" cell/kg

Ducommun et al. 2002

Fluorescence

Fluorescence of cell culture fluid, correlated with intracellular NAD(P)H

Cells in suspension

na2)

Akhnoukh et al. 1996; Junker et al. 1994; Siano and Mutharasan 1991

In situ microscopy

Cell counting, via a CCD camera, in an in situ flow-through chamber

Cells in suspension

1-

10 X 106 cell/mL

Joeris et al. 2002

Acoustic resonance

Acoustic resonance of cell culture fluid, correlated

Cells in suspension

1-

-8 X 106 cell/mL

Kilburn et al. 1989

densitometry

with the total cell concentration

and immobilized

Complex permittivity

Complex permittivity measurement of 3D

Cells in microporous

0.4-1.8 X 106 cells

Bagnaninchi et al. 2003

microporous scaffold, correlated with the total

scaffold

volume of cells inside

Optical waveguide

Refractive index of a cell layer, correlated with the

Tissue culture

>1000 cells

Hug et al. 2001; Hug et al. 2002

lightmode

number of cells per unit surface area and the

spectroscopy

cell-surface interaction

Indirect method

Table 13.1 Continued

Indirect method

Measured/calculated

Range of biomass

variable (s)

Biomass-related parameter

Application(s)

concentration^1

References

Glucose uptake or

Concentration of metabolically-active cells

Cells in suspension.

1-20 X 105 cell/mL

Ducommun et al. 2001; Ducommun et

uptake rate

on microcarriers

1-8 X 1010 cell/kg

al. 2002; Ozturk et al. 1997b; Pelletier

and in a packed bed

et al. 1994; Rodrigues et al. 1999

Lactate production

Concentration of metabolically-active cells

Cells in a packed bed

1-8 X 1010 cell/kg

Ducommun et al. 2002; Ozturk et al.

1997b

Oxygen uptake or

Concentration of metabolically-active cells

Cells in suspension

1-20 X 105 cell/mL

Ducommun et al. 2001; Kamen et al.

uptake rate

and on

1996; Ozturk et al. 1997a; Tatiraju

microcarriers

et al. 1999

Oxygen uptake rate

Energy production rate (mol ATP/(m3 s))

Cells in suspension^"

1-7 X 106 cell/mL

Dorresteijn et al. 1996

combined with

corresponding to the volumetric biomass activity

1-20 X 105 cell/mL

Eyer and Heinzle 1996

lactate evolution rate

Carbon dioxide

Concentration of metabolically active cells

Cells in suspension'"

1-10 X 106 cell/mL

Kamen et al. 1996

evolution rate

Recombinant protein

Concentration of metabolically active cells

Cells in suspension

1-20 X 105 cell/mL

Ducommun et al. 2001; Ducommun et al.

production

and in a packed bed

1-8 X 1010 cell/kg

2002

Redox potential

Viable cell density (via empirical correlation)

Cells in suspension'"

1-20 X 105 cell/mL

Eyer and Heinzle 1996

1) The values give only an order of magnitude of the possible range of measurement. Different methods to determine this range, particularly the detection limit, have been applied by the various authors, and the reported experiments did not necessarily cover the whole possible range of biomass concentration.

2) No direct correlation between biomass concentration and fluorescence was reported in the cited references.

3) This technique could likely be applicable to immobilized cells, too, although it has not been reported in the cited references.

1) The values give only an order of magnitude of the possible range of measurement. Different methods to determine this range, particularly the detection limit, have been applied by the various authors, and the reported experiments did not necessarily cover the whole possible range of biomass concentration.

2) No direct correlation between biomass concentration and fluorescence was reported in the cited references.

3) This technique could likely be applicable to immobilized cells, too, although it has not been reported in the cited references.

Dielectric spectroscopy is a direct monitoring technique that is used increasingly in animal cell cultures. It is based on the measurement of dielectric permittivity in the bioreactor under an alternating electrical field at a fixed frequency, in the range of 0.1 - 10 MHz. Living cells, i.e. those with an intact plasma membrane, act as small capacitors and thus affect the overall signal. For animal cells in suspension, the overall capacitance, after subtraction of a background signal due to media components, is approximately proportional to the product of the cell concentration, the cell radius to the fourth power, and the cell capacitance per unit of membrane area. Dielectric spectroscopy thus detects only 'viable' cells; solid particles or lysed cells do not get polarized (Noll & Biselli 1998). The viable cell concentration may, however, differ from that determined by a dye-exclusion method, since the segregating principle between viable and non-viable cells is different. Cell densities as low as 1-2 X 105 cell/ml can be measured accurately provided that smoothing techniques are applied to the capacitance signal (Guan et al. 1998). Careful calibration, using cells and media representative of the culture conditions, must also be performed. The main advantages of dielectric spectroscopy are simplicity and robustness, in addition to the possibility of measuring viable cells. Furthermore, it can also be used to monitor cells growing on macroporous microcarriers and in packed-bed bioreactors (Ducommun et al. 2002; Noll & Biselli 1998), for which no other satisfactory on-line technique is available. The main problem with dielectric spectroscopy is that any significant change in media composition, cell size and cell-specific capacitance can affect the signal (Figure 13.3). The effect of cell size, however, can be corrected by an additional (off-line) measurement of this variable (e.g. electronic particle counting, flow cytometry). Alternatively, the influence of cell size and specific capacitance on the signal can provide additional information on the cell physiology, which can be used for process control. For instance Zeiser et al. (1999) have shown that the size of insect cells increased significantly after viral infection, such that capacitance measurement was a very promising technique for monitoring and controlling the infection process. Noll and Biselli (1998) observed that in a continuous fluidized-bed reactor with cells growing on macroporous carriers, the cell size decreased significantly during the culture as a result of the limited space available in the pores. The glutamine consumption per unit of capacitance, however, reached a constant value. The capacitance signal could hence be used to adjust the feed rate of glutamine and thus to control accurately the concentration of this species. By measuring capacitance at several frequencies, Cannizzaro et al. (2003) obtained additional information on the metabolic state of a culture and estimated the cell size on line.

Measurement of fluorescence in the cell culture fluid has also been used in an attempt to monitor biomass; the classical technique actually detects intracellular fluorophores, i.e. mostly NADH and NADPH, at excitation and emission wavelengths of 360 nm and 450 nm respectively. Care must be taken that interfering parameters such as temperature, pH, pO2 and agitation are kept constant. For determination of biomass from the fluorescence signal, the cellular content of the intracellular fluorophores must of course be constant; this is only the case during particular culture phases or conditions (Olsson & Nielsen 1997). The signal is also strongly affected by media components ('inner-filter effects'), making its interpretation difficult. Fluorescence should thus be viewed as a complementary technique, to use with another biomass measurement method, for an on-line characterization of the cell metabolic state (for instance, Akhnoukh et al. 1996; Siano & Mutharasan 1991). The original technique was recently improved by in situ scanning, i.e. by measuring fluorescence over a spectrum of excitation and emission wavelengths. This permits the separate quantification of several intra- and extra-cellular fluorophores. By selecting the appropriate fluorophores, a more reliable correlation with biomass may be achieved. A further improvement has been reported, with various microorganisms, where scanning fluorescence and turbidity measurements were performed at the same time with the same probe (Marose et al. 1999). The instrument could potentially be used with animal cells, too.

A—

Q.

/ «>

*

\

h

vy

Bioreactor Turbidity Cell Distribution

0.15

0.05

Figure 13.3 Monitoring of biomass during a batch cultivation of hybridoma cells in suspension. (a) Capacitance (measured here off line) (A), cell population density (•) and viability (□); these last two variables were determined by cell counting, using a dye-exclusion method. (b) Cell-specific capacitance (A), cell population density (•) and viability (□). The cell-specific capacitance (i.e. the capacitance per 105 viable cells) can be considered as a good approximation of the capacitance per unit of membrane area since the size distribution of viable cells was essentially constant (not shown). At first sight, there seems to be a good correlation between capacitance and cell density (a). However the highest capacitance was reached about 24 h before the highest cell concentration. This can be explained by the variation of the cell-specific capacitance, which reached a maximum during the second half of the exponential growth phase and then decreased (b). (Reproduced with permission from Noll and Biselli (1998).)

0.25

0.20

0.15

0.00

"100 80

-100 80

e 60

0.05

Figure 13.3 Monitoring of biomass during a batch cultivation of hybridoma cells in suspension. (a) Capacitance (measured here off line) (A), cell population density (•) and viability (□); these last two variables were determined by cell counting, using a dye-exclusion method. (b) Cell-specific capacitance (A), cell population density (•) and viability (□). The cell-specific capacitance (i.e. the capacitance per 105 viable cells) can be considered as a good approximation of the capacitance per unit of membrane area since the size distribution of viable cells was essentially constant (not shown). At first sight, there seems to be a good correlation between capacitance and cell density (a). However the highest capacitance was reached about 24 h before the highest cell concentration. This can be explained by the variation of the cell-specific capacitance, which reached a maximum during the second half of the exponential growth phase and then decreased (b). (Reproduced with permission from Noll and Biselli (1998).)

There are several other on-line techniques available for biomass, which are, however, used only for very specific applications or are not mature enough yet to be widely used in process monitoring. These include in situ microscopy, acoustic resonance densitometry, complex permittivity for cells growing in microporous scaffolds, and refractive index measurements in tissue cultures. Some key references for these techniques are given in Table 13.1.

13.2.3.3 Indirect method

Indirect techniques are based on the on-line measurement of other variables in the culture, from which the biomass or a related estimator is calculated using models of various complexities. This combination of a real sensor with an algorithm is often referred to as a software sensor (Cheruy 1997). This approach is particularly useful when a direct method cannot easily be applied, for example with immobilized or aggregated cells.

The most common indirect techniques consist of measuring one or several metabolites (glucose, lactate, glutamine, oxygen, carbon dioxide, etc.), continuously or at high frequency. By calculating the production or uptake rate of the metabolite and knowing the corresponding specific rate (from previous experiments), one can estimate the biomass. Alternatively, the biomass can be determined from the consumed or produced amount of the metabolite and from a previously determined yield of biomass on this metabolite. Several examples are given in Table 13.1, and a general review can be found in Konstantinov et al. (1994a). In some of the listed references, only the principle of the technique is presented and the metabolite was actually not measured on line; a proper on-line measurement using one of the techniques discussed below (Section 13.2.4) would, however, make the technique truly on-line for biomass. The main difficulty with indirect techniques is that specific rates and yields are rarely constant over the whole culture duration; in many cases, they are not even constant during the exponential phase. Ducommun et al. (2001) proposed an interesting approach to circumvent this problem. In all cases, however, indirect techniques are only valid for the culture phase and conditions where the specific rate or yield has been determined, and cannot be extrapolated outside these limits.

Another type of indirect technique is the determination of a biomass-related variable instead of the biomass itself. Dorresteijn et al. (1996), for instance, estimated on line the biomass activity, defined as the volumetric energy production rate, from the oxygen uptake rate and the lactic acid production rate. Although this approach does not give an estimate of the biomass per se, it can provide very valuable information on its activity; for control purposes, the biomass activity can actually be more relevant than the biomass amount. This technique can also detect changes in the cell metabolism, such as the onset of the plateau phase, before they become visible through cell counts.

Healthy Chemistry For Optimal Health

Healthy Chemistry For Optimal Health

Thousands Have Used Chemicals To Improve Their Medical Condition. This Book Is one Of The Most Valuable Resources In The World When It Comes To Chemicals. Not All Chemicals Are Harmful For Your Body – Find Out Those That Helps To Maintain Your Health.

Get My Free Ebook


Responses

  • Ellie
    Is temperature probe oresent in bioreactor?
    1 year ago

Post a comment