Microbiological Process Discussion

Analytical Method for Calculating Heat Sterilization Times

F. H. DEINDOERFER
AND

A. E. HUMPHREY

The School of Chemical Engineering, University of Pennsylvania, Philadelphia, Pennsylvania Received for publication December 1, 1958

The determination of conditions necessary to adequately heat sterilize various media is a problem of considerable importance in the fermentation and food industries, or other such areas where a lack of sterile techniques preclude satisfactory performance. Until recently, however, practical ways to handle this problem were either wholly empirical ones, or theoretical ones which depended on some simplifying assumptions regarding the shape of the heating curve. Along the latter line is the development of design procedures for the sterilization of canned foods by Ball (1923) and Levine (1956), among others. A theoretical, yet practical and completely general, method for calculating heat sterilization times for fermentation media (as well as other liquid media), based on the thermal-death kinetics of bacterial spores and the actual heating curve, was proposed by Deindoerfer (1957). For batch sterilizations, the method involved a graphical integration of the specific reaction rate for thermal spore destruction along the heating curve. All that was required for this integration was a knowledge of the temperature dependence of the specific reaction rate. For isothermal portions of the heating curve a simple analytical integration was possible, as the specific reaction rate remained constant. In certain types of continuous sterilizers this simple integration was sufficient to determine the time-temperature relationship for any desired degree of sterilization. Humphrey (1957) was successful in analytically integrating the specific reaction rate over certain nonisothermal periods of a batch sterilization. Also, the integrated solutions for several temperature-time profiles of an element of medium flowing through continuous sterilizers were presented by Deindoerfer and Humphrey (1958) in a discussion of the principles entering into the design of these units. This paper reviews briefly the underlying theory of heat sterilization and develops, using this theory and common expressions for heat transfer in various equipment, the analytical integrations mentioned above. The practical solution of many sterilization problems can be carried out directly by employing the equations developed. Values of difficult to compute exponential
256

integral functions occasionally needed for particular solutions of these equations are also included in this work.
THEORY Thermal destruction of bacterial spores may be correlated by apparent first-order reaction kinetics. The rate of destruction at a particular temperature is mathematically represented by equation 1.

d=

-kN

(1)

See the nomenclature at the end of this report for an explanation of any unfamiliar symbols. Sterilizations are designed to reduce the viable spore population from its initial value to some predetermined level adequate for the degree of sterilization desired. An expression of this objective results from the integration of equation 1 over a particular time interval.
No
I

(2)

In this expression, V, the design criterion is a measure of the size of the job to be accomplished. The solution of the integral on the right hand side of equation 2 leads to useful expressions for evaluating and predicting the performance of a sterilizing unit. A confronting problem exists, however, in expressing k as an explicit function of t. This problem will be handled for several types of sterilizers in subsequent developments through the kinetic relationship of k with absolute temperature, and the relationship of temperature, in turn, with time, through heat transfer rate and heat balance equations. It is assumed that the relationship between the specific reaction rate for spore destruction and absolute temperature follows the familiar Arrhenius equation.'
k
=

Be-IRT

(3)

The same type of relationship can be obtained by

1 This appears a realistic assumption based on isothermal experiments. Whether it is applicable to the dynamic case where temperature changes rapidly over a time interval can be questioned. Research is currently under way to establish the validity of this assumption.

1959]

HEAT STERILIZATION

257

applying the theory of absolute reaction rates to the thermal destruction process. k =T geAS*IRe-AH*IRT (4) Equation 4 sheds more light on the significance of the constant B in equation 3 since, for all practical purposes, equation 4 can be reduced to an expression identical to equation 3. Application of absolute reaction rate theory to the thermal denaturation of biological substances is discussed by Johnson et al. (1954). They, and Pollard (1953), visualize thermal inactivation resulting from the intramolecular breaking of bonds in activated molecules, thus rearranging the molecular structure and resulting in a change in biological activity. More recently Charm (1958) suggested that activated water molecules striking sensitive areas of the spore cause its inactivation. Based on this model he derives an equation similar to equation 3. Substitution of the Arrhenius expression for k (equation 3) into equation 2 yields the integral which will be evaluated for a number of sterilization cases.
B t

B. Continuous sterilization 1. Constant rate of heat flow a. Constant energy loss (figure 2a) b. Nonisothermal heat source or sink with equal and countercurrent mass flow (figure 2b) 2. Changing rate of heat flow a. Isothermal heat source or sink (figure 2c, d) Situations which can be treated as constant rate of heat flow cases include the batch heating of media by

a)

b)

sT
STEAM SPARGING

ElEC I
I'

ELECTRICAL HEATI NG
a)

B V =

IR

dt

(5)

c)

The use of equation 5 and its analytical integration requires a tacit assumption to be made. The Arrhenius equation must represent the data over the temperature range of the sterilization. In other words, y must remain constant. As a matter of convenience in determining a value for V, it is also assumed that the entire bacterial population consists of spores of the design species. Although this latter assumption is incorrect, it offers a logical basis for design with an appropriate safety factor if the most heat resistant spores in the medium are used as the design species.
STERILIZATION CASES Various common methods of sterilization that will be considered here are listed below.2 Applications are illustrated in figures 1 and 2. A. Batch sterilization 1. Constant rate of heat flow a. Constant rate of addition to medium mass (figure la) b. No change in medium mass (figure lb) 2. Changing rate of heat flow a. Isothermal heat source or sink (figure lc) b. Nonisothermal heat source or sink (figure ld)
Another common method of sterilizing media in fermentors of pilot plant size is to simultaneously sparge steam into the vessel and into the vessel jacket or coils. This leads to a heating curve which does not permit analytical integrations of the type that will be illustrated. The sterilization time for this case can be calculated best by the graphical procedure described by Deindoerfer (1957).
2

]
SC

F
HW

CW
COLD WATER COOLING

STEAM HEATING

CODE: S-STEAM SC-STEAM CONDENSATE E-ELECTRICAL ENERGY CW-COLD WATER NW-NOT WATER

Figure 1. Methods of heating and cooling during batch sterilizations.

a)
NM,.

b)
CW

NW
DOUBLE-PIPE EXCHANGER
a

(i,
NMN LONG HOLDING COIL
c)
iCHANGER
NM

CW
(r) gEXCH J1GER/ NW CW

| PLATE EXCHANGER

"

IliX X V SC DOUIILE-PIPE EXCHANGR

CM
t t

s |-HII

d)
CW

N HM

~~~~~~PLATE ~~~~ECHANGR

.~~~~~~~~~

~ ZZ-~--

COIL

IMMERSED
EXCHAGR

X:~
CW

SPIRAL

EXCHANGER

SC

SC

CODE: CM-COLD MEDIA NM-HOT S-STEAM SC-STEAM CONDENSATE

MEDIA CW-COLD WATER HW-HOT WATER

Figure 2. Methods of heating and cooling during continuous sterilizations.

258

F. H. DEINDOERFER AND A. E. HUMPHREY

[VOL. 7

direct steam sparging, or by electrical heaters, and the continuous cooling of media due to energy losses in a flow-type sterilizer. Cases where media are heated by a chamber or coils containing constant pressure condensing steam, or cooled by a constant temperature water bath, fall in the category of changing rate of heat flow from or to isothermal sources or sinks. Media which are heated or cooled by sources or sinks, which themselves cool or heat as heat is transferred, constitute the changing rate of heat flow, nonisothermal source or sink category. The complexity of the general countercurrent flowtype sterilizer case does not permit an analytical solution by the procedures outlined in this paper, except for the special condition of a countercurrent source or sink which is equal in mass flow rate and heat capacity to the medium being heated or cooled. This condition, however, reduces the situation to a case of constant rate of heat flow.
TYPES OF TEMPERATURE-TIME PROFILES For batch sterilizations the temperature-time profile is simply the heating curve of the medium in the par-

ticular vessel. The entire contents of the sterilizer is always at a common, but usually varying, temperature. For continuous sterilizers, corresponding curves can be constructed by visualizing the temperature change in an element of medium as it flows through the sterilizer. Equations describing the temperature change with time for the sterilization cases under consideration can be derived from heat balance and heat transfer rate equations.3 Several equations are worked out by Kern (1950); others can be derived in a similar manner. The derived equations for the cases under consideration illustrated in figures 1 and 2 are tabulated in table 1. Notice that all these cases reduce to one of three equations. They are either linear, and of the form T = 3(1 + Kt) (6) or exponential of the form
(7) Another assumption made is that the over-all heat transfer coefficient remains constant. Although it varies with temperature slightly, the variation is small enough so that an average coefficient is sufficient in these equations.
=

3(1

+ be-Kt)

TABLE 1

Temnperata re-time profiles of variouis portions of sterilization cycles

Portion and Type of Cycle (Sterilizer)
Temperature-Time Profile
b

Batch heating with constant rate of steam addition into medium

Hyperbolic
T

(figure la)
Batch heating with constant rate of heat flow (no change in medium mass) (figure lb) Batch heating using isothermal heat source (figure Ic) Linear

TI

To

0 1

1 +

+ 1(s/M) t
q

sI/McTo

McTO
q

hsTo T
T T0
TH

T = To + T = T01+

q1
~TH

To
H

McTo

Exponential
T
=

T= TH k

TH

T1 + T

e(-UAIMC)t\ e

UA
Mc

,TH

T C= T,o1 ld) Cooling due to constant rate of en- Linear ergy loss (figure 2a) T =To T=
Flow heating using isothermal heat source (figure 2c)

Batch cooling using continuous nonisothermal heat sink (figure

Exponential

oT

+ - T0

(-WCIMd(1-e-UAIwc)t ej-

To

(1 Mc

-e

e-UAIWC)

77cT

Ok+~WC To
-

(1 +

WcTo
To - TH UA
WC

-qTo

Exponential 7' = TH (1 + T

TH TH

e(UAIWC) tA
~

TH
T

/TH
To- TC

Flow cooling using isothermal heat sink (figure 2d)

Exponential
T
=

Tc 1 + TO -TC e(-UAIWc)t
T
Tc

UA

TC

Flow heating using countercurrent Linear heat source of equal flow rate and heat capacity (figure 2b) Flow cooling using countercurrent Linear heat sink of equal flow rate and heat capacity * Let r
=
-

/ 1

AT UA T Wc t) AT UA

AT UA TC Wc AT UA

T To y H=y

VTHoWC

THo WC

TH0

195-9 ]
or hyperbolic of the form

HEAT STERILIZATION

259

where, by definition,
0o
-x

T = 3 1 + Kt i (8) \1, + rtJ where 3, b, K, anid r are general parameters characteristic of each individual case as described in table 1. These temperature-time profiles (equatioiis 6 to 8) are illustrated in figure 3 for media being heated and cooled. ANALYTICAL INTEGRATION Substitution of each of the general temperature-time profiles (equations 6, 7, and 8) into equation 5 results in integrals which can be solved analytically. The integrations are carried out as follows:
In the

E2(z) = ]

zx2

dx

(13)

Exponential temperature-time profile:

t

V7

e-c/lbKt dt
=

(14)

Change variables by letting x

dx
=

1 - abeKt,then
dt

(1

abKe Kt
+

be-Kt)2

and

dx
-

Kx(a

x)

following integrations,
t

let

=J
(9)

WVhen t =0, x = 1 for t, then

Ba
al

Substitute the new variable x

Linear temperature-time profile:

V
=

(1+be- Kt)

-x

Je-al(l+Kt) dt
a

=K

dx

a/ (1+b)

x(a - x)

(15)

Change variable by letting x = 1

dx =

K then
=--

a dx Kx2 When = 0, x = a. Substitute the new variable x for t, then

aK dt

Using partial fractions, equation 15 can be written B= a/(1+be -Kt) -x dx + B Idx eX V K (/l1+b) X + K a/ (1+b) a - x

ja/(+b6e-Kt)

(1+ Kt) 2

and dt

Change variables in the second integral by letting y = x - a, then

x=y+a and dx=dy

Substitute the new- variable y for x, then
B a/ (1+be-Kt) e-x
K
/

Bsa K J0

Ba

a/(1+Kt)

-x

__ ez2dx x2

(10)

Equation 10 can be written

(l+b)

dx

= Ja X2 dx + x2 Ks

K 0(1+Kt)

X2

dx (11)

Bea

a/ (I+be-K t)-a -y

(16)
dy

K KN Jae a/(1+b)-a

The integrals in equation 11 are second-order exponential integrals which have been numerically evaluated for various values of their lower limits, or arguments. Rewritten in common mathematical notation
V = K[E2(1 + Kt)- E2(a)]
HEATING CURVES COOLING CURVES

Equation 16 can be written

V=K a/(1+b)
00

e-x

~dx X
Be a
K

Ja/(1+be-Kt)
oo

o0

-dx
X

e-X

(12)

e-

dy

(17)

'a/(1+b)-a y

e-Y Be-a 0x + BK f0dy

Ja(1+be-Kt)a

L.

c.

All the integrals in equation 17 are first-order exponential integrals which have been numerically evaluated for various values of their lower limits, or arguments. Thus
B E1 a K L I\ + b/
-

CL

;' L.

E1

I+b
(

K
t - TIME
-

[El
-

-a)
?be-Kt
a)]

(18)

MIN.

Figure 3. Temperature-time profiles for heating and cooling portions of sterilization cycles.

E1

(1

260
where, by definition,

~~~F.H.

DEINDOERFER AND A. E. HUMPHREY[vL7

[VOL. 7

can

then be written

El (z)

fe-dx

(19)

BaKe"

eCd
BaKe-m r0e
(23)

Hyperbolic temperature-time profile:

For this case, let p
be written
T=
=

K +

r.

Then

equation

can

Ja~(l+rt)/(l+pt)-rn
The

integrals

in

equation
which
can

23

are

second-order expo-

3(

1 + Pt\
1 +rt~

nential

integrals,

be rewritten

=BaKe"'
Also,
V
=

--a(l+rt)I(1(+p
rti
Pt'
=

t)

dt

(20)
where

2a

(1

Pt)

m},

E2(a
13.

in)]

Change variables by

1 + 1 +

E2(z)

is

as

defined in

equation

then

When the individual characteristics for the various sterilization

cases
are

substituted
r

for

the

general
and 24,

=x
r

aind

dt

-px
=

-p) (r -px) I
new

(r

dx

parameters
these
for t,

a,

b,

K,

and

into

equations 12, 18,

equations

become the
cases.

particular design equations

These
sum-

When t
then

0,

1. Substitute the

variable

for

the

respective

marized in table 2. The listed in table 1.

equations are parametric expressions

were

~~(r
Change variables again, letting
a

px)

(21)
EVALUATION
OF

EXPONENTIAL INTEGRAL FuNCTIONS

(r
p

px),

The first and second order exponential integral functions, appearing in equations 12, 18, and 24 and defined in

equations
and
as

13 and

19, have been numerically

then
2

evaluated

graphically
as

represented

for

positive
as

arguments
r

large

33 and

negative arguments

y2=a (r -px)',
p
a

large
p

as

-21 in reactor handbooks edited

by Rockwell
appear up to 15.

(1956) and Hogerton and Grass (1955). They

tabulated in table 3 for
Above

and

positive arguments
the

this
a

argument,

following approximation

dx

dy
a

yields

satisfactory

value:

Substitute the
B (r

new

variable y for x, then

p)a-aGrIp ae(1+rt)/I(+pt)-arfp e-'

/p

E. (z)
(22)

(25)
in many

~y
r.

2-dy

Thus, for large arguments

sterilization
22

as

are

common
can

ar

Let

mn

and recall that p

p

K +

Equation

operations, the integral

be evaluated

by simple computation. Usually, too, where the func-

TABLE 2

Design equations for various temperature-time profiles

of Profile Type (Sterilizer)

Design Equation

Linear (figure lb; 2a, b)

BaF
K L

-IE2

1 + Kt/

a\I-

E2(a)I
a+
a

Exponential (figure ic, d; 2c, d)

KL\a

b,Beb-

LE~

-E(1
Hyperbolic (figure la)
BaKe-

be-Kt

E{(

+rtrt.

1959]

HEAT STERILIZATION

261

tions appear in a design equation as a difference, one is sufficiently larger than the other so that the smaller value can be neglected. A rule of thumb to follow here is that, whenever one argument is larger than another by six or more, its functional value will be small enough to neglect.
APPLICATION OF DESIGN EQUATIONS To illustrate the facility of the equations developed, the following example of a batch sterilization is cited.
TABLE 3 Exponential integral functions*
z

El(z)

E2(z)

Ei (z)

E2(z)

0.1 0.2

0.3 0.4 0.5 0.6 0.7 0.8

0.9

7.2 7.23 1.823 7.4 2.87 1.223 7.6 9.06 X 10-1 1.563 9.74 X 10-1 7.8 7.02 8.0 6.52 5.60
4.54 3.74 3.11 2.60 2.19
4.61 3.35 2.51 1.916 1.485 8.2 8.4 8.6 8.8 9.0

9.22X 10-6 1.150 7.36 8.96X 10-6 5.89 6.99 4.71 5.46 3.77 4.27

1.0

3.02 2.42 1.936 1.552 1.245

3.34 2.62 2.05 1.610 1.265

X

1.2 1.4 1.6 1.8 2.0

2.2 2.4

9.26 X 10- 9.2 9.99 X 10-6 9.95 1.584 9.4 8.02 7.82 5.99 1.162 9.6 6.44 6.16 8.63 X 10-2 3.99 9.8 5.17 2.71 4.85 6.47 10.0 4.16 1.877 3.83 4.89

10-v

2.6 2.8 3.0

3.2 3.4 3.6 3.8 4.0 4.2 4.4 4.6 4.8 5.0

3.72 2.84 2.19 1.686 1.305

1.317 9.36 X 6.72 4.86 3.55

10-i 10.4

10.2 3.34 2.69 10.6 2.16 10.8 1.740 11.0 1.400

3.02 2.39 1.887 1.492 1.180

11.2 2.60 1.013 11.4 7.89 X 10-' 1.925 11.6 1.430 6.16 11.8 1.067 4.82 8.00 X 10-1 12.0 3.78

1.127 9.35 X 10-1 9.08 X 10-7 7.40 7.31 5.87 5.89 4.65 4.75 3.69

2.97 2.34 1.841 1.453 1.148

9.09 X

6.02 4.54 3.44 2.62 1.993

10-4

12.2 12.4 12.6 12.8 13.0

3.83 3.08 2.49 2.01 1.622

2.93 1.848 1.469 1.165 1.025

The use of the equations in designing sterilization conditions in continuous sterilizers has also been described by the authors (1958). Example. Large industrial fermentors containing raw medium often are sterilized by passing steam through the air sparger until a desired sterilization temperature is reached. The medium and fermentor are maintained at this temperature for a prescribed amount of time, and then cooled by passing water through coils located within the fermentor. Obviously, some contribution to the sterilization occurs during both heating and cooling the fermentor. The problem then is one of assessing the sterilization accomplished during these portions of the sterilizing cycle and calculating the necessary holding time at so-called sterilization temperature required to complete the desired sterilization. Consider this problem where the following conditions prevail. 1. The fermentor contains 12,000 gallons of a raw medium which in periodic laboratory checks has consistently shown a bacterial count in the neighborhood of 20 X 109 cells per gal. It is desired to reduce this population to such an extent that the chance for a contaminant surviving the sterilization is 1 in 1000. 2. During heating, 50 psig saturated steam is passed into the fermentor at a rate of 200 lb per min, until the temperature reaches 250 F, the desired sterilization temperature. The medium is initially at 130 F. The enthalpy of the steam relative to 130 F water is 1091 BTU per lb. 3. In cooling the fermentor, 4000 lb per min of 50 F water passes through coils until a process temperature of 85 F is reached. The coils have a heat transfer area of 400 ft2 and for this operation the average over-all heat transfer coefficient for cooling is 120 BTU per hr X ft2 X OF. 4. The most heat resistant bacterial spores in the medium are characterized by an Arrhenius coefficient of 1 X 1036.2 sec-' and an activitation energy of 67,700 cal per gmol for thermal destruction. The design criterion can be calculated from condition 1 using equation 2. No V = lnN
In

(2 X

5.2 5.4 5.6 5.8 6.0

6.2 6.4 6.6 6.8

7.20 5.71 4.53

3.60

13.2 1.309 1.523 9.30 X 10-9 13.4 1.057 1.166 7.40 8.95X 10- 13.5 8.53 X 10-s 5.89 6.88 13.8 6.89 4.69 14.0 5.57 5.30 3.74
4.10 3.17 2.45 1.903 1.479
14.2 4.50 14.4 3.63 14.6 2.94

1010 cells/gal)(1.2 X 104 gal) (1 X 10-3 cell)

= 400

2.86 2.28
1.816 1.448

7.0

1.155

14.8 2.37 15.0 1.919

2.98 2.38 1.896 1.513 1.207

The constant addition of steam to the medium results in a hyperbolic temperature-time heating profile. Using the equation for this case listed in table 2, it is easily shown that 62 min heating time are required to heat the medium from 130 to 250 F. The sterilization design equation for a hyperbolic profile, listed in table 3, is shown below.
V=

*Adapted from Hogerton and Grass (1955).

BaKe [E2

1 +

rt)

-E2(a-m)]

26;2

F. H. DEINDOERFER AND A. E. HUIMPHREY

[VOL. 7

The factors in this e(quationi that are known include:

B
=

1036.2 see-1

6)

X 1037.2 Mill-W

The lowest of these is evaluated using the approximation of equation 25. Since 67.7 > 50.0 + 6.0, E2 (67.7) is not significant and need not be evaluated.

t = 62 min

The remaining factors are calculated as follows:

.a -

E2 (50.0) = 7.10 X 10-25

The extent of sterilization during heatinig is

6= 7,700 cal/gmol R = 1.10 cal/gmol X R 3 = 7'o = 130 F = 590CR

(67,700 cal/gmol)
a
H. K

(1.0 c(al/gmol
hs

'1t) (590R)

04.3

V = (8.72 X 10252)(7.10 X 10-25) = 9).8 or alnmost 25 per cent of the desired sterilization. Coolinig of the medium by cold water in coils (batch cooling using a continuous nonisothermal heat sink) leads to an exponential cooling curve. Under the conditions described, it will take 4 hr to cool the fermentor from 250 to 85 F. The design equation for the exponentital coolinig curve, listed in table 3, is shown below.
K [ I' (I + b)
IE (1 + be-Kt)]

h = 1091 BTU/lb s = 200 lb /mim 111 = (12,000 gal)(8.34 lb/gal) = 100,000 lb

c = I BTU/lb X OR

_BK [1l (1 + b-a The various factors in this are:

1 (1 + be-t-a)] e(quation that are knowi

(1091 BTU/lb) (200 lb/min) (100,000 lb) (1 BTU/lb X OR) (590011) =.3.70 X 10-3 min1
.
=
ar

B = 6 X 1037.2 miln4
t = 240 min

The remiainiing factors are calculated as follows:

IJA

i. K

-c

(I- C)

(200 lb/mmi) = 0Xi-3 mm (100,000 lb) 1!1 p K + r = 3.70 X 10-3min-'+ 2.0 X 10-3min= 5.70 X 10 3 mill-1 (104.3) (2.0 X lO3mill'),, m = =6.6 (5.70 X 10-3 min-')
S

w = 4000 b/min = 24 X 104 lb/hr Ur = 120 BTU/hr X ft2 X OF

400 ft2

K-

(4000 lb/min) (1 BTU/lb X OF) (100,000 lb) (1 BTU/lb X OF)

(1 -

-n

C-36fi'

1.26 X 10-16

e(24 X 1(4 lb/hr) (1 BTU/I

(120 BTU/hr X ft2 X OF) (400 ft2) ) X 'F))

Then,
BaKep2

K = 7.24 X 10-3 Mmil-1

(6 X 1037.2 min 1) (104.3) (3.70 X 10-3 min-') (1.26 X 10-16)

(5.70 X 10-3 min-'
=
)2

ii. a=

__

8.72 X 10-252

/1

+ rt I +pt/

To = 50 F = 510 OR (67,700 cal/gmol)128

(1.10 cal/gmol X 0R)(5100R) = 120.8
=

86. + (2.0 X 103minv')(62min) L1 + (5.7 X 10-3 min-') (62 min) The exponential integral functions are, therefore,

104.3

e-a =e -120.8

3.16X 10

53

iii. b =

-:, T T,o

F2

a-m
E2 (a
-

F2

(86.6- 36.6)
-

E2(50.0)

m)

E2(104.3

36.6) E2(67.7)

To= 250 F = 710 OR 710 OR - 510 OR = b10= 0.392

-

1959]
Then,
B K 6 X 10372 minm 7.25 X 10-3 min-

HEAT STERILIZATION

263

8
8

392
-53 26 103) = 2.62 X 1o= 0.069

Be`a01. Ke = (8.28 X
be-Kt
0392

10392)(3.16

12.8

e(7.25Xlo-3min-l)(240min)

The exponential integral functions are, therefore:

1 +bJ El1+0392)
=

the basic assumption in their development is that spore destruction rates can be correlated by an Arrheniustype relationship over the temperature range of the sterilization. This assumption is believed valid on the basis of the success of similar though approximate design procedures employed for heat sterilization calculations in the food industry. A suggested safety factor is introduced when the entire contaminant population is assumed to consist of the most heat resistant spore species.
a = A =

El(86.8)
E1(- 34.0)
=

E ( 1 13.0) (4 8 El=) (1 + be-K) = El 1 l130 (l+ 0.069)

El (1 + b -a)

=
=

E1(86.8

120.8)

b =
B = c =

E1 (1 +

- a)

E1(113.0 - 120.8)

El(-7.8)

Since 113.0 > 86.8 + 6.0 and -7.8 > -34.0 + 6.0, the function of the larger argument of each of these pairs will not be significant in the sterilization calculation. The other functions are evaluated using equation 25. E1 (86.8) = 2.01 X 10-40 E1 (-34.0) = -1.82 X 1013 The extent of sterilization during cooling is V = (8.28 X 10392)(2.01 X 10-40) - (2.62 X 10128)( 1.82 X 1013) = 2.6 + 7.5 = 10.1 or slightly more than 25 per cent of the sterilization. Now since
VTotal
=

g =
h =

k = K =

NOMENCLATURE parameter in design equations, equal to,u/R3, dimensionless surface area across which heat transfer occurs during sterilization, ft2 temperature ratio in design equations, defined in table 1, dimensionless constant in Arrhenius equation, sec-' specific heat of medium, sources and sinks, BTU/lb X OF constant in absolute rate theory equation, sec-' X 0Rankine-I enthalpy of steam relative to raw medium temperature, BTU/lb specific reaction rate for thermal spore destruction, sec-1 time parameter in design equations, defined in table

1, secm = parameter in design equation, equal to ar/p, dimen-

Vheating + Vholding + Vcooling

VHolding =40.0 - 9.8 - 10.1 = 20.1 At 250 F, the velocity constant for thermal death is 1.83 min71. The required time at so-called sterilization temperature is therefore

-kc

20.1 1 1.83 min-'-1.0m

Thus, the sterilization time needed is shortened considerably (in this example almost 50 per cent) by long heating and cooling periods in the sterilization cycle.
SUMMARY The equations developed in this paper make it analytically possible to calculate the degree of sterilization accomplished during portions of sterilization cycles where temperature varies. Their incorporation in

sterilization design procedures permits a simplified and rational approach to calculating the degree of sterilizain tion inthe over-all process. One should be aware that

sionless M = initial mass of medium in batch sterilizer, lb N, No= number of viable spores, number of viable spores initially present p = time parameter in design equation, equal to K + r, sec-' q = rate of heat transfer, BTU/sec r= time parameter in design equation, equal to s/M, sec-' R = universal gas constant, cal/gmol X R s = steam mass flow rate, lb/sec t= time of heat exposure, sec T, Tc, To,, TH, THO, TO = absolute temperature of medium, heat sink, heat sink (initial), heat source, heat source (initial) and medium (initial), respectively, 'Rankine U = over-all heat transfer coefficient, BTU/sec X ft2 X 'F w = coolant mass flow rate, lb/sec W = mass of flowing medium in contact with surface area A in sterilizer, lb =H* standard activation energy for thermal spore destruction, cal/gmol ,AS*= entropy of activation for thermal spore destruction, cal/gmol X OR = temperature parameter in design equations, defined in table 1, OR ,u= activation energy for thermal spore destruction in Arrhenius equation, cal/gmol V = design criterion for sterilization, equal to ln No/N, dimensionless El(z), E2(z), En(z) = exponential integrals of argument z, of first, second and the n-th order, respectively. The use of sec as the unit time measure is arbitrary. Min or hr can also be used. Care should be exercised, however, to maintain consistency when employing units of time measure.

264

F. H. DEINDOERFER AND A. E. HUNIPHREY

[VOL. 7

REFERENCES BALL, C. 0. 1923 Thermal process time for canned food. Btull. Nat]. Research Council (U. S.), 7, 1-76. CHARM, S. E. 1958 The kinetics of bacterial inactivation by heat. Food Technol., 12, 4-8. DEINDOERFER, F. H. 1957 Caleulation of heat sterilization times for fermentation media. Appl. Microbiol., 5, 221228. J)EINDOERFER, F. H. AND HUMPHREY, A. E. 1958 Principles in the design of continuous sterilizers. Fermentation Suib-

division, American Chemical Society, 134th Meetinig, Chicago, Illinois, September 10, 1958. HOGERTON, J. F. AND GRASS, R. C. (Editors) 1955 Reactor handbook: physics, pp. 686-692. McGraw-Hill Book Co., New York, New York.

HUMPHREY, A. E. 1957 D)ynamics of sterilization, Ch. E. Seminar. University of Pennsylvania, Philadelphia, Pennsylvania, October 7, 1957. JOHNSON, F. H., EYRING, H., AND POLLISAR, M. J. 1954 'he kinetic basis of molecular biology, pp. 187-285. John Wiley & Sons, New York, New York. KERN, 1). Q. 1950 Process heat transfer, pp. 626-635. McGraw-Hill Book Co., New York, New York. LEVINE, S. 1956 Determination of the thermal death rate of bacteria. Food Research, 21, 295-301. POILLARD, E. C. 1953 The physics of viru1lses, pp. 103-109. Academic 1'ress, Inc., New York, New York. ROCKWELL, T. (Edelitor) 1956 Reactor shielding design manuai, Pp. 372-384. U. S. Atomic Energy Commission, Washington, 1). C.

Microbiological Process Discussion

Principles in the Design of Continuous Sterilizers'

F. H.
DEINDOERFER AND

A. E. HUMPHREY

The School of Chemical Engineer ing, University of I'ennsylvania, Phila(lelphia, Pennsylvania Received for publication D)ecember 11, 1958

Properly designed continuous sterilization offers a method for overcoming undesirable destruction of nutritive quality and formation of toxic substances in fermentation media, two consequences often associated with batch sterilization. Also, operation of a pure culture fermentation process on a continuous basis requires a continuous supply of sterilized medium. Although this medium can be sterilized batchwise intermittently in alternate tanks in quantities large enough to permit an uninterrupted supply, the ideal method from an operational viewpoint is continuous sterilization. Because it often results in process improvement and operational advantages, continuous sterilization is finding increased interest in the fermentation industry. A number of different types of continuous sterilizers have been described in the literature. From a sterilization point of view they differ mainly in their heating and cooling characteristics. From an operational viewpoint, they differ in their control stability, ease of manipulation, and operational maintainability. Each type is used, of course, where its particular advantages are most suitable. The purpose of this paper is to review the various types of continuous sterilizers, and, based on their temperature characteristics, to illustrate the use of recently developed analytical design methods for
I Presented before the Division of Agricultural and Food Chemistry, 134th Meeting, American Chemical Society, Chicago, Illinois, September 10, 1958.

calculating the time needed to achieve the sterilization re(Iliirement dictated by a process.
CONTINUOUS STERILIZERS USED IN FERMENTATION PROCESSES Althoiugh conltinluous sterilizationi has been mentioned in connection with fermentation processes in a number of review articles in such a way that one implies it is the chief method of sterilization in the industry, not many direct references to its use in specific fermentation processes are available. Some actual applicationis for 12 different processes, cited in the literature, are summarized in table 1. Included in this summary are brief descriptions of the types of sterilizers employed in ealch application.
TYPES OF STERILIZERS AND THEIR CHARACTERISTICS A continuous sterilizer consists of three main sections. They are (a) a heating sectioni, (b) a holding section, and (c) a cooling section. Sterilizers can be classified by the mode of flow of nutrien-t medium and the manner of heat tranlsfer in each sectioni. All of the sterilizers described in table 1 reduce to at least one of each of the three inain sections illustrated in figure 1. Stirred-tank heating and holding sections shall not be conisidered in the enisuing discussion since very long reteintion times are required in these units to achieve reasonable degrees of sterilization. Stirred-tank heating