Microbiological Process Discussion

Bioengineering and FermentationELMER L. GADEN, JR.

Department of Cheemical Engineering, Columbia University, New York, New Y'ork

Received for publication June 29, 1959

One thing at least can be said with certainty about"bioengineering" or "biochemical engineering"--itmeans many things to many people. The origins ofthese terms are indeed difficult to trace. Whatevertheir earliest use, however, one of the first important,published records was an editorial in the periodicalChemical Engineering for May 1947 (Anon., 1947),which said in part:

"This is inot a plea for another branch of engineering. Thereare too many already. Btut there is an important need in thechemical industry today for engineers with somewhat broaderscientific training than in chemistry, physics and mathematics.And this need will continue to grow as we come nearer to under-standing and influencing the processes of our very existence.We are referring, of course, to the increasingly important roleof the biological sciences.

"In the past five years the pharmaceutical industries havecontributed more to chemical progress, and in turn have beenbenefited more by chemical engineering progress, than everbefore in their long history.... They have drawn heavily onthe best of our technology and have often surprised themselveswith the resultant savings in time, cost and quality. Theywant now to apply the same methods and equipment to otherproducts, but they tell us they are handicapped by lack ofengineering manpower with fundamental knowledge and ex-perience in biology-particularly bacteriology and biochem-istry. "

Shortly thereafter the University of Wisconsin gavenotice that it had established an optional biochemicalengineering curriculum within its undergraduate chem-ical engineering program (Hougen, 1947). At about thesame time, interest in biochemical engineering wasfurther stimulated by an article entitled "A Case Studyin Biochemical Engineering" (Kirkpatrick, 1947), re-porting the 1947 Award for Chemical EngineeringAchievement to Merck & Co., Inc., for the commercialdevelopment of streptomycin.Two things are evident here. First, biochemical en-

gineering was considered by its early protagonists tobe a variation, pure and simple, on the chemical en-gineering theme; second, these same people wereapparently unaware of the facts of life in the fermenta-tion industry, at least. In fact, these promotional efforts,

l Based on a lecture presented at the Antibiotics Symposium,Pragtue, Czechoslovakia.

however well intentioned, created misunderstandingswhich, unfortunately, still exist.Whatever the difficulties in defining and delineating

bioengineering, one thing at least is reasonably wellagreed upon; it is closely related to chemical engineer-ing. Perhaps a better picture of the true nature ofbioengineering can then be gained from a closer look atthe chemical engineering "sire."

Chemical Engineering and Bioengineering

First, it should be recognized that chemical engineer-ing- or any branch of engineering for that matter-does not comprise any distinct body of knowledge, likethat normally associated with the various sciences-physics, chemistry, or biology, for instance. Rather itembodies a general attitude and approach to practicalproblems which recognizes the need for getting some-thing done with reasonable expedition even in the ab-sence of complete information.

Engineering, like medicine, is a practical art. It isfirmly based on the constantly expanding pool of scien-tific knowledge and, in fact, frequently contributes tothis pool through its associated research activities.Nevertheless, the basic engineering function is to pro-vide solutions to the practical problems of technology.To accomplish these ends, each branch of engineering

has evolved its own peculiar techniques its "bag oftricks." These methods are for the most part empirical,based on rationalization from experience (includingexperiments) and practice rather than on theoreticalconsiderations. To be sure, these empirical tools areconstantly referred to and tested for validity againstfundamental concepts. Furthermore, increased gen-eralization based on greater understanding is constantlysought as a means for simplifying existing methods andmaking them more readily useful.

This is particularly true of chemical process designand evaluation,; the primary business of chemical en-gineers. We are still very far from an adequate under-standing of most of the operations involved and somust depend on largely empirical approaches. Thedesign of agitation systems, so important in fermenta-tion, is an excellent example. With virtually no detailed

123

on March 25, 2020 by guest

http://aem.asm

.org/D

ownloaded from

E. L. GADEN, JR.

knowledge of the mechanisms of impeller behavior andturbulence generation, empirical correlations have beenestablished which enable us to size agitators satis-factorily.

Correlations, the most familiar tools of chemicalengineers, are really substitutes for knowledge. Theyprovide a means for solving practical design problemswithout knowing what is really going on. Exact math-ematical analysis is often either impossible, because therequired knowledge of mechanisms is not a hand, or tooexpensive and difficult to carry out. Its direct role inprocess design is therefore limited.

Chemical engineers are, by definition, the engineersof the chemical process industries, whatever their orig-inal degrees. A look at any successful chemical processcompany will generally demonstrate this quite clearly.Those who best fill the engineering function, particu-larly process development and design, are those whohave been selected according to interest and ability-not according to degrees. Of course one hopes that thecurriculum followed in school, and hence the degree ob-tained, will reasonably reflect interest and ability butwe cannot assume that this is always so.To extend this concept, bioengineering is simply the

application of these same techniques, plus some specialones, to the problems encountered in a particular groupof chemical industries, those dealing with biochemicaland microbiological processes.

Chemical processes are concerned with making sub-stances from other substances rather than with shapingsubstances into things. The biochemical-microbiologicalprocess industries are no different in principle; theonly real distinction is in the origin and nature of thematerials dealt with. Their most common characteris-tic is molecular complexity and they are usually ob-tained from some natural source rather than by directconversion or synthesis from simpler raw materials.Even this delineation is more a matter of practicalconvenience than principle, very much like similardivisions which have grown up between the plastics,petrochemical, nuclear, and other branches of the chem-ical process industries. Each makes use of the generalapproaches to process design and evaluation whichhave been developed, modified more or less to meet itsparticular requirements.

Bioengineering and FermentationThis bioengineering approach has been most fruit-

fully applied to microbiological processes-fermenta-tion. In an excellent discussion of the engineering view-point in fermentation, Warner et al. (1954) haveemphasized the dominant role of the biochemist-micro-biologist in establishing the fundamental componentsof the fermentation process organism, substrate, andprocess conditions. In maximizing process efficiency and

providing for operational economy and reliability, theengineer's contribution is secondary but vital.

In another paper, Gaden (1955a) analyzes fermenta-tion in terms analogous to those commonly used forother chemical processes. The basic considerations inprocess design stoichiometry, energy relationships,kinetics, and equilibrium are equally applicable tofermentation; unfortunately they are not so often easilyidentified and evaluated. Similar views have beenoffered by Hartley and Warner (1958).The bioengineering approach to fermentation has

been significant only for the last 15 years or so. To besure, engineers mostly chemical have always beenengaged in commercial fermentation operations. Untilthe 1940's, however, their role was largely confined tothe purely mechanical aspects of plant and equipmentdesign and operation. Matters of process design werealmost exclusively the concern of fermentation tech-nologists, generally microbiologists or biochemists.

Perhaps the easiest way to assess and illustrate therole of bioengineering in fermentation technology is tofirst summarize its contributions over the last 2 decades.Then a specific problem, scale-up of fermentation proc-esses, will be examined in detail to illustrate both theapplication and the limitations of the bioengineeringapproach.The new question, then, is simply this: "What bio-

engineering problems actually exist in fermentationtechnology and what real progress has been made to-ward their solution?"When we try to analyze this question in a reasonably

dispassionate manner, one fact becomes abundantlyclear; the total bioengineering contribution to currentfermentation practice is really very small. Furthermore,there is no indication that it will become significantlygreater in the near future. This does not mean that ourefforts are unimportant or valueless. Quite the con-trary; in many cases bioengineering developments havehad and will have profound effects but these must al-ways be viewed in, and not out of, perspective.

This situation is no one's fault. It is simply the re-sult of certain inherent features of the fermentationindustry as it has developed and as it now exists. Al-though there is no fundamental difference betweenchemical processes employing natural enzyme catalystselaborated by living cells and more conventional types,the patterns of process development are distinctly dif-ferent.The heart of any chemical process is the reaction

system itself. Other factors may have greater economicweight in particular cases but the chemical changesbeing effected provide the reason for existence. For anynew process, the key considerations are (a) the basicreactor system to be employed and the range of (b)physical, and (c) chemical variables involved.

[VOL. 8124

on March 25, 2020 by guest

http://aem.asm

.org/D

ownloaded from

BIOENGINEERING AND FERMENTATION

In the development of a catalytic, gas-phase conver-sion, for example, a great variety of reaction systemsand conditions may reasonably be considered: fixed orfluid beds, high or low temperatures and a wide range ofpressures, flow rates and reaction mixtures. For fermen-tation, however, the choice is much different, in someways more limited and in others almost limitless. Toillustrate:

Reaction systems. Virtually all modern aerobic proc-esses employ a single reactor type; the stirred, baffledtank with gas sparging beneath the impeller. There are,of course, limited examples of other types; fixed bedsare used in a few cases (vinegar generators, tricklingfilters) and one may view the continous medium steri-lizer as a differential, tubular reactor.

Physical variables. Temperature, between rathernarrow limits, and the intensity of agitation (bothmechanical stirring and air sparging) are the primaryphysical variables in aerobic fermentation systems.Fluid properties of the mash, for example, viscosity,density, and surface tension, as well as agitator design,power input and air rate are the factors which determineagitation intensity. Pressure, in the moderate rangeordinarily employed, seems to have no effect on fermen-tation reactions.

Chemical variables. Fermentation media are usuallyso complex that it is hard to consider chemical variablesin the same sense as for other chemical processes. Twobasic factors, at least, are recognizable though; first,the composition of the medium at the beginning of thecell-substrate interaction and, second, deliberate addi-tions during the interaction-nutrients, precursors and,of course, oxygen. Although the effects of certain chemi-cal factors, glucose concentration for instance, arebroadly understood in some fermentation reactions,virtually nothing is known about the myriad othercomposition variables. Process media formulations aresimply arrived at by extensive laboratory trial.

Fermentation process development is then largely amatter of finding, by experiment, the most productiveand economical organism-medium combination and ofreproducing, in larger capacity equipment, the optimumconditions for cell-medium interaction. The over-allprocess scheme (flowsheet) and reactor system (fer-mentor) remain virtually the same; only conditions in-side the reactor change.To emphasize the peculiar nature of fermentation

process development, one need only look at the typicalfermentation plant. Design details and control methodsdiffer, of course, but the reaction systems employed arevirtually identical. Increased capacity is obtained byadding duplicate units of the same type, and develop-ment of an entirely new process scheme is rarely under-taken. Recognizing this, many people use the phrase

process improvement to describe the primary activityof development groups in the fermentation industry.The inherent features of fermentation processes there-

fore limit considerably the normal activities of chemicalengineering, the design of new processes and plants. Itis only in this light that the bioengineering contributioncan be realistically evaluated

Bioengineering Problems and Progress

Bioengineering contributions to fermentation tech-nology can be looked at in many different ways. We cango through the characteristic fermentation processflowsheet and look at the main stages: (a) medium prep-aration and sterilization, (b) inoculum preparation, (c)reaction (fermentation), and (d) pretreatment for re-covery (lysis, filtration, and so forth). Alternatively wecan adopt a unit operations approach and collectivelyexamine all activities which have a common basis, heatsterilization of media, aseptic transfer of fluids, masstransfer (aeration) and so forth.

Regardless of the approach taken, it rapidly becomesclear that some subjects, aeration for example, havebeen studied to death while others have hardly beenglanced at. Table 1 is a summary, admittedly quitesubjective, of problems and progress in bioengineering.Parts of the table, especially the column on practicalutility, are dependent on information given the authorby people in industry and so cannot always be properlyreferenced. Furthermore, the introduction or simpleadaption to fermentation practice of equipment anddevices used elsewhere, with little or no change, is notincluded here, even though such mechanical improve-ments may have great practical significance.

Scale-Up"Scale-up" is an inherent part of process develop-

ment, in fact the terms are virtually synonymous formany people. Scale-up has been successfully achievedwhen yields and productivities, previously demon-strated onl a small scale, have been produced in largercapacity units. This basic definition holds equally wellfor microbiological processes, though the problems en-countered are markedly different.Two basic approaches to scale-up are available; first,

extrapolation of model experiments based on the prin-ciples of similitude and, second, mathematical analysisof the complete (or controlling) mechanism (Johnstoneand Thring, 1957; Metzner and Pigford, 1958).While the second of these has unlimited potential

value, it also has serious limitations in practice. Oftenthe relationships are too involved to permit rigorousdefinition or the resultant mathematical expressionsare too complex for economical solution, even with com-puting equipment. Electrical or mechanical analogsmay sometimes be used to overcome the second of these

1960] 125

on March 25, 2020 by guest

http://aem.asm

.org/D

ownloaded from

TABLE 1Bioengineering problems and progress

Development/Study Purpose Utility in Practice

Medium sterili- Calculation methods for batch ster- Improvement of batch process Nil (apparently not used in indus-zation ilization (Deindoerfer, 1957). scale-up. try).

Equipment for continuous steriliza- a) Improve process scale-up by Being introduced in many plants buttion: providing uniform medium com- experience is variable; equipmenta) retention tubes. position. and operational problems stillb) plate heat exchangers (Pfeifer b) Reduce turnover time in pro- exist.

and Vojnovich, 1952; Whit- duction fermentors.marsh, 1954). c) Improve productivity.

Air sterilization Study of the factors affecting effi- a) Elucidate the mechanism of fil- Reliable filters of greatly reducedciencies of fibrous filters (Ter- ter action. volume and lower pressure dropjesen and Cherry, 1947; Decker et b) Provide rational design method have been developed and used in-al., 1954; Humphrey and Gaden, for air filters. dustrially.1955).

Design and testing methods for fi- Reduce lead on filters or act as pri- Use variable; all compressor sys-brous filters (McDaniel and Long, mary "sterilizer" for some appli- tems introduce some heating.1954; Gaden and Humphrey, 1956; cations.Maxon and Gaden, 1956).

Heat sterilization of air (Stark andPohler, 1950).

Process design Fermentation process kinetics (Ga- a) Elucidate general kinetic pat- Nil.den, 1955b; Luedeking and Piret, terns in fermentation processes.1959). b) Analyze kinetics of batch proc-

esses.

Penicillin process kinetics (Stefa- Examine effects of process variables Nil.niak et al., 1946; Brown andPeter- on rates of penicillin biosynthe-son, 1950; Calam et al., 1951; Hos- sis.ler and Johnson, 1953; Owens andJohnson, 1955).

Theory of continuous cell propaga- Design of steady-state continuous Continuous processes for cell propa-tion (Herbert, 1959). propagators for cells. gation and "simple" (type I) fer-

mentations.

Continuous, multistage product Design of continuous processes for Nil.fermentations (Bartlett and Ger- type II and III (biosynthesis)hardt, 1959; Sikyta, 1959). processes.

Aeration-agita- Oxygen absorption measurements: a) Study effects of operating vari- Power input and air rate are mosttion a) Sulfite method (Cooper et al., ables (gas rate, agitator speed, important factors in aeration;

1944; Schultz and Gaden, 1956; fluid properties, etc.) on oxygen agitator and sparger design notPhilips and Johnson, 1959). absorption. critical. Flat-blade turbine, open-

b) Polarographic method (Hix- b) Comparison of aeration-agita- pipe sparger adequate.son and Gaden, 1950; Wise, tion systems.1951; Bartholomew et al.,1950a, b; Chain and Gualandi,1954).

Performance tests (Chain et al.,1952; Friedman and Lightfoot,1957; Elsworth et al., 1957).

Aeration tests in product fermenta- Examine effects of aeration-agita- a) Minimum power-air rate levelstions (Bartholomew et al., 1950b; tion on productivity. for penicillin established.Karow et al., 1953; Maxon and b) Aeration not as critical in actino-Johnson, 1953). mycete fermentations.

Cell-fluid sepa- Study of centrifugation (Ambler, Develop theory for centrifuge Used for scale-up of cell and virusrations 1959; Patrick and Freeman, 1959). scale-up. propagation processes.

Froth flotation of cells (Boyles and Development of a new technique. Nil.Lincoln, 1958; Guadin et al.,1960a, b).

126

on March 25, 2020 by guest

http://aem.asm

.org/D

ownloaded from

BIOENGINEERING AND FERMENTATION

limitations, but their application to microbiologicalprocesses is still a long way off. This leaves us withscale-up from model experiments by the principles ofsimilitude as the main technique in practice.To scale-up by similitude, one must first establish

some functional relationship between the various di-mensionless parameters which can be used to char-acterize the system. A typical example, often referredto in fermentation studies, is the correlation for powerrequired in an agitated vessel (Rushton et al., 1950).

Np = K(NRe)m(NFr)n (1)

The three dimensionless groups are:

N= power number = PG,_p ~~~~PN3 D5_ Reynolds number _ ND2 pNRe- (for agitation) -

DN2NFr = Froude number = Gc

All are compounded from the primary system variables,impeller diameter (D), speed (N), fluid density (p),and viscosity (,u) (Gc is the standard dimensional con-stant from the laws of motion).

Values of the constant, K, and the exponents, m andn, have been evaluated experimentally for many dif-ferent mixing systems (Rushton et al., 1950), includingbaffled and unbaffled (vortex) tanks equipped withpropellers, on and off center, and a variety of paddleand turbine-type agitators. On this basis scale-up ofpower requirements is readily achieved, providing thevessels involved are geometrically similar.Power requirement, however, is only one aspect of

scale-up in an agitated reaction vessel. One must alsoknow the relationships existing between agitation andthe various rate processes, physical and chemical, whichtake place in the fluid mass.Rushton (1951) has proposed a scale-up method for

situations where the "process result," heat or masstransfer rate for example, can be related to the govern-ing variables. In such a case this process result, 4, ex-pressed in the form of a dimensionless group, is somefunction of the Reynolds number for agitation:

0 = kNxe

The power, x, may be determined from data obtainedin small-scale experiments and used to scale up theprocess by the following relationships:

N2= N1(D1jD2)(2x-l)lx (2)

and

P2= P1(D1/D2)5-3(2x-l)Ix (3)

where N2 and P2 are the impeller speed and power re-

sion is D2. In other words, D2/D1 is the scale-up ratioand the volumes are as (D2/D1)3.The power ratio, P2/P1, will be equal to one only

when x = 0.75, and so Rushton (1951) stresses thedangers in the common practice of scaling up on an

equal power per unit volume basis without knowing therelationship between process result and agitation in-tensity, that is, "x".At the same time experience has shown that in many

common agitation situations, x is near enough to 0.75,usually in the range 0.6 to 0.8, to make scale-up byequal power per unit volume reasonably accurate(Jordon, 1955). In the case of oxygen absorption insulfite solution, for example, Solomons and Perkin (1958)found x = 0.78. A 2: 1 scale-up, based on equation (2)above, was, as expected, good but not perfect. Solo-mons (1958) has critically examined some of the prob-lems in this scale-up method and stresses the impor-tance of the absolute values of impeller speed and airrate, particularly where more than one phase is in-volved, for example, aeration of fermentation broths.

Scale- Up in Fermentation

Although scale-up problems in the fermentationindustry have often been discussed, very little hasbeen done about them. Specific proposals for scale-upmethods, with supporting data, are very few indeed.Furthermore, even these do not deal with the realproblem, scale-up of the fermentation process. Ratherthey treat the scale-up of the fermentation reactor,more specifically its aeration and oxygen supplycapabilities, alone.At first hand this may seem a trivial matter, but it

is certainly not. The general inadequacy of scale-upmethods for microbiological processes is a direct resultof their failure to consider the process as a whole. Toillustrate this point, let us briefly review the methodswhich have been proposed.

Scale-Up by Aeration-Agitation

As a result of their studies on aeration and oxygen

transfer in penicillin and streptomycin fermentations(Bartholomew, 1950 a, b; and Karow et al., 1953)proposed a general scale-up procedure for aerobicfermentations. Theirs was the first specific recommen-

dation to appear in the open literature and mostsubsequent suggestions are simply modifications of it.The basis for their scheme, and for those which havefollowed, is the promise that:

"Cellular respiration, and hence antibiotic production, de-pends upon the maintenance of an oxygen concentration at thecell wall such that respiration rate is independent of oxygenconcentration-that is, a zero order rate.... Antibiotic yieldthus ultimately depends upon the rate of oxygen transfer to the

quired in the larger tank whose characteristic dimen-

1960] 127

cells. "

on March 25, 2020 by guest

http://aem.asm

.org/D

ownloaded from

E. L. GADEN, JR.

According to this view, scale-up of the fermentationreaction system is then only a matter of scaling up therate of oxygen absorption in the agitated, aeratedfermentor. This underlying assumption is, however,questionable at best anid at worst-false; certainlyit is not supported by any of the admittedly limitedkinetic observations made so far (Gaden, 1959). For themoment, however, let us examine the scale-up methodswhich are based on it.Karow et al. (1953) based their scale-up procedure

on three premises: (a) that antibiotic yield dependsupon the rate of oxygen supply to the cells (stated indetail above), (b) that oxygen transfer rates (andtherefore yields) depend upon the scale and intensity ofturbulence, and (c) that absorption- of oxygen into theliquid, and not transport of oxygen within the medium-cell suspensioni, is the limiting rate.The rate of oxygen absorption is then simply ex-

pressed as:

Rate of absorptioni = KLa(C* - CL)

where:

KLa = absorption coefficienta = interfacial area

C* = saturation oxygen concentrationiCL = actual dissolved oxygen concentration

The product KLa, an absorption capacity coefficient,cannot ordinarily be separated further because theinterfacial area is usually unknown. It is commonlyused to characterize the absorption efficiency of anagitation-aeration system (Finn, 1954) and Karowet al. (1953) used it as the primary scale-up parameter.For constant fluid properties, this capacity coefficientis a function of agitator power (P) and the superficialair velocity (V1,) through the fermentor (Cooper et al.,1944).To scale-up a fermentation system by this method,

one first determines the relationship between yieldand/or productivity (the so-called process result) andKLa in small vessels. Data on KLa must be then ob-tained for the various vessel sizes involved in thescale-up. Because of the difficulties inherent in makingsuch measurements at the plant level, Karow et al.(1953) obtained their values for 200-, 10,000-, and15,000-gallon tanks by extrapolating the data fromCooper et al. (1944), an admittedly inexact but muchmore convenient scheme (figure 1).To scale-up, they simply selected, on the basis of

experience, an air rate sufficient to supply the oxygenrequirements of the organism without excessive foam-ing difficulties, practically speaking the highest ratethat can be held in the tank. With the air rate fixed,the power input required to provide the desired KLavalue can then be established.Wegrich and Shurter (1953) have reported the

scale-up of penicillin fermentations from 2,000- to24,000-gallon tanks by a modification of the methodjust discussed. Their scale-up variables are the same-power input and superficial air volocity-but they didnot employ the oxygen absorption coefficient, KLa, asthe scale-up parameter. Instead they simply applied apressure correction to the air velocity, because of theincreased fluid head in the taller, plant-scale fermen-tors, and scaled up by providing equal power input perunit volume and superficial air velocity (corrected)at each level.

Scale-up data of this sort, in terms of a specificprocess result like penicillin productivity, are scarceindeed. Although the general pattern (figure 1) seemsgood, it is still hard to accept this as real proof of thevalidity of the method. It is just reasonable to say,simply, that increased agitation has provided relativelycomplete mixing of the fermentation batch, permittingthe biochemical mechanisms involved to operate freeof gross diffusional control.

Finally, there is the matter of the funidamentalvalidity of this approach to fermentation scale-up. Itrests, as we have seen, in the premise that productformation is dependent upon the rate of oxygen supplyto the cells (Karow et al., 1953). While this is probablya fair assumption for many microbiological processes

O 100 200 300 400 500 600 700

ABSORPTION COEFFICIENT - KLaFigure 1. Scale-up correlation for penicillin fermentation

[VOL. 8128

on March 25, 2020 by guest

http://aem.asm

.org/D

ownloaded from

BIOENGINEERING AND FERMENTATION

the propagation of yeast and aerobic bacteria and theoxidation of sugars to organic acids for example- it isevidently not so for biosynthetic processes, antibioticproduction in particular.

Direct experimental evidence to show that no clearrelationship exists between respiration and penicillinsynthesis in Penicillium chrysogenum has been given byCalam et al. (1951) and many other indications of theindependence of respiratory and antibiotic-producingmechanisms in the mold have been noted (Gaden, 1959).Finn (1954) has cited a number of other microbiologicalprocesses in which the same situation prevails. On theother hand, nothing has been offered to support thebasic assumption underlying scale-up by this method.

Scale-up on the basis of equivalent oxygen absorptionis certainly convenient but its validity is very much inquestion. There is always a tendency to overevaluatefactors which can be measured and to avoid those whichcannot. This may very well be the case in the ap-proaches to fermentation scale-up which have beenused so far.

Fermentation Process Scale-Up-CriticismThe real problem in fermentation process scale-up is

the complete reproduction, in large capacity equipment,of those conditions for cell-medium interaction whichhave been established as optimum by small volumeexperiments. Many variables affect this interactionand oxygen supply is only one of them. Generallyspeaking, the most critical factors are considered to be:

140

120 Am

100

80

P.C

Inoculum: type, age, and amount.Medium: initial composition, pH, redox potential;

changes during sterilization; and additions duringfermentation.

Conditions: temperature, pressure, oxygen supply,and agitation/mixing.

Of course, the scale-up methods discussed earlier allassume the duplication of inoculum and mediumconditions and of temperature. In practice, however,this assumption is easier made than fulfilled. Emery(1955), for example, has summarized the many complexand ill-defined microbiological problems which interferewith scale-up. His conclusion, certainly a legitimateone, is that trial and error is still the primary scale-uptool for fermentation processes.

Probably the most striking-and common-violationof scale-up principles is found in the batch sterilizationof medium. A sterilization format, say 15 min at 120 C,is established by laboratory flask experiments in whichstandard autoclaving procedures are employed. Thisformat is then followed in successive stage of develop-ment, stirred glass jars, pilot plant, and even plant,fermentors. But heating and cooling rates, and there-fore rates of temperature rise and fall, change radicallyfrom one equipment volume to the next. As a resultthe total heat input, and consequently the degree ofchemical change which occurs, increases greatly for thesame nominal sterilization procedure. In figure 2 theactual time-temperature relationships for 100- and1300-gallon pilot plant fermentors are compared for a

40 80 120

TIME -minutesFigure B. Batch sterilization of streptomycin medium

1960] 129

on March 25, 2020 by guest

http://aem.asm

.org/D

ownloaded from

E. L. GADEN, JR.

nominal 15 min, 120 C sterilization. The areas underthe curves are roughly proportional to heat input; theimmense disparity is obvious.The same problem is encountered in batch cooking

of canned foods and empirical methods have been setup to correct for variations in heat transfer rates(Stewart and Clark, 1949). These can readily be appliedto fermentation processes but so far there has beenlittle interest shown in this possibility. In any eventcontinuous sterilization, when properly practiced,provides for exact reproduction of starting mediumquality at all levels of operation.Many factors affect the performance of chemical

process. In fermentation the number of thesebecomes inordinately large because of the chemicalcomplexity of the system. Since we are largely unableto explain the relationships between these variablesand product yields, scale-up of fermentation processeshas been based so far on physical measurements,particularly oxygen transfer rates.

This approach appears to work, in a rough way, foraerobic antibiotic fermentations but its essentialvalidity is doubtful. The apparent scale-ups achievedin this manner may simply be fortuitous. More likely,still, is the possibility that gross, external diffusionalresistances have been minimized, permitting the cell-medium interaction to proceed under biochemicalcontrol.The real key to successful scale-up of fermentation

processes lies in a more complete understanding of thereactions taking place. Fermentation is, above all, achemical process, although an extremely complex one.When the chemical mechanisms controlling the forma-tion of desired products are reasonably understood,scale-up on a rational, but undoubtedly empirical,basis will be possible.

CONCLUSIONS

It seems quite clear that the bulk of bioengineeringwork so far lies in the areas of (a) calibration andmeasurement and (b) equipment development. Theprocess design contributions, interestingly, are almostall from the biochemist-microbiologist, not the chemicalengineer.The real nature of the engineering function has been

emphasized here. It is not simply mathematics andmechanical gadgetry, a view unfortunately held bymany microbiologists. At the same time engineers mustrecognize their extreme dependence on full and completeunderstanding of the biochemical system with whichthey must work.The key to future progress in fermentation is a more

detailed knowledge of (a) the biochemical mechanismsinvolved in growth and product formation and (b) thefactors which influence the organism's ability to cata-lyze these reactions. It follows that there will be a

preponderance of biochemist-microbiologists, undoubt-edly possessing increasing skills in physical chemistryand mathematics, in fermentation process develop-ment.The biochemical engineer's role will be to translate

this increased understanding by empirical means inmost cases into increased productivities. Towardthis end he can profitably spend more time on empiricalanalysis of process kinetics and improved methods forexperinmental design and analysis, particular computersfor data analysis (digital) and process stimulation(analog) and scale-up techniques.

REFERENCES

Anonymous 1947 The case for biochemical engineering.Chem. Eng., 54, 106.

AMBLER, C. M. 1959 The theory of scaling up laboratorydata for the sedimentation type centrifuge. J. Biochem.Microbiol. Technol. & Engr., 1, 185-205.

BARTHOLOMEW, W. H., KAROW, E. O., SFAT, M. R., AND WIL-HELM, R. H. 1950a Oxygen transfer and agitation insubmerged fermentations. Mass transfer of oxygen in sub-merged fermentation of Streptomyces griseus. Ind. Eng.Chem., 42, 1801-1809.

BARTHOLOMEW, W. H., KAROW, E. O., SFAT, M. R., AND WIL-HELM, R. H 1950b Oxygen transfer and agitation in stub-merged fermentations. Effect of air flow and agitationrates upon fermentation of Penicillium chrysogenum andStreptomyces griseus. Ind. Eng. Chem., 42, 1810-1815.

BARTLETT, M. C. AND GERHARDT, P. 1959 Continuous anti-biotic fermentation. J. Biochem. Microbiol. Technol. &Engr., 1 (in press).

BOYLES, W. A. AND LINCOLN, R. E. 1958 Separation andconcentration of bacterial spores and vegetative cells byfoam flotation. Appl. Microbiol., 6, 327-334.

BROWN, W. E. AND PETERSON, W. H. 1950 Factors affectingproduction of penicillin in semi-pilot plant equipment.Ind. Eng. Chem., 42, 1769-1774.

CALAMI, C. T., DRIVER, N., AND BOWERS, R. H. 1951 Studiesin the production of penicillin, respiration and growth ofPenicillium chrysogenumn in submerged culture, in relationto agitation and oxygen transfer. J. Appl. Chem., 1, 209-216.

CHAIN, E. B., PALADINO, S., CALLOW, D. S., UGOLINI, F., ANDVAN DER SLUIS, J. 1952 Studies on aeration. I. Bull.World Health Organization, 6, 73-97.

CHAIN, E. B. AND GUALANDI, G. 1954 Aeration studies. II.Rend. ist. super. sanita; 17, 5-60.

COOPER, C. M., FERSTROM, G.A., AND MILLER, S. A. 1944Performance of agitated gas-liquid contactors. Ind. Eng.Chem., 36, 504-509.

DECKER, H. M., HARSTAD, J. B., PIPER, F. J., AND WILSON,M. E. 1954 Filtration of microorganisms from air byglass fiber media. Heating, Piping and Air Conditioning,23, 155-158.

DEINDOERFER, F. H. 1957 Calculation of heat sterilizationtimes for fermentation media. Appl. Microbiol., 5, 221-228.

ELSWORTH, R., WILLIAMS, V., AND HARRIS-SMITH, R. 1957A systematic assessment of dissolved oxygen supply. J.Appl. Chem., 7, 261-268.

EMERY, W. B. 1955 Bacteriological problems in translatingfrom small equipment to the industrial scale. Chem. &Ind., 1956, 236-240.

130 [VOL. 8

on March 25, 2020 by guest

http://aem.asm

.org/D

ownloaded from

BIOENGINEERING AND FERMENTATION1

FINN, R. K. 1954 Agitation-aeration in the laboratory andin industry. Bacteriol. Revs., 18, 254-274.

FRIEDMAN, A. M. AND LIGHTFOOT, E. N., JR. 1957 Oxygenabsorption in agitated tanks. Ind. Eng. Chem., 48, 1227-1230.

GADEN, E. L., JR. 1955a Fermentation-a chemical engi-neering approach. Chem. Eng. Prog., 51, 540-543.

GIADEN, E. L., JR. 1955b Fermentation kinetics and produc-tivity. Chem. & Ind., 1955, 154-159.

GlADEN, E. L., JR., AND HUMPHREY, A. E. 1956 Fibrouisfilters for air sterilization. Design procedure. Ind. Eng.Chem., 48, 2172-2176.

GADEN, E. L., JR. 1959 Fermentation process kinetics. J.Biochem. i\Iierobiol. Technol. & Engr. 1, (in press).

GAUDIN, A. M., MULAR, A. L., AND O'CONNOR, R. F. 1960aSeparation of microorganisms by flotation. I. Develop-ment and evaluation of assay procedures. Appl. Micro-biol., 8, 84-90.

GAUDIN, A. M., MULAR, A. L., AND O'CONNOR, R. F. 1960bSeparation of microorganisms by flotation. II. Flotation ofspores of Bacillus suibtilis var. niger. Appl. Microbiol.,8, 91-97.

HARTLEY, H. AND WARNER, F. E. 1958 Biochemical engineer-ing, pp. 9-36. Edited by R. Steel. The Macmillan Co.,New York, New York.

HERBERT, 1). 1959 Recent progr-ess in microbiology, pp. 381-396. Edited by G. Tunevall. Almqvist and Wiksell, Stock-holm, Sweden.

HIXSON, A. W. AND GADEN, E. L. JR. 1950 Oxygen transfer insubmerged fermentation. Ind. Eng. Chem., 42, 1792-1801.

HOSLER, P. AND JOHNSON, M. J. 1953 Pilot plant data forpenicillin production on synthetic media. Ind. Eng.Chem., 45, 871-874.

HOUGEN, 0. A. 1947 For biochemical engineers. Chem. Eng.,54, 202-204.

HUMPHREY, A. E. AND GADEN, E. L., JR. 1955 Air steriliza-tion by fibrous media. lnd. Eng. Chem., 47, 924-930.

JOHNSTONE, R. E. AND THRING, M. W. 1957 Pilot plantsmodels and scale-utp methods in chemical engineering, pp.5-26. MIcGraw-Hill Book Co., Inc., New York, New York.

JORDAN, D. G. 1955 Chemical pilot plant practice, pp. 36-47.Interscience Publishers, Inc., New York, New York.

KAROW, E. O., BARTHOLOMEW, W. H., AND SFAT, M. R. 1953Oxygen transfer and agitation in submerged fermenta-tions. Agr. Food. Chem., 1, 302-306.

KIRKPATRICK, S. D. 1947 A case study in biochemical en-gineering. Chem. Eng., 54, 94-101.

LUEDEKING, R. AND PIRET, E. 1959 A kinetic study of thelactic acid fermentation; batch process at controlled pH.J. Biochem. Microbiol. Technol. & Engr., 1 (in press).

MAXON, W. D. AND GADEN, E. L., JR. 1956 Fibrous filters forair sterilization; experimental studies with a pilot-scalefilter. Ind. Eng. Chem., 48, 2177-2179.

MAXON, W. D. AND JOHNSON, MI. J. 1953 Aeration studies on

propagation of baker 's yeast. Ind. Eng. Chem., 45,2554-2560.

MCDANIEL, L. E. AND LONG, R. A. 1954 Laboratory equip-ment for testing filtering materials for air sterilization.Appl. Microbiol., 2, 240-242.

MIETZNER, A. B. AND PIGFORD, R. L. 1958 Scale-up in prac-tice, pp. 16-47. Edited by R. Fleming. Reinhold Publish-ing Co., New York, New York.

OWENS, S. P. AND JOHNSON, M. J. 1955 The effect of tem-perature changes on the production of penicillin by Peni-cillium chrysogenum W49-133. Appl. Microbiol., 3, 375-379.

PATRICK, W. C. AND FREEMAN, R. R. 1959 Batch centrifugescale-up for virus processing. J. Biochem. Microbiol.Technol. & Engr., 1, 207-215.

PFEIFER, V. F. AND VOJNOVICH, C. 1952 Sterilization ofmedia in biochemical processes. Ind. Eng. Chem., 44,1940-1946.

PHILLIPS, D. H. AND JOHNSON, M. J. 1959 Oxygen transferin agitated vessels. Ind. Eng. Chem., 51, 83-88.

RUSHTON, J. H. 1951 The use of pilot plant mixing data.Chem. Eng. Prog., 47, 485-488.

RUSHTON, J. H., COSTICH, E. W., AND EVERETT, H. J. 1950Power characteristics of mixing impellers. Chem. Eng.Prog., 46, 395-404, 467-476.

SCHULTZ, J. S. AND GADEN, E. L., JR. 1956 Sulfite oxidationas a measure of aeration effectiveness. Ind. Eng. Chem.,48, 2209-2212.

SIKYTA, B., DOSKOCIL, J., AND KASPAROVA, J. 1959 Continu-ous streptomycin fermentation. J. Biochem. Microbiol.Technol. Engr., 1 (in press).

SOILOMONS, G. L. 1958 Oxygen transfer in laboratory fer-menters. J. Appl. Chem. (London), 8, 445--448.

SOLOMONS, G. L. AND PERKIN, M. P. 1958 The measurementand mechanism of oxygen transfer in submerged culture.J. Appl. Chem. (London), 8, 251-259.

STARK, W. H. AND POHLER, G. 1I. 1950 Sterile air for in-dustrial fermentations. Ind. Eng. Chem., 42, 1789-1792.

STEFANIAK, J. J., GAILEY, F. B., JARVIS, F. G., AND JOHNSON,M. J. 1946 The effect of environmental conditions onpenicillin fermentations with Penicillium chrysogenumX-1612. J. Bacteriol., 52, 119-127.

STEWART, J. A. AND CLARK, B. S. 1949 The canned food refer-ence manual, 3rd ed., American Can Co., New York, NewYork.

TERJESEN, S. G. AND CHERRY, G. B. 1947 The removal ofmicroorganisms from air by filtration. Trans. Inst. Chem.Engrs. (London), 25, 89-96.

WARNER, F. E., COOK, A. M., AND TRAIN, D. 1954 Engineer-ing and microbiological processes. Chem. & Ind., 1954,114-119.

WEGRICH, 0. G. AND SHURTER, R. A., JR. 1953 Developmentof a typical aerobic fermentation. Ind. Eng. Chem., 45,1153-1160.

WHITMARSH, J. M. 1954 Continuous sterilization of fermen-tation media. J. Appl. Bacteriol., 17, 27-30.

WISE, W. S. 1951 The measurement of the aeration of cul-ture media. J. Glen. Microbiol., 5, 167-177.

1960] 131

on March 25, 2020 by guest

http://aem.asm

.org/D

ownloaded from