Fermenter and Bioreactor Design
||Growth rate constants
||Activation energies for growth rate
||Specific growth rate
||Maximum specific growth rate
||Volumetric oxygen transfer rate
||Viscosity constant (non-Newtonian)
||Monod equation constant
||Carbon source molecular weight
||Viscosity constant (non-Newtonian)
||Agitator power number (ungassed)
||Gassed agitator power
||Ungassed agitator power
||Volumetric gas flow rate
||Oxygen transfer rate
||Fermenter liquid volume
||Superficial gas velocity
||Rate of product formation per unit cell mass
||Cellular yield per substrate mass consumed
||Viscosity of water
Fermentation processes are generally understood to be those caused by the actions of micro-organisms, but there is some argument as to whether the term should refer only to the actions of live cells, or be extended to include the actions of extracted enzymes (the biocatalysts which govern the chemical processes within cells). However, since the requirements of live micro-organisms are markedly different from those of enzymes, it is convenient to separate the design of fermenters from the design of enzyme reactors which are discussed in detail in a later chapter.
The metabolism of micro-organisms is a complex system in which a series of chemical processes are carried out, usually simultaneously, within the cell confines. Each reaction is catalyzed by a specific enzyme, and the rate of each step is governed by the cell's production of the various enzymes. By the use of triggers to promote and control the rate of enzyme production, the organism can control the different stages in equilibrium and can also respond to changes in its environment. Industrial fermentations take advantage of a particular part of the cell's metabolic process, and to this end a successful and economic fermentation process requires not only an efficient micro-organism and a suitable growth medium, but also a fermenter in which the optimum conditions for metabolism exist.
It is essential to recognize that, whereas many of the principles of chemical reactor design apply to fermenters, there are certain fundamental differences. Although some of the major advantages quoted for fermentation over 'conventional' chemical reactions are the moderate conditions of temperature, pressure and pH, an organism is normally only viable within a narrow range of conditions. More importantly, if different conditions exist within the fermenter, whether due to poor mixing, recycle loops or rigid flow conditions, then the metabolism may be altered due to the, cell's response to the changing environment. A typical example would be a poorly aerated zone in a loop fermenter, if the organism is forced to 'hold its breath' for too long in the unaerated zone then the cells will become overstressed and stop production.
A further difference is the need to maintain sterile conditions within the fermenter. Whereas the main concerns in a chemical reactor are to prevent materials from leaking, a major concern of industrial fermentations is to prevent competing organisms from entering the fermenter since the contaminating organisms will consume part of the substrate to reduce the yield and may also contaminate the product. Containment of the fermenter system is only necessary in a small proportion of fermentations which utilise hazardous organisms.
The concentration of micro-organisms in a fermenter is usually low; typically cell weight may comprise only 10-15% of the fermenter contents, and the product weight is a small fraction of this. A typical industrial fermentation process comprises the different activities shown in Fig. 1. It is important that the fermenter design is not considered in isolation but as part of the whole system. The capital cost of the fermenter and its associated equipment will usually be only 40-50% of the total plant cost: the downstream separation recycle and feed preparation processes will make up the remainder. There is considerable scope, especially in large-scale processes, for optimization between the fermenter conditions and the cost of the downstream and recycle processes.
The oldest uses of fermentation have been for the production of foodstuffs, and it is only in the recent past that fermentation has been utilized to make antibiotics and chemicals. An essential characteristic of these 'traditional' industries is that the flavour and appearance of the final product must be consistent from day to day, despite seasonal changes to the raw materials. For this reason, industries still rely largely on human judgement to control the fermentation, since analysis and control of the number of minor components that constitute 'flavour' are virtually impossible by other means. Fermenter design in the traditional industries has changed to allow more hygienic operation and more efficient practices, but batch operation is usually retained in order to permit control of the final product.
In large-scale processes, the use of continuous fermentation results in significant economies in capital, labour and operating costs compared to batch operation. The trend will be to the use of continuous processes for new bulk foodstuffs especially where the product must compete with existing products manufactured by other methods and flavour is not critical (e.g. SCP). Better understanding of the mechanisms involved will also allow some existing batch processes (such as beer production) to be carried out continuously.
Fig. 1. Typical fermentation proces.
The primary aim of an industrial fermenter system is to provide the optimum conditions for the micro-organism, and a thorough understanding of the requirements and limitations of the organism is essential to the design of the fermenter. Much of the development effort that goes into the scale-up of new microbial processes is aimed at gaining an understanding of the ideal conditions for the system; this is further discussed in Section 5.
Although most fermentations are still performed in stirred, air sparged tank fermenters, the wide range of designs that have been utilised or proposed for different fermentations arise from the need to satisfy the needs of the organisms as economically as possible. It is therefore useful to discuss these requirements before considering both the range of fermenter designs available and the methods used to design the systems.
Unlike a chemical reaction, where the reaction rate is a direct function of the concentrations of each of the reactants and products. in microbial growth the relationships are more complex. The basic relationship between growth rate and concentration of each substrate is given in the Monod equation:2
where G is the specific growth rate; Gmax is the maximum specific growth rate; S is the substrate concentration; Ks is a constant, equal to the concentration when specific growth rate is half the maximum rate.
This expression is shown graphically in Fig. 2(a) and it is apparent that, when S ] 10Ks, the specific growth rate becomes virtually constant. The value of Ks for growth of S. cerevisiae in glucose is about 25 mg/litre and therefore the specific growth rate is close to the maximum when the concentration is above about 0.25 g/litre.
At higher concentrations of nutrient the organism usually suffers inhibition and the specific growth rate is reduced, resulting in a growth rate curve similar to Fig. 2(b). In some cases, inhibition is caused purely by the increasing osmotic pressure within the cell, but high concentrations of substrate may also have toxic effects on the organism or may trigger the organism to use a different metabolic route. The Monod model does not always predict the behaviour of an organism at low substrate concentrations, since low nutrient concentrations may also trigger a different metabolic path.
A further complication when using commercial substrates such as molasses and saccharified starches is that most organisms will consume one growth nutrient in preference to another, and will not start to utilise the second nutrient until the first has been depleted. While this would merely lengthen the cycle time for a batch fermentation, in a continuous fermenter the high loss of substrate may have serious economic implications. For this reason, most Zymomonas strains are unsuitable for economic ethanol production on mixed substrates, since they can only utilise a portion of the substrates.
The restraints on nutrient levels outlined above mean that optimisation of the medium is essential for economic fermentation. A high proportion of batch fermentations use the fed-batch technique, where nutrient is continually added during the course of the fermentation. This maintains the nutrient concentration at the optimum level and results in high specific growth rates and improved yields. In continuous fermenters, careful attention may be needed in the distribution of the substrate feed so as to avoid local high concentrations if the organism is sensitive to these.
The vast majority of current fermentation processes are aerobic in nature, and utilise oxygen to generate energy for growth, but micro-organisms in submerged culture are unable to use gaseous oxygen and can only absorb dissolved oxygen. Since oxygen is considerably less soluble in water than the other nutrients required by the organism (8 mg/litre at 30Âº C compared to 10000-1000000 mg/litre for most nutrient salts), mass transfer of oxygen into the broth is normally the rate-limiting step in fermentations.
An expression for the oxygen required for cell growth was derived by Mateles3 as:
where Y is the cellular yield/substrate used; Mw is the carbon source molecular weight; O'. C', H', N' are fractions present in the cell; O.C. H, N are numbers of atoms/substrate molecule. Because the oxygen demand is inversely proportional to the cellular yield, it is apparent that the oxygen requirement of the fermentation is a good measure of the efficiency of substrate conversion.
The important measurement of oxygen availability to the organism is the dissolved oxygen tension (DOT), expressed either as mgO2,/litre or as a percentage of oxygen saturation. The relationship between DOT and cell growth rate, shown in Fig. 2 (c), is similar to the Monod relationship: above a critical DOT the growth rate and formation of products are independent of the DOT. This critical DOT is normally low and will typically be less than 10% saturation. Below the critical DOT the growth rate falls rapidly.
This behaviour reflects the mechanism for oxygen transfer, which has been found by numerous workers to be limited by transfer through the liquid film. If insufficient oxygen is available, the cell will absorb it from the substrate as rapidly as it is transferred from the gas phase, since little driving force is required for the transfer from liquid to the cell. The DOT in the substrate will therefore remain low. Immediately the rate of oxygen transferred into the liquor exceeds the cell's demand, then the DOT will rise significantly. As with nutrient concentration, the DOT may affect the metabolic pathway chosen by the organism, and the influence of both oxyen and glucose concentrations on yeast fermentation is well document. At low glucose concentrations and high DOT, aerobic fermentation takes place resulting in cell growth, while at higher sugar concentrations the 'Crabtree effect' takes place even at high DOT and ethanol is preferentially produced.
The optimal growth temperature arises from a balance between the growth rate of the organism, which is enhanced by increasing temperature, and the death rate which is also increased at higher temperatures. Both rates can be represented by the Arrhenius equation and so the overall growth rate becomes:
where M is the cell mass; A, A' are constants for growth and death; E, E' are activation energies for growth and death; T is the absolute temperature. A typical plot of the overall growth rate is shown in Fig. 2 (d).
A complication in determining the optimum fermenter conditions is that the temperature at which maximum growth occurs may not be the same as that at which the best yield of product is obtained. In addition, increasing temperature will reduce the lag time in batch fermentation but may also increase the mutation rate of an organism.
Heat is evolved from all stages of fermentation processes, but the amount evolved depends on the balance between cell growth and cell maintenance requirements. Under conditions of cell growth. 50-60% of the energy available is converted to heat, while during cell maintenance virtually all the available combustion energy is released as heat to the environment.6 This heat must be efficiently removed from the fermenter in order to maintain suitable conditions for fermentation; the mechanisms for heat removal are discussed.
It has been shown that heat evolution is closely related to the efficiency of conversion of nutrients to cell mass. Because the rate of oxygen uptake increases with the degree of oxidation (and so the heat evolution), the heat evolution can be calculated from the oxygen uptake rate by the correlation developed by Cooney
Qf = 0.12 Q (O2)
where Qf is the heat evolution (kcal/litre-h); Q (O2) is the oxygen uptake rate (mmol/litre-h). This correlation has been shown to be valid for a number of organisms and substrates, and is useful for initial design estimates. During the scale-up of a microbial process, the heat evolution should be confirmed from practical measurements.
A serious engineering problem is caused by the moderate temperatures required for most fermentations because the fermenter temperature is usually close to, or even below, that of the available cooling, water supply. This results in the need for excessively large heat transfer areas in commercial fermenters, and refrigeration is frequently required. The potential economic benefit in developing thermophilic organisms for many commercial fermentations is great, and could be essential in developing new fermentation routes to existing products.
As with temperature, most organisms will only retain viability within a limited range of pH. Ideal conditions for growth are at around pH 7 for most organisms, but cells will remain viable over a range of several pH units. The control of optimum pH is essential, either by buffering (on laboratory scale) or by a pH control system. In general, yeasts are more resistant to acid conditions than bacteria and other organisms, and this can be used to protect appropriate yeast fermentations from bacterial infection.
Specialized organisms may operate either at high acidity (e.g. lactic acid bacteria) or at high alkalinity. Several organisms are triggered to produce different metabolites at different pH conditions, and in these cases extremely close pH control is required to optimise the product yield.
The effect of the fermentation broth rheology is often critical in fermenter design, and the maximum cell concentration achievable in a fermenter may be limited by the resulting viscosity.
The effect of cell concentration on viscosity varies depending on the form of the organism the viscosity of mycelial broths is significantly higher than those containing more spherical yeast or bacterial cells. At low concentrations, viscosity is roughly proportional to concentration, but at higher cell concentrations the viscosity rapidly increases.
Fermentation broths are highly non-Newtonian (i.e. the apparent viscosity is dependent on the shear rate), and most broths approximate to pseudo-plastic behaviour where:
where t is the shear stress; g is the shear rate; ma is the apparent viscosity; K, n are non-Newtonian viscosity constants.
As a fermentation progresses, the values of K and n would typically vary as shown in Fig. 3(a). As the viscosity increases, the transfer rate of both oxygen and nutrient to the cell will be reduced. The heat transfer rate will also fall with increased viscosity, resulting in the need for larger areas of cooling surfaces. In a highly non-Newtonian viscous broth, the degree of agitation decreases rapidly away from the agitator; see Fig. 3(b). Thus, aeration may only occur in the zone local to the impeller, and the rate of heat transfer through the vessel wall will be drastically reduced. All of these effects have a detrimental effect on fermentation, and fermenters to handle viscous broths require different designs for aeration and agitation from those handling more fluid broths.
There are a number of additional constraints that must be considered in the development of a fermentation process. In deep fermenters or those which are pressurized to enhance oxygen transfer, the effects of pressure may become limiting. The cause may be either the increased hydrostatic pressure on the cell causing damage, or alternatively due to carbon dioxide inhibition. Carbon dioxide will inhibit most organisms, and at increased pressures the solubility of CO2, is increased until it eventually reaches an inhibitory level.
The effect of shear on an organism is related to the cell shape: bacteria and yeast tend to be more resistant to shear than mycelia. As the shear rate is increased in a mycelial fermentation, the productivity may initially increase as the filaments become separated and mass transfer is improved. Eventually, increasing shear will damage the cells and their viability will rapidly fall. The effect of shear is particularly severe on animal tissue cultures, where the cells are easily damaged by high-speed agitation, and special low-shear fermenter designs are required. Shear damage can be caused both by high-speed agitators and by external recycle pumps.
The mutation rate of an organism, especially a genetically engineered one, may limit the length of time for which a fermentation may be operated before the cell population reverts to a less favoured variant. If an organism is not stable for greater than 3 months, then continuous fermentation is unlikely to be economic.
In many processes, the economics may be improved by recycling unconsumed substrate from the downstream recovery systems. However, trace nutrients and by-products may have an inhibitory effect on the organism, and this will limit the amount of substrate that can be recovered by recycle from the downstream separation stages.
The nature of the required product and its relationship to the growth phase of the micro-organism also have a fundamental bearing on the design of the fermenter system. As shown in Table 1, the required product may be contained within the cell or secreted from the cell, and may be formed either as part of the growth process or as a secondary metabolite.
The majority of fermentation products are retained within the cell by the organism, and sufficient cells must be grown to contain the yield of product. Unless the product is the whole cell (e.g. SCP) then a significant proportion of the substrate must be used in producing 'by-product' cellular material; indeed the nutrients needed for cell growth may be greatly different from those needed for production of the required product.
Production of intracellular product is usually an advantage in product recovery, because as the cells will not be recycled to the fermenter they may be separated from the broth in non-sterile equipment following the fermentation, and the cost of this initial product concentration equipment is relatively low; see Fig. 4 (a). If substrate is to be recycled to improve the process economics, then this would need to be re-sterilized.
Fig. 4. (a) Fermenter system: intracellular product, (b) Fermenter system: extra-cellular product.
A minority of fermentation products are secreted by the organism, either because they are a waste product (ethanol) or because their role is to influence the environment of the organism (antibiotics). In this case, the product must be recovered from the broth after removal of the cells, and the opportunity exists to recycle cells back to the fermenter; see Fig. 4(b). This not only reduces the amount of nutrient required for cell growth, but also reduces the size of the fermenter since the cellular concentration (and so the volumetric yield) may be maintained at a high level.
Since the live cells may not be sterilised before return to the fermenter, the separation equipment must be operated aseptically and the cost of the separation equipment will increase significantly. Genetic selection of an organism which will secrete the product rather than retain it within the cell may in the future bring significant improvements in production costs of bulk products as better sterile separation equipment (e.g. ultrafiltration systems) is developed.
Secondary metabolites are materials that are not needed for growth of the micro-organism and are usually produced after the growth phase of the organism. Most secondary products are antibiotics and mycotoxins, and formation of the secondary metabolite requires deficiency of a key nutrient as well as the presence of a precursor to trigger production. The fermentation must therefore be carried out in two distinct phases, and the optimum conditions determined to maximize production of the required compound in the second phase.
Similar considerations apply to a number of other fermentation processes (such as ethanol and citric acid production) where, although a primary metabolite is formed, the product is generated after growth of the organism has ceased. In both cases, cellular growth is only encouraged to produce enough cells for efficient formation of the secondary product.
The range of fermenter designs that have been either successfully used or proposed reflects the large number of factors that can affect the design. Selection depends on factors such as the biological constraints of the organism, the scale of production, the level of technology available, economics and often the range of products that are to be considered.
Because of the many different forms of fermenters and the way in which they are operated, it is necessary to consider the design from a number of different aspects such as batch/continuous operation, the method of agitation, the use of free microbial flocs or immobilized systems, and whether the fermentation is aerobic or anaerobic.
For example, many baker's yeast processes are batch, aerobic fermentations involving free organisms in a stirred tank fermenter, while some waste treatment plants using nitrifying bacteria have anaerobic continuous systems with immobilised organisms, agitated in a fluidized bed. A summary of some of the different fermentation methods and fermenters used in the food industry is given in Table 2 to show the range of possibilities that have been used to date.
It is apparent from the table that the vast majority of commercial fermenters are stirred vessels containing free aerobic organisms and are batch operated. The reasons for this are discussed in the following sections, but as fermentation technology improves a wider range of designs will become used.
The design of fermenter is also affected by whether the commercial plant is to be dedicated for a single product or to be adaptable for multi-product-operation. A batch, stirred tank fermenter will almost always be used for a multi-product plant since the design is adaptable for different conditions.
Bulk products will normally be produced in single-product plants where the fermenter will be designed specifically to provide the most efficient oxygen transfer, volumetric efficiency and heat removal for that process in order to minimise the production costs. In these cases, other designs of fermenter will often be more suitable. Good examples of this trend are the development of the air-lift fermenter for SCP production, and the various fermenter designs in use for fuel alcohol production.
In batch fermentation, conditions within the fermenter change during the fermentation cycle, with the product and cell concentrations increasing as the substrate is depleted; see Fig. 5(a).
Traditionally, fermentations have been carried out in batch operation and this design has many advantages which are still valid and lead to its continued use in many applications.
- Each batch retains its discrete identity, and this is essential for pharmaceuticals and other products which require FDA-type batch logging. Each batch can also be individually controlled to take account of variations in feeds or other conditions.
- The fermenter and substrate can be sterilised between batches, ensuring that unwanted organisms are minimised. Once the organism has been established in the fermenter, the effect of infections is usually small since a competing organism has little time to become dominant. In addition, the loss of an occasional batch through contamination may be less expensive than the effects of contamination of a continuous fermenter, where the system would need to be drained, sterilised, re-inoculated and restarted.
- Conditions within the fermenter can be changed during the time course of the fermentation for production of e.g. secondary metabolites and other growth-unrelated products such as enzymes.
- Scale-up from laboratory and pilot-plant scale is relatively straightforward.
- Some suitable batch fermentations may be carried out using unsophisticated technology, and without continuous supervision. For this reason, batch plants are used for much fuel alcohol production in countries where controlled operation under hygienic conditions is not feasible.
- Production may be easily adjusted to suit demand, availability of seasonal feedstocks etc. The same fermenters may often be used for different products.
- Fed batch operation may be used to maintain the concentration of critical nutrients within acceptable limits by addition of substrate throughout the fermentation rather than solely at the start.
A major disadvantage is that only a proportion of a batch fermentation cycle is available for fermentation; a significant part of the cycle is taken up by sterilization, filling and emptying, and inoculation. In addition to the main fermenter (s). a seed fermenter is also required to grow the inoculum, and this normally operates at a higher level of sterility than the production vessel.
In a continuous fermentation, the conditions in the fermenter remain constant during operation (apart from the start-up period) so that fermentation is controlled with the same cell and substrate concentrations, temperature etc. throughout see; Fig. 5 (b). Continuous operation is normally chosen for production of low value, high volume products because of the cost benefits. However, there are a number of other potential advantages to be gained by its use. The main advantages are listed below.
- The productivity based on fermenter volume is better since the fermenter can operate at a constant high cell concentration and peak fermentation rate. In addition, there is no downtime between batches as in a batch fermentation.
- The fermenter conditions can be optimized to maintain the best conditions for yield throughout operation.
- Higher productivity results in reduced capital costs since fewer fermenters are required, and the ancillary seed fermenter can be small since it is only required for occasional start-up duty.
- Labour costs are reduced since there is no labour-intensive turnaround of fermenters between batches.
- Control of the fermenter is simplified since it operates at steady-state conditions.
There are, however, disadvantages inherent in continuous operation which may make it inappropriate for some fermentations:
- 'Wash out' of the fermenter will occur if the rate of cell removal is greater than that of cell growth. In a continuous stirred tank fermenter (CSTF) the minimum fermenter volume is limited by this effect, while in a plug flow type fermenter cells must always be recycled to the inlet in order to maintain a high cell concentration.
- The cell population has a spread of ages, all of which operate in the same environment. If a product is required that is not produced in the growth phase, then a second fermenter stage will be required, operating at the necessary conditions for product formation.
- Continuous operation is only economic if the period of continuous operation can be extended for at least several months. This requires that the organism be stable over the period (i.e. unlikely to mutate or degenerate) and that sterile conditions are maintained for the same lengthy period.
The increased use of continuous fermentation has led to improvements in methods for aeration and heat transfer so that high fermentation rates can be used, and to the maintenance of aseptic conditions to enable monoseptic operation over extended periods.
Most fermenters employ free organisms which are suspended in the fermentation broth either as individual cells or as flocs.
The different forms of fermenters used for free-floating organisms (either) individually or as does differ mainly in the means used for aeration and agitation (fig. 6). The most commonly used fermenter from, is the stirred vessel with air introduced via a sparger under the agitator. To achieve the best transfer rate of oxygen into the broth, a radial flow turbine agitator is normally used and there may be two or three impellers on a single shaft in tall fermenters. The fermenter is filled with the jacket and/or internal coils for removing the heat involved from fermentation. Both top and bottom entry agitators and commonly used.
The major advantages of stirred fermenters are that the scale-up methods are fairly well understood and scale up is relatively simple from laboratory equipment especially if a specialists agitator manufacturer is involved during development. The fermenter can also be readily adapted for multi product use. The stirred fermenter design is less suitable for high viscosity broth. (fig. 3 (b) and even both broths of normal fiscosity the circulation pattern within the fermenter leads to significant changes in conditions throughout the fermenter.
Many manufactures are now offering mixed axial/radial turbines or better hydrodynamically shaped agitators to improve the mixing efficiency within the fermenter. These designs can lead to significant improvements in fermentation efficiency by ensuring that fermentation condition that are constant throughout the fermenter. Because heat must be removed throughout a vessel jacket or internal coils, fermenters above 120-150 m3 require an additional external cooling circuit.
Air-agitated fermenters use an excess of air fed into the vessel through air spargers covering the whole of the base area. The method was used historically for production of baker's yeast, and is still utilised for some mycelial fermentations in which a conventional impeller would be infficient due to the high broth viscosity. Normally, air agitation is considerably less efficient than a stirred tank and the cost of the large volumes of sterile air required is prohibitive. Foaming is also a problem.
In some simpler fermenters, the agitation caused by the carbon dioxide evolved in the fermentation provides an adequate degree of agitation; this is used in beer fermenters. The intensity of agitation increases according to the underlying depth of fermenting broth, so that agitation is more intense at the top than the bottom, and this encourages the use of tall fermenters.
Various configurations have been proposed to improve the efficiency of the stirred fermenter by including draft tubes or similar devices to give a pumped internal circulation. The best known is the Waldhorf fermenter developed for baker's yeast production in which the flow rate past the air sparger is increased by placing the draft tube around the agitator. The pumped circulation ensures that once the broth has been aerated it will not immediately be drawn back into the aeration zone but will circulate round the vessel until the oxygen has been depleted. In this, way the organism is subjected to a cyclic variation of dissolved oxygen tension (and, in a continuous fermenter, nutrient level) and the design must ensure that the critical level is always exceeded.
At high fermentation rates, the area required for cooling may be greater than that available from the combination of fermenter jacket and internal coils, and the broth must be circulated via an external cooler. In this case the pumped recycle may be utilised for agitating the vessel and for aeration. However, the use of an external circulation loop introduces new problems in maintaining sterile conditions in the pump, heat exchanger and piping of the outside loop.
The air-lift fermenter creates an internal recirculation system by efficient use of sparged air. The fermenter is a tall vessel divided internally into a riser and downcomer section, usually by a concentric tube. Air is sparged into the base of the riser section, and the decrease in density caused by the gas-liquid mixture results in circulation of the liquid. Cooling is by internal coils, and internal baffles are fitted to improve mass transfer. The high rate of internal circulation allows the substrate concentration to be maintained at the optimum level. However, the downcomer is unaerated and care must be taken to avoid oxygen starvation in this zone.
By using the power available from air compression, the air-lift fermenter can achieve the high energy input required for aeration and mass transfer in large fermenters without the problems associated with massive agitators. Air-lift fermenters are most economical at larger sizes, and have been installed up to 1500 m3 (for SCP production).
Because micro-organisms are small and have a density close to that of water, it is difficult to retain them preferentially within the fermenter. The equipment for separation from the outgoing broth must be kept sterile and is large in relation to the weight of cells to be returned. Immobilised systems are utilised to overcome this problem, and allow the product yield to be increased compared to the cellular growth. Immobilised systems are also used to retain enzymes (which have been extracted from the micro-organism) within a fermenter system.
Immobilised cell, systems are not economic for batch fermentation with live cells but the high value of enzymes encourages the use of immobilisation to recycle enzymes in batch systems. The ideal applications are in processes such as waste treatment, citric acid production and ethanol production where the product is formed under conditions different from those needed for growth, and research has been carried out into the use of immobilised systems for production of beer, wine and various solvents.
Immobilisation is carried out separately from the main fermenter and involves either absorbing the live cells onto the surface of an inert carrier or combining them within a permeable support such as silica or polysaccharide gels. Alternatively, in some systems which are only mildly agitated (e.g. waste treatment), the presence of an inert support within the vessel enables a film of micro-organisms to form which attains an equilibrium between growth of new cells and removal of dead cells.
Various fermenter configurations can be used to take advantage of immobilised cells, and the main forms are shown in Fig. 7. Most are variations of the packed-bed fermenter in which the immobilised cells are supported on a fixed bed which may be of random packing material, fixed grids, fibres or sheets. Substrate is passed over the bed in plug flow and the concentration of substrate and product change is shown in Fig. 5 (c), from a high substrate concentration at the inlet to high product level at the outlet. Substrate may be recycled if necessary to improve efficiency and reduce the bed height.
Other examples of packed beds are packed tower fermenters, trickle bed fermenters (in which the flow is downwards over the packing) and various reactors in which the immobilised cells are held within piping lengths. Potential problems arising in the design are the need to replace any support medium that is lost by attrition, and the limited aeration level that can be used before the bed floods or becomes disturbed. In tower fermenters, an added limitation is the amount of disturbance caused by carbon dioxide evolution which increases as the fermenter height is increased.
Fluidised beds have the advantage that the bed is highly agitated, which improves both heat and mass transfer, and allows a higher aeration rate. If a separator is incorporated in the outlet, then loss of immobilised cells can be minimised and it is feasible to replace continuously any support that is lost. In this configuration, the fermenter kinetics are between those of a CSTF and a packed bed; agitation is sufficient to mix partially the substrate within the vessel but there is still a concentration gradient between inlet and outlet.
Considerable research work is being carried out into the use of immobilised systems since they offer the major advantages of retaining micro-organisms within the fermenter and of separating cell growth from product yield. The main areas of research are into methods of immobilisation and into optimizing, the conditions for fermentation. It is apparent that most organisms remain viable at much lower growth rates when supported in an immobilised system, and that the growth characteristics are changed from those in free flocs. The mass transfer rate of nutrients into the cell and waste products from the cell is also critical and all these factors affect the optimum fermenter conditions. Another important area of current research is into suitable support materials and methods of immobilisation.
Totally anaerobic fermentations are generally used only in waste treatment, and the best known are fermenters for production of methane from waste products. These fermenters contain a mixture of bacteria which operate symbiotically; the methane-producing bacteria metabolize fatty acids produced by other bacteria in the digester to obtain energy for growth. The growth rate of anaerobic organisms is far slower than for aerobic organisms (typically 0.1-0.2/day compared to 0.1-0.2/h) and this leads to larger fermenters (or digesters). To maintain a high cell concentration, either sludge from the outlet is settled and returned to the digester, or immobilised systems such as packed or fluidised beds are used.
Few plants have been built since they require a high capital investment and depend for economic operation on the value of the methane by-product. The main applications are for high volume wastes with relatively low concentrations of fermentable material, such as from distilleries, breweries and paper plants.
Although the majority of commercial fermenters are conventional stirred vessels, a wide range of configurations has been proposed and tested, usually at laboratory or pilot scale, and some of these designs will be used commercially. The aims of these novel designs are variously to improve oxygen transfer, to increase the cell concentration, to give better or more suitable agitation or heat transfer, and to improve sterility. Reviews of different systems have been published.
Much work is being carried out into immobilised systems to increase cell concentration without an external separation system. In addition, different designs have been proposed for mammalian cell culture where the requirements are rather different from those of micro-organisms and much lower shear rates are required. Another approach to increasing cell concentration is the use of membrane fermenters which use filtration membranes in the form of sheets or hollow fibres to separate secreted product from cellular material. Fermenter systems with gravity settlers are used with flocculating yeasts in ethanol production to enable a high degree of cell recycle under sterile conditions.
Various forms of air-lift fermenter have been suggested which aim at increased oxygen transfer efficiency at a high level of sterility. Another fermenter design giving the same benefits is the high turbulence plunging jet in which aeration is achieved by the action of a coherent jet of recycled broth.
The information needed to design a fermenter system for a new process is normally obtained by scale-up from small-scale trials. Although a considerable number of correlations exist it is necessary to carry out experimental work to generate accurate scale-up methods and to optimise the fermentation at successive scales of equipment. In addition, confidence in the scale-up techniques is acquired at each successive stage, and any pitfalls are hopefully discovered before erection of a full-scale plant.
The number of stages in the development of a process from the laboratory to full-scale production will depend on the amount of data necessary to ensure confidence in the design and costing of the production unit. In general, the importance of obtaining scale-up information is increased by the use of continuous processes, novel organisms, above average size fermenters, and by the use of fermenters other than 'conventional' agitated vessels. Depending on the size of the final production fermenters, four stages of scale-up may be required following initial laboratory shake-flask experiments, and these are outlined in Table 3. The size of each stage is about an order of magnitude greater than the preceding stage.
At the initial laboratory phase, the main aims are to collect kinetic data, to arrive at the optimum fermenter conditions for nutrient, temperature, pH etc., and to determine the materials balance. The final development of the organism is normally carried out at the same time. It is necessary to develop the organism, the fermenter and the overall process design together to ensure not only an economic engineering design but also that research is carried out into the most important aspects. It is frequently the case in the development of fermentation processes that research work is centred solely on the fermenter performance without considering the effects on upstream and downstream processing. The features of a hypothetical ideal organism given in Table 4 are not achieved by any current commercial micro-organism, but the list shows the number of factors that contribute to a successful commercial fermentation process.
Several laboratory fermenters are required to investigate the range of fermenter conditions and obtain the necessary data, and the work may still continue at least throughout the pilot plant operation, since the cost of the equipment and the raw materials is relatively low. The next stage of development is pilot plant operation, where the aims are to obtain engineering data for aeration, agitation and heat transfer, to investigate alternative fermenter configurations, and to develop the necessary methods for feed preparation and product recovery.
The Ideal Organism
High product concentration
High conversion efficiency
Low sterility needs
Tolerant of substrate and oxygen variations
Ferments above 45Â°C
Insensitive to salt level
Operation of the semi-scale plant is intended to prove that the overall process works, including any recycle streams and the maintenance of product quality and asepsis. Further optimisation will be carried out at this stage and the plant will be expected to produce commercial grade product for extensive testing. In many cases, the initial production will also be carried out on this plant.
Apart from the microbiological aspects (e.g. organism, substrate optimisation), the main engineering data obtained for scale-up of a typical aerobic fermentation relate to fermentation kinetics; oxygen transfer rate and agitation, and heat transfer.
A knowledge of the kinetics of a fermentation is necessary to size the fermenter and its associated equipment, and this information is normally obtained from laboratory experimentation with 1-3 litre fermenters. In batch fermentations, the kinetic model provides information to predict the rate of cell mass or product generation, while in continuous fermentation it will predict the rate of product formation under given conditions. The optimum fermenter size and fermentation conditions can then be calculated for the highest or most economic yield of product by modelling the fermenter system. In a continuous fermenter system, such an economic model might include the capital cost of fermenters with the associated equipment for sterilisation, aeration and initial product recovery, the cost of utilities for sterilisation, air compression, cooling etc., the cost of substrate and nutrients, and manpower.
Since all fermentations involve an interrelated series of enzymic reactions, the kinetic model will usually be greatly simplified compared to the actual series of reactions that occur. The form of kinetics for product formation will vary dependent on whether the product is cellular material, primary metabolite or secondary metabolite.
The oxygen demand for established fermentations may be well documented or alternatively an initial estimate may be predicted by the Mateles equation given in Section 2.2. However, in the development of a new process it is necessary to determine the oxygen requirement experimentally by materials balance techniques before designing the fermenter aeration and agitation.
Basis of Scale-up
Much work has been carried out into predicting the rate of oxygen transfer into stirred fermenters, but most of these correlations are based on experiments in which the mass transfer is determined by chemical methods such as sulphite oxidation (which do not necessarily predict the rate in a biological system). Since the oxygen transfer rate to a micro- organism is affected by the presence of salts, proteins, surfactant etc. in the broth, and also by the rate at which the organism can utilise the oxygen, most production fermenters are designed by scaling up from pilot-scale units, based on the existing correlations given below, and modified to reflect the actual experimental behaviour. Little oxygen transfer data is available from full-scale plants, primarily because the information is considered commercially secret.
Several different scale-up methods are available, although design for constant KLa (volumetric oxygen transfer rate) is most commonly used for yeast and bacterial fermentations. Because aeration and agitation are interdependent, the best scale-up method may depend on factors such as the viscosity of the broth, the resistance to shear of the organism and the importance of constant nutrient levels. In a large fermenter, it is not possible to design for a range of possible agitation levels, so the accuracy of the scale-up method must be confirmed in the progression from pilot plant 10 semi-scale operation.
Constant power per unit volume. One of the simplest methods of scale-up is to maintain constant power per unit volume, and this is certainly the simplest approximation to use for an initial estimate. For oxygen-limited fermentations, typical power inputs are 2k W/m3 for both biomass and penicillin production.
Generally the geometric proportions of the fermenter and impeller are kept constant, in which case the impeller speed is given by:
N1 = N2 (D2/D1)2/3
In initial studies, the air rate may be scaled-up based on the flow per unit volume (VVM), and this method is normally used to scale-up oxygen rates from laboratory scale experiments where power requirements cannot be measured. Typical air rates are between 0-5 and 1.0 vvm for a wide range of aerobic fermentations.
Constant KLa. The most commonly used scale-up for fermentations limited by oxygen transfer is on the basis of constant KLa or constant oxygen transfer rate. This will achieve the same growth rate in the scaled-up fermenter and ensure similar kinetics. The method requires that the power required to achieve a given KLa can be predicted for the fermentation system, and many correlations have been proposed. Because of the non-Newtonian nature of most fermentation broths and the effects of surface tension and solution composition on the oxygen transfer rate, it is recommended that the correlations are confirmed or modified by experimental work. Several different methods are available and include sulphite absorption tests, degassing methods, oxygen uptake and measurement of an overall mass balance.4 If possible, it is best to carry out trials on the organism itself rather than by sulphite absorption since this will give a more accurate estimate.
Typical predictions for KLa are of a similar form, but the exponents vary depending on the scale of operation, the nature of the substrate and its viscosity, i.e.
In a fermenter system the power delivered by an agitator will differ depending on whether the vessel is gassed (or aerated) or ungassed. The ungassed power input is needed to size the agitator motor for start-up conditions such as substrate preparation or sterilisation, and also in many calculations as the basis of calculating the gassed power input.
The ungassed power input for a broth which approaches Newtonian behaviour can be estimated using the equation suggested by Rushton:22
where Np (= Qa/ND3) is the ungassed power number and is approximately 5.0-6.0 for turbulent flow using a six-bladed turbine.
In broths containing mycelia or concentrated yeast the non-Newtonian character has a significant effect on the power input: a modified Reynolds number is proposed by Wang.
Typically, the power input from a gassed impeller is less than 2/3 of the power from an ungassed impeller, and several different correction methods have been propo sed. The equation of Michel and Miller predicts for Newtonian liquids:
Alternative fermenter designs, such as air-lift fermenters, require similar methods for scale-up, but there is considerably less data available for mass transfer estimates and experimental confirmation of the scale-up method becomes more important.
Constant shear. If the micro-organism is shear-sensitive, the shear force should be maintained constant during scale-up. The maximum shear force is proportional to the tip speed and so:
N2 = N1 (D2/D1)
It is possible to maintain constant tip speed and either constant power per volume or constant KLa, although geometric similarity cannot also be maintained for the latter.
If the culture is found to be damaged by shear, it is important to establish whether the average shear force is more important. The average shear is proportional to the agitator speed, and therefore reduces if the agitator is scaled up using constant power per volume or constant KLa.
Alternatively, some antibiotic fermentations require that a minimum shear be exceeded and hence use fast, small-diameter impellers. The correlations for KLa given above are used to estimate the necessary impeller power input.
The fermenter contents must be cooled to remove both the heat produced by metabolism of the organism and also the heat input from agitation. Cooling is usually achieved through a water jacket on the fermenter, and this may be supplemented by internal cooling coils. In general, the mixing provided for aeration will ensure a good heat transfer coefficient, but in viscous or non-Newtonian broths the agitation at the wall may be severely reduced. The heat transfer coefficient can be estimated using methods available for chemical reactors of the form:
where Nu. Re and Pr are the Nusselt, Reynolds and Prandtl numbers.
As the size of the fermenter is increased, the area available for heat transfer through jacket and coils decreases relative to the fermenter volume and eventually becomes insufficient. For this reason, jacketed fermenters are generally limited to below 150 m3 and larger fermenters require a heat exchanger operating on an external pumped circuit. This exchanger may be a plate exchanger if only a moderate level of asepsis is needed, but if high sterility is required then exchanger, pump and pipework must all be designed to high standards.
Care is needed in the design of the external cooling system to ensure that the organism is not overcooled, that the organism is not damaged by the shear action of the pump and that its viability is not reduced by removal from the fermenter for too long, due to loss of oxygen, substrate, etc.
The introduction of new biochemical processes results in additional safety hazards to those met in the 'conventional' food processing or chemical industries due to the handling of live organisms. However, once these hazards have been understood and overcome, the overall level of safety in biotechnology processes is very likely to be greater than that in the chemical industry. Fermentation processes are carried out in dilute solutions at low temperatures and pressures, and the major hazards associated with the chemical industry (high temperatures and pressures, flammable materials, exothermic or explosive reactions and production of toxic by-products) are usually absent.
The potential hazards that can result from the use of biochemical processes are the risk of exposure to live cells, and the lesser risk associated with exposure to dead organisms or fermentation products. Although dead cellular material and products are not normally toxic, they may present a health hazard to operators if they are not correctly handled. Biologically active molecules such as antibiotics may cause allergic reactions while the health hazards arising from cell debris are similar to those from naturally occurring protein dusts. These hazards normally occur in the separation processes downstream of the fermenter, and the necessary degree of protection corresponds to accepted good manufacturing practice in the food and pharmaceutical industries.
The risks associated with live cells are present in the fermenter and its associated equipment. Unlike the risks caused by exposure to chemicals, there is little quantitative data on the acceptable level of exposure to different micro-organisms and the concept of TLV (threshold limit value) used to determine the protection needed for personnel from chemical processes is not yet applicable. Undoubtedly quantitative data will be developed as more operating experience is gained in fermentation processes but it will take some considerable time before reliable data are available. In the meantime, other methods have been evolved to determine the level of protection that is required for biochemical processes.
There are three broad categories of micro-organism, based on the understanding of the potential risk due to the organism
- A number of micro-organisms are currently in use in conventional food processes. Examples of these are yeasts for beer and baking, and bacteria for dairy products. The hazards associated with these organisms are generally low, and well understood, and acceptable levels of hygiene and protection have been evolved within the industry.
- New fermentation processes will make use of genetically engineered organisms which are tailored for efficient production either by classical genetic methods or by the more recent procedures of genetic manipulation. The potential risks associated with these organisms have to be determined by experience, and precautions are required until an organism is proved to be safe. The possible risks may be due to the host organism, modification of the organism by mutation or other unknown characteristics.
- Pathogenic organisms have direct toxic effects and are highly unlikely to be used in food manufacturing processes. However, they may be used in laboratory experiments to develop suitable organisms for food-related products.
In the mid-1970s, there was considerable concern both by the public and parts of the scientific community about the dangers of genetic engineering. The GMAG (Genetic Manipulation Advisory Group) was formed to report on the risks and to advise on the level of containment and protection required when handling recombinant organisms. A similar body. ACDP (Advisory Committee on Dangerous Pathogens), had been formed earlier to advise on the risks associated with handling pathogens. Both GMAG and ACDP categorised microorganisms according to their perceived hazard, and laid down containment levels for handling each category under laboratory conditions. In assessing the risk cateogry, GMAG considered the ability of a recombinant organism to cause damage to the environment (i.e. to humans, other animals and to plant life). The extent of the perceived hazard depends on the organism's ability to infect a host, its capability to produce a toxin and finally whether the organism can survive and produce toxins under the new conditions. The ACDP categorisation relates to individual identified pathogens which are assessed by an expert committee.
Operation of large-scale commercial fermentation processes is recognized to be different from laboratory work, and GMAG published a report advising on the different procedures and precautions required for industrial processes involving recombinant organisms. In the laboratory, the work is experimental and the GMAG report stated: 'Large scale work differs from most laboratory work involving recombinant DNA technology in that organisms used and product (if any) isolated are completely defined prior to the start of development or production'. Although the higher number of cells present in industrial scale fermenters may increase the occurrence of alternate strains by mutation or reversion, the economics of a commercial fermentation usually depends on the use of a stable organism.
Subsequently, the responsibility for much of the work of GMAG (including regulating the safety of industrial fermentation processes) has passed to the H&SE (Health and Safety Executive). H&SE guidelines for industrial scale processes will eventually be published, but their role in regulating the safety of biotechnology processes has been clearly explained. It is likely that three risk categories will be proposed for industrial processes instead of the four categories used in laboratory work. These are shown in Table 5 with the equivalent existing laboratory codes for comparison. In practice, the majority of commercial processes will be included in category 1. The cost of including extensive containment facilities would make the cost of commercial fermentation plant unacceplably high except for high-value products which can be produced in small, laboratory scale equipment.
Where the organism has been identified as presenting a low risk then it is expected that new fermentation processes may proceed without restrictions. As in all other areas of industrial safety, the onus is on the operating company to ensure safe operating conditions 'so far as is reasonably practicable', and to show to regulatory authorities that the process and operating procedures are adequate. This is best achieved by consultation with the H&SE from an early stage.
STERILE DESIGN AND CONTAINMENT
Both the level of sterility and the degree of containment necessary for a fermentation process vary over a wide range. Strictly speaking, sterility is an absolute term and a sterile environment is one in which no viable organisms exist. However, it is convenient to refer to levels of sterility to represent the probability of organisms passing into the process from the environment.
The reason for maintaining sterility is to protect the fermentation from extraneous organisms which may compete with the preferred microorganism to reduce the yield, or which may produce unwanted or toxic products. Whereas the containment level is determined by the characteristics of the organism, the level of sterility depends on process economics, the characteristics of the nutrients, the organism and the product.
Continuous processes tend to require a higher level of sterility compared to batch processes, since the process economics usually require operation for a long period between shutdowns. With batch processes, the cost of achieving a high level of sterility may be more than the value of a small proportion of batches lost through contamination. Obviously, many products such as those for the health care and pharmaceutical industries, and foodstuffs such as SCP, must be protected against contamination, and this increases the required level of sterility of the fermenter and its associated equipment. Research and development fermentations must also be carried out under highly sterile conditions in order to obtain representative and reproducible data, even if the resulting commercial process is expected to be run under only nominally sterile conditions.
Figure 8 indicates how the requirements of containment and sterility vary between different categories of fermentation process, with the axes showing increasing requirements of each. The longer-established processes can operate under less rigorous standards of process protection, and it is apparent that (apart from R&D operation) the value of the product increases with increasing levels of either sterility or containment. If a new fermentation process is to be competitive with existing comparable products, it must not require a significantly higher level of sterility or containment than its competitors since the cost of achieving increased levels of either is high. It is therefore essential to define the correct level of barrier between the fermentation process and the environment that is needed to achieve both safe and satisfactory operation before designing the fermentation equipment.
Fig. 9. Hazard pathways in a fermenter system.
It is also useful to consider a unified approach whereby similar engineering methods can be used to achieve both sterile barriers and containment in order to reduce the overall cost of achieving these aims. A diagrammatic stirred fermenter is used in Fig. 9 to show the 'hazard pathways' typically present in a fermenter. Similar pathways exist in the feed preparation prior to the fermenter and in the downstream separations. The use of different symbols to show where a risk to sterile operation or containment exists highlights the pathways where a common protection can be provided. The degree of protection necessary at each pathway is determined by the relative degrees of sterility and containment required, and by the available engineering.
An example is the principle of design of a sampling system. For the maintenance of hygienic conditions, steaming the sample point to atmosphere is often adequate, while if aseptic conditions are required then the connecting system between the fermenter and sample holder must be sterilised immediately before sampling (Fig. 10). However, if containment is also required, then in addition the sample transfer line must be drained and sterilised (to kill any organisms remaining in the system) before the sample bottle is removed. The same engineering methods can therefore be used at the appropriate standards to achieve a sterile barrier and containment.
In the operation of a sterile fermentation system there are three distinct sets of techniques to be applied. The fermenter and its associated equipment must be pre-sterilised, the feeds to the fermenter must be made sterile, and finally the sterile barrier between the fermenter and the environment must be maintained. It has been stressed that different levels of sterility and containment require differing standards of engineering. Table 6 summarises the engineering solutions necessary for these different hazard levels.
The most commonly used agent for sterilisation is wet steam, and exposure to wet steam at 120Â°C (1 bar g) for 30 min is sufficient to ensure asepsis. It is essential that, the steam condenses on all parts of the equipment and can drain freely away to avoid non-sterile areas being left in the system. For this reason both the design of the fermentation equipment and the procedures for pre-sterilisation are critical if high levels of sterility are required. Every part of the fermenter system within the sterile barrier must be pre-sterilised to avoid contamination.
All equipment must be free-draining, and this requires careful design not only of fermenter internals such as baffles, supports, agitators etc., but also of piping systems and valve orientation. The surfaces must also be smooth and crevice-free to ensure that deposits can be washed off prior to sterilisation, without leaving pockets of contamination. Although the finish of welds is important to avoid crevices (100% radiography is required for asepsis), it is not necessary to specify highly polished surfaces on vessels since it has been shown that a lesser quality of finish is adequate if spray jets are used for the initial cleaning. Welded joints should be used, as far as is possible, for piping and fittings to avoid crevices.
Not only must equipment be free-draining to ensure asepsis, but every part must be raised to the condensing temperature. Care is needed to ensure that the insulation is adequate to avoid cold spots, especially where heat sinks may be present in the equipment. All the fermenter systems must be designed for full vacuum conditions if steam sterilisation is used, to avoid the possibility of imploding the vessels.
The sequence of pre-cleaning and pre-sterilisation for a complex fermentation system will take a considerable tme, and for a batch system this will lead to a high proportion of downtime between batches. Initially the equipment must be cleaned using a dilute caustic detergent to remove deposits, and then rinsed. The steam pre-sterilisation must then be carried out in a predefined sequence to ensure that no unsterilised zones are left between valves or due to blanketing by inert gas or condensate pools. Because of the critical requirements and the complex sequencing required in pre-sterilisation of some systems, microprocessors are necessary to control the operation and provide the correct interlocks.
Lower levels of sterility can be achieved by less rigorous methods. In the brewing industry, the use of cold caustic solutions is adequate to provide the required hygienic standards. This results in significant reductions in the cost of vessels compared to those needed for steam sterilisation since the fermenters are not designed for vacuum duty and lagging is not necessary. It is evident that, with caustic cleaning, the detergent solution must reach all areas and the system must drain freely. For lower sterile levels, hygienic equipment such as dairy couplings for pipes and fittings are appropriate.
The purpose of sterilizing the incoming raw materials is to prevent the ingress of competing organisms. A few fermenter systems are self-sterilizing and can operate successfully without the need to sterilise the feeds (e.g. some continuous fuel alcohol processes where ethanol inhibits growth of other organisms), but these are the exception. Sterilisation can be achieved either by moist heat or by filtration; the use of heat is both more economical and more generally effective.
As with pre-sterilisation, the feed must be held at the sterilizing temperature for a period of time to ensure that unwanted organisms are killed. Because the destruction rate depends both on the resistance of the infecting organisms and the sterilisation temperature, both the sterilisation temperature and time can be varied to achieve the required degree of sterilisation.
In small batch operated fermenters, the liquid feeds may be sterilised in situ. The fermenter is first pre-sterilised and then, after introduction of the feeds, the contents are raised to the required temperature and held before cooling to the fermentation temperature. Because of the time taken for heating and cooling, and the added complexity of the vessel design, this method is normally only used for fermenters below 1000 liters. The feeds to most continuous processes and larger batch fermenters are sterilised continuously before entry using a series of heat exchangers and a holding vessel. The energy needed to sterilise the feeds for a large continuous process is considerable, and significant cost savings can be made by optimising the system and using heat recovery.
Heat sterilisation cannot be used for heat-sensitive feeds or for gaseous feeds such as air, and these feeds are sterilised by filtration to between 0.2 and 0.5 mm depending on the degree of sterility required. Depth filters are relatively low cost items and rely on filtering through a thickness of medium such as glass wool to remove particles. The filters are not absolute devices and a proportion of small particles will pass through the filter. Because particles can both migrate and grow through the filter, the efficiency falls with time and the filters require re-sterilizing at intervals. A membrane filter ensures a sharper separation of particles by use of a thin membrane with fine pores. To avoid blinding, a series of filters are employed, making the best use of the filtration area by removing larger particles in the initial stages and, in large installations, optimisation of this arrangement can lead to large reductions in cost. Significant improvements are being made to designs of membrane filters for sterilisation, especially in the ability to steam-sterilise and in reliability; these filters usually offer a more effective and economic design.
Where containment is required, similar methods to those described above can be used. Liquid products may be killed-in-place in the fermenter (using live steam or biocides) before harvesting, or heat-sterilised continuously after leaving the fermenter.
In order to maintain a high level of sterility within a fermenter system, every boundary between the fermenter and the environment must be protected. It is important to minimise the number of entries to the fermenter because of the cost and complexity of providing suitable barriers, and the possibility of their failure.
Values. For the highest sterility and containment levels every pipework connection to the fermenter must be protected by a combination of values incorporating a live-steam barrier. The overriding consideration when specifying values for the most stringent sterile levels in the absence of non-draining pockets, crevices and voids which cannot be steam-sterilised. Values for high levels of sterility employ either steam-purged glands or are totally enclosed and crevice-free (e.g. special diaphragm valve designs). For lower levels of sterility and containment engineering, special designs of ball or diaphragm values suffice.
Sterile barriers for out-flowing streams can take several forms but are mainly based on forward velocity or a combination of forward velocity and a valve combination. A system based solely on velocity is only permissible for the lower levels of hazard, since the barrier does not work when flow ceases, and unforeseen flow upsets can allow contamination.
The requirements for sampling are more critical, since not only must the sterile barrier be effective, but a representative and viable sample must be obtained.
Safety relief valves, which do not provide an effective biological barrier, are not allowed for sterile applications, and bursting discs incorporating stainless steel diaphragms should be used. In containment cases, the relief must be led to a secondary vessel, designed to contain and ultimately process the vented material.
Rotary seals. Sterile rotating seals are needed both for agitator shafts and pumps within the sterile environment, and either double mechanical or labyrinth seals are suitable. Moist steam is used as the sterilizing medium between the seals for high-risk applications, but hot water or biocide is acceptable for lower risks. If possible, the use of seals should be avoided where the risk is high, and magnetic couplings or alternative designs should be used.
Vent gases. For sterile design the concern is to stop contamination. Depending on the scale and aeration rate, either a sub-micron filter or a sterile flow barrier in the vent line is used. One particular problem with vent filters on multipurpose fermenters is the risk of batch-to-batch contamination, and a policy of changing or steam-sterilizing the vent filter has to be adopted in these cases. Where a lower sterile level is needed, adequate protection can be provided by a 'Pasteur bend' or similar vent pipe. This relies on a constricted 180Â° bend to provide both a gravity and a velocity barrier to organisms entering via the vent line.
For vent gas containment, the highest standard is to pre-filter the gases to knock out spray and then use two 0-5 mm filters in series or to pass the gas stream to an incinerator.
On large-scale units, the size and cost of adequate filters and fans (to overcome the pressure drop) may make filtration uneconomic and an incinerator the only suitable means. At lower hazard levels only one filter is needed, or caustic scrubbing may be suitable.
Enclosures. The barriers discussed so far have all been primary barriers to separate the fermenter from its surrounding environment. The use of secondary enclosures (providing a high degree of ventilation to local areas of risk) and tertiary enclosures (enclosing the process area in a specially ventilated building or room) is well understood for small-scale sterile operations in the pharmaceutical industry. In this context, sterile cabinets are used to provide secondary enclosure, and these are sited within "clean rooms', to provide air at a high level of purity around critical sterile operations. Within fermentation halls, special building finishes and standards are used to discourage the growth of unwanted organisms.
The critical difference between the types of enclosure for sterility and containment is the air flow pattern. For sterile engineering the air flow is outwards, in order to protect the contents by an elevated pressure in the enclosure, and the inlet air is passed through sterile filters. The converse is used for containment. If a fermentation requires both precautions containment will take precedence and the enclosure will be at negative pressure, but include sub-micron HEP A filtration of both inlet and outlet air. Also in this case the secondary enclosure would be enclosed in a 'clean' tertiary enclosure maintained at a positive pressure with respect to atmosphere. The cost of providing sterile or containment enclosures for other than high-value products would be prohibitive.
It is probable that over the next decade at least, a high proportion of commercial fermentations will still be carried out in conventional, stirred tank fermers. The reasons have been discussed earlier and are both practical, and economic. The equipment will allow most fermentations to be carried out economically and the fermenter system can be designed using existing scale-up techniques, allowing a new process to be developed in the shortest time and at minimum cost. Improvements are likely in such aspects of the detailed design as impeller design (to achieve better oxygen transfer and mixing), arrangements for sterile and containment barriers, and in the development of process control.
Alternative fermenter designs will tend to be used in processes for bulk products, where the product cost is crucial to the competitiveness of the product and it is necessary for the fermenter to operate at the best possible efficiency. This incentive is not normally present for high value products where the overriding consideration is to get the product on the market as quickly as possible. For these products, alternative designs will only be used where necessary to satisfy the demands of the organism.
The airlift fermenter is likely to find increased applications for production "of "bulk products, since it has several advantages for large scale, aerated fermentations. The construction is suitable for large volume fermenters and the oxygen transfer efficiency is high. In addition, because the power for agitation and aeration is supplied to the air compressor, it is possible to use alternative sources of power (other than direct electric drive) depending on local availability and costs.
Potentially important advances could be made in the use of immobilised fermenter systems. These offer the potential both to decouple cellular growth from production and expression of the desired product, and the means to retain a high cellular concentration in the fermenter. The conversion efficiency will therefore increase (since substrate is not 'wasted' on cell growth) and the cost of downstream separation equipment will be reduced, both of which have a significant effect on the economics of most processes.
Finally, there is considerable scope for improving the characteristics of fermentation micro-organizations by genetic engineering, and it is certain that improved strains will continue to be developed to make many fermentation processes more economic. The most significant improvements would be the development of thermophilic organisms, and those having high tolerance of feed and product concentrations.