Ckmird Engiwering Science. Vol. 43. No. 5. pp. 1195-l 198, 1988
00
Printed in Great Britain.
Effects of intraparticle diffusion
(First received 27 June
In the literature there are many papers in which catalyst
deactivation and intraparticle resistance effects are both
investigated. However, the interaction of these effects has not
yet been discussed in details. In association with process
design optimization, more attention should be paid to
changes in deactivation kinetics caused by intraparticle
diffusion effects. As a rule, precautions are taken against
diffusion limitations which are considered as leading to a
disadvantageous retardation of the processes. Being reason-
able for reactions over slowly decaying catalysts, this pre-
sumption is not always correct in cases when the catalyst
undergoes fast deactivation. The intraparticle diffusion ef-
fects actually cause a decrease in rate of the main reaction but
may, at the same time, retard the catalyst deactivation. The
superposition of these opposite retardation effects may result
in their mutual compensation. Moreover, the effective rate of
the process affected by intraparticle resistance may come to a
point when it even exceeds the rate of reaction on the decaying
catalyst in the kinetic region. The problem of predicting the
influence of intraparticle diffusion effects on the processes
accompanied by catalyst deactivation becomes still more
comlicated because of the considerable variety of deactivation
phenomena. Best et al. (1971) and Rest and Wojciechowski
(1976) investigated the influence of diffusion on the kinetics of
gas oil cracking over decaying zeolite catalysts on the basis of
the catalyst aging theory (Wojciechowski. 1968) where the
catalyst activity is postulated to be a function of the “catalyst
time-on-stream” only. On simulating different types of the
deactivation law and considering the fraction of unblocked
active sites independent of the conversion degree, these
authors have come to the conclusion that in certain cases
increasing diffusion limitations results in less rapid aging of
the catalyst. There is good agreement between this result and
the concept that increased intraparticle diffusion resistance
would hinder the penetration of poisons in the interior of the
catalyst particles and corresponds to the case when deacti-
vation is caused by an agent present in the initial mixture. In a
previous paper (Kumbilieva er al., 1987) we have performed a
theoretical analysis of how the combination of intraparticle
resistance and deactivation would influence the kinetics of
processes where the decrease in catalyst activity is an explicit
function of the degree of conversion. This type of the
deactivation function covers a wide range of experimentally
observed catalyst behaviours and is especially appropriate for
phenomena of consecutive deactivation. A typical example in
this respect is the deactivation due to consecutive coke
formation, i.e. the case when the concentration of coke
precursors depends on the degree of conversion. The analysis
of different approximations of the deactivation law expressed
in terms of the conversion degree led us to the conclusion that
in some cases the rate of the catalytic reaction under diffusion
limitations, upon attaining a definite “threshold” degree of
conversion, exceeds the rate in the kinetic region. This result is
not unexpected in view of the fact that the higher rate of the
main reaction in the kinetic region leads to a higher concen-
CES43;5-N
000%2509/88 S3.00+ 0.
0 1988 Rrgsmon Press pk
resistance on catalytic reactions over decaying
catalysts
1986; accepted in revised form 30 June 1987)
tration of coke precursors, which, on their part, accelerate the
deactivation compared to that in the diffusion region. On the
other hand, the lower rate of catalyst deactivation in the
diffusion region could prolong the operation time of the
catalyst, These theoretical results show the necessity of taking
into account the possibility of enhancing the efficiency of
some deactivation-accompanied processes by provoking
intraparticle diffusion resistance control on the process.
Anyhow, reliable conclusions on the influence of diffusion
resistance in each specific catalytic process should be based on
models including detailed kinetic description of both main
and the deactivation reaction.
In the present work we shall apply this approach for the
investigation of the influence of intraparticle diffusion on the
dehydrogenation of isoamylenes to isoprene over a commer-
cial catalyst. This reaction is accompanied by catalyst deacti-
vation as a result of coke formation mainly due to the
isoprene. Here we shall use for thedeactivation kinetics not an
approximation of the deactivation law but an experimentally
found equation for the rate of coke formation. The industrial
process is realized in cycles, Each cycle consists of a work
period r, and a rest period r,,, the latter being needed for
catalyst regeneration and/or feed switching. It is obvious that
the duration of r should be determined by the permissible
range of activity decrease caused by coke formation.
The process kinetics is studied in details in the kinetic
region at atmospheric pressure over a temperature range of
85G960 K (Butkovsky et al., 1977ac). According to the
kinetic equations given in the above papers. the rate of
isoprene formation r and the rate of coke formation p can be
expressed in terms of the conversion and the amount of
deposited coke as follows:
k,P,(l ~ x,)
(1)
IR= (l+C+k,P,x,)
k, + (La - kdx,
PR = pO
(1 + C + k,P,x,)
+(k,+(k,-kk,)x,).C.exp(-CC)
(2)
(1 + C + kzP”x.) 1
where PO is the initial partial pressure of isoamylenes, x, is the
degree of conversion on the catalyst pellets, C is the amount of
coke deposited per gram catalyst, L is the length of the
reactor, and k,, k,, k,, k,, k, and k, are constants whose values
are described by Butkovsky et al.
To elucidate the effect of intraparticle diffusion resistance,
we analysed the changes of isoprene and coke formation with
increasing size of the catalyst pellets. The analysis was carried
out on the basis of the quasi-steady-state assumptions
(Kiperman, 1964) and the quasihomogeneous model of the
process in the pellet pores. To derive the expressions for the
rates of the reactions affected by intraparticle diffusion
limitations we used the Pshezhetsky-Rubinstein formula
(Pshezhetsky and Rubinstein, 1946). which correlates the
1195
1196 Shorter Communications

reaction rate under diffusion limitations with the intrinsic
reaction rate:
,=;JF
where R is the ratio of the volume of the catalyst pellet to its
external surface, D is a diffusion coefficient, and cj is the
concentration of the jth reagent. The subscripts “s” and “eq”
denote variables for the surface and center of the catalyst
pellet, respectively.
In accordance with formula (3), the functions defining the
rates of the main and the deactivation reactions in the
intraparticle resistance region take the form:
- -
r = J~J’IPo
d
R :s-=a
(S, = 20 min) and in 60min in the intraparticle diffusion
region (T., = 60 min). Therefore, the transition of the process
to the region of intraparticle ditrusion resistance will lead to
an increase in the useful working time of the catalyst. The
(3)
plots in Fig. 3 show the increase with time of the yield of
isoprene in the kinetic and in the diffusion resistance regions
for ~~ = 20 min, TV= 60 min, and r0 = 10 min. Comparing
the curves in Figs 2 and 3 shows the advantages of work in the
diffusion-controlled region: after a certain time interval which
is short compared to the operation time of the reactor, the
yield of isoprene obtained in the diffusion resistance region
exceeds that under kinetic control.

x.
UX, C) dx
CONCLUSIONS
The results obtained indicate that the proposed approach
can be profitable for the design of processes over decaying

J,
-=k
,/ 2Dd’o catalysts. In cases when intraparticle diffusion resistance
Pd = Pr(x, C)dx (5)
R -% proves to be advantageous, this should be taken into account
where D, and D, are the effective dift’usion coefficients of the in process design with a view of provoking, if possible,
diffusion resistance control on the process, using sufficiently
isoamylenes and of the coke precursors, respectively.
In order to examine the effects of intraparticle diffusion large pellets, suitable pore sizes etc.
resistance and to compare the efficiency of work in the kinetic Since catalytic processes are usually accompanied by
catalyst deactivation, the proposed approach, which is ap-
and the diffusion resistance regions, the proceeding of the
reaction in an isothermal plug flow reactor was simulated by plicable for a large class of processes, would be useful for the
study of analogous reactions.
means of a numerical solution of eqs (1) and (2) for the kinetic
region and of eqs (4) and (5) for the region of diffusion
control. The values of the parameters of the process were kept K. KUMBILIEVA+
close to those corresponding to the industrial conditions: M. M. KOSTJUKOVSKY:
temperature T = 953 K, PO = 0.04 atm, space velocity Y
= 250 h-‘, reactor length L = 1.5 m. s. L. ~l’,“z,“,“,“,‘~
The criterion proposed by Hudgins (1968) for the arbitrary tlnstitute of Kinetics and Catalysis
form of the kinetic equation was applied to evaluate the
Bulgarian Academy of Sciences
ul. Academic Bonchev bloc 11
diffusion influence. In agreement with this criterion, the
conditions of intraparticle resistance control of the process Sofia 1040, Bulgaria
can be expressed by the relationship: fzelinsky Institute of Organic Chemistry
U.S.S.R. Academy of Sciences
R J-&Q Leninsky prospect 47
(6) Moscow, U.S.S.R.
-rd
D, -=l (drrjdx)’
It was the size of the catalyst pellet that was considered to be
the factor determining the region of control, i.e. the transition
of the process from the kinetic region to the region of NOTATION
intraparticle resistance was realized by increasing the pellet C concentration of coke, mg per gram catalyst
size only, the other operating conditions remaining D,D,,Dz effective diffusion coefficients, sm2/s
unchanged. k,, k,, k,,
The results obtained show how the superposition of two L k,, kc+ constants
opposite effects influences the process. At first the isoprene L length of the reactor
formation rate in the kinetic region considerably exceeds the r reaction rate
rate of the reaction under intraparticle diffusion control. t time
However, the decrease of the reaction rate with time is T temperature
smoother in the region of intraparticle diffusion resistance x degree of conversion
due to the diffusion-induced retardation of deactivation. As a
result, at a given moment the process rate under diffusion Greek letters
control becomes higher than the rate in the kinetic region. P rate of coke formation
Some of the results calculated for values of the parameters: P work period of an operation cycle
DI = 0.004 sm*/s, Dz = 0.0001 sm’/s and sizes of the pellets 50 rest period of an operation cycle
0.05 sm in the kinetic region against 1.5 sm in the region of
intraparticle diffusion resistance, are plotted in Figs l-3. Subscripts
The variation of exit conversion with time under both d diffusion region
regimes of work is shown in Fig. 1. Obviously, after a certain k kinetic region
period, the conversion achieved under intraparticle resistance eq equilibrium value
conditions exceeds that in the kinetic region. s surface of the catalyst pellet
Figure 2 illustrates the increase of isoprene yield with time
for equal work periods of the catalyst in the two regions, i.e. T&
= TV= 20 min. r0 = 10 min. The calculated amount of coke
deposited by the end of the work cycle was about 25 mg per REFERENCES
gram catalyst in the kinetic region, and about 20 mg per gram Best, D., Pachovsky, R., and Wojciechowski, B. W., 1971,
in the diffusion region. These results show a possibility of Diffusion effects of gas oil cracking. Can. J. them. Engng 49,
prolonging the useful working time of the catalyst. As was 809-S 12.
mentioned above, the work period should be determined as
the time of drop of the conversion below a certain value. Thus,
if the value of 0.2 is assumed to be the lower boundary of
conversion, i$ will be reached in 20 min in the kinetic region §To whom correspondence should be addressed.
 Shorter Communications 1197

L--
\
%

.5u

3 90,. f
L
‘\
.sa- ‘Z
9 ---_
2 ---_ ---___ ---_ --a_
-----__
20..
E

Fig. 1. Variation of the exit conversion with time under kinetic (-) and diffusion resistance (- - - --)
control of work.

Fig. 2. Increase of isoprene yield with time in the kinetic (-) and in the diffusion resistance (- - - - - -)
regions for *L = ~~ = 20 min ~~ = 10 min.

Best, D. and Wojcicchowki, B. W., 1976, An experimental isoprene considering the kinetics of coke formation. Kinet.
technique for determining the significance of intraparticle Katal. 18, 1463-1473.
diffusion on an aging catalyst. Can. 1. them. Engng 54, Butkovsky, M. E., Gaidai, N. A., Kol’tsov, N. I., Kiperman,
197-202. S. L., Kotel’nikov, G. R., Shishkin, A. N. and Strunnikova,
Butbovsky, M. E., Gaidai, N. A., Kiperman. S. L. and L. Y., 1977~. Kinetics of catalytic dehydrogenation of
Kotel’nikov, G. R., 1977a, Kinetics of catalytic dehydro- isoamylenes under stationary conditions. Neftekhimiya 17,
genation of isoamylenes to isoprene under unstationary 203-211.
conditions. Nefbkhimiya 17, 371-375. Hudgins, R. R., 1968, A general criterion for absence of
Gaidai, N. A., Butovsky, M. E., Kol’tsov, N. I., Kiperman, S. diffusion control in an isothermal catalyst pellet. Chem.
L. and K4tel’nikov, G. R., 1977b, Analysis of the kinetic Engng Sci. 23, 93-94.
model of the process of dehydrogenation of isoamylenes to Kiperman, S. L., 1964, InZroduction to Kinetics of
 Shorter Communications

Fig. 3. increase of isoprene yield with time in the kinetic (-----) and in the diffusion resistance (- - - - -)
regions for: rk = 20 min, 5, = 60 min. r0 = 10 min.

Heterogeneous Catalytic Reactions. Nauka, Moscow. Pshezhetskii, S. and Kubinstein, R., 1946, Kinetics of hetero-
Kumbilieva, K., Kostyukovsky, M. M., Petrov, L. and geneous reaction on porous catalysts. Zh.#z. Khim. 20,
Kipcrman, S. L., 1987, Joint effects ofcatalyst deactivation 207-2 10.
and intraparticle diffusion resistance. In Catalyst Wojcicchowski, B. W., 1968, A theoretical treatment of
Denctioation 1987, pp. 381-389. Elsevier, Amsterdam. catalyst decay. Can. J. them. Engng 46, 48-53.

Chpnricol Engineerrng Science. Vol. 43. No. 5, pp. 1198-I 199. 1988. CKW-2509/88 53.00 + 0.00
Printed in Great Britain Pergamon Press plc

A modified one-point collocation method for diffusion-reaction problems

(First received 15 October 1986; accepted in revised form 25 September 1987)

A one-point collocation method [Villadsen and Michelsen a=~,=1 at x=x,=1

(1978)] is often useful in revealing the main features of the
du
solution to a boundary value problem as well as in developing -=O at x=0.
and
a graphical solution of the problem. It is the purpose of this dx
note to introduce a modification to the one-point collocation
method which may yield somewhat higher accuracy for the
Make the following substitutions
profile than the standard one-point collocation.
y = x2
PROBLEM STATEMENT
P = rb2
As a simple illustration we consider a reaction occurring in
a catalyst slab: to obtain

(1) 4y$+2% = pR(u). (2)