Studies on Immobilized Saccharomyces cerevkiae.

I. Analysis of Continuous Rapid Ethanol Fermentation

in Immobilized Cell Reactor

R. D. TYAGI and T. K. GHOSE,* Biochemical Engineering Research

Centre, Indian Institute of Technology, Delhi, Hauz Khas. New Delhi,
11 0016 India

summary
Rapid fermentation of cane molasses into ethanol has been studied in batch, continuous (free-
cell and cell-immobilized systems) by a strain of Saccharomyces cerevisiue at temperature 3OoC
and pH 5.0. The maximum productivity of ethanol obtained in immobilized system was 28.6 g
L-' h-'. The cells were immobilized by natural mode on a carrier of natural origin and retention
of 0.132 g cells/g carrier was achieved. The immobilized-cell column was operated continuously at
steady state over a period of 35 days. Based on the parameter data monitored from the system,
mathematical analysis has been made and rate equations proposed, and the values of specific pro-
ductivity of ethanol and specific growth rate for immobilized cells computed. It has been
established that immobilized cells exhibit higher specific rate of ethanol formation compared to
free cells but the specific growth rate appears to be comparatively low. The yield of ethanol in the
immobilized-cell system is also higher than in the free-cell system.

INTRODUCTION

Recently, increasing attention has been paid to the utilization of whole-cell

immobilization rather than the immobilized enzymes because the use of the
latter is linked to single and non-cofactor reactions because continuous
regeneration of cofactors is, as yet, an unsolved problem (except ATP). In this
system extraction and purification of enzyme from microbial cells could be
removed but sequential reactions are easily achieved.
Compared to chemical systems, biological reactions necessarily require large
reactor volume. In a CSTR with increased flow rate and hence dilution rate,
the cell concentration and product (ethanol) concentration in the outgoing
stream decrease whereas the substrate concentration increases. 1-4 When dilu-
tion rate exceeds the value of pm for yeast cells, washout of the culture occurs.
If the system is operated at optimum productivity, it will lead to substrate

*A11 correspondence should be addressed to this author. Present address: DCpartement de

Gene Biologique. 1'UniversitC de Technologie de Compikgne, B. P. 233, 60206 Compikgne Cedex,
France.

Biotechnology and Bioengineering, Vol. M I V , Pp. 781-795 (1982)

0 1982 John Wiley & Sons, Inc. CCC 0006-3592/82/040781- 15$01.50
782 TYAGI AND CHOSE

losses in the outgoing stream, and as a result, a multistage system must be

employed.
In a CSTR with a cell recycle system, higher biomass, product concentra-
tions, and productivity can be achieved. In such a process studied by Ghose
and Tyagi,'g2 the system was operated successfully at a dilution rate (0.5-0.7
h-l) higher than p,,, for yeast (0.4 h-'). At higher flow rates the liquid velocity
inside the settler exceeds the settling velocity of yeast, which results in some
cells being lost in the ovefflow stream that is generated during fermentation.
Consequently, washout of the cells from the fermentor takes place.
In an immobilized-cell reactor, dilution rates higher than the washout rate
can be achieved with higher p r o d u c t i v i t i e ~This
. ~ ~ ~is due to the persistence of
improved mass transfer properties resulting from factors such as lower fluid
viscosity and higher nutrient concentration at the liquid-solid interface. 7,8
Moreover, in this system other additional advantages are reduced reactor size
and minimum product inhibition; problems such as cell separation and recycle
difficulties are avoided.
During the course of this work, ethanol fermentation of molasses has been
studied in batch, continuous (free-cell systems), and in a whole-cell immobi-
lized column. An assessment of various systems from the ethanol productivity
point of view has been made. Moreover, the mathematical equations used to
describe the kinetics of ethanol fermentation of bagasse hydrolysate also have
been applied to the ethanol fermentation of molasses and the degree of
substrate inhibition assessed. Further, these equations have also been applied
to the whole-cell immobilization column, and the values of specific ethanol
productivity and specific growth rate of immobilized cells are estimated.

MATERIALS AND METHODS

Organism
Saccharomyces cerevisiae NRRL-Y-132 was obtained from U.S. Depart-
ment of Agriculture (Peoria, IL).

Medium
The medium used for the production of cells under aerobic conditions as
well as for fermentation was as prescribed by Bose and G h ~ s eIt. ~consists of
molasses (360 g), urea (1.08 g), M g S 0 4 - 7 H 2 0(0.3 g), and phosphoric acid
(0.3 mL), at pH 5.0, and a temperature of 30°C.

Inoculum
The culture was maintained on malt extract-glucose-peptone (MYGP)
slants. Cells were grown aerobically in shake flasks incubated in a rotary
shaker at 30°C and harvested at the stationary phase. After 12 h the cells were
inoculated aseptically into the main fermentor (for the free system). For im-
mobilization, the cells were centrifuged to produce a thick suspension.
 IMMOBILIZED S. CEREVISLAE. I 783

Immobilization
The method of immobilization of yeast cells has been discussed in Indian
Patent Application No. 35/De1/80.

Pretreatment of Molasses
Cane molasses was diluted with tap water to double its volume. Then it was
heated at 100°C for 0.5 h and allowed to cool to room temperature. Following
centrifugation it was diluted again to desired sugar concentrations. The re-
quired nutrients were added in proper proportion and were adjusted to pH 5.0
with concentrated sulfuric acid.

EXPERIMENTAL

Batch and continuous experiments (free-cell system) with molasses were car-
ried out according to a procedure adopted in our earlier reports. 1-3
Continuous experiments (immobilized cells) were carried out in a glass col-
umn (1.3 m ht., 14 cm i.d., having ports at 8, 22, 37, 50, 64,77, 91, and 105
cm from bottom; see Fig. 1). The exact procedures followed in the present
studies have all been reported by Ghose and Bandy~padhyay.~
Analytical methods including assay of cell mass, reducing sugar (RS), and
ethanol also have been reported earlier. 1-5

RESULTS AND DISCUSSION

Free-Cell System
Batch experiments were carried out with different RS (100-180 g L-l) of
molasses at constant level of inoculum and under identical conditions. Table I
represents the salient data for different levels of sugar concentration. It
demonstrates that higher sugar'concentrations impose inhibiting effects on cell
growth, ethanol production, and substrate utilization, a result in agreement
with the previous findings on bagasse h y d r ~ l y s a t e .The
~ , ~ rates of sugar con-
sumption and alcohol production were higher during the deacceleration of
growth. Moreover, the extent of substrate inhibition on molasses is higher than
on cellulose hydrolysate. The results of single-stage continuous culture at feed
sugar of 180 g L-' are shown in Figure 2. It demonstrates that by increasing
the dilution rates the ethanol and cell mass concentrations in the outflow
stream are decreased while the RS concentration increased. The maximum
ethanol productivity (3.35 g L-' h-l) was obtained at a 0.65 h-l dilution rate.

Immobilized-Cell System
The experiments were carried out at various residence time with RS concen-
tration of 150 g L-l. The results obtained are illustrated in Figures 3-5. All
fermentable sugars were consumed at a residence time of 2.67 h or more. Fur-
784 TYAGI AND GHOSE

[O27 I
IMMOBILIZE0 BED

SAMPLING PORT

NEOPRENE RUBB
SEPTUM

PERFORATED
STAINLESS STEEL
SUPPORT

11 t
RESERVOIR
t COLLECTOR

Fig. 1. The whole-cell immobilized reactor system.

ther decrease in residence time resulted in loss of fermentable sugar in the

oufflow stream. Figure 8 represents various yield data and productivities in the
outflow of the immobilized reactor. Ethanol yield was found to 96% of
theoretical. The yield of cells (0.02 g g-*) reduced markedly, which is in agree-
ment with the work of Ghose and Bandy~padhyay.~

TABLE I
Batch Data on Various Sugar Concentrations

Time of Cell yield Ethanol Specific

RS fermentation Yx/s yield Yp/s growth rate p
(g L-') (h) (g g - l ) (g g-l)

100 27 0.088 0.46 0.13

140 32 0.087 0.47 0.102
180 38 0.089 0.473 0.09
 IMMOBILIZED S. CEREWSIAE. I 785

JL.0

- 3.0

-
I
c
r
1-
- 2.0 c"
0
n

- 1.0

-0
OlLUTlON RATE ,h-'

Fig. 2. Fermentation of molasses in a CSTR (free-cell system). ( A ) RS; ( 0 )PD;(I)X ;(1)

P.

160

1LO

120

100

7- 80
m
m'

60

40

20

10 20 30 LO 50 60 70 80 90 100 110
HEIGHT, Cm

Fig. 3. Variation of sugar concentration with reactor height. Feed of RS--150 g L-'. 7: (0)
2.07; (A) 2.67; ( z )3.7; ( X ) 6.3; ( z )8.3 h.
786 TYAGI AND CHOSE

I I L
0 10 20 30 LO 50 60 70 80 90 W 110 133
HEIGHT, Cm

Fig. 4. Variation of F + OH concentration with reactor height. Symbols as in Fig. 3.

The yields of cells as well as product were found to remain constant irrespec-
tive of residence time, column height, and ethanol concentration (Fig. 6). The
gas holdup (19.5%)was determined by draining the column content at steady
state (Fig. 7). The maximum productivity obtained within the range of dilu-
tion rate studied was found as 28.6 g L-l h-l. Various productivities are com-
pared in Table 11.

HEIGHT, Cm

Fig. 5. Variation of cell mass with reactor height. Symbols as in Fig. 3, except (0)7 = 2.07 h.
 IMMOBILIZED S. CEREVISZAE. I 787

06 - - 0.06
A,. A YPIS
L l = - - u a a
7m 0 4 - -aor;b
a
m
-
m'
- 2
v)

Po2-a 6 0 ; i
YXlS
; 0-OD2

I I I 9

Fig. 6. Variation of yield factors in immobilized reactor at various residence times.

8
a
0
ta 3c
Y
a

21
n
a
0
6
I
11

VI
4
U
L
03 0.2 0-3 01 0.5 0.6
DILUTION RATE 5'
Fig. 7. Variation of gas holdup in the immobilized reactor at various residence times.

TABLE I1
Productivity Ethanol in Various Systems

Productivity of ethane1
Process Substrate (g L-' h-')

Batch molasses 2.0 (excluding the down-

time of fermentor)
Continuous (free-cell system) molasses 3.35
Continuous (immobilized) (ref. 5) molasses 24.9
Continuous (immobilized) molasses 28.6
Continuous (cell recycle) (ref. 2) bagasse hydrolysate 32.0
788 TYAGI AND GHOSE

t70

30-60

-so-; - 30
ol

I
n
'
a
20-LO '
a 28

I
0.1
3.5
I
0.2
30
I
0.3
2.5
I
04
2 0
I
0-5
12.0

DILUTION RATE, h-'

Fig. 8. Variations of various yields and productivities in the outflow stream of the immobilized
reactor. Symbols as in Fig. 2: ( 0 )XD.

MATHEMATICAL ANALYSIS

Fi-ee-Cell System
The combined effects of the inhibition imposed by product (ethanol) and
high substrate concentration during ethanol fermentation of bagasse
hydrolysate for the strain of yeast used has been described by Ghose and
T ~ a g iby
~ .the
~ following equations:

= l / x d r / d t = pm(1 P/P,,,)
p -
( S + K , + S2/K,o

( - d S / d t ) = l/Yp/S
d P / d t = l/Yx/sd x / d t (3)
These equations were applied to account for the kinetics of batch and con-
tinuous ethanol fermentation of molasses. The constants p,, v,, y ~ /YxIs, ~ ,
K,,and K i were determined experimentally for molasses medium.
Equations (1)-(3)were solved for the values of P and S corresponding to
various periods of batch fermentation at various concentrations of substrate.
The results of these computations are illustrated in Figures 9 and 10. The data
points represent the batch fermentation of molasses at various RS concentra-
tion whereas the solid lines represent the computed data. The following
numerical values of the constants were studied: p,,, = 0.25 h-'; K , = 0.476 g
 IMMOBILIZED S. CEREVZSZAE. I 789

,
TIME h

Fig. 9. Simulated plot for S versus the batch process (free-cell system). RS: 100 g L-I.

11

11

c Free cell system

1-
m
d

1 L
4 8 12 16 20 24 28 32 36 40
TIME, h

Fig. 10. Simulated plot for P versus the batch process.

790 TYAGI AND GHOSE

L-'; P , = 87.0 g L-'; Xo = 0.5 g L-'; Y,,, = 1.0 h-'; K,' = 0.666 g L-1;
Ph = 114; Yp/s = 0.47 g g-'; Yx/s = 0.09 g g-'.
The applicability of eqs. (1)-(3) was also verified for single-stage continuous
fermentation of molasses and the results of these computations are shown in
Table 111. The values of w and w' in molasses as compared to cellulose
hydrolysate3r4(o = 427.5; w' = 455.0) are characteristic of higher inhibitory
effects.

Immobilized-Cell System
The equations used to describe the kinetics of ethanol fermentation of
molasses in a free system were applied to the immobilized-cell system. Con-
sider a small element of immobilized bed between the cross section of I and
(I -I- dl) and taking material balance for cell mass and product (Fig. l l ) ,
input (bulk flow + dispersion) + generation
= output (bulk flow + dispersion) + accumulation (4)
Table IV represents the summary of expressions for materials due to various
transport properties inside the differential element.
Since at steady state the accumulation is zero,
P p a 4-D a ( d P / d t ) , -k ',,a dl = Pl+dl da -k Da(dP/dt),+dl (5)
and
xl ua -k Da(dx/dtIl -k r, a dl = X / + d l ua -k Da(dx/dt)l+dl (6)
Rearranging eqs. (5) and (6) and taking limit as dl - 0,
+
D(d2P/d12) u(dP/dl) - rp = 0 (7)
D(d2x/d12)-k u(dx/dl) - rx = 0 (8)

TABLE 111
Comparison of the Values of P Obtained
from Experiment and Computed by the Kinetic
Model

Theoretical value Experimental values

of P of P (from Fig. 5)
(g L-1) (g L r L )

10 11
20 20.5
30 29.0
40 41.5
50 49.0
60 61.0
70 70.5
75 76.0
 IMMOBILIZED S. CEREWSZAE. I 791

Fig. 11. An element of the immobilized-cell column.

Setting z = 1/L, p = PIP;, a n d y = x/x,, where z is the dimensionless

height of column; p is the dimensionless product concentration; and y =
dimensionless biomass concentration, eqs. (7)and (8) will become:
+
(D/Lu)P;(d2P/dz2)
p;(dp/dz) - r,,r = 0 (9)
+ x,(dy/dz)
(D/Lu)X,(d2y/dz2) - rX7= 0 (10)
The dispersion number D/Lu was determined experimentally by tracer tech-
nique and its value was found to be 0.04.
The rate of product formation and rate of growth can be considered as,
rate of product formation (r,,)
= rate of product formation in free space ( r p f )
+ rate of product formation on carrier ( r p c ) (11)
rate of growth (I;)= rate of growth in free space (r,.,)
+ rate of growth on carrier (r.J (12)
Rate of growth and rate of product formation can be represented in a similar
fashion to the equations applicable to a free-cell system. Therefore, eqs. (2)
and (3) define r,., and r p f ,respectively. Since the cell retained per unit weight
of camer is constant throughout the column irrespective of the liquid velocity,
the rate expressions for product formation and biomass growth can be written:

r,, = v;(x + ax’)(1 - PIPA) ( S + K i +S S2/K;W’ ) (13)

S
r, = &(x + P x ’ ) (1 - P/P,,,)
( S 4-K , + S 2 / K , w
where v A is the maximum specific productivity of immobilized cells (h-l); pA
is the maximum specific growth rate of the immobilized cells (h-l); x’ is the
concentration of immobilized cells per unit volume of free space (g L-l);
a = v;/v,; = Qp,. Now, if E is the fraction of working space excluding
gas holdup, and $, is the fraction of gas holdup inside the column, then 1 -
(E+ t ) is the fraction of volume occupied by the inerts inside the column;
compute that the total volume of the column is 17 L; the volume occupied by
792 TYAGI AND GHOSE

TABLE IV
Summary of Expressions of Materials

Transport system Product formation Cell mass formation

Input bulk flow

Input dispersion
Output bulk flow
Output dispersion
Generation
Accumulation (steady state)

the carrier is 7.1 L; the fraction of the column volume occupied by water is
0.417; and the volume occupied by gas holdup is 3.26 L. Therefore,

€ =
17 - (7.1 + 3.26) = 0.39
17
and
$, = 3.26/17 = 0.1917
Now, if dVis the elemental volume; d W is the weight of inert in the element;
x, is the weight of cells per unit weight of carrier (g g-l), then dVis the volume
of working space inside the element; x, d W is the weight of cells on the carrier
in the element; and x, dW/dV = concentration of immobilized cells per unit
volume of free space = x ' . If (it is assumed) that the carrier packing remains
uniform throughout the column, then
total weight of carrier
dW/dV= = 1000/17
total volume of column
= 58.82 mg d-',
and from experimental results, x, = 0.132 g g-'; therefore,
x ' = x, dW/e d V = (0.132/0.39) X 58.82 = 20 mg d-'
Since cell-yield factor YXls(0.02) and product-yield factor Yp/s(0.49) remain
constant at different heights of the column, varying reducing sugar concentra-
tions, and at various residence times, eq. (3) can be used to correlate the yield
factors for the whole-cell immobilized column. Equations (9), (lo), and (3)
then were solved numerically by trial and error using the four-point
Runge-Kutta method to determine the values of constants a and p. The initial
conditions were z = 0, p = 0, y = 0, and various values of dp/dz and dy/dz
were assumed at z = 0 for the best fit of the experimental data at various
values of a and p. The results of these computations are shown in Figures 12
and 13. The solid lines represent the results of computations whereas the data
points represent experimental results. The values of 01 and /3 which gave the
best fit of data points are 1.4 and 0.32. The values of specific productivity of
 IMMOBILIZED S. CEREWSZAE. I 793

z
Fig. 12. Plot ofy vs. Z . p/pm = 0.32; dy/dt = 4.0.

z
Fig. 13. Plot of P VS. Z.
794 TYAGI AND GHOSE

ethanol and specific growth rate for immobilized cells thus can be calculated.
These values were determined as 1.4 and 0.08 h-l, respectively.

CONCLUSIONS

The effects of ethanol on growth and product formation rates in free as well
as immobilized cells have been established. The degree of substrate inhibition
for molasses medium was found to be higher than the ethanol fermentation of
bagasse hydrolysate. The maximum productivity of 28.6 g L-' h-' was ob-
tained in the immobilized process. Still higher productivities can be achieved
by increasing the dilution rate, but it results in the loss of fermentable sugar in
the outflow stream. The viability of yeast at the outlet of the column was main-
tained above 92% for a continuous 35-day steady-state conversion runs. The
reactor performance was found very steady and required no extra cell input. In
such a system a recycling problem also is avoided. A dilution rate of nearly two
times the dilution rate for washout in the free-cell system was used. The
specific growth rate and specific rate of product formation for immobilized
cells were computed mathematically and it was established that growth of im-
mobilized cells is only 32% that of free cells. However, specific ethanol pro-
ductivity was 14 times higher for immobilized cells compared to the free cells.
The overall yield of product was higher for immobilized cells. But yield of cells
was found to reduce markedly.
In spite of better specific productivity and yield of ethanol, the immobilized
system has its own limitations. For example, a) the physiological state of
microorganisms is difficult to control, and b) a large amount of column volume
is occupied by gas holdup (ca. 19%) and the carrier (41.7%). Further studies
on the physiology of the system and ways to reduce the gas holdup and to in-
crease the retention of the cells on the carrier are in progress. There are also
many opportunities to investigate carriers with large bulk densities so as to
minimize the volume occupied in the column.

Nomenclatore
cross section area (cm2)
diffusion coefficient (cm' h-I)
flow rate (L h-')
saturation constant for growth (g L-')
saturation constant for product formation (g L-')
arbitrary length of column (cm)
total length of column (cm)
ethanol concentration (g L-'1
ethanol concentration above which cells do not grow (g L-I)
ethanol concentration above which cells do not produce ethanol (g L-])
ethanol concentration at length, I , of the column (g L-I)
+
ethanol concentration at length, I df,of the column (g L-')
dimensionless ethanol concentration
rate of ethanol formation (g L-' h - l )
 IMMOBILIZED S. CEJ7EWSL4E. I 795

rate of ethanol formation by immobilized cell (g L-' h-I)

rate of ethanol formation by free cells (g L-' h-I)
rate of biomass growth (g L-I h-I)
rate of biomass growth for immobilized cells (g L-' h-')
rate of biomass growth for free cells (g L-' h-')
substrate concentration (g L-I)
fermentation period (h)
velocity of liquid inside the column (cm h-')
volume of reactor (L)
biomass concentration (g L-I)
biomass concentration at zero time (g L-')
biomass concentration at length I of the coiumn (g L-I)
+
biomass concentration at length, 1 dl, of the column (g L - I )
maximum cell concentration obtainable from a given quantity of substrate (g L - I )
immobilized-cell concentration per unit voiume of free space (g L-')
weight of cells retained per unit weight of camer (g g-'1
yield of biomass based on fermentable sugar (g g-')
yield of product based on fermentable sugar (g g-I)
biomass concentration (dimensionless)
height of the column (dimensionless)

Greek
P specific growth rate (h-')
Pm maximum specific growth rate (h-')
Pm maximum specific growth rate of immobilized cells (h-I)
Y specific ethanol productivity (h-')
vnl maximum specific ethanol productivity (h - I )
VA maximum specific ethanol productivity of immobilized cells (h-')
w degree of substrate inhibition for growth
w' degree of substrate inhibition for ethanol formation
€ fraction of working space inside the column
t fraction of gas holdup inside the column
7 residence time (h)

References
1. R. D. Tyagi and T. K. Ghose, Proceedings ofthe Bioconversion Symposium, T. K. Ghose,
Ed. (IIT, Delhi, 1977), p. 585.
2. T. K. Ghose and R. D. Tyagi, Biotechnol. Bioeng., 21, 1387 (1979).
3. R. D. Tyagi and T. K. Ghose, Biotechnol. Bioeng., 21, 1401 (1980).
4. R. D. Tyagi and T. K. Ghose, Biotechnol. Bioeng.. 22, 1907 (1980).
5. T. K. Ghose and K. K. Bandyopadhyay, Biotechnol. Bioeng.. 22, 1489 (1980).
6. R. D. Tyagi and T. K. Ghose, private communication.
7. A. LeDuy, A. A. Marsan, and B. Coupal, Biotechnol. Bioeng.. 16, 61 (1974).
8. R. Hattori, J. Gen. Microbiol., 18, 319 (1972).
9. K. Bose and T. K. Ghose, Process Biochem.. 8, 23 (1973).

Accepted for Publication June 30, 1981