You are on page 1of 12

Biochemical Engineering Journal 27 (2005) 167–178

Effects of simultaneous internal and external mass transfer and


product inhibition on immobilized enzyme-catalyzed reactor
Ali E. AL-Muftah a,∗ , Ibrahim M. Abu-Reesh b
a Petroleum Engineering Department, Bahrain Petroleum Company (BAPCO), Awali, P.O. Box 25504, PETEX, Bahrain
b Department of Chemical Engineering, King Fahd University of Petroleum & Minerals (KFUPM), Dhahran 31261, Saudi Arabia

Received 13 February 2005; received in revised form 23 July 2005; accepted 2 August 2005

Abstract

A mathematical model has been developed for predicting the performance and simulation of a packed bed immobilized enzyme reactor
performing lactose hydrolysis, which follows Michaelis–Menten kinetics with competitive product (galactose) inhibition. The performance
characteristics of a packed bed immobilized enzyme reactor have been analyzed taking into account the simultaneous effects of internal and
external mass transfer limitations. The model design equations are then solved by the method of weighted residuals such as Galerkin’s method
and orthogonal collocation on finite elements.
The effects of simultaneous internal and external mass transfer coupled with product inhibition have been studied and their effects were
shown to reduce internal effectiveness factor. The effects of product inhibition have been investigated at different operating conditions
correlated at different regimes using dimensionless βxo (St, Bi, θ, φ). Product inhibition was shown to reduce substrate conversion, and to
decrease effectiveness factor when βs > βxo ; however, it increases internal effectiveness factor when βs < βxo . The effectiveness factor is found
to be independent of product inhibition at crossover point at which βxo is defined. Effects of St and Bi have been investigated at different
kinetic regimes and the results show their effects have a strong dependence on kinetic parameters θ, γ (i.e. Km /Kp ) and βxo . The dimensionless
residence time at crossover point, βxo , has been correlated with kinetic and mass transfer parameters.
© 2005 Elsevier B.V. All rights reserved.

Keywords: Biocatalysis; Enzyme bioreactors; Immobilisation; Immobilised enzymes; Product inhibition; Internal mass transfer

1. Introduction The amount of lactose produced annually from whey


is about 3.3 million tonnes [3]. It is produced as cheese
Lactose is a disaccharide that occurs naturally in both whey, which is the liquid, separated after milk coagulation.
human and cows milk which accounts for 40% of milk It represents about 90% of the milk volume. The disposal
solids. It is widely used in baking and commercial infant- of whey is considered a serious pollution problem facing
milk formulas. The hydrolysis of lactose, the sugar of milk to dairy industry because of its high pollutant content (COD of
glucose and galactose has received much attention in recent about 70,000 ppm). Acid hydrolysis of lactose is not favor-
years [1,2]. It is used for production of low lactose milk for able because of color formation and fouling of ion exchange
consumers that suffer from lactose deficiency (70% of the resins used in processing. A better alternative is the use of
world population is lactose deficient, Carrara and Rubiolo enzymatic method. Enzymatic lactose hydrolysis is carried
[3]). The hydrolysis product is sweeter and more soluble and out by adding ␤-galactosidase commonly known as lactase
biodegradable than lactose and can be used in further biotech- to milk, skim milk or whey to hydrolyze lactose prior to pas-
nological processes. teurization. Lactase is commercially available and used in
large-scale processes. One problem associated with the use
of lactase is that complete hydrolysis is difficult to achieve
∗ Corresponding author. because of product (galactose) inhibition and production of
E-mail address: ali almuftah@bapco.net (A.E. AL-Muftah). isomer of lactose, allolactose. Several microbial sources of

1369-703X/$ – see front matter © 2005 Elsevier B.V. All rights reserved.
doi:10.1016/j.bej.2005.08.026
168 A.E. AL-Muftah, I.M. Abu-Reesh / Biochemical Engineering Journal 27 (2005) 167–178

Nomenclature
Pb dimensionless product concentration in the
A external surface of support per unit volume of bulk liquid
reactor Pb0 dimensionless product concentration at the
CP product concentration in an immobilized reactor inlet
enzyme support particle Pe Peclet number
CP  average product concentration in an immobi- R̃ dimensionless reaction rate in an immobilized
lized enzyme support particle enzyme support particle
CPb product concentration in the bulk liquid (reac- S dimensionless substrate concentration in an
tor phase) immobilized enzyme support particle
CPb0 , CP0 product concentration at reactor inlet S dimensionless average substrate concentration
CS substrate concentration in an immobilized in an immobilized enzyme support particle
enzyme support particle Sb dimensionless substrate concentration in bulk
CS  average substrate concentration in an immobi- liquid
lized enzyme support particle Sb0 dimensionless substrate concentration at reac-
CSb substrate concentration in the bulk liquid tor inlet
CSb0 , CS0 substrate concentration at reactor inlet St Stanton number
DSp, DPp effective substrate and product diffusivity in X fractional substrate conversion
an immobilized enzyme support particle Y reactor yield
DSz , DPz effective substrate and product axial disper-
sion coefficient Greek symbols
Ke reaction equilibrium constant αz , αP effective diffusivity ratio of substrate and prod-
KL mass transfer coefficient uct in axial and interparticle, respectively
KL a volumetric mass transfer coefficient β, βs dimensionless residence modulus
KLS , KLP mass transfer coefficient in substrate and ε reactor voidage
product side, respectively φ Thiele modulus
Km intrinsic Michaelis–Menten constant γ dimensionless inhibition modulus
Kp product inhibition constant η internal effectiveness factor
L length of the reactor ηE external effectiveness factor
r radial coordinate of distance in an immobilized θ dimensionless Michaelis–Menten constant
enzyme support particle τ dimensionless time
R̃b dimensionless reaction rate at the surface of ξ dimensionless radial coordinate
the spherical particles ζ dimensionless axial coordinate
RP local product production rate per unit of cat-
alytic particle volume
RP  average product production rate ␤-galctosidase and reactor types have been used for the pur-
RS local substrate consumption rate per unit of pose of economic production of low lactose milk. Lactose
catalytic particle volume hydrolysis in plug flow reactor gives higher conversion com-
RS  average substrate consumption rate pared to continuous stirred tank reactor although the latter
t time inside reactor has good mixing and lower construction cost.
u superfacial fluid phase velocity inside the reac-
tor 1.1. Kinetics of lactose hydrolysis
vmax maximum reaction rate per unit of catalytic
particle volume Kinetics of lactose hydrolysis has been studied extensively
x, z reactor radial and axial coordinate in the literature [1,2,4]. Michaelis–Menten model with com-
petitive product inhibition by galactose is widely used to
Dimensionless variables
describe the hydrolysis [1]. Different types of bioreactor [5,6]
Bi Biot number
and biocatalyst [1,2,7,8] have been investigated for lactose
Da Damkohler number
hydrolysis.
KE inverse of the equilibrium constant
P dimensionless product concentration in an
immobilized enzyme support particle 1.2. Modeling immobilized enzyme reactor
P dimensionless average product concentration
in an immobilized enzyme support particle Enzyme immobilization offers a number of advantages
over enzymes in suspension. Immobilization permits the
reuse of the enzyme and may provide a better environment
A.E. AL-Muftah, I.M. Abu-Reesh / Biochemical Engineering Journal 27 (2005) 167–178 169

for catalyst activity and also it reduces the cost of down reactor) is shown to be superior in predicting the perfor-
stream processing in addition to good product quality. These mance of packed bed reactor for isomerization of glucose
factors make it widely used in industry where many enzyme- to fructose described by Michaelis–Menten [10,20]. Further-
catalyzed reactions are of industrial interest. more, dynamic behavior of plug flow reactor was studied and
Continuous processes with immobilized enzymes can be the kinetics parameters were estimated by fitting the exper-
carried out in different types of reactor. Lortie and Pelletier imental data to satisfy the dynamic response RTD curve.
[9] have shown that plug-flow reactor model with external Michaelis–Menten kinetics is considered in experimental val-
mass transfer resistance can represent adequately a fixed bed idation of PFR [21,22] and in spiral reactor [6]. Abu-Reesh
immobilized enzyme reactor for moderate or low dispersion. [11] developed a general dimensionless model for predicting
However, dispersed plug flow reactor model is superior in the steady state performance of immobilized dispersed plug-
predicting the performance of packed bed reactor for isomer- flow reactor performing reversible Michaelis–Menten kinet-
ization of glucose to fructose [10]. Furthermore, a compre- ics. The effects of dimensionless parameters of Damkohler
hensive model for a general rate expression such as reversible number (Da), Stanton number (St), Peclet number (Pe), the
Michaelis–Menten kinetics was developed by Abu-Reesh equilibrium constant and input substrate concentration were
[11] neglecting internal mass transfer resistance. A fluidized studied parametrically. Abu-Reesh [11] found that conver-
bed reactor model taking into account the reversibility of the sion is almost complete for high Da and St number especially
reaction, inhibition by substrate and products or diffusional in plug flow reactor which gives higher conversion compared
limitations was developed [12,13]. to other reactor models. Furthermore, it is found that substrate
Among these reactors, packed bed enzyme-catalyzed reac- conversion increases with increasing substrate external diffu-
tors have promising applications in many biochemical pro- sion (i.e., decreasing diffusion resistance) and residence time.
cesses [11,14,15] and biological processes [16] having advan- Moreover, the higher the St number, the higher the maximum
tages of longer solid retention times and ease of operation and conversion that can be achieved. The effect of the equilibrium
relatively high conversion rates. constant on reactor performance was also studied. Carrara et
In spite of the well-established industrial application of al. [23] studied the behavior of fixed bed reactor considering
packed bed immobilized enzyme reactors, little effort has steady-state conditions and external mass transfer resistance
been made toward mathematical modeling of such reactor in the fluid around spherical catalyst particles. Their results
with kinetics other than Micaelis–Menten equation. In most showed the importance of hydrodynamic and kinetic reaction
of the cases, enzyme immobilization is accompanied by mass parameters for error reduction in the prediction of experimen-
transfer limitation. Different factors have to be taken into tal behavior.
consideration in modeling of immobilized enzyme reactors: When enzyme is attached to a porous carrier matrix the
internal mass-transfer limitations have a great influence on
(1) Mode of operation, whether the steady state or transient.
the intrinsic kinetics. It is necessary to develop compre-
(2) Mass transfer limitations (external, internal and simulta-
hensive models that quantitatively account for the internal
neous transfer).
diffusional effects in addition to external one. The inter-
(3) Resistance of the membrane to any transport processes.
nal diffusional limitations with external diffusional effects
(4) The kinetics of enzyme catalyzed reaction.
can be quantified through the use of an effectiveness factor,
(5) The axial dispersion effects.
η, or apparent kinetic parameters using the idealized plug
(6) Types of reactors: packed bed, CSTR, hollow fiber biore-
flow reactor assuming Michaelis–Menten kinetics [24–30].
actor (HFBR) or fluidized bed reactor (FBR).
Isothermal, steady state and omission of the axial dispersion
(7) The stability of enzyme and effect of temperature on the
term were assumed in these analyses. The Biot number, effec-
enzyme activity.
tiveness factor and Thiele modulus were studied numerically
(8) Heat transfer effects.
using Taylor expansion and orthogonal collocation methods.
Quantitative knowledge of the effect of these factors on the Theoretical analysis was incorporated with experimental data
reactor performance and simulation is required for efficient to correlate mass transfer coefficient [26].
design of immobilized enzyme reactor. Several isothermal Many researchers considered coupled internal and exter-
steady state models have been considered using one or more nal diffusional limitation [12,31–37]. Coupled internal and
of these phenomena in various combinations. external diffusional limitation were considered in develop-
External mass transfer limitation is shown to have sig- ment of a general CSTR model in which the effect of mem-
nificant effect on the performance of immobilized enzyme brane diffusional resistance and Biot number were taken into
reactor [17] using reversible enzyme reactions [18]. Analyti- account for prediction the effectiveness factor of an encapsu-
cal solution was given by Carrara and Rubiolo [4] and tested lated enzyme particle [38]. Also, analytical solution of effec-
with experimental setup, for evaluation of mass transfer coef- tiveness factor was developed for Michaelis–Menten kinetics.
ficient and conversion. However, Kobayashi and Moo-Young Bódalo et al. [12] and Manjon et al. [39] considered the exter-
[19] were the first to apply the dispersion model to immo- nal and internal diffusional limitations and their model was
bilized enzyme reactor. Dispersed plug flow reactor model solved numerically for reversible Michaelis–Menten kinet-
(taking into account the effect of axial dispersion on flow ics with competitive product inhibition in a fluidized bed
170 A.E. AL-Muftah, I.M. Abu-Reesh / Biochemical Engineering Journal 27 (2005) 167–178

reactor (FBR). They found a good agreement between the (7) Assume no enzyme deactivation.
model predictions experimental data obtained from operating (8) The enzyme can catalyze a specific reaction according
a FBR containing ␤-galactosidase covalently immobilized in to reversible Michaelis–Menten kinetics.
Chromosorb-W. Xiu et al. [40] developed a model for immo- (9) The enzymatic reaction is monosubstrate and yields
bilized enzyme catalyzed kinetic resolution of racemate in a only one product.
fixed-bed reactor. They studied the effects of mass transfer (10) Ficks law can model the substrate and product diffusion
limitations, competitive substrate inhibition and deactivation inside the catalytic particle. The effective diffusivity
of immobilized enzyme. does not change throughout the particles and is inde-
Since the effects of various phenomena such as axial pendent of the concentration.
dispersion, external mass transfer resistance with reversible
Michaelis–Menten kinetics will be strongly coupled, mak- 2.1.2. Mass balance on the reactor
ing necessary prediction of reactor performance difficult. Under these assumptions, the differential mass balance in
Most of the published work assume either first order or the liquid bulk phase for the substrate and the product can be
Michaelis–Menten kinetics with an external or internal mass written respectively as,
transfer. Many enzyme-catalyzed reactions of industrial inter-
est do not follow this kinetics such as lactose hydrolysis by ∂CSb ∂2 CSb ∂CSb
ε = DSz ε −u
the enzyme lactase, which is modeled in the literature using ∂t ∂z2 ∂z
Michaelis–Menten model with competitive product inhibi- −(1 − ε)KL a (CSb − CS |r=R ) (1)
tion by galactose [1]. This kinetic equation is also able to
explain Michaelis–Menten kinetics when the product con-
centration is very low. ∂CPb ∂2 CPb ∂CPb
Therefore, the objective of this work is to study the steady ε = DPz ε −u
∂t ∂z 2 ∂z
state performance of immobilized enzyme reactor perform-
ing lactose hydrolysis at various operational conditions. The −(1 − ε)KL a (CPb − CP |r=R ) (2)
effect of internal and external mass transfer limitation, axial
where ε is the reactor voidage, DSz , DPz are the effective
dispersion and kinetic parameters on reactor performance
substrate and product axial diffusivity (or axial dispersion
will be studied. Lactose conversion and the effectiveness fac-
coefficients) and KL a is the overall mass transfer coefficient.
tor will be used as performance criteria for the immobilized
The initial condition for reactor start-up contains the
enzyme reactor. The effect of galactose inhibition on reactor
immobilized enzyme particles suspended in a solution with-
performance will also be examined.
out substrate or product. At t = 0+ , the substrate and product
are continuously pumped into and out of the reactor at con-
stant rate,
2. Mathematical analysis
at z = 0− , CSb = 0.
2.1. Internal and external mass transfer—reaction
Eqs. (1) and (2) are subjected to the following boundary
model
conditions requiring continuity of fluxes at both ends of the
reactor [41] for both substrate and product
Consider a packed bed immobilized enzyme reactor of

length, L, fluid velocity, u, where CS0 and CP0 are the substrate DSz ε ∂CSb 
at z = 0+ , CSb |z=0+ = CSb |z=0− + ,
and product inlet concentrations, respectively. u ∂z z=0+

2.1.1. Assumptions DPz ε ∂CPb 
CPb |z=0+ = CPb |z=0− + (3)
The mathematical model that describe the behavior of a u ∂z z=0+
packed bed immobilized enzyme reactor has been formulated
 
using the following assumptions: ∂CSb  ∂CPb 
at z = L, = = 0. (4)
(1) Isothermal packed bed immobilized enzyme reactor. ∂z z=L ∂z z=L
(2) Enzyme activity is uniform throughout the particle.
The substrate and product mass balance equations in
(3) The enzyme is immobilized evenly inside porous spher-
immobilized enzyme particles on a porous spherical parti-
ical particles, which are uniformly packed in the reactor.
cle support is given as:
(4) The convective velocity is uniform.
 
(5) The hydrodynamics of the fluid bed is described by the ∂CS 1 ∂ ∂CS
= DSp 2 r2 − RS (5)
dispersed plug-flow model. ∂t r ∂r ∂r
(6) Pressure drop across the reactor and radial concentra-  
tion gradient in the bulk fluid phase are assumed to be ∂CP 1 ∂ 2 ∂CP
= DPp 2 r + RP (6)
negligible. ∂t r ∂r ∂r
A.E. AL-Muftah, I.M. Abu-Reesh / Biochemical Engineering Journal 27 (2005) 167–178 171

where, DSp , DPp are the effective substrate and product inter- Table 1
particle diffusivity. Dimensionless parameters
1. Kinetic parameters
Michaelis modulus θ = CKm
S0
2.1.3. Kinetic equation of lactose hydrolysis Product inhibition γ = CKS0P
Consider the general mechanism of reversible modulus
Michaelis–Menten kinetics. The reversible equation is 2. Internal mass transfer parameters
 ε 
St Lε DSp
able to explain irreversible Michaelis–Menten when Dimensionless residence βs = Bi 1−ε = u R2
(Ke → ∞ and Kp → ∞) and competitive inhibition by a time, βs
Da St Da vmax R2
product when (Ke → ∞). For example, enzymatic lactose Thiele modulus, φ φ2 = βs = θβs = Km DSp
DSp
hydrolysis has been modeled in the literature using soluble Diffusivity ratio in the αp = DPp
β-galactosidase [1]. Michaelis–Menten model with com- internal sites of the
enzyme particles
petitive product inhibition by galactose is widely used to
describe lactose hydrolysis. The hydrolysis rate is given by
3. External mass transfer parameters  
Stanton number, St St = βs Bi 1−ε = (1 − ε)KL a Lu
(Ke → ∞): Da
ε
φ2
 ε
 ε vmax
Damkohler number, Da Da = θSt = Bi 1−ε = 1−ε Km KL a
θφ2 Lε vmax
CS − CP /Ke Modified Damkohler, Da = θDa St = = Bi St u CS0
RS = RP = vmax (7) Da
Km (1 + CP /Kp ) + CS 4. Simultaneous internal and external mass transfer parameters
Substrate, product Biot BiS = KDLS R , BiP = KDLPPpR
where, R(CS , CP ) = reaction rate, mol/l h; CS = substrate (lac- number
Sp

tose) concentration, mol/l. CP = product (galactose) concen- 5. Axial dispersion on external fluid side
tration, mol/l. Km = apparent Michaelis–Menten constant, Peclet number Pe = DLuε
Sz
mol/l. Kp = inhibition constant, mol/l. vmax = apparent maxi- Diffusivity ratio with αz = D Sz
DPz
mum reaction rate, mol/l h. respect to the axial
position
The rate constants vmax , Km and Kp depend on temperature
according to Arrhenius relationship [1].
The initial and boundary conditions are taken as,
The model can be described by the following dimension-
at t = 0, CS = CP = 0 less parameters
 
∂CS  ∂CP  (8) Km CS0 1 vmax R2
at r = 0, = =0
∂r r=0 ∂r r=0 θ=
CS0
, γ=
KP
, KE =
Ke
, φ2 =
Km DSp
,

 DSp Lε vmax Km
∂CS  αp = , Da = , R̃s = Rs ,
at r = R, DSp = KLS (CSb − CS |r=R ) , DPp u CS0 vmax
∂r r=R Lu DSz L
 Pe = , αz = , St = (1 − ε)KL a ,
∂CP  DSz ε DPz u
DPp = KLP (CPb − CP |r=R ) . (9)
∂r r=R KLP R KLS R
BiP = , BiS = .
DPp DSp

2.1.4. Governing equations in dimensionless form Consequently, Eqs. (1)–(9) can be written in dimension-
Eqs. (1)–(9) can be reduced to the corresponding dimen- less form as follows,
sionless forms by introducing the following dimensionless
∂Sb 1 ∂ 2 Sb ∂Sb  
parameters (Table 1). Da = − − St S b − S| ξ=1 (12a)
Substrate and product concentration variables ∂τ Pe ∂ζ 2 ∂ζ

CS CS CSb ∂Pb 1 ∂ 2 Pb ∂Pb  


S= = , Sb = , Da = − − St P b − P| ξ=1 (12b)
CS0 CSb |z=0− CS0 ∂τ αz Pe ∂ζ 2 ∂ζ
CP CPb with the following boundary conditions
P= , Pb = . (10)
CS0 CS0 
1 ∂Sb 
at ζ = 0+ , Sb |ζ=0+ = 1 + ,
Dimensionless axial, radial coordinate variables and Pe ∂ζ ζ=0+

dimensionless residence time 1 ∂Pb 
Pb |ζ=0+ = Pb |ζ=0− + (13)
z r vmax t αz Pe ∂ζ ζ=0+
ζ= , ξ= , τ= . (11)
L R CS0
172 A.E. AL-Muftah, I.M. Abu-Reesh / Biochemical Engineering Journal 27 (2005) 167–178

 
∂Sb  ∂Pb  if all enzyme molecules inside the particle were exposed to
at ζ = 1, = = 0. (14)
∂ζ ζ=1 ∂ζ ζ=1 the same substrate concentration as that at the surface, i.e.,
in the absence of diffusional effects. The bulk reaction rate
Mass balance for the substrate in immobilized enzyme is the reaction rate at the bulk concentrations. Alternatively,
particles supported on porous spherical particles is given by, the effectiveness factor may be defined as the ratio of the
  average reaction rate to the average bulk reaction rate, Rb .
∂S 1 ∂ ∂S
Da = βS 2 ξ2 − Da R̃S (15a) This relation is expressed by the following equation:
∂τ ξ ∂ξ ∂ξ
  1
∂S 1 1 ∂ ∂S RS  R̃S  3 0 R̃S ξ 2 dξ
or, = 2 2 ξ2 − R̃S (15b) η= = = . (23)
∂τ θφ ξ ∂ξ ∂ξ RS (CSR , CPR ) R̃b (Sb , Pb ) R̃b

and similarly the product concentration in immobilized 2.2.3. Fractional conversion and yield
enzyme particles is given by, Conversion is a convenient variable and it is often used in
  place of concentration in engineering work. It is defined as
∂P 1 1 ∂ 2 ∂P
= ξ + R̃P . (16) the ratio between the total moles of substrate converted into
∂τ αp θφ2 ξ 2 ∂ξ ∂ξ
product and the total moles of substrate fed into the reactor
Dimensionless reversible Michaelis–Menten kinetics can per unit time for a continuous reactor.
be written as: The following apparent conversions for the substrate and
S − KE P product can also be defined:
R̃S = R̃P = . (17)
θ(1 + γP) + S CSb
X=1− = 1 − Sb . (24)
Eqs. (15) and (16) are subject to the following boundary con- CSb0
ditions: Yield is defined as the ratio of the substrate converted to
 
∂S  ∂P  the maximum amount that could be converted during one
at ξ = 0, = =0 (18)
∂ξ  ξ=0 ∂ξ ξ=0
residence time. Yield is used to measure the efficiency of
enzyme utilization. It can be expressed as:
 CS0 − CSL 1 − Sb,L 1 − Sb,L
∂S   
Yield, Y = = =
at ξ = 1,  = BiS Sb − S|ξ=1 , vmax εL/u Da θDa St
∂ξ ξ=1
  
∂P    1 − Sb,L Bi 1−ε
= αp BiP Pb − P|ξ=1 . = = 2 X. (25)

∂ξ ξ=1
(19) θβS φ2 θφ St ε

2.2. Criteria for reactor performance


3. Results and discussions
2.2.1. Average concentrations and average reaction rate
The calculations of the substrate and product profile along The objectives of this study are to investigate the character-
the radial and axial coordinates were obtained from the solu- istics and behavior of packed bed immobilized enzyme reac-
tion of the diffusion reaction equations. The average concen- tors. These objectives can be achieved explicitly by studying
trations in the spherical particle can be obtained from the the effects of kinetics parameters, internal and external mass
following expressions transfer parameters, effects of reactor hydrodynamics and the
effects of simultaneous diffusional effects.
 R  1
3 Reactor conversion, yield and internal effectiveness fac-
CS  = 2 CS r dr or C = 3
2
S ξ 2 dξ (20)
R 0 0
tor are calculated as a function of dimensionless parameters
 R  1 Thiele modulus, φ, St, Pe, θ, and γ. Peclet number, Pe, which
3 measures the degree of dispersion is assumed to be 2.0 for
CP  = 2 CP r2 dr or P = 3 Pξ 2 dξ. (21)
R 0 0 DPFR model, since at higher Pe numbers the reactor perfor-
The average reaction rates are defined as: mance tends to approach the PFR model. In order to evaluate
the performance of packed bed immobilized enzyme reactors,
 R  1
3 numerical values of the process parameters were obtained
R̃S  = 2 RS r dr = 3
2
R̃S ξ 2 dξ. (22)
R 0 0 from the literature [3,4,11,12,14,16,18,22,23,40]. Since a
wide range of such values have been reported, the range of
2.2.2. Effectiveness factor dimensionless Michaelis modulus, θ, and Stanton number, St
The internal diffusional limitations can be quantitatively used in this simulation study of conversion, yield and internal
expressed by the effectiveness factor, η, defined as the ratio of effectiveness factor, was varied from 0.1 to 10. Similarly, the
the average reaction rate to the rate which would be obtained range of product inhibition modulus, γ, used in this analysis
A.E. AL-Muftah, I.M. Abu-Reesh / Biochemical Engineering Journal 27 (2005) 167–178 173

varies from γ = 0, corresponding to Michaelis–Menten kinet-


ics, to γ = 10. The developed model of immobilized enzyme
reactor has been analyzed by taking into account the effect
of kinetic parameters (θ, γ, KE ); external mass transfer limi-
tations (St, Da) and/or internal mass transfer limitation (φ,
αp ) and axial dispersion on external fluid side character-
ized by Pe and αz which represent reactor hydrodynamic.
In addition, Biot number represents simultaneous interaction
between internal and external mass transfer.
The differential mass balance equations have been written
and normalized for both substrate and product in both external
bulk phase and microenvironment phase (inside IME parti-
cles). The normalized PDEs are then solved using orthogonal
collocation on finite elements (OCFE) and Galerkin’s method
[42–46]. The number of collocation points and number of Fig. 2. Effects of Biot number on internal effectiveness factor as a function
finite elements were chosen to give satisfactory convergence. of Thiele modulus with θ = 1, γ = 1; Pe = 2.0 and St = 1.0.
Numerical simulation was performed to analyze the effect of
transport and kinetic parameters on the reactor performance. takes places when the process is reaction rate controlled. At
The model equations were solved using the orthogonal col- extremely high Thiele modulus, however, all curves of dif-
location method. Twelve internal collocation points were ferent Bi asymptotically approach constant value where the
chosen for the reactor bed axial direction and five collocation system becomes diffusion rate controlled. It also shows that
points were used inside enzyme particle. It is found that with the substrate conversion increases as a result of decreasing
these points a good accuracy can be obtained compared to Bi when the process is kinetically controlled. This relation
those results with 15 collocation points. becomes independent of Bi when the process is mass trans-
fer controlled. However, a weak relation is shown when both
intraparticle diffusion and kinetic processes are dominating
3.1. Effects of Bi and φ
(i.e., in the mixed regime).
Lower Biot number indicates the presence of strong exter-
Biot number is defined as the ratio of intraparticle diffu-
nal mass transfer resistance and hence both internal and exter-
sion resistance to external mass transfer resistance [47–48].
nal mass transfer resistances are important for the determina-
The effects of Biot number on the substrate conversion will
tion of substrate conversion. As the Biot number increases the
be studied with intraparticle limitations in Section 3.1. More-
external mass transfer resistance decreases in its importance.
over, effects of Bi and external mass transfer limitations with
This occurs because the microenvironment substrate conver-
kinetic parameters θ and γ will be discussed in Sections 3.2
sion approaches the bulk phase concentration and thereby
and 3.3, respectively.
the external mass transfer disappears. The effects of Bi are
Fig. 1 demonstrates quantitatively the effects of Biot num-
dominant in the reaction-controlled regime. However, the
ber on the substrate conversion as a function of Thiele mod-
conversion is independent of Biot number when the process
ulus with θ = 1, γ = 1, Pe = 2.0 and St = 1.0. At low Thiele
is diffusion rate.
modulus, the effect of increasing Biot number on the substrate
Fig. 2 shows the effects of Biot number on the internal
conversion is shown to reduce the substrate conversion. This
effectiveness factor as a function of Thiele modulus with
θ = γ = 1, Pe = 2.0 and St = 1.0. The trend of effectiveness fac-
tor versus Thiele modulus is shown to be a function of Biot
number. The effectiveness factor approaches unity and it is
independent of Bi when the process is kinetically controlled.
Conversely, the effectiveness factor increases with increasing
Biot number in the reaction controlled.

3.2. Effects of Bi, θ and St

Fig. 3 shows quantitatively the effects of Michaelis mod-


ulus, θ, and Stanton number on the substrate conversion as a
function of Biot number with γ = 1.0, φ = 2.0 and Pe = 2.0. It
shows that the conversion decreases upon increasing Biot
number at a given St. A further decrease in Biot number
Fig. 1. Effects of Biot number on reactor conversion as a function of Thiele results in the substrate conversion becoming independent
modulus with θ = 1, γ = 1; Pe = 2.0 and St = 1.0. of Michaelis modulus and the substrate conversion asymp-
174 A.E. AL-Muftah, I.M. Abu-Reesh / Biochemical Engineering Journal 27 (2005) 167–178

Fig. 3. Effects of St number on reactor conversion as a function of Biot Fig. 5. Effects of Biot number on internal effectiveness factor as a function
number with θ = 0.1, 1, 10, γ = 1, φ = 2.0 and Pe = 2.0. of Stanton number with γ = 1, φ = 2.0 and Pe = 2.0 for θ = 0.1, 1, 10.

totically approaches constant value. Also, it depends on St


• An increase in St reduces the effectiveness factor when
because the diffusional resistance is dominating factors in
θγ < 1 and increases the effectiveness factor when θ␥ > 1.
determing the substrate conversion. When the external mass
However, the effectiveness factor is independent of St when
transfer is very low (i.e., St = 0.1), the substrate conversion
θγ = 1.
is very low compared to higher St. In addition, the effect of
• As can be seen from Fig. 4 for a given Bi, increase St from
Michaelis modulus is negligible for lower Biot number.
θ = 0.1 and 1.0, causes the effectiveness factor to asymp-
The effects of mass transfer limitations were measured
totically approach a constant value at certain St. Above this
quantitatively by internal effectiveness factor. These effects
value of St, the effectiveness factor is only a function of Bi
are shown in Fig. 4, which illustrates the effects of Biot
regardless of St and θ. Similarly, for the case where θ = 10
number and St on the effectiveness factor for three differ-
and 1.0 the two curves approaches each other at a tangent
ent Michaelis modulus θ = 0.1, 1, 10 and γ = 1.0. Although
point (i.e., cusp point). This cusp point forms earlier for
the substrate conversion increases with decreasing Biot num-
the case of θ = 10 compared to θ = 0.1.
ber, it is found in Fig. 4 that the effectiveness factor increases
upon increasing Bi up to certain value above which no further It can be concluded that the effectiveness factor is only a
change is detected in effectiveness factor. In this regime, the function of θ when the process is kinetically controlled at very
reaction rate approaches the intrinsic value where the pro- high Bi and extremely low St. In this regime, upon increasing
cess is kinetically controlled. It is noticed that the increase St, the effectiveness factor becomes independent of St and
in St in the reaction rate regime will be approached at lower can be a function of Bi and θ. However, at extremely high
Bi. Therefore, at very high Bi, the effectiveness factor is a St, the effectiveness factor becomes independent of both St
function of Michaelis modulus as shown in Fig. 5. and θ and is characterized by Bi when the process is mass
The following results can be drawn from Figs. 4–6: transfer controlled. Between these two limiting cases, the
• At very low St, the effectiveness factor is characterized by effectiveness factor is a function of θ, Bi and St while other
Michaelis modulus and Bi, which is independent of St. parameters remains constant.

3.3. Effects of Bi, γ and St

A similar study was carried out to study the effects of


Biot number for different product inhibition modulus, γ. The
simulation results are shown in Fig. 7 for substrate conver-
sion. Figs. 8–10 illustrate the effects of Biot number on the
effectiveness factor.
Fig. 7 shows the effects of γ on the substrate conversion
for varying Biot number and Stanton number with φ = 2.0,
θ = 1.0, Pe = 2.0. As discussed earlier, it can be concluded
that the substrate conversion increases with decreasing Biot
number. The effect of product inhibition reduces the substrate
conversion, and more significant effects occurs when both
Fig. 4. Effects of θ on internal effectiveness factor for varying Biot number kinetic and mass transfer limitations are dominating factor,
and Stanton number with θ = 0.1, 1, 10, γ = 1, φ = 2.0 and Pe = 2.0. i.e., Bi lies in the mixed regime.
A.E. AL-Muftah, I.M. Abu-Reesh / Biochemical Engineering Journal 27 (2005) 167–178 175

Fig. 6. Effects of Stanton number on internal effectiveness factor as a function of Biot number for θ = 0.1, 1, 10, γ = 1, φ = 2.0 and Pe = 2.0.

Fig. 8 shows the effects of γ on the internal effectiveness


factor for varying Biot number and St with θ = 1.0, φ = 2.0
and Pe = 2.0. The trend can be characterized by three dis-
tinct regimes. First, the process is kinetically controlled when
the mass transfer resistances are negligible at very high Biot
number and low St. In this regime, the effectiveness fac-
tor is independent of both Bi and St but is a function of γ
where the product inhibition reduces the effectiveness fac-
tor. Second, the kinetic and mass transfer are controlling the
process when the effectiveness factor varies with St and it
depends on γ and Bi. In this regime, the effectiveness factor
increases with increasing St when θγ > 1, with decreasing St
at θγ > 1 and it is independent of St when θγ = 1. Third, the
Stanton number is extremely high, when the effectiveness
Fig. 7. Effects of St number on reactor conversion as a function of Biot
number with γ = 0, 1, 10, θ = 1, φ = 2.0 and Pe = 2.0.

Fig. 8. Effects of product inhibition on internal effectiveness factor for vary- Fig. 9. Effects of product inhibition on internal effectiveness factor with
ing Biot and Stanton number with θ = 1, φ = 2.0 and Pe = 2.0. Pe = 2.0 by varying St and φ at θ = 5.0, Bi = 0.1.
176 A.E. AL-Muftah, I.M. Abu-Reesh / Biochemical Engineering Journal 27 (2005) 167–178

effects of product inhibition form a crossover points trajec-


tory in St–φ at given Bi = 0.1. Therefore, the βs at crossover
points can be given by Eq. (26) at θ = 1.0,
6.0
βxo ≈ . (26)
φxo
2

The following equation can be correlated from Table 2


which can be used to approximate the βs at crossover point
that relates to θ and φ,
Stxo 6.0
βxo = ≈ 1/3 2 . (27)
Bixo θ φxo
Thus, the crossover points found in Fig. 10 can also be
Fig. 10. Effects of product inhibition ␥ on internal effectiveness factor for approximated by Eq. (27) at βxo = 10.
varying Michaelis and Thiele modulus for γ = 0, 1, 10; Pe = 2.0, Bi = 0.1 and
St = 1.0. 0.776
φxo ≈ . (28)
θ 1/6
factor asymptotically reaches a constant value regardless of Thus, the mixed regime can be characterized by βxo
St but depends on γ. It is shown that the product inhibition at which the effectiveness factor βs > βxo , whereas, when
favor the effectiveness factor in this regime. Therefore, the βs < βxo the effectiveness factor decreases as St increases.
effectiveness factor is independent of St, Bi when the mass Therefore, when βs < βxo , the reaction kinetic factors are
transfer is dominating factor in determining the effectiveness more significant where the effect of product inhibition has
factor. It can be concluded that the effect of product inhi- reduced the effectiveness factor by reducing reaction rate.
bition reduces the effectiveness factor when the process is However, when βs < ␤xo , the mass transfer limitations are
reaction rate controlled but it favors the effectiveness factor dominant and the product inhibition favors the internal effec-
when the process is mass transfer controlled. Although the tiveness factor since it reduces rate of mass transfer.
internal effectiveness factor is a function of γ in these two
regimes, the effectiveness factor is independent of both St and
Bi as shown in Fig. 8. 4. Conclusions
In the mixed regime, however, the effectiveness factor
increases with increasing Biot number and the trend is pro- A mathematical model for a packed bed immobi-
portional when θγ > 1. Whereas when θγ < 1, the trend is lized enzyme reactor has been developed considering
inversely proportional. Due to these two opposite trends Michaelis–Menten kinetics with competitive product inhi-
shown in the effect of γ and St on η, it is found that effective- bition. The effects of intraparticle diffusion, external mass
ness factor is independent of γ at certain values of St and Bi transfer, axial dispersion and kinetic parameters have been
at crossover point. It is evident from Fig. 8 that the crossover taken into consideration in the model. The relevant equations
points trajectory can be represented by dimensionless param- were solved by the method of orthogonal collocation on finite
eters at crossover point βxo = Stxo /Bixo = 1.5, at θ = 1.0 and elements and Galerkin’s method. The performance of packed
φ = 2.0 that satisfied all crossover points trajectory. bed immobilized enzyme reactor has been investigated para-
Further to investigate the effects of θ and φ on the crossover metrically for various operational parameters. The effects of
points trajectory, a similar study was conducted for different θ, γ, φ, St and Bi have been identified quantitatively on the
combinations of θ and φ and the simulation results are sum- substrate conversion and internal effectiveness factor.
marized in Table 2. The following conclusions were drawn from simulation
Furthermore, Fig. 9 illustrates the effects of product inhi- results:
bition on the internal effectiveness factor as a function of St
and φ with θ = 1.0, Bi = 0.1 and Pe = 2.0. As can be seen, the (1) Intraparticle diffusion resistance, external mass transfer
resistances and axial dispersion were shown to reduce
Table 2 internal effectiveness factor.
Calculated βxo from different simulation results (2) Product inhibition was shown to reduce substrate conver-
Parameters θ xo φxo βxo sion, and to decrease effectiveness factor when βs > βxo ;
Fig. 8: θ = 1, φ = 2.0, Pe = 2.0 1.0 2.0 1.5
however, it increases effectiveness factor when βs < βxo .
Fig. 9: Pe = 2.0; φ = 2.0, θ = 0.5 5.0 2.0 0.95 The effectiveness factor is found to be independent of
θ = 1, Pe = 2.0, φ = 0.5 1.0 0.5 20 product inhibition at crossover point at which βxo is
θ = 1, Pe = 2.0, φ = 4.0 1.0 4.0 0.4 defined, where βxo is a function of St, Bi, θ and φ.
Pe = 2.0, φ = 2.0, θ = 0.1 0.1 2.0 6.0 (3) Effect of Stanton number was shown to reduce the inter-
Pe = 2.0, φ = 2.0, θ = 5.0 0.5 2.0 2.1
nal effectiveness factor when θγ < 1 (i.e., Km < Kp ), but it
A.E. AL-Muftah, I.M. Abu-Reesh / Biochemical Engineering Journal 27 (2005) 167–178 177

favors the effectiveness factor when θγ > 1. Due to these References


opposite trends, the effectiveness factor has been found
[1] Santos, M. Ladero, F. Garcia-Ochoa, Kinetic modeling of lac-
to be independent of St at θγ = 1.
tose hydrolysis by a ␤-galactosidase from Kluyveromices fragilis,
(4) The effectiveness factor is only a function of θ when Enzyme Microb. Technol. 22 (1998) 558–567.
the process is kinetically controlled at very high Bi [2] I. Petzelbauer, B. Nidetzky, D. Haltrich, K. Kulbe, Development
and extremely low St. In this regime, upon increas- of an Ultra-High-Temperature process for the Enzymatic hydrolysis
ing St, the effectiveness factor becomes independent of lactose. I. The properties of two thermostable ␤-galactosidase,
Biotechnol. Bioeng. 64 (1999) 322–332.
of St and can be a function of Bi and θ. However, at
[3] C.R. Carrara, A.C. Rubiolo, Determination of kinetics parameters for
extremely high St, the effectiveness factor becomes inde- free and immobilized ␤-galactosidase, Process Biochem. 31 (1995)
pendent of both St and θ and is characterized by Bi 243–248.
when the process is mass transfer controlled. Between [4] C.R. Carrara, A.C. Rubiolo, A method for evaluating lactose hydrol-
these two limiting cases, the effectiveness factor is a ysis in a fixed bed reactor with ␤-Galactosidase immobilized on
chitson, Chem. Eng. J. 65 (1997) 93–98.
function of θ, Bi and St while other parameters remains
[5] S.T. Yang, M.R. Okos, A new graphical method for determining
constant. parameters in Michaelis–Menten-type kinetics for enzymatic lactose
(5) The effect of product inhibition reduces the effectiveness hydrolysis, Biotechnol. Bioeng. 34 (1989) 763–773.
factor when the process is reaction rate controlled but it [6] A.P. Bakken, C.G. Hill, C.H. Amundson, Hydrolysis of lactose in
favors the effectiveness factor when the process is mass skim milk by immobilized ␤-galactosidase in a spiral flow reactor,
Biotechnol. Bioeng. 33 (1989) 1249–1257.
transfer controlled. Although the internal effectiveness
[7] N. Papayannakos, G. Markas, D. Kekos, Studies on modeling
factor is a function of γ in these two regimes, the effec- and simulation of lactose hydrolysis by free and immobilized ␤-
tiveness factor is independent of both St and Bi. galactosidace from Aspergillus niger, Chem. Eng. J. 52 (1993)
(6) In the mixed regime, however, the effectiveness factor B1–B12.
increases with increasing Biot number and the trend [8] M. Tomaska, M. Stredansky, A. Tomaskova, E. Sturdik, Lac-
tose hydrolysis in aqueous two-phase system by whole cell ␤-
is proportional when θγ > 1. Whereas when θγ < 1, the
galactosidase of Kluyveromyces marxianus, Bioprocess Eng. 12
trend is inversely proportional. (1995) 17–20.
(7) Due to existence of two opposite trends shown in the [9] R. Lortie, D. Pelletier, Comparison between dispersion and plug
effect of γ and St on η, it is found that effectiveness fac- flow models for fixed-bed enzyme reactors, AIChE J. 38 (1992)
tor is independent of γ at certain values of St and Bi 1477–1480.
[10] C.B. Ching, Y.Y. Ho, Flow dynamics of immobilized enzyme reac-
at crossover point. The crossover points trajectory can
tors, Appl. Microbiol. Biotechnol. 20 (1984) 303–309.
be represented by dimensionless parameters at crossover [11] I.M. Abu-Reesh, Predicting the performance of immobilized enzyme
point βxo = Stxo /Bixo and can be approximated by the fol- reactors using reversible Michaelis–Menten kinetics, Bioprocess Eng.
lowing correlation: 17 (1997) 131–137.
[12] A. Bódalo, E. Gómez, J.L. Gomez, J. Bastida, M.F. Maximo, Flu-
Stxo 6.0 idized bed reactors operating with immobilized enzyme systems:
βxo = ≈ 1/3 2 . Design model and its experimental verification, Enzyme Microb.
Bixo θ φxo Technol. 17 (1995) 915–922.
[13] T. Kisser, G.A. Oertzen, W. Bauer, Modeling of a fluidized bed
bioreactor for immobilized enzyme (part 1), Chem. Eng. Technol.
Therefore, the mixed regime can be characterized by βxo 13 (1990) 20–26.
at which the effectiveness factor βs > βxo , whereas, when [14] M.M. Hassan, M. Atiqullah, S.A. Beg, M. Chowdhury, Analysis
βs < βxo the effectiveness factor decreases as St increases. of nonisothermal tubular reactor packed with immobilized enzyme
systems, Biochem. Eng. J. 58 (1995) 275–283.
Therefore, when βs < βxo , the reaction kinetic factors are [15] M.M. Hassan, M. Atiqullah, S.A. Beg, M.H.M. Chowdhury, Effects
more significant where the effect of product inhibition of enzyme microcapsule shape on the performance of a nonisother-
has reduced the effectiveness factor by reducing reaction mal packed-bed tubular reactor, J. Chem. Technol. 66 (1996) 41–
rate. However, when βs < βxo , the mass transfer limita- 55.
tions are dominant and the product inhibition favors the [16] M.M. Hassan, S.A. Beg, Theoretical analysis of a packed bed bio-
logical reactor for various reaction kinetics, Chem. Eng. J. 36 (1987)
internal effectiveness factor since it reduces rate of mass 15–27.
transfer. [17] D. Vasic-Racki, N. Pavlovic, S. Cizmek, M. Drazic, B. Husadzic,
Development of reactor model for glucose isomerization catalyzed
by whole-cell immobilized glucose isomerase, Bioprocess Eng. 7
(1991) 183–187.
Acknowledgments [18] Y.H. Park, M.H. Han, H.K. Rhee, Effect of external mass transfer
in a packed bed reactor system for a reversible enzyme reaction, J.
The authors gratefully thank Prof. Abdullah Shaikh, chair- Chem. Technol. Biotechnol. 34B (1984) 57–69.
man of Chemical Engineering Department, and Dr. Kevin F. [19] T. Kobayashi, M. Moo-Young, Backmixing and mass transfer in
the design of immobilized -enzyme reactors, Biotechnol. Bioeng. 13
Laughlin for their technical support during the study. This
(1971) 893–910.
work was funded by Scientific Research Department at King [20] C.B. Ching, K.H. Chu, Modeling of a fixed bed and a fluidized bed
Fahd University of Petroleum & Minerals (KFUPM), Saudi immobilized enzyme reactor, Appl. Microbiol. Biotechnol. 29 (1988)
Arabia, through grant ENZYME project. 316–322.
178 A.E. AL-Muftah, I.M. Abu-Reesh / Biochemical Engineering Journal 27 (2005) 167–178

[21] R. Lortie, D. Thomas, Heterogeneous one-dimensional model for [33] D.D. Do, D.S. Clark, J.E. Bailey, Modeling enzyme immobilization
fixed bed enzyme reactors, Biotechnol. Bioeng. (1986) 1256– in porous solid supports, Biotechnol. Bioeng. 24 (1982) 1527.
1260. [34] D.S. Clark, J.E. Bailey, Structure-function relationships in immobi-
[22] R. Lortie, Evaluation of the performance of immobilized enzyme lized chymotrypsin catalysis, Biotechnol. Bioeng. 25 (1983) 1027.
reactors with Michaelis–Menten kinetics, J. Chem. Technol. Biotech- [35] D.S. Clark, J.E. Bailey, Deactivation kinetics of immobilized ␣-
nol. 60 (1994) 189–193. chymotrypsin subpopulations, Biotechnol. Bioeng. 26 (1984) 1090.
[23] C.R. Carrara, E.J. Mammarela, A.C. Rubiolo, Prediction of the fixed- [36] D.S. Clark, J.E. Bailey, D.D. Do, A mathematical model for
bed reactor behavior using dispersion and plug flow models with restricted diffusion effects on macromolecule impregnation in porous
different kinetics for immobilized enzyme, Chem. Eng. J. 92 (1–3) supports, Biotechnol. Bioeng. 27 (1985) 208.
(2002) 123–129. [37] R. Shyan, B. Davidsson, W.R. Vieth, Mass transfer and biochemical
[24] W.N. Marrazzo, B.J. McCoy, Enzyme immobilized in a packed bed reaction in enzyme membrane reactor system 2 expanded analysis
reactor: kinetic parameters and mass transfer effects, Biotechnol. for single enzyme system, Chem. Eng. Sci. 30 (1975) 669–678.
Bioeng. 23 (1981) 1237–1254. [38] S.H. Lin, The effects of mass transfer resistances on encapsulated
[25] F. Shiraishi, S. Fujiwara, An efficient method for solving two-point enzymatic reaction systems, Chem. Eng. J. 17 (1979) 55–61.
boundary value problems with extremely high accuracy, J. Chem. [39] A. Manjon, J.L. Iborra, J.L. Gomez, E. Gómez, J. Bastida, A.
Eng. Jpn. 29 (1996) 88–94. Bódalo, Evaluation of effectiveness factor along immobilized enzyme
[26] F. Shiraishi, T. Hasegawa, H. Nagasue, Numerical solution of two- fixed-bed reactors: Design of a reactor with naringinase covalently
point boundary value problem by combined Taylor series method immobilized into glycophase coated porous class, Biotechnol. Bio-
with a technique for rapidly selecting suitable step sizes, J. Chem. eng. 30 (1987) 491–497.
Eng. Jpn. 28 (1995) 306–315. [40] G.-H. Xiu, L. Jiang, L. Ping, Mass-transfer limitations for immobi-
[27] F. Shiraishi, T. Hasegawa, S. Kasai, N. Makishita, H. Miyakawa, lized enzyme-catalyzed kinetic resolution of Racemate in a fixed-bed
Characteristic of apparent kinetics parameters in a packed bed immo- reactor, Biotechnol. Bioeng. 74 (2001) 29–39.
bilized enzyme reactor, Chem. Eng. Sci. 51 (1996) 2847–2852. [41] P.V. Danckwerts, Continuous flow systems, Distribution of residence
[28] F. Shiraishi, H. Miyakawa, T. Hasegawa, S. Kasai, A computational times, Chem. Eng. Sci. 2 (1953) 1–18.
method for determination of the mass transfer coefficient in packed [42] J. Villadsen, M.L. Michelsen, Solution of differential equation by
bed immobilized enzyme reactors, J. Chem. Technol. 66 (1996) polynomial approximation, in: International Series in Physical and
405–413. Chemical Engineering Series, Prentice-Hall, Englewood Cliffs, NJ,
[29] H. Miyakawa, H. Nagasue, F. Shiraishi, A highly accurate numerical 1978.
method for calculating apparent kinetics parameters of immobilized [43] B.A. Finlayson, Nonlinear Analysis in Chemical Engineering,
enzyme reactions: 1. Theory, Biochem. Eng. J. 3 (1999) 91–101. McGraw-Hill, New York, 1980.
[30] H. Miyakawa, H. Nagasue, F. Shiraishi, A highly accurate numerical [44] R.A. Adomaitis, Y.-H. Lin, A technique for accurate collocation
method for calculating apparent kinetics parameters of immobilized residual calculations, Chem. Eng. J. 71 (1998) 127–134.
enzyme reactions: 2. Accuracies of calculated values, Biochem. Eng. [45] R.A. Adomaitis, Objects for MWR, Comput. Chem. Eng. 26 (2002)
J. 3 (1999) 103–111. 981–998.
[31] A. Bódalo, J.L. Gomez, E. Gómez, J. Bastida, J.L. Iborra, A. Manjon, [46] Y.-H. Lin, H.-Y. Chang, R.A. Adomaitis, MWRtools: a library for
Analysis of diffusion effects of immobilized enzymes on porous weighted residual method calculations, Comput. Chem. Eng. 23
supports with reversible Michaelis–Menten kinetics, Enzyme Microb. (1999) 1041–1061.
Technol. 8 (1986) 433–438. [47] H.W. Blanch, D.S. Clark, Biochemical Engineering, Marcel Dekker,
[32] A. Bódalo, E. Gómez, J.L. Gomez, J. Bastida, M.F. Maximo, F. Dia, Inc., New York, 1997.
A comparison of different methods of ␤-galactosidase immobiliza- [48] J.E. Bailey, D.E. Ollis, Biochemical Engineering Fundamentals, sec-
tion, Process Biochem. 26 (1991) 349–353. ond ed., McGraw-Hill, New York, 1986.

You might also like