Sensors may be categorized as on-line and off-line sensors. On-line sensors are preferred since they provide process information quickly without any disruption in the process and any sampling and cultivating delays, and fewer human errors, and allow for arbitrary frequencies of measurement. Off-line analysis techniques are used because of the difficulty and expense of developing sterilizable probes or constructing a sampling system for some process variables and product properties.
Sensors must have several characteristics that must meet the specifications for use in a particular application [366, 439, 475]:
Accuracy is the degree of conformity to standard when the device is operated under specified conditions. This is typically described in terms of maximum percentage of deviation expected based on a full-scale reading on the device specification sheet.
Precision (Repeatability) is the exactness with which a measuring instrument repeats indications when it measures the same property under the same conditions. Sensors display a drift in time which can be corrected by periodic calibration.
Range is the difference between the minimum and the maximum values of the sensor output in the intended operating limits. It is essential that accuracy and precision improve as the range is reduced, which implies that a small range would be preferred. However, the range must be large enough to span the expected variation of the process variable under typical operational conditions, including disturbances and set point changes.
Durability refers to the endurance of a sensor under the exposure to different operational conditions (pH, temperature, acidity). Since most of the industrial scale cultivations require extensive periods of operation time for completion (2-20 days), the sensor response should be stable for extended periods.
Reliability is the degree of how well a sensor maintains both precision and accuracy over its expected lifetime. Reliability is a function of the failure rate, failure type, ease of maintenance, and robustness of the sensor.
Response Time is the time it takes for the sensor output to reach its final value. It indicates how quickly the sensor will respond to changes in the environment. This parameter indicates the speed of the sensor and must be compared with the speed of the process.
Sensor technology is a rapidly growing field that has significant potential to improve the operation, reliability, serviceability, and utility of many engineering systems. Advances in chemistry, biochemistry, materials science and engineering have accelerated the development of new and more capable sensors. However, as the complexity and the capabilities of the sensors increase, there is a significant amount of associated cost. Cost may be a critical consideration in the selection of a sensor and the way a measurement should be made (on-line or off-line). There may always be a trade-off between a high-cost on-line frequent measurement or a relatively low-cost off-line infrequent measurement.
On-line sensors are crucial for monitoring and controlling a process for its safe and optimal performance. These instruments can be classified as
1. sensors that do not come in contact with the cultivation broth, (e.g., in a thermocouple)
2. in-situ sensors that are immersed directly into the cultivation broth and hence are in contact with it (e.g., pH meter, dissolved oxygen probe and level sensor).
3. other sensors, such as tachometer and rotameter.
When the sensors/probes directly come in contact with the cultivation broth, one potential problem is to maintain aseptic conditions. Under these conditions, probe should be sterilizable and should be placed in a way so as to avoid any possible leakage from/ to the bioreactor through the connections. The seal is usually accomplished by elastomer "O" rings that also provide an easy insertion of the probe.
The location of the sensor in the fermenter is very important since the contents of the bioreactor are usually heterogeneous. As a result, the measurements of variables that are critical for control action will be dependent on the location of the sensor. Conventionally, sensors are placed in the midsection of the vessel, though placement somewhere else may also be considered depending on the design of the bioreactor. A sensor should be placed in a region with sufficient turbulence to maintain the surface of the sensor clean and avoid build-up of material on it. Besides corrupting the sensor output, such build-up may lead to fouling of the sensor.
In the absence of in-situ sensors, on-line analysis of medium components is preferred. The main idea is to sample the medium automatically by collecting it in a loop that has a relatively small volume compared to the cultivation broth and to analyze it. Automatic sampling can be performed in two ways: (1) direct withdrawal of sample by using a syringe or a catheter into a loop, (2) use of a membrane module that separates the sample from the tip of the sensor. After collecting the sample, microbial activity has to be stopped to achieve precise measurement by either adding a deactivating agent or by cooling the sample . Direct sampling allows for the measurement of the biomass and intracellular and extracellular components. The membrane modules used can be categorized either as membranes placed in a recycle loop connected to the bioreactor or in-situ membrane modules. The membrane may be replaced during the operation without any interruption of the process if it is placed in a recycle loop.
Flow Injection Analysis (FIA) has proven to be a valuable tool for on-line measurement of medium components due to its high speed (frequent analysis ability), good precision, and reliability. For the same purpose, other analytical systems are also used such as Mass Spectrometer (MS), High Pressure Liquid Chromatography (HPLC), Gas Chromatography (GC), with somewhat less efficiency. Among the in-situ sensors, Pt-resistance thermometers are commonly used for temperature measurement. Temperature control is typically implemented by manipulating flow rate of coolant circulating in coils if the temperature exceeds the control limits and by steam injection if the temperature goes below the minimum acceptable limit. For pH measurement, glass electrodes are used and the pH is regulated by the addition of acid or alkali. Dissolved Oxygen (DO2) is measured by Pt, Ag/AgCl, Ag and Pb sensors. They could be either polarographic which are expensive but reliable or galvanic types. DO2 is kept within the desired control limits by changing the agitator speed, inlet air flow rate and gas composition. The level of foam is determined by using conductance or capacitance sensors that trigger the addition of aliquots of antifoaming agent when there is excessive amount of foam formation.
Agitation speed and its power requirement are measured by a tachometer and watt-meter, respectively. Variable speed drives perform the control action. Air flow rate is measured by rotameters and mass flow meters and regulated by flow control valves. The pressure built inside the bioreactor is measured by spring and oil-filled diaphragms and regulated by pressure control valves. Feed flow rate is measured by electro-magnetic flow meters, vortex devices and electronic balances, it is controlled by upstream flow control valve and peristaltic pumps. On-line gas analysis (O2 and CO2) is performed by gas analyzers (paramagnetic and infrared, respectively) and by mass spectrometer.
Off-line analysis becomes a viable option especially when there is a need to measure a large number of medium components in order to improve the understanding of the process. Disadvantages of off-line analysis include in frequent and time-delayed process information. This may be caused by the speed of the instrument or preliminary treatment of the sample. In some cases, the amount of sample necessary may be sufficiently high to cause a volume change in the bioreactor if sampling is done frequently. In a typical cultivation, substrates, precursors, intermediate metabolites and products are measured in addition to other biomass related material (biotic phase) and other components of the cell-free medium (abiotic phase).
Dry weight and optical density measurements are used to determine the mass of the cells. For homogeneous cell populations, usually optical density is correlated with the weight of the sample. Microscopy and plate counts are used to measure the number of cell colonies present in an aliquot of cell suspension and on an agar-plate, respectively. Coulter counter provides a means for counting the number of particles passing through an orifice hence giving size distribution. But, this instrument is very expensive and has a limited usage due to the problems associated with small cells and inability to measure fungal organisms. Flow cytometry is used to determine the protein, DNA and RNA contents of the biotic phase. Although this is a very powerful technique, it can only be applied to unicellular cultures. There are also chemical methods, such as enzymatic assays and colorimetric analysis for the measurement of these compounds, but some of them may be quite labor intensive. Image analysis systems are very useful especially in performing detailed morphological studies.
For the measurement of medium components, HPLC, being less selective, offers a wide range of advantages over FIA. GC is another widely used instrument with limited capacity. For certain components such as glucose, lactate and ethanol, analyzers specifically designed for these components are also available (e.g., YSI Glucose Analyzer). Physical properties of the cultivation medium such as viscosity, density and turbidity are also measured off-line in most of the cultivations.
The interested readers may find detailed information about sensors in many references [366, 396, 424, 439, 475].
Data collection and process control in most modern fermentation systems are computer-based. The computer is interfaced to process sensors by analog to digital converters and to final control elements such as control valves with digital to analog converters (Figure 3.1). This interface provides a link between hardware signals and software variables. Some analyzers may have their own microprocessors to refine and interpret data that they col-
Was this article helpful?