Authors
1 Department of Biotechnology, Faculty of Science, University of Tehran, P.O. Box 1198-16765, Tehran, I.R. Iran
2 Department of Chemical Engineering, Faculty of Engineering, Tarbiat Modarres University, P.O. Box 14155-4843, Tehran, I.R. Iran
Abstract
Keywords
INTRODUCTION
The batch aerobic process has been used successfully
for the production of single cell protein (SCP) from
cheese whey using the yeast Trichosporon sp. The
genus Trichosporon represents a taxon comprising
microorganisms with a unique set of enzyme capabilities for aerobic biodegradation of diverse organic compounds including cheese whey (Spanning and Neujahr,
1990; Gholson and Gough, 1980). Cheese whey fermentation for the production of single cell protein SCP
using the yeast Trichosporon sp. can be described as a
biochemical reaction of cells and lactose to produce
microbial cells as the main product. Microbial growth
kinetics, i.e., the relationship between the specific
growth rate (µ) of a microbial population and the substrate concentration (s), is an important tool in microbiology and biotechnology. Traditional kinetics is
based on the assumption that a single compound (e.g.
lactose in this present model) is controlling the rate of
growth of a microbial cell.
For design, operation and control purposes, an
accurate simulation of a reactor performance is essential (Camarasa et al., 2001). The production process
can be performed in different kinds of vessels including stirred tank reactors (STRs) and airlift reactors.
Various models are used for simulating the performance of airlift reactors depending on whether flow is
close to plug or mixed. The tanks-in-series models can
be used for any extent of mixing in a reactor. It is a
combination of a series of theoretical and well-mixed
reactors. The tanks-in-series model provides a set of
first order differential equations, which can be solved
using rather simple numerical techniques
(Levenshpiel, 1999).
Znad et al., (2004) applied the tanks-in-series
model for mathematical modeling of the unsteady per
formance of a semi-batch operation in an internal loop
airlift bioreactor for production of gluconic acid by
fermentation. Zuo et al. (2006) used a modified tanksin-series model to describe the cultivation of
Acetobacter xylinum for bacterial cellulose production
in a modified airlift reactor with wire-mesh draft tubes.
In this work, cheese whey was used as a substrate for
biomass production in a stirred tank bioreactor and in
an external airlift bioreactor. A mathematical model
based on a tanks-in-series model without back-flow
has been used to simulate the production of SCP in the
external airlift bioreactor under an unsteady condition
and without oxygen limitation.
MATERIALS AND METHODS
Microorganism and cultivation: After sampling
from several cheese making plants, the favored
microorganism was selected and enriched. The selection was based on the microorganisms’ ability for
chemical oxygen demand (COD) reduction and cell
dry weight (CDW) production. During this procedure,
Trichosporon sp. was selected as the best microorganism for this purpose (Shafaghi, 2000).
The yeasts were maintained on potato dextrose agar
(PDA) slants. The cultures were incubated at 30ºC for
24h and thereafter preserved at 4ºC. For preparing the
culture medium, fresh cheese whey was adjusted to pH
4.5 and then boiled at 100ºC for 15 min. After cooling,
denatured proteins were separated by filtration. For
removing smaller proteins, ultrafiltration was carried
out. The resulting cheese whey was green to yellowish
in color and was stored at 4ºC until further use.
Ammonium sulfate was added as nitrogen source, pH
value adjusted to 3.5 and finally sterilized in an autoclave. In order to prepare the seed culture, 350 ml of
ultra-filtrated cheese whey was transferred to 1-liter
flasks and sterilized at 12ºC for 15 min. 10-20 ml of
sterile medium was transferred to the slants. The cell
suspension was then added to the flasks which were
incubated at 30ºC, with shaking at 200 rpm for 24h.
The prepared seed culture was used for inoculation in
bioreactors.
Batch fermentation in the stirred tank reactor: A 2-
liter stirred tank bioreactor (INFORS, Switzerland)
with a working volume of 1 liter was employed in this
experiment. The bioreactor and all accessories (mixing
system, tubing, etc.) were sterilized in an autoclave
before use. 900 ml of sterilized cheese whey medium
was transferred to the bioreactor which was then inoculated with 100 ml of microbial suspension. The optimum culture condition was determined using several
experiments (Shafaghi, 2000). In each optimizing
experiment all of the variables (pH, temperature, agitation rate, and aeration rate), except one, were kept
constant. The initial value of each constant variable
was selected according to references. The bioreactor
was maintained in optimum operational conditions:
30ºC, pH 3.5, aeration 2 v.v.m, and 800 rpm. After
inoculation the bioreactor was set to work for 24h and
samples were collected every 2 h.
Batch fermentation in the external-loop airlift
bioreactor (ELAB): The external airlift reactor used
in this study was made of glass. It was 1500 mm in
height with a 100 mm diameter riser, and a 50 mm
diameter downcomer. The gas sparger in the airlift was
located just above the pipe connecting the riser and the
downcomer. It consisted of a cross with 32 holes,
each 1 mm in diameter, in a triangular arrangement.
Figure 1 shows a schematic of the external loop air lift
reactor used in this work.
The bioreactor and all accessories (mixing system,
tubing, etc.) were sterilized for one hour by steam.
Temperature and pH as effective factors in microbial
growth and metabolism were set at optima of 30ºC and
3.5, respectively. On the other hand, after choosing the
sparger, several experiments with cheese whey as
medium, without the microorganism, were carried out
to optimize the ratio of downcomer’s diameter to
riser’s diameter, liquid level in the gas separator, and
aeration rate. The optimum aeration rate, temperature,
liquid level in the gas separator, and pH were 2.5
v.v.m, 30ºC, 3, and 3.5, respectively. The basic parameters of the ELAB is given in Table 1.
Analytical methods: Samples were collected every 2
h during a 20 h fermentation period in the STR and 24
h in the airlift during batch culture and were then analyzed. In order to measure the amount of biomass,
samples were centrifuged at 1900×g for 15 min; supernatants were transferred to other tubes for the purpose
of measuring lactose. After washing the biomass with
Ringer serum, biomass was precipitated for a second
time before measuring dry weight. The results are displayed in Table 2. Lactose concentration was determined by the Somogyi-Nelson method (Shafaghi
2000).
Model development
Tanks-in-series model for external-loop airlift reactor: In this simulation, the mixing characteristics are
described by a tanks-in-series model. In the tanks-inseries model, the flow in the airlift bioreactor is considered as flow through a series of equally sized, wellmixed stirred stages or tanks and the parameter
describing non-ideal flow is the number of stages. The
mixing characteristics of the riser, downcomer, top and
bottom sections in the airlift bioreactors are different
(Verlaan et al., 1989). For example, in the computer
simulation model of Merchuk and Stein (1981), the
mixing characteristics in the riser and the downcomer
were postulated as plug flow and the head space was
considered to be well mixed. An extension for the
incorporation of micro-mixing effects into the model
can be carried out by introducing back-flow and lateral-flow ( Zuo et al., 2006). In our model, the bottom
(i=1) and top (i=N/2) sections are treated as wellmixed stages. The riser and the downcomer top sections, with i = 2, …, N/2-1 and i = N/2+1,…N respectively, are described as tanks-in-series. Also it is possible to modify the number of stages in each part.
At the top section, most of the gas bubbles passing
upward in the riser disengage and only the rest is
entrained downward by liquid recirculation into the
downcomer. On the other hand, the flow in the downcomer is almost single-phase and relatively well
defined. Therefore, the backmixing in the downcomer
is neglected. It has been assumed that the oxygen concentrations in the gas phase are uniform and that there
is no oxygen limitation for cells during SCP production. Consequently the oxygen balance is not taken
into account; this is due to the premise that the fluid
was saturated with oxygen. At the bottom section, the
gas feed and the recycle flow from the downcomer are
introduced. It is assumed that the fermentation has a
good temperature control and the temperature is constant. Consequently, in this study, energy balances are
not taken into account, as well as in the work of
Luttman et al. (1983), Kanai et al. (1996), Znad et al.
(2004).
The tanks-in-series model without back-flow provides simultaneous first order ordinary differential
equations, which are material balances of the microorganism and substrate in hypothetical well-mixed tanks
or stages. The unsteady state material balances of these
components can be written as follows:
For the microorganism (biomass), x, substrate (lactose), s:
Bottom section (i = 1):
dx/dt = Ql(xN - xi) / (vb(1 - εgr)) + µmsixi / (Ksxi + si) (1)
ds/dt = Ql(sN - si) / (vb(1 - εgr)) - µmsixi / (Ksxi + si) - λxi (2)
Riser section (i = 2, . . ., N/2-1):
dx/dt = Ql(xi-1-xi)(N/2 - 2) / (vr(1 - εgr))+ µmsixi/(Ksxi + si) (3)
ds/dt = Ql(si-1 - si)(N/2 - 2)/(vr(1 - εgr)) - µmsixi / (Ksxi + si) - λxi (4)
Top section (i = N/2)
dx/dt = Ql(xi-1- xi) / (vt(1 - εgr)) + µm sixi / (Ksxi + si) (5)
ds/dt = Ql(si-1- si) / (vt(1 - εgr)) - µm sixi / (Ksxi + si) - λxi (6)
Downcomer section (i = N/2 + 1, . . ., N)
dx/dt = Ql(xi-1 - xi)(N/2) / (vd(1 - εgd)) + µm sixi / (Ksxi + si) (7)
ds/dt = Ql(si-1 - si)(N/2) / (vd(1 - εgd)) - µm sixi/(Ksxi + si) - λxi (8)
Kinetic model:The kinetic model presented by Ghaly
et al. (2004) has been used in this simulation to describe the biomass production from cheese whey:
dxi/dt = µixi (9)
dsi/dt = -γ(dxi/dt ) – λxi (10)
where the specific growth rate is defined by:
µi = µmsi / (Ksx