Bioprocessing

Avoid the Pitfalls of

Bioprocess Development

Mukesh Doble Use these tips to successfully handle the

India Institute of Technology
challenges that may arise during
the design and scaleup of various
bioprocess operations.

M
any correlations and relationships have been changes in the many biochemical steps functioning in con-
developed for the scaleup of chemical processes, cert e.g., microorganisms or enzymes may lose their
but these equations are not always applicable to activity, and biomass quantity may decrease if the operating
their biochemical or bioprocessing counterparts. The differ- conditions drift outside of the optimum range.
ences between these engineering disciplines lie not in the One factor that must be monitored continuously in the
principles of unit operations and unit processes, but in the production of biomass is pH. Good mixing is essential to
nature of living systems. The development, scaleup and prevent the medium from being affected by pH perturba-
optimization of the complicated biochemical activities of tions that may result from intermittent pH-control
microorganisms, including yeasts, bacteria, algae, and actions. In large-scale fermenters, the lack of a complete-
molds, and of isolated animal and plant cell systems, pres- ly well-mixed environment may lead to pockets with
ents engineering challenges that are sophisticated and diffi- varying pH values. For example, during the use of
cult. This article addresses a variety of issues that must be Bacillus thuringiensis for thuringiensin production, pH
carefully considered during a myriad of bioprocess opera- must be maintained between 7 and 8.5. Any variation
tions and discusses ways to handle them successfully. from a pH of 7.0 could decrease the biomass and product
Since it is beyond the scope of this article to cover every concentration. (However, short periods of pH fluctuation
system comprehensively, the author cites practical examples have a minor effect on thuringiensin production and bio-
from the literature that demonstrate successful techniques mass concentration.)
for a given scenario. Additional references may be obtained Taking a closer look, the organism usually produces
by contacting the author directly. biomass and a metabolic acidic waste during the log
growth phase. The acids are used as a carbon source after
1. Identify the operating regime glucose has been consumed. Alkali is added to maintain
Processes that use isolated cells have a lot in common the pH of the broth. Once the organism enters the station-
with those based on microorganisms. But unlike chemicals, ary phase, biomass production slows down considerably.
the commercial use of microbes, cells or enzymes derived At the proper operating conditions, the biomass yield is
from cells is restricted to the conditions at which these sys- approximately 92%, with a final pH of around 8.5 (1).
tems can function. The vast majority of microorganisms In general, an organisms activity is depicted graphical-
operate in mild conditions, but some thrive at temperatures ly as a bell-shaped curve, where the maximum point
above the boiling point of water or at pH values far from reflects a change in pH. Although the organisms rate of
neutrality. The performance of a bioprocess can suffer from activity increases with increasing system temperature (per

34 www.cepmagazine.org August 2006 CEP

 1.6
Nomenclature

Biomass Productivity, dx
a = ratio of the specific growth rate of the variant to that of the
parent organism 1.2
CL = concentration of CO2 in the liquid phase
CF = fraction of CO2 entering the reactor feed gas
D = dilution rate, 1/s 0.8
F = liquid flowrate to the reactor, L/s
f = fraction of recycle to reactor
H = Henrys law constant
0.4
Ks = equilibrium constant in Monod equation, concentration units
k1, k2 = forward and reverse rate constants for cell movement from
suspended to adhered state, 1/s
ms = maintenance coefficient, g glucose per g dry cell mass per s 0
Ng = number of generations 0 0.2 0.4 0.6 0.8 1
P = reactor pressure, Pa
[P] = product concentration, g/L
Dilution Rate, D
Q = air flowrate, L/s
R = universal gas constant, 8.314 J/mol-K

Figure. 1. Relationship between dilution rate and biomass productivity
[S] = substrate concentration, g/L in a Monod chemostat system.
s = substrate concentration inside the reactor or in exit stream, g/L
so = substrate concentration in the feed, g/L equation (dimensionless); Y is the yield factor (g of cells pro-
T = reaction temperature, C
t = time, s duced per g of substrate consumed); and so is the substrate
V = reactor volume, L concentration in the feed.
X = final number of cells, on a dry basis, g Generally, the concentration of biomass entering the reac-
Xo = number of cells in the original inoculum
tor is assumed to be zero. This leads to a new relationship for
x = amount of suspended biomass, g
x1 = amount of adhered biomass, g the steady-state cell productivity:
xm = concentration of the variant species, g/L
xn = concentration of the parent species, g/L Dx = DY(so DKs/( m D)) (4)
Yp = grams of product formed per gram of cells formed
Y, Ys = cell yield coefficient or yield factor = g cells produced
per g substrate consumed
As D is increased, the cell productivity also increases,
Ysp = yield coefficient for the production of lactic acid, reaching a maximum of m. At a certain value of D, the
g lactic acid per g glucose substrate is not being utilized, and both the cell growth
rate and cell productivity drop to zero the latter taking
Greek Letters
= ratio of specific growth rates, dimensionless a sharper drop (Figure 1). This condition is called wash
= rate of appearance of variants, 1/s out. Knowledge of the wash-out flowrate is necessary in
= viscosity of the solution, g/cm-s order to ensure that the reactor is operated below this
= Monod exponential growth rate, 1/s
value and that it can still maintain high cell productivity.
m = maximum growth rate, 1/s

2. Effect of shear on cells

the the Arrhenius equation), an increase in temperature In general, whole cell systems that are used to produce
above 70oC may result in a loss of activity. biomass have higher yields in an environment where the
A model for a continuously stirred tank reactor in shear forces are more uniform, such as in an airlift bioreac-
which a microorganism produces biomass (the concentra- tor. In a stirred-tank bioreactor, shear forces will be consider-
tion of which is x, measured in g/L) by consuming sub- ably higher near the impeller than in other parts of the reac-
strate (concentration noted as s; g/L) is written as: tor. High-molecular-weight substances such as dextran,
which increase the viscosity of the medium, may be used to
dx/dt = Dx + x (1) protect the cells against damage by absorbing kinetic energy.
Let us consider the effect of shear on various types of
ds/dt = D(so s) Yx (2) cell cultures. Single plant and animal cells are larger than
microbial cells and are easily damaged by shear due to
= ms / (Ks + s) (3) intense agitation; hence, stirred tank bioreactor designs
tend not to work well with animal- and plant-cell systems.
where D is the dilution rate (flowrate/volume; 1/s); is the Recommended instead are airlift reactors. Bacterial cells
Monod exponential growth rate (1/s); m is the maximum are better able than filamentous fungi to survive in the
growth rate; Ks is the equilibrium constant in the Monod shearing environment of a sparged, stirred-tank reactor
CEP www.cepmagazine.org August 2006 35
Bioprocessing

because the cell walls of filamentous fungi are composed and reverse rate constants of k1 and k2 respectively, then:
of chitin, which is not as strong as the peptidoglycan bag dx/dt = Dx k1x + k2x + x (5)
that protects bacterial cells.
Animal cells do not have cell walls and are particularly dx1/dt = k1x k2x + x1 (6)
vulnerable to shear forces. However, these cells are also
are very small and can slip between eddies, which shield At steady state:
them from shear forces. Nevertheless, an axial flow pat-
tern is preferred for animal cell systems. The impeller-to- D = (1+ x1/x) (7)
tank diameter ratio should be higher than typical values if
microcarrier cultures are used. Equation 7 suggests that as the ratio of the concentration
Gas bubbles may damage cells with membranes in two of adhered biomass to suspended biomass increases, the
ways: When bubbles collapse at the surface of the liquid, wash out flowrate increases by a factor of (1+ x1/x).
cells trapped in the wake of the bubble are subjected to In a cylindrical vessel, the surface area of the cylindri-
relatively high stresses, which can rupture the cell mem- cal section is proportional to the volume expressed as
brane. Also, when bubbles or foams surrounding a cell 1/V1/3, indicating higher amounts of cell adhesion are like-
move in different directions or at different speeds, the ly at smaller rather than at larger-scale volumes. Figure 2
trapped cells are dragged along, and the membrane may depicts the fraction effect of cell adhesion to the walls as a
be torn or sheared apart. Minimizing the formation of function of V (assuming =1).
foam layers, for instance, by increasing the diameter of Thus, the effect of cell adhesion on D will be more
the disengagement zone, or by using slow aeration rates, predominant in lab-scale vessels. Conversely, if lab-scale
will keep bubble damage in check. Also, the addition of vessels are used for the estimation of operating conditions
surface-active agents protects the cells from breakage by or for the design of a large-scale vessel, one should over-
either making them slippery enough to escape from high estimate the dilution rate and wash out flowrate, and
forces, or by increasing the viscosity of the medium, underestimate the reactor size for a desired cell produc-
thereby facilitating its absorption of the kinetic energy of tivity (4).
the cells.
In a similar manner, the microbial cells mycelial pel- 4. Fluid viscosity (non-Newtonian behavior)
lets and hyphae may be damaged by the agitation caused Dispersed filamentous broths demonstrate non-
by sparging and mixing. During the submerged cultivation Newtonian behavior i.e., a decrease in viscosity with
of Aspergillus awamori, for instance, the change in pellet increasing shear rate. Also, the apparent viscosity is nor-
size and pellet porosity were found to be inversely propor- mally quite high. The power law is used to describe the
tional to the specific energy dissipation rate. In fact, a total rheological properties of such fermentation broths.
breakup of pellets may occur in the case of hollow, cen- When the fermentation medium contains a polymeric
trally autolyzed aggregates (2). substrate such as starch, the apparent viscosity will
Mycelial cells also produce an extracellular matrix of decrease during batch fermentation due to its enzymatic
polysaccharides, which causes cells to clump under condi- degradation and consumption, and a change from non-
tions of shear stress. This increases the overall metabolite Newtonian to Newtonian behavior can occur. On the other
production. However, given the strong tendency of the hand, when a polymer is produced during batch fermenta-
cells to attach to something, the fragments of hyphae from tion, the viscosity of the broth will increase with an
the outer pellet surface may serve as new centers for bio- increase in broth concentration.
mass growth. A power law correlation has been proposed for the vis-
cosity of mold-containing solutions as a function of the
3. Cell adhesion to walls biomass dry weight (X), where a ranges from 1.1 to 2.65:
In some fermentation systems, such as those using fila-
mentous fungi, a small part of the culture grows on the Xa (8)
inner surface of the reactor. Hydrophobicity of the bacterial
cell surface is one of the most important factors that govern Another power law equation correlates yield stress with
the mechanism of this bacterial adhesion (3). If the concen- mycelial concentration, where the exponent a ranges from
trations of suspended and adhering biomass are x and x1 2.3 to 2.5 (5).
respectively, and if there is a first-order exchange of materi- Oxygen transfer and CO2 removal rates are affected by
al between suspended and adhering biomass with forward viscosity. Hence, they are important during scaleup. As vis-

36 www.cepmagazine.org August 2006 CEP

 CH4 + 1.618 O2 + 0.0728 NH3 >
6 0.3642 CH1.8 O0.5 N0.2 + 0.6358 CO2 +1.782 H2O
5
If the heat generation rate is balanced with the heat
4
Wash Out

removal rate, cell productivity will decrease from a value

3 of 6 g/m3h to 0.04 kg/m3h if the reaction is scaled from a
2 1 L to 50 m3 cylindrical vessel, assuming heat removal is
from the sides of the vessel. Hydrocarbons produce more
1
heat than partially oxygenated species. Hence, fermenta-
0 tions of the former demand greater heat transfer rates than
0 20 40 60 80 100
microbial consumption of carbohydrate at the same mass
Percent of Cells Adhering to Surface, % rate (for example, methane generates 57 times more heat
than malate, acetate or glucose equivalents). Extended

Figure 2. Effect of cell adhesion to the walls of a vessel on the wash heating times for sterilization, and long cooling times due
out ratio when = 1. to insufficient heat-transfer area, can also lead to destruc-
tion of thermo-liable compounds.
cosity increases, diffusivity decreases, which releases volu-
metric mass transfer. Usually, the effect of viscosity on vol- 6. Concentration of substrate added,
umetric mass transfer ranges from a = 0.28 to 0.60. If flow and inhibition
is in the transition region, a = 1.03. The breakage of a bub- In fed-batch or extended fed-batch fermentation
ble also depends on viscosity and liquid turbulence (2). processes, substrate or substrates are added to the contents
Biomass in the form of pellets imparts Newtonian prop- of the reactor at a predetermined rate. Generally the con-
erties to the fermentation broth. Mold suspension of centration of the substrate in the feed is kept very high,
Absidia corymbifera in filamentous and pelleted growth while its concentration in the reactor volume is very low
forms exhibits different flow behavior (6). As cited in the to avoid unnecessarily increasing the volume of the reac-
literature, filamentous mycelial suspensions showed non- tor contents with dilute feed.
Newtonian behavior and were correlated by a pseudoplas- Substrate inhibition at high concentrations of sugar is
tic model. Newtonian behavior was observed with pelleted observed in many biochemical systems. The sugar concen-
mycelial suspensions, but at higher pellet concentrations, tration in the feed, typically 100300 g/L, is maintained
the rheological behavior of the broths became pseudoplas- below 20 mg/L to avoid the formation of undesired
tic. Non-Newtonian behavior was also observed during byproducts. Mixing such a concentrated sugar solution
antibiotics production using filamentous fungi. into a medium with a 10,000-times-lower sugar concentra-
tion, and avoiding contact with cells containing pockets of
solution greater than 50100 mg/L, is a challenge.
5. Heat removal and activity/growth rates Generally, mixing time in a vessel is inversely propor-
In the case of exothermic growth, biomass production tional to the agitator speed, and, for equal mass-transfer
rates are tightly coupled with heat-removal rates, since rates at all scales, mixing time will be proportional to ves-
uncontrolled temperature increases could lead to loss of sel size as V2/9. Scaleup by a factor of 10,000 will lead to a
activity of the microorganism. If the reactor is assumed to about an 8 times increase in mixing time. Also, a high-vis-
be cylindrical and heat is removed from the surface of the cosity medium slows down the mixing times. Hence, the
cylindrical portion, one could derive from the heat balance microorganism should be able to survive in a highly con-
that cell productivity is proportional to 1/V1/3. As shown in centrated environment for longer time durations with
Figure 3, scaling up by a factor of 10,000 will lead to a increasing scale of operation. So as the scale of operation
decrease in the cell productivity by a factor of 21.5 increases, the problem becomes more acute. Optimal con-
(assuming the proportionality constant equals 1). Hence, struction of feed ports and ideal mechanical agitator
large-scale vessels will require extra heat-transfer area to designs to improve mixing have to be attempted in large-
make up for large decreases in cell productivity. scale operations. Similar issues are faced during the addi-
Aerobic production of single cell protein from natural tion of concentrated acids or alkalis for pH maintenance.
gas using methylococus capsultus is represented by the If in aerobic yeast fermentation, oxygen tension in
following stoichiometry and has a heat-generation rate is some areas of the reactor is too low, or the sugar concen-
79.6 MJ per kg protein produced (4). tration is higher than about 50 mg/L, undesired byproduct

CEP August 2006 www.cepmagazine.org 37

Bioprocessing

ethanol is formed. During fermentation of Lactococcus

lactis at high sugar concentrations, lactic acid is the only 100

Percent of Cell Productivity, %

product formed. At low sugar concentrations, a mixture of
several acids and ethanol is formed. 80
Mixing time that is shorter than substrate or oxygen
60
depletion time is crucial for substrates and oxygen that are
are added continuously or intermittently. If at any point 40
the cells have depleted their critical substrates or are
starved of oxygen, an irreversible apoptosis may be trig- 20
gered. The mixing time (which depends on the agitator
0
and reactor dimensions and the rpm of the agitator) has to
1 10 100 1,000 10,000
be shorter than the time the cells take to deplete oxygen to
ScaleUp Ratio
the critical value (which depends on the rate of consump-
tion of oxygen by the cells).

Figure 3. Effect of change in heat transfer area on volume
during scaleup on percentage cell productivity.
7. Product separation an integrated approach
An integrated product-removal strategy could offer sev- the design is simple and there are no moving parts.
eral advantages, including better economics. Continuous Typically, they are specified when the gas-to-liquid mass-
product removal, as in the case of alcohol fermentation, transfer rate is not limiting, or when the power require-
prevents deactivation of microorganisms. Such strategy is ments for mechanically agitated reactors are high due to
also ideal for organisms that exhibit product inhibition. the large reactor size required.
Cell separation and recirculation of the liquid leads to a In special cases, if fouling and pressure drop are con-
higher wash out flowrate, as given below: trollable, the organism may be immobilized or grown on
structured packing or monolith, providing a packed-bed
D = (1+f) (9) configuration. Such reactors are mainly used for biotreat-
ment of waste gases and liquids in an environmental-treat-
where f is the fraction recycled. Hence, cell removal fol- ment process.
lowed by recycling may be useful for processing larger Experiments conducted for gluconic acid production
quantities of material. with Aspergillus niger proved equal product generation in
Membrane bioreactors retain the biomass and allow airlift and stirred-tank reactors (7), while cellulase pro-
only the liquid products to pass through the separation duction by A. fumigatus was better in an airlift bioreactor.
medium. Submerged membrane bioreactors, in which the Airlifts are appropriate for moderate plant cell densities,
membrane is submerged inside the reactant medium, can i.e., l015 g/L by dry weight, and are considered unsuit-
selectively retain liquid products containing the biomass. able for density more than 20 g/L due to problems related
Such reactors have good application possibilities in efflu- to mixing and oxygen transfer. A very wide variation in
ent treatment. neomycin production was observed with a mutant strain
of Streptomyces marinensis NUV 5 when the reaction was
carried out in stirred-tank, packed-bed, fluidized-bed and
8. Reactor selection airlift reactors. Productivities were 7,135, 1,353, 6,923
Generally, a stirred-tank design, which features unifor- and 7,435 mg/L, respectively. The problem of oxygen
mity of concentrations and temperatures within the reac- transfer in a highly viscous broth during penicillin pro-
tor, is preferred as a laboratory chemostat. But at an duction with filamentous fungus P. chrysogenum was
industrial scale, reactor selection is based on a plethora of solved by growing it in a tower-loop reactor (rather than
operating conditions. in a conventional stirred-tank bioreactor), which led to
A mechanically stirred tank is preferred for larger-scale high specific productivities (8).
operations, and together with the semi-batch (fed-batch) Reactor selection is also based on the reaction kinetics.
reactor, it has the advantage of operating at a low substrate A stirred-tank reactor is suitable for substrate-inhibition
level, which is usually necessary for expression of the reactions, while a plug-flow design is suitable for
desired proteins. Gas-agitated reactors, such as bubble Michaelis-Menten kinetics, as well as for product-inhibi-
columns and airlift reactors, are often preferred because tion reactions. A plug-flow design is preferred over agitat-

38 www.cepmagazine.org August 2006 CEP

 Table 1. Scaleup criteria in fermentation Industries (9). keep the CO2 concentration in the exit gas stream low, so
as to achieve a large driving force for its desorption from
Scaleup Criterion Percentage of Industries
the liquid medium. Excessive CO2 concentration in the
Using this Criterion
medium increases osmolality (th\e toxic level of CO2 is
Power per unit volume 30% 1520%). Excess CO2 also necessitates the addition of
Gas-liquid mass-transfer 30% larger amount of base to the medium to maintain the pH.
coefficient To strip off CO2 sufficiently, a fast air flowrate must be
Tip velocity 20% used, which also keeps the CO2 concentration in the exit air
Dissolved O2 concentration 20% low, thereby achieving a large driving force. In a high-den-
sity culture with roughly 107 cells/mL, O2 uptake and CO2
removal rates can be approximately 5 mmol/L-h. The latter
ed reactor designs for first- and second-order reactions. The is defined as:
type of organism and reaction medium also influences the
reactor configuration. CO2 removal rate = (Q/V)(P/RT)(HCL CF) (10)

where H is Henrys law constant; CL and CF are the concen-

9. Agitator scaleup trations of CO2 in the liquid phase and in the reactor feed
Several scaleup techniques are practiced by fermentation gas; R is the gas constant; P the reactor pressure; and T is
industries for the design of agitators. These include constant the temperature. Generally, CF is roughly zero. The gas
power per liquid volume, constant mass-transfer coefficient, velocity in large-scale operations is kept low to avoid foam-
constant tip velocity and constant dissolved oxygen concen- ing of the liquid, which can lead to a decrease in the coeffi-
tration in the broth. Table 1 lists the percentage of fermenta- cient Q/V from small to large-scale systems, and hence an
tion industries using these criteria (9). increase in CL.
Scaleup methodology based on the energy dissipation A high partial pressure of CO2 during the batch growth
rate or equal impeller tip speed were not valid as a measure of A. niger increases the mean hyphal length and mean
of hyphal damage and production rate for the growth of branch length, inhibits penicillin production, and lowers the
Penicillin chrysogenum. However, the results were success- apparent viscosity. Thus, better oxygen transfer is seen in
fully correlated based on the circulation rate through the suspensions of P. chrysogenum due to enhanced chitin syn-
impeller zone. Justen, et al. (10) arrived at similar conclu- thesis (12). An excess of chitin, a component of the cell
sions based on a study that used fermenters of different wall, increases the plasticity of the cell wall and, conse-
sizes with various types of impellers. Scaleup of linolenic quently, the rheology of the fermentation broth.
acid production with Mortierella ramanniana in the pellet-
ed morphological form was successfully carried out assum-
ing equal impeller-tip-speed velocities at both pilot and 11. Agitation vs. changes in cell morphology
production scales. During the continuous cultivation of the bacterium
Corynebacterium glutamicum in a stirred bioreactor agitat-
ed at 410 and 1200 rpm and aerated at 1 and 3 vvm, cell
10. Aeration rates and CO2 accumulation growth profiles remained unaltered and no cell membrane
Necessary oxygen is supplied to microorganisms by aer- damage was observed. At high agitation and aeration inten-
ation. Hence, the design of a bioreactor for proper oxygen sities, cells were smaller and irregularly shaped, due to the
uptake is crucial. Improper aeration could lead to depletion uneven accumulation of cell wall components, probably to
of cells. Parameters to be considered during scaleup are protect their cells against high shear forces. Morphological
reactor height-to-diameter ratio, total gas flowing upwards differences of A. niger were observed in a pilot-scale
and the rate at which oxygen is transferred. If aeration is stirred tank bioreactor. It was found that low shear rates
not proper, air pockets may develop around the impeller resulted in the formation of looser pellets, while higher
blades (11). Microbes will occasionally pass through zones shear, coupled with larger-diameter cells, led to compact
of low oxygen concentration, and the time required to pass spherical pellets (12).
through low-oxygen-concentration zones increases with The length of a mycelium and the number of actively
reactor scaleup. growing tips (by branching) increases exponentially under
Aeration is needed not only for cell sustenance, but also unconstricted conditions. But high shear results in fragmenta-
to drive CO2 from the medium to keep toxicity low and to tion of the mycelium; pieces of hyphae from the outer pellet

CEP August 2006 www.cepmagazine.org 39

Bioprocessing

surface serve as new centers for biomass growth. This phe- clearance is high, settling of the biomass is observed,
nomena was observed during a submerged cultivation. At whereas if the agitator is located too low, axial mixing is
times, hyphal damage can also lead to cytoplasm release. poor. In such cases, multiple axial stirrers may be required.
The pellet size and porosity of Aspergillus awamori were Similarly, the location of the concentric tube in an airlift
found to be inversely proportional to the specific energy dis- reactor from the bottom is found to affect cell growth and
sipation rate. Specific growth rates and the overall biomass the production of ginseng saponin and polysaccharide dur-
concentrations of Penicillium chrysogenum have increased ing fed-batch cultivation of Panux notoginseng cells. At
with agitation intensity, while the specific product formation small bottom clearances, gas-liquid mass transfer and oxy-
was generally higher at lower impeller speeds. gen transfer rates were limiting the extent of the reaction. At
large clearances, cell sedimentation was observed (13).

12. Knowledge of microbiology and enzymology

In the cell growth process, the reactor size affects both 14. Substrate consumption, biomass
the number of generations required to achieve the final bio- production and product formation
mass concentration and also the concentration of the vari- A simple model relating substrate consumption, bio-
ant species. The number of generations, Ng, required to mass production and product formation is shown below.
achieve a final biomass concentration X starting from a Some fermentation systems follow this type of behavior:
known number of cells in original inoculum, Xo, is a func-
tion of the final fermenter volume, V, under exponential d[S]/dt = (1/Ys)d[X]/dt (13)
growth conditions (12):
d[P]/dt = (1/Yp)d[X]/dt (14)
Ng = 1.44 (ln V + ln X ln Xo) (11)
where, [S] and [P] are the concentrations of substrate and
The rate of appearance of variants, expressed as the product, respectively; Ys = grams of cells produced per gram of
number of variants produced per genome per generation substrate consumed; and Yp = grams of product formed per
and the fraction of parent organisms after Ng generations, gram of cells formed. The yield coefficients are considered to
will be: be constant during the course of the process (often called a
growth-related process). Gluconic acid production from glu-
xn/(xn + xm) = ( + 1)/ cose by Aspergillus niger is a typical example of such behav-
[( 1) + 2Ng ( + 1)] (12) ior. However, simple behaviors are not always observed. For
example, citric acid production exhibits varying values of Ys.
Refer to the nomenclature table at the beginning of this Product formation may exhibit a time delay, and Yp may
article for definitions of all terms. appear to be zero; maximum product is formed when the
Since the cellular system is very complex, the cells may amount of biomass is low.
ultimately adapt to the environment. However, they may Cell growth and product formation oscillate during batch
respond very slowly as well as unpredictably. The response production of penicillin from sugar under the continuous addi-
time will be a combination of widely different time constants tion of glucose. Oscillatory kinetic behavior of NADH2 in
of the various subsystems, the longest of which may be sever- intact yeast cells (Saccharomyces carlsbergensis) under anaer-
al hours. When cells are grown at a dilution rate close to wash obic conditions during the uptake of glucose has also been
out, they slowly adapt to this condition, increasing their m by observed. NADH2 acts as a hydrogen donor in glucose
genetic changes. Similarly, reduction in glucose levels for a metabolism, carrying hydrogen back and forth between reac-
short period of time may induce a starvation reaction, which tions catalyzed by glyceraldehydes, 3-phosphate dehydroge-
leads to loss of plasmids a genetic response that is irre- nease, and alcohol dehydrogenase.
versible. Transient reactor dynamics can be modeled, but not In other cases, Lactobacillus plantarum grown in media
much is known about the regulatory network of microorgan- containing glucose and yeast extract exhibited a varying yield
isms. A wild-type Aspergillus strain used in the production of factor, depending upon the environment. The maximum
-amylase is repressed by high glucose concentrations (15). growth yield based on yeast extract consumption was 0.5 g dry
weight/g, whereas in the yeast-extract-limited culture, the
13. Reactor bottom clearance growth yield increased with the dilution rate (14).
Agitator location in the reactor bottom affects mixing The depletion of fermentable sugar [S] (g/L) and its conver-
time as well as gas-liquid mass transfer coefficient. If the sion into lactic acid [L] (g/L) may be modeled as:

40 www.cepmagazine.org August 2006 CEP

 Literature Cited
1. Wong, I., et al., Fermentation Scaleup for Production of Antigen K88 Information Science Application in Bioprocess Engineering, Moreira,
Expressed in E.coli, Process Biochemistry, 38 (9), pp. 12951299 (Apr. A.R., and K. K.Wallace, eds., Kluwer Academic Publishers, Dordrecht,
2003). Netherlands, pp. 125142 (1996).

2. Kilonzo, P. M., and A. Margaritis, The Effects of Non-Newtonian 10. Justen, P., et al., Dependence of Mycelial Morphology on Impeller
Fermentation Broth Viscosity and Small Bubble Segregation on Oxygen Type and Agitation Intensity, Biotechnol. Bioeng, 52, pp. 672685
Mass Transfer in Gas-lift Bioreactors: A Critical Review, Biochemical (1996).
Engineering Journal), 17 (4), pp 2740 (2004)
11. Asenjo, J. A., and J. C. Merchuk, Bioreactor System Design,
3. Daffonchio, D., Thaveesri, J., and Verstraete, W., Contact Angle Marcel Dekker, Inc., New York (1994).
Measurement and Cell Hydrophobicity of Granular Sludge from Upflow
Anaerobic Sludge Bed Reactors, Applied and Environmental 12. Znidarsic, P., and A. Pavko, The Morphology of Filamentous
Microbiology, 61 (10), pp. 36763680 (Oct. 1995). Fungi in Submerged Cultivations as a Bioprocess Parameter, Food.
Technol. Biotechnol, 39 (3), pp. 237252 (2001).
4. Leiba, T. M., et al., Bioreactors: A Chemical Engineering
Perspective, Chemical Engineering Science, 56, pp. 54855497 (2001). 13. Hu, W., and J. J. Zhong, Effect of Bottom Clearance on
Performance of Airlift Bioreactor in High-Density Culture of Panax
5. Roels, J. A., et al., The Rheology of Mycelial Broths, Biotechnol. notoginseng Cells, Journal of Bioscience and Bioengineering, 92 (4),
Bioeng, 16, pp. 181208 (1974). pp. 389392 (2001)

6. Kim, J. H., et al., Comparative Study on Rheological Properties of 14. Kask , S., et al., A Study of Growth Characteristics and Nutrient
Mycelial Broth in Filamentous and Pelleted Forms, Eur. J. Appl. Consumption of Lactobacillus plantarum in A-Stat Culture, Antonie Van
Microbiol. Biotechnol, 18, pp. 1116 (1983). Leeuwenhoek, 75 (4) pp. 309320 (1999).

7. Trager, M., et al., Comparison of Direct Glucose Oxidation by 15. Leroy, F., and L. De Vuyst, Growth of the Bacteriocin-Producing
Gluconobacter oxydans Subsp. Suboxydans and Aspergillus niger in a Lactobacillus sakei Strain CTC 494 in MRS Broth Is Strongly Reduced
Pilot Scale Airlift Reactor, J. Ferment. Bioeng, 68, pp. 274281 (1992). Due to Nutrient Exhaustion: A Nutrient Depletion Model for the Growth
of Lactic Acid Bacteria, Applied and Environmental Microbiology, 67
8. Pollard, D. J., et al., Saccharomyces cerevisiae Fermentations in a (10), pp. 44074413 (Oct. 2001)
Pilot Scale Airlift Bioreactor: Comparison of Air Sparger
Configurations, Bioprocess Engineering, 15, pp. 279288 (1996). 16. Ndon, U. J., and A. A. Randall, Periodic Aerated Treatment and
in-situ Bioremediation Strategies for Polyhalogenated Compounds, Wat.
9. Mavituna, F., Strategies for Bioreactor Scale-up, in Computer and Res., 33 (11), pp. 27152720 (1999).

d[S]/dt = (1/Ys)d[X]/dt + ms[X] (15) reactor is operated in cycles consisting of various phases of
operation, such as fill/draw, react, settle, discharge and idle.
d[P]/dt = (1/Ysp)d[S]/dt (16) The reaction may consist of aeration and/or non-aeration phas-
es. The discharge step consists of partial discharge.
where Ys is the cell yield coefficient (in grams of cell dry mass It has been observed that during the treatment of dye and textile
per gram of glucose); ms is the maintenance coefficient (in wastes, an SBR performs better than simple batch or continuous
grams of glucose per gram of cell dry mass per hour); and Ysp is reactors with respect to decreasing the chemical and biological
the yield coefficient for the production of lactic acid (g of lactic oxygen demands. Degradation of polyhalogenated compounds has
acid per g of glucose). According to Ndon, et al., for L. sakei been achieved by alternating between long anaerobic conditions,
CTC 494, Ys and ms depend on the temperature and pH. Ysp is interposed by short low-oxygen tension, aerobic/methanotrophic
constant and equal to 1.0 g/g (15). mini-conditions. Continuous or intermittent mixing is
performed to allow the settling of biomass (16). CEP

15. Dynamic operations MUKESH DOBLE is an associate professor in the department of

Generally, fermenters are operated under batch or fed- biotechnology at the India Institute of Technology (IIT;
Madras, Chennai-600036, India; Tel: +91-44-2257-8243; E-
batch mode, where the substrate concentration decreases and mail: [email protected]). Prior to his work at IIT, Doble
the product concentration increases with time, leading to held positions at the technology centers of ICI India and GE
continuous variation in the reaction environment with time. India, totaling 20 years. Doble has written more than 90
technical papers and co-authored three books , including
This results in unpredictable behavior, unlike a steady-state Homogeneous Catalysis: Mechanisms and Industrial
operating environment. Applications, John Wiley; and Biotransformations and
For biological wastewater treatment applications, the Bioprocesses, Marcell Decker. He holds B.Tech and M.Tech
sequential batch reactor (SBR) is gaining popularity as an alter- degrees in chemical engineering from IIT Madras, as well as
a Ph.D. from the University of Aston in Birmingham, U.K.
native to conventional treatment techniques, due to its simplici- Doble is a member of AIChE and the Indian Institute of
ty, flexibility of operation and cost effectiveness. This batch Chemical Engineers.

CEP www.cepmagazine.org August 2006 41