Process Design And Optimisation

There are a great number of factors which can influence the performance of separation processes. These include the properties of the target compound, the numbers and properties of the contaminants, the nature of the separation process and the environment in which it is to be operated. The exact effects of these parameters on the outcome of the process is often difficult to predict/ This is particularly so for the isolation of biological components of biomass as the starting materials are complex mixtures which can vary in character depending on the conditions under which the organism was maintained and harvested. Biological processes can rarely be defined absolutely, nor controlled with precision. Consequently, the optimisation of purification procedures is largely empirical. The knowledge and experience of the technologist, combined with trial and error, defines the optimum methods and conditions for purification in the laboratory. These are then scaled-up via pilot trials to production level. Frequently, the isolation of the target will involve a series of steps which are optimised individually. The types of unit process which are selected will be determined by factors such as the equipment available and economic constraints in addition to separation performance.

A more rational approach to the design of separation strategies and the optimisation of performance may be aided by a better understanding of the scientific and engineering issues involved. Mathematical models have been constructed which describe various separation techniques including liquid chromatography (Bellot & Condoret, 1991) and extraction in aqueous two-phase systems (Baskir & Hatton, 1989). However, those researchers with an academic interest in the theory of separations do not tend to work with complex multi-component systems of the type relevant to many industrial biotechnological processes (Wang, 1990). A rational approach is therefore limited by the lack of an in-depth understanding of the nature of complex biological feedstocks and their behaviour during separation processes. "Expert systems" seek to combine databases of knowledge with mathematical models to obtain rapid answers to complex problems. A computer algorithm contains a set of rules which define a decision making process and the database provides the information from which the decision is made. The aim is to select the best solution to a problem out of many alternatives in a manner which simulates the rationalization of a human expert. The nature of expert systems and their potential in process control in areas such as the food industry are discussed by Shinskey (1989), Whitney (1989), Aarts et al (1990) and Konstantinov & Yoshida (1992). The possibility that this technique may be applied to the design of large-scale separation process was considered by Asenjo (1990). He concluded that there was insufficient information available for the construction of an expert system for highresolution separation operations, but that a hybrid approach combining algorithmic and heuristic methods may be feasible. He is seeking to obtain the information required for a separations database by the characterisation of biological products and the components of fermentation streams (Asenjo & Leser, 1993).

It may be that biological materials are too complex and variable for the exact values of process parameters to be determined using mathematical models. Although the processes may be reduced to a series of relatively simple equations, the outcome may be subject to a significant dependency on starting conditions. Small errors in the initial estimations or measurements thus accumulate and are amplified through the series of calculations. Through the processes of culture inoculation, cell growth, harvesting and multiple-step purification, the actual results may therefore diverge significantly from those predicted. However, mathematical models can place theoretical constraints on the operation of individual methods which will help to define their scope and limitations. This will ease process design and scale-up and will also provide insight into the function of separation apparatus which will contribute to improvements in equipment design and operation.

It is the aim of a process engineer to design a purification scheme which will achieve the desired product specification within the limits imposed by the available resources. Often, it will be necessary that downstream processing costs are kept to a minimum, particularly if the product has a low market value. A certain amount of the product is usually lost during each process operation, either by impairment of biological activity or by loss of material to waste. The use of as few process steps as possible is therefore required for maximum recovery to be achieved. To minimise operational costs, it is also advantageous that each unit process is of short duration, thus expending minimum man-hours and energy. If the material is more concentrated, then smaller volume plant is required and costs per unit product are reduced. It is perhaps obvious that each unit process should also operate as efficiently as possible. In addition, it is becoming increasingly apparent that the optimisation of individual operations is only one aspect of efficient process design and operation. In a series of separations, the performance of each unit will be dependent on the characteristics of the material produced by its predecessor. Overall process efficiency is therefore also dependent on the integration of unit operations into a coherent process stream (Fish & Lilly, 1984).

Various developments in separation technology have made a contribution to process integration. Advances in genetic manipulation have made it possible to consider product isolation at the earliest stages of process development. A purification strategy can thus be built-in to the product, as was discussed in section 3 of this chapter. Other means of integrating production and purification processes have also been devised. Immobilized biocatalysts are an example of this as they allow catalyst reuse while overcoming the need to remove them from the product further downstream.

Bioreactor configurations have been developed which combine a separation element with product biosynthesis such that the product is rapidly removed from the reactor as it is liberated by the cell. This is particularly advantageous for fermentations which are limited by end-product inhibition, such as the production of organic acids and alcohols. Rapid separation of the product from the fermentation medium overcomes inhibition and thus increases productivity. Some recent examples of these techniques from the scientific literature are listed in Table 6. For reviews of integrated production and recovery systems, see Daugulis (1988) and Groot et al (1992).

Combined production/purification strategies have also been developed for secreted enzymes. Aqueous two-phase systems, in which the organism and the extracellular protein have different partition coefficients, have been described for the production/extraction of enzymes. Examples include the production of alkaline protease by Bacillus licheniformis (Lee & Chang, 1990), and the production of a-amylase by Bacillus amyloliquefaciens (Park &

Wang, 1991). Technology now exists for the commercial production of antibodies on a continuous basis using animal cell cultures in hollow-fibre membrane reactors (Knight, 1988; Griffith, 1992). The cells are retained by the membrane while the antibodies pass through in a continuous flow, thus combining production and separation. Grandies et al (1991) described an automated process in which the antibodies were recovered from the membrane permeate in an integrated affinity chromatography purification loop.

Table 6

Production of organic acids & alcohols by extractive fermentation

Product

Organism

Extraction method

Reference

Ethanol

Butanol

Propionic & acetic acid

Propionic acid

Acetonobutylic acid

Lactic acid

Kluyveromyces fragilis

Saccharomyces bayanus

Saccharomyces cerevisiae

Clostridium acetobutylicum

Propionibacterium acidipropionici

Clostridium acetobutylicum

Lactobacillus delbreuckii membrane distillation Udroit et al (1989)

liquid membrane solvent extraction filtration/solvent extraction membrane assisted solvent extraction electrodialysis solvent extraction ultrafiltration/ distillation solvent extraction solvent extraction ion-exchange

Christen et al

Kang et al (1990)

Chang et al (1992)

Weier et al (1992)

Minier et al (1990)

Yabannavar & Wang (1991)

Seevaratnam et al

Srivastava et al

Continuous operations enhance the possibilities of integrated processing as the characteristics of the product are relatively constant with time. In addition, they facilitate real-time monitoring and control of product parameters so that the process may be fine-tuned during the run. Continuous production methods are now well established through the use of immobilized cell and enzyme systems. Continuous centrifugation and filtration plant are also in common use for low-resolution separations. These may now be complemented by the development of continuous high-resolution separation techniques. Examples include liquidliquid extractions (Papamichael et al, 1992) and continuous recycling affinity separations (Hughes & Charm, 1979; Burns & Graves, 1985; Pungor et al, 1987; Niven & Scurlock, 1993). Continuous elution in column chromatography can also be simulated by using a series of synchronised chromatography columns, such that one column is always eluting the product while the others are at various stages in the chromatography process (Nicoud, 1992).

The monitoring and control of biotechnological processes may be greatly aided by the development of biosensors. These utilise a biological component such as cells or enzymes which cause a biochemical reaction in the presence of a particular compound in the medium. This is combined with a transducer which produces a measurable electrical signal in response to the biochemical event. It is thus possible to exploit the specificity of biological interactions to produce detectors which measure the presence of selected compounds in a complex mixture. Various systems have been developed for monitoring compounds such as lactate, glucose and ethanol, some of which are available commercially. Readers are referred to Brooks et al (1991) for a general review, and to Wagner & Schmid (1990) for a review of their application in food analysis. The ability to assay rapidly biological chemicals without the need for lengthy separation and analytical procedures has great potential for the automation of continuous processes. The control of these systems can therefore be based on real-time measurements of the product species, rather than on non-specific parameters such as UV adsorption or pH.

0 0

Post a comment