(Oxygen Transfer and Agitation in Submerged Fermentations)
EFFECT OF AIR FLOW AND AGITATION RATES UPON FERMENTATION OF Penicillium chrysogenum AND Streptomyces griseus An experimental
I n the turbulent region of flow in unbaffled tanks, exponent n on the Reynolds number group is about -0.15. Thus the power in unbaffled tanks in the turbulent region is expressed as: HP = Kp0.SN2.~Dl.7pQ.1K (3)
study has been made of the effects of air flow and agitation variables on mycelial growth, sugar utilization, and relative antibiotic formation in the fermentation of strains of Penicillium chrysogenum and Streptomyces griseus. Experiments were performed in 5Biter laboratory scale fermentors which were designed to provide a rapid and economic means for investigation of &mentation variables. The experimental results are discussed in terms of the diffusion theory presented in the preceding paper. The techniques employed have been used with success in translation among laboratory scale, pilot plant, and factory fermentors.
In a gas dispersion unit, the power absorbed by the impeller decreases as air is introduced under the impeller. This drop is explained by decrease in apparent density of medium in contact with the agitator due to mixing of gas with liquid (6). As air is blown through a fermentor it does work upon the liquid, and this may be computed through Bernoulli’s theorem of mechanical energy balance:
x1 + PlVl
T
HE previous paper ( 8 ) waa concerned with principles and
+ 2guz+ -_?
l2 PdV
=
xz + PZVZ +.E + w (4) 29
where
mechanisms steps for mass transfer of oxygen in the submerged fermentation of Streptomyces griseus. The present paper deals with an experimental study of the etlect of air flow and agitation variables on mycelial growth, mgar utilization, and antibiotic formation in the fermentation of strains of Pencillium chrysogenum and Streptomyces griseus. The oxygen transfer principles developed in the previous paper serve as a means of explaining and unifying the results. Although the experiments reported were performed in laboratory fermentors, the design of these units as well as the procedures of malysis has permitted quantitative translation of results among laboratory, pilot plant, and factory scale fermentation. The organisms used, although of interest in themselves, were employed in a supporting role to the main objective, which was a study of the interplay of biological-engineering variables on mycelial growth and biosynthesis. Other organisms probably could have been used as well.
XI, X Z = static head in terms of gas phase, feet P1,PZ = absolute static pressure, pcmnds per square foot VI, VS = specific volume of air, cubic feet per pound
U,,U2
b5
= !inear velocity of gas flow, feet per second gravitational constant, 32.2 feet per sec.2 mechanical work done by air on fermentor liquid
In Equation 4 , PlVl - PzV, = 0 because air may be considered a perfect gas. XZ - X,*O, because the static head of gas is numerically very small. @/2g likewise equals 0 because the linear velocity 0; the gas a t the top of the fermentor is negligibly small. The work done by air on the fermentor charge therefore is :
w=
29
+J;”
PdV
(5)
or, after integration,
AGITATION PRINCIPLES
Agitation of the broth in tank fermentors derives from me&anica] power and from the air expanding and rising through the units. An extensive review of technology of agitation and mixing is given by Rushton (9). The White (11) equation for power absorption in a mechanically agitated vessel is given as:
where
and aeration The power absorption from sources, represented through Equations 2 and 6 , is important as an over-all index of agitation behavior. Because the mechanical power absorbed varies as the cube of number of revolutions per minute and power from flowing air is independent of the agitator speed, the sum of both will depend upon rate of agitation, and the contribution from air will diminish relatively as the rotational speed increases. The importance of the air contribution increases with increasing air flow, greater liquid head, and an increase in linear air velocity through a sparger orifice.
p
EXPERIMENTAL
H P = KpN8D6
(l)
HP = horsepower K a n d n = constants = liquid density N = impeller speed D = impeller diameter y = liquid viscosity, all in consistent units
fn,
the turbulent
of flow in fully baffled tanks, the e +
ponent, n, on the modified Reynolds number,
*,F
becomes 0,
and the power absorbed is stated by Equation 2. H P = KpN4D’
(2)
Laboratory deep tank fermentations were performed in 5-liter glass and stainless steel fermentors ( I ) , which were 6 inches in diameter, 12 inches tall, fully baffled, impeller-agitated, and prowas 3.2 vided with an air sparger. The charge generally liters. Agitator speed, aeration rate, and sparger type (constricted pipe and sintered) were varied. Shake flask control fermentations were performed in 250-ml. Erlenmeyer flasks on rctating shakers turning a t 220 r.p.m. on a 1.5-inch diameter throw. Within any fermentation series the following variables were maintained constant: type, amount, and age of inoculum; com1810
1811
INDUSTRIAL AND ENGINEERING CHEMISTRY
September 1950
position of medium; temperature; total pressure and ox gen partial pressure. Mechanical agitation, air flow rate, a n 8 type of sparger were altered between experiments. Dependent variables ordinarily measured a t time intervals were: residual sugar contf?r..t, antibiotic activity, m celial weight, pH, diesolved oxygen, and oxygen saturation levef Auxiliary measurements were performed to determine mechanical power and the s cific oxy6en absorption coefficient, kd. The power contributedry air flowing through the fermentors was computed.
remains substantially unchanged between fermentations, but imposed variables such as air flow rate and degree of agitation affect relative amounts and times within the framework of the pattern. The course of fermentation for penicillin has been found typical also for streptomycin. The mycelial weight curve is observed to rise with time to 8 maximum. Sugar disappearance and penicillin formation-mea are interrelated. An induction period is mted during which the sugar concentration remains practically constant and penicillin formation has not yet begun. Beyond the induction period, sugar utilization progresses a t a uniform rate toward a zero value. In the same interval the concentration of penicillin is noted to rise. Saturation levels for oxygen are seen to decrease slightly with time, presumably because of change in soluble components in the medium. Dissolved oxygen levels pass through a minimum value,
MER- CONSTRICTED PIPE
FERMENTATION TIME-HOURS
Figure 1. Course of Penicillin Fermentation in Laboratory Fermentor
a Organisms used were a mutant strain of Penicillium chrysogenum and a mutant strain of Streptomyces griseus developed in these laboratories. The medium used with the former organism was similar to the corn stee lactose medium described by Gailey et a/. (7) with a penicillin &recursor added, The medium used with Streptomyces griseus was similar to the soybean medium described by Dulaney et al. (3). The temperature for penicillin fermentations was 24' C. and that for stre tomycin was 27' C. The total preasure was constant a t 1atrnospEere. Cell dry weights were determined gravimetrically. Cell weights determined when other insoluble solids were present were obtained by a differential centrifugation technique. Sugar concentrations were determined by the Somogyi (10) method upon broths clarified by zinc hydroxide preci itation. Penicillin was assayed by the method of Foster and Vifmdruff (6)and streptomycm, by a cup assay method (4). Temperature, H, air flow rate, and agitator speed were determined by standarb)procedures. The dissolved oxygen level during fermentation was measured by the amDerometric techniaue (2). - ka . . Procedures for determining are given in (3). Mechanical wwer absorbed bv the medium in Miter fermentors was measured with a torque-"table dynamometer such as that described by Hixson and Luedeke (8). Routine precautions were taken to ensure reliability of the fermentation data. Sterility control checks were maintained, and shake flask control fermentations were rformed on the inoculum used and on fermentor samples obtainerafter inoculation. The standard deviation for antibiotic eld varied between *7% for penicillin fermentations performe$under optimum conditions and * 17y0 a t limiting conditions with respect to oxygen supply for the organisms, The standard deviation of streptomycin fermentations varied from * 13% in the plateau region of activity to t 100% at limiting conditions. ~
RESULTS
Course of Typical Fermentation. The pattern of events, characteristic of the fermentations, may be described through the time variations of sugar concentration, antibiotic activity, mycelial weight, dissolved oxygen, and saturation oxygen concentration. These variables are presented for a penicilli fermentation in Figure 1. The antibiotic production data given in this and subsequent figures concerned with penicillin are plotted on the same relative scale. The absolute activity values a t optimum conditions are at an acceptable level. The sequence of events
1.00
e
-i
5
I 1. o 2 0 0 W
u)
0 I-
o
4 -I
FERMENTATION TIME-HRS.
Figure 2. Influence of Agitation on Couroe of Penicillin Fermentation
Wpe-
Effect of Agitator Speed on Course of Fermentation. gree of agitation was varied in a series of penicillin f e r T ? & p n s by changing the rotational speed of the agitator. Typie#.&irvw are illustrated in Figure 2 with the effect of agitation ra&$own upon oxygen concentrations, mycelial weight, sugar content, and penicillin formation. The normal pattern for fermentations maintained and increasing rate of agitation is noted to increase penicillin and mold formation. Gross sugar utilization is noted to be approximately the same at the high values of agitation, 375 and 560 r.p.m. Normally, a decrease in agitation to 250 r.p.m leads to a slower rate of sugar consumption. The sugar utilirs tion a t 190 r.p.m. is abnormally high, leading to an early e& haustion of this component. At this low agitation speed the mycelia tended to settle and form clumps in the fermentor antt the character of the fermentation a t the bottom of the fermentor is suspected to have changed toward the anaerobic. Effect of Air Flow Rate on Course of a Fermentation. !&ir flow rate was varied from 1to 6 liters per minute in a set of ;penicillin fermentations. The agitator speed was 375 r.p.m. The effect of air flow (Figure 3) is not so marked as that shawn:for agitation. Mycelial weights do not vary significantly with air flow rates at this agitation speed. High air flows resulltiin-aimilsr patterns of penicillin production and sugar utilizahian; .at ithe
1812
INDUSTRIAL AND ENGINEERING CHEMISTRY
Antibiotic Yield Related to Power. STREPTOMYCIN. Maximum yields of streptomycin were measured at different agitator speeds and a t a constant air flow of 6 liters per minute. The variation of antibiotic yield with agitator speed and total horsepower absorbed per gallon is shown in Figure 5 . Streptomycin production, represented on a relative scale, rises rapidly with increase in power to a broad maximum, ultimately decreasing at high agitator speed. The productivity with sintered spargers is greater than in comparable experiments with constricted pipe spargers.
AGITATORS-TWO 4-BLAOE IMPELLERS SPARQER-CONSTRICTED PIPE BATCH VOLUME-3.2 LITERS AGITATOR SPEED- 375 RPM AIR FLOW-S LITERS/MIN.-
3
I,
Vol. 42, No. 9
--------
I
)r)
2 5 0 40
8
30
I
0
E
PENICILLIN
FERMENTATION TIME HRS.
n
'8 8
z g a
Figure 3. Influence of Air Flow on Course of Penicillin Fermentation
5 a m
5 4
a
3
$
2
i \
0: W
g0
1 lowest air rate penicillin production and sugar consumption rates W are lower. Aeration and agitation rates are interdependent in u) their effect on antibiotic production-for example, a t 560 r.p.m. 0.5 I 100 200 300 500 1000 over a range of air flows from 0.2 to 6 liters per minute, total sugar utilization, penicillin production, and growth are almost identical AGITATOR SPEED RPM over the range and are but little higher than those at an air Figure 4. Horsepower per Gallon flow of 6 liters per minute and 375 r.p.m. (Figure 3). Absorbed from Agitators us. R.P.M. Air flow and agitator speed have similar qualitative effects in Conditions. 6-inch diameter ( 5 1 l / j a inch inside diameter) X 12-inch fermentor. streptomycin fermentations. two flat-blade impellers, 3-inch diamete; Measurement of Power. Although rotational speed of imwith four blades; four baffles, one-tenth tank diameter- constricted pipe sparger; pellers may be used as an index of relative agitation intensity in a ch&e of 3.2 litera given fermentor, a general means of characterizing degree of agitation is by the power absorbed per unit volume of fermentor liquid. In fully baffled fermentors with similar types of impellers, unit power absorption serves as a satisfactory basis for Table I. Calculated Work Done by Air in Laboratory Fermentor comparing and translating among widely varying scales of operation. Batch volume, 0.85 gallon Operating ressure, atmospheric In the 5-liter laboratory fermentors direct mechanical power Constricteisparger orifice, 0.0625-inch diameter was measured using a torque-table dynamometer. MeasureSintered sparger, Type H, 5 micron pores ments were performed with water, corn steep medium, and soyHorsepower/Gallon Absorbed Air Flow bean meal medium. The experimental variables were revoluConstricted Sintered Liters/min. Feetlhoura pipe sparger sparger tions per minute of the impeller, type of sparger, and rate of air 1 12 0.4 x 10-4 0.6 x 10-4 input. Figure 4 presents mechanical power absorption results 3 36 1 . 5 X 10-4 2.0 x 10-4 6 72 5 . 0 X lo-' 6.5 X 10-4 when a constricted pipe sparger was used for air introduction. Similar data were obtained using a sintered sparger. There is no E Superficial velocity a t sparger level. significant difference in power absorbed by water and by the uninoculated mediums, as can be seen from the fact that the experimental points for the Table 11. Operating Conditions latter are in agreement with the plotted lines rlir Total, Bubble Relative which represent the data for water. From the CondiFlow, HB./C:RI. Type of Holdup, Penioillin slope of the curves it is determined that the horsetion R.P.M. L./Rlin. X IO3 Sparper /id X 101 % Activity power varies as the third power of agitator speed, 1 250 6 1.4 Con0.3 4 0.33 stricted in agreement with general experience in fully baf2 250 6 1.3 Sintered 1.5 7 0.53 3 375 6 2.4 Con0.66 6 0.86 Red tanks (Equation 2). stricted The work done by the air flowing and expanding 4 561 6 6.8 Con0.96 10 0.94 stricted in a fermentor was computed by Bernoulli's energy 5 561 0.2 12.4 Con0.25 2 0.83 stricted balance (Equation 6 ) . Table I presents computed 6 561 0s 12.5 Con0.04 Sot 0.23 power introduced by air for 5-liter fermentors stricted measurable operating in the range from 0 to 6 liters per niina Air across surface. ut.e air flow.
6
-
INDUSTRIAL AND ENGINEERING CHEMISTRY
September 1950
d
9
I
I
0
E i A
0.75 SPARGER
0.50
W
>
; 0.25
"oat
300
200
100
,----&----,
tIU
E
I
I
SINTERED SPARGER
1.00
z 0
I
-
AGITATOR SPEED R PM
AGITATOR SPEED- RPY I b 200 300 400 SO0 600 700 800 I
1813 500
400
I
I
S I N T E Y D SPAR-
-2
\
l 0.75
w
-
SPARGER
0.50
4
4
d W K
-I
W
a
600
0.25
C
CONSTRICTED PIPE SPARGER IIS 2 SINTERED SPARGER TOTAL HPIQAL. ABSORBED x 103
64b""
1.2
PENICILLIN.Similar experiments were performed with penicillin. Rate of agitation was varied from 190 to 660 r.p.m. (Figure 6). A rapid increase in penicillin productivity with increme in power absorption is observed, reaching a plateau which is common for both types of sparger. The sintered sparger, as in the case of streptomycin, maintained activity over a wider range of rate of agitation than did the constricted pipe sparger. Antibiotic Yield and Interdependence of Power and Air Flow Rate. STREPTOMYCIN. Interesting results are observed in Figure 7 when both air flow rate and agitator speed are varied. At 376 r.p.m. with the constricted pipe sparger there is a pronounced drop in productivity at a critical air flow range of 2 to 4 liters per minute volume flow, which corresponds to a 24 to 48 feet per hour superficial linear velocity. An activity level of the order of one tenth and less is obtained below t,he critical air flow rate, and peak activities are obtained at air flows exceeding
s
1.00
RPM
SYMBOL
375 560 375
d-
560
2
4
3
5 6 7
SINTERED SPARGER TOTAL HPIGAL. ABSORBED X IO'
Medium, soybean. Inoculum, S. griaeua, 5 % vegetative. Air flow, 6 liters per minute
SPAROER CONSTRICTEO CONSTRICTED SINTEREO SINTERED
1,s 2
I
Figure 5. Streptomycin Activity US. Agitator Speed in 5-Liter Fermentors
1
Figure 6. Penicillin Activity US. Agitator Speed in %Liter Fermentors Medlum, corn steep-lactose. Inoculum, P. chryaoAir flow, 6 liters per minute
genum, 8 % vegetative.
the minimum requirements. At the critical air velocity of 3 liters per minute, three fermentations gave acceptable normal yields, and three fermentations gave values a t about one tenth the normal level. Data obtained at 560 r.p.m. are also plotted. At this rate of agitation the critical condition was not observed and high activity was maintained to low air flow rates. Thus, the degree of mechanical agitation considerably affects the air flow requirements of the system. Sintered spargers were substituted for the constricted pipe spargers for fermentations at 375 r.p.m., in which air flow wm varied from 1 to 6 liters per minute. The results of these runs are plotted also on Figure 7. High activity levels were maintained over the range studied, showing that the design of the sparger can influence the results to as great an extent as increase in mechanical energy input. Productivity data for streptomycin in Figures 5 and 7 are on the same relative scale. PENICILLIN. The variation of penicillin production as a function of air flowing through a constricted pipe sparger was studied at 375 and 560 r.p.m. over a range of gas flow from 0.5 to 6 liters
1.00
-
I
I
I
I
5
6
560 RPM
-
z
=
Eio.
0.76
0.50
0.50
c Q)
0
-AIR INTRODUCED OVER T O P OF L I Q U I D
d A I R FLOW- LITERSIMINUTE
Figure 7. Effect of Agitator Speed and Sparger on Streptomycin Activity Medium, soybean.
Inoculum, vegetative
S. griseus, 5%
0 0
0 1
2
AIR FLOW
3
4
- LITERS/MINUTE
Figure 8. Effect of Air Flow and Agitator Speed on Penicillin Activity Medi ,r,corn " ' P e p . Inoculum, P. chrym-genu n, 8 % * egetative. Sparger, con+tri ted pi.>@
INDUSTRIAL AND ENGINEERING CHEMISTRY
1814 z 0
--
-
-
CONSTANT AIRFLOW CONSTANT RPM
-
!i1.0-
15 TOTAL HP/GAL.x IO3
Figure 9.
Penicillin Productivity as Function of Agitation and Air Flow 'Air introduced above liquid
per minute. The data are plotted on Figure 8. At a 375 r.p.m. agitator speed, the penicillin production increases gradually to maximum values as the air flow rate is varied from 1 UP to 6 liters per minute. At a 560 r.p.m. agitator speed, the penicillin production is essentially constant at a high activity level over a range of air flowfrom 0.5 to 6.0 liters per minute. When air is introduced above the liquid rather than through the sparger, there is a sharp drop in productivity.
g
I.0-
z
-
-
-
oxygen partial pressure a t the cell, p o will lie in the range in which the cell oxygen uptake rate is independent of oxygen concentration. Under this condition, aeration variables will have little effect upon cell processes. However, if either of the conductances, k d or kd', has low values, the average oxygen concentration at cell sites will drop to low values a t which the cell uptake rate departs from zero order and cell processes are impaired. Experimentally, it would be desirable to relate biosynthesis and cell growth rates with the mechanism steps associated with the coefficients, k d and k d ' . Of these two, however, only k d can be evaluated independently. Its value was found to vary greatly with changing air flow rate and agitation rate. Indirect experiments showed that the clump conductance, kd', could be increased by more intense agitation, although its absolute value was not determined. It was decided therefore to relate experimental results directly with the measured k d and indirectly with kd' through power per unit volume upon which the latter depends. Figures 9 and 10 present relative penicillin formation as functions of total horsepower per gallon and k d , respectively. Figure 11 shows ultimate mycelial weight as a dependent of total horsepower per gallon. Parameters in these curves are air flow rates and rotational speed. Figure 10 represents activity data for a constricted pipe sparger. At any given rate of agitation, the mechanical power absorption from the impeller increases as the air flow rate is decreased because of changes in effective density of the liquid. Above about 250 r.p.m. in this system, the increase more than compensates for the concurrent loss in power derived from flowing air. If the data are examined at constant air flow rates, a minimum power is required for satisfactory antibiotic production a t any air flow, and the smallest power requirement results from the highest air flow rate used in these studies (72 feet per hour). Low air flow rates may be compensated for by greater power input through increased agitator speed. On the other hand, as mechanical power input is decreased, high air flow can no longer compensate and production falls sharply. The type of sparger used can shift the curves, but the broad pattern remains the same. Curves of similar nature have been developed for streptomycin. Comparison of Figures 9 and 10 shuws that high values of activity, relatively constant over a range, are obtained when k d and the power input per unit volume (associated with kd') possess substantial values. Penicillin formation is noted to diminish whenever either variable drops to low values. Estremes of op-
.
-
Vol. 42, No. 9
_
'
--.-
I
zI -
I
I
I
I
'
I
'
J
I
I
'
CONSTANT AIRFLOW CONSTANT RPM 561 RPM
I-
I 15-
IO k.,x
w
-
w
-
IO4
Figure 10. Penicillin Activity 0s.
kd
ha determined by amperometric method o n uninocu-
lated. corn steep medium. Data for constructed pipe sparger. Air flowsymbols a s in Figure 9
LINEAR AIRFLOW
DISCUSSION
Oxygen transfer principles presented in the preceding paper ( 2 ) provide a basis for the discussion of the effects of air flow and agitation rates on antibiotic formation and cell growth. The effective oxygen concentration a t the cell walls, pa, was postulated to depend upon conductances, k d , associated with bubbles or air-liquid interfaces, followed by diffusive penetration into cell clumps, kd'. Alternatively, direct contact diffusion, k d " , may take place between cells and air interfaces which tends to diminish the effect of cell clumps as diffusion barriers. Conductances k d and kd' are in series, and for moderate values of these constants, the
-
n
a 1
-
5
0
o*
0 -
-
x
-
72 F T / H O U R A-36 " 0 - 1 2 " 0 - 6 ' 4 2.4 ' 0
~
Figure 11.
"
'
~
~
"
"
5 IO TOTAL HP/GAL.X 103
"
"
Maximum Cell Weight as Function of Agitation and Air Flow *Air introduced above liquid
15
September 1950
INDUSTRIAL AND ENGINEERING CHEMISTRY
erating conditions (taken from Figures 9 and 10 except for condition 2) are recorded in Table I1 for purposes of discussion. Condition 1represents low agitation rate or power absorption. Although air flow is a t a high rate, nevertheless the value of k d is low and activity is low. Condition 2 represents a marked improvement in k d at the same power input through change in the sparger to form smaller bubbles with consequent incrime in penicillin production, but apparently with a still insufficient value of kd’ to permit highest antibiotic levels. When mechanical agitation intensity is increased greatly as in conditions 4,6, and 6, both k d and kd’ are favorably affected and only when air flow through the broth is eliminated entirely does a limiting condition appear. The effect of agitation and air flow upon cell growth (Figure 11) is qualitatively the same as upon penicillin formation. These results show that an oxygen-saturated system is obtained with approximately constant rates of cell growth and antibiotic production when the threshold values of k d and of power absorption have been achieved. The combinations of air flow
1815
rates and mechanical power input to achieve or exceed these threshold levels form the basis for economic process design. LITERATURE CITED
(1) Bartholomew, W. H., Harow, E. 0.. and Stat, M. R.. IND.ENG. CEEM.,42, 1827 (1960). (2) Bartholomew, W. H., Karow, E. O., Sfat, M. R., and Wilhelm, R. H., Ibid., 42,1801 (1960). (3) Dulaney, E. L., Ruger, M., and Hlavao, C., MycoZo&, 41, 388 (1949). (4) Federal Regia&, Title 21, Part I, Section 141 (June 3.1948). (6) Foster,J. W., and WoodruE, H. B., J . Bact., 47,43 (1944). (6) Fouat, H. C., Mack, D. E., and Rushton, J. H., IND. ENQ. CHEY.,36, 617 (1944). (7) Gailey, L. B., Stafaniak, J. J., Olson, B. H., and Johnson, M. J., J . Bad., 52,129 (1946). (8) Hixson, A. W., and Luedeke, V. D.. IND.ENG.CHEX., 29, 927 (1937). (9) Ruehton, J. H., Can. Chem. P r o c e s s Z ~ .30,56 , (1946). (10) Somogyi, M.,J.BioZ. Chem., 117,771 (1937). (11) White, A. M., and Brenner, E., Trans. Am. Inst. C h a . Enqra., 30,585 (1934). RECEXV~D April 21, 1960.
Continuous Fermentation Cycle Times PREDICTION FROM GROWTH CURVE ANALYSIS S. L. ADAMS AND R. E. HUNGATE State College of Warhington, Pultman, Warh.
R
A
method has been devised for the accurate prediction Characteristiw of continICHARDS (6, 6 ) hm ofthe minimum time requid for the continuournfermenuous fermentcation under a shown that when the medium is maintained conon variety of operating conditation of a particularsubtrate. This method is b& the analysis of growth curves constructed from data from tions can be Predicted from stant by Preventing the the periodic analysisof batch fermentations. Continuous relatively Simple cumulation of toxic Products, fermentation cycle times and the level at which continumentsonbatchfermentations Yemt P O W 8 at a amstant ous fermentations ehould be operated can be dmlated Fromdata On cell PoPulation rate. Gray ( 4 ) also States that as long as the nutrients, from growth am-.me ofthe hm and sugar concentration a t substrate, andmetabolicprodvarious intervals during the test& by continuous fermentation of Subsbat-. ucts in the external environbatch process, it is a simple ment are held a t a satisfactory matter to construct a “growth and constant level a cell population multiplies at a constant rate. curve” and a “sugar utilaation curve” by plotting the observed In a continuous fermentation an equilibrium is established in data against time. which the concentrations of nutrients, substrate, and metabolic The rate of growth or fermentation-i.e., sugar utilization-at products are maintained constant by the addition of fresh medium any level will be the slope of the curve a t that point. Because maintenance of cell population is prerequisite to maintaining ferand the removal of an equal volume of fermented or partially fermented medium. mentation, the yeast growth curve has been considered the basic The position of this equilibrium defines the degree of conone on which rates must be calculated. The rate of feed must tinuous fermentation-that is, the differencein concentration of equal the rate of growth to maintain the cell population at the fermentable material between the feed and withdrawn medium. selected level. Thus with a high feed withdrawal rate only a small percentage of In the work reported here growth curves were established for a the substrate is fermented, whereas with lower feed rata a higher variety of substrates of economic importance. Theoretical conpercentage of the available carbohydrate is utilLed. tinuous fermentation cycle times were calculated from these Previous investigators in the field of contiiuous fermentation curves and tested experimentally by continuous fermentations. have employed purely empirical methods to determine the MATERIALS AND METHODS optimum amount of fresh medium that can be fed to a given volume of fermenting material per unit time. The usual practice The fermentation mediums employed in these experimentswere is to vary the amount of this feed over a wide range and to deterpear, apple, and cherry juice, and maraschino cherry brine. This mine for each amount the sugar concentration or yeast count that choice of mediums wae based on the desirability of applying the 1p is maintained. After a long seriesof such experiments the optimum sults of these experiments to the development of a method for the operating conditiotm are chosen. Should the operating conditions utilimtion of cannery fruit wastes ( I , 8). appear impractical for the problem at hand, further investigation The juices were prepared by wrapping the fruit in c h m l o t h is necessary in order to improve the medium or operating condifruit press. UnlesA otherwise atated, the and preseing in a -11 tions. Such studies, when conducted by trial and error, require juices were fermented without dilution or supplement. an excessive outlay of time and research effort. Sugar was determined by the method of Shaffer and Somogyi (7). In order to determine total reducing sugar, aliquots were in1 Present addrean, Joseph E.Seagram & SOM. Ina., Loukville 1, Xy.
