Professional Documents
Culture Documents
K.R. Westerterp*
Received22 February1995;accepted16June1995
Abstract
The kinetics of the selectiveoxidation of ethenein air over an industrial silver on a-alumina catalyst were studied. Special
attention waspaid to the influenceof the reaction products on the reaction ratesof epoxidation and completecombustion.Kinetic
data were obtained in two different types of internal recycle reactor and in a cooled tubular reactor, and were fitted separately
to severalreaction rate expressionsbasedon different kinetic models.A Langmuir-Hinshelwood mechanism,in which adsorbed
ethene reacts with adsorbed molecular oxygen, was chosen as the best kinetic model. The reaction products compete for
adsorptionon the active sitesand reducethe ratesof both reactions.Carbon dioxide enhancesthe selectivity towardsetheneoxide,
whereaswater hasalmostno influenceon the selectivity. The fitting of the three individual data setsobtainedin the three reactors
resultsin accurate,but different, reaction rate expressions,whereasthe fitting of the three data setssimultaneouslyresultsin less
accurate reaction rate expressions.The systematicdeviations found may be explained, to someextent, by differencesin the
operating regimesin each reactor. The main reasonfor the deviations is probably the different catalyst activities in the three
reactors causedby poisoning.The effect of the addition of products to the feed on the behaviour of the cooled tubular reactor
can be describedreasonablywell by a mathematicalmodel in which the kinetic equationsobtained in the laboratory reactors are
incorporated. The resultsof thesesimulationsare sensitiveto the kinetics used.
Keywords:
1. Introduction
CH
2 4
2HO
2
Fig. 1.Reactionscheme
for theoxidationof ethene.AH, = - 105kJ
(mol E)-I; AH,= - 1323 kJ (mol E)-; AH3= - 1218kJ (mol
E)-.
108
the published data differ becausethe catalyst formulation and/or operating conditions are not comparable.
Therefore we have been forced to study the kinetics
ourselves.
Previously, we have reported a significant influence
of the reaction products, CO,, HZ0 and EO, on the
reaction rates [4]. In this study, kinetic data measured
in a tubular reactor and data from two recycle reactors
are used to determine the kinetics; special attention is
paid to the influence of the reaction products. The
influence of the reaction products is also studied by
their addition to the feed in a cooled tubular reactor;
the effect of the products on the behaviour of the
reactor is determined both experimentally and theoretically.
2. Literature
Silver is a good catalyst for ethene epoxidation because it forms weak bonds with ethene and oxygen. A
typical catalyst consistsof silver, promoted by oxidized
alkali and alkaline earth metals such as potassium,
barium and caesium, dispersed on a highly pure a-alumina carrier. In the literature, there is no agreement on
the mechanismsof the catalysed reactions. The discussion is focused on the forms of oxygen present on the
catalyst surface and which form is responsible for epoxidation and/or total combustion. Further matters which
remain to be solved are whether ethene reacts in the gas
phase or in an adsorbed activated state, the way in
which the reaction products are adsorbed on the catalyst surface and their influence on the reaction rates.
2.1. Mechanism of ethene epoxidation
E.P.S.
Schouten
et al. / Chemical
Engineering
2.2.2. Water
and Processing
35 (1996)
109
107-120
Table 1
Influence of reaction products on the epoxidation and total combustion reactions as reported in the literature. Symbol on left-hand side
refers to epoxidation and symbol on right-hand side to total combusion: + , rate increases; = , no effect; - , rate decreases; - - , rate
decreases so strongly that selectivity is also affected
Product
EO
Murray [36]
Schultze and Thiele [26]
Hayes [27]
Nault et al. [30]
Mikami et al. 1281
Metcalf and Harriot 1321
Force and Bell [19]
Temkin [5]
Bremer [13]
Griffe et al. [22]
Stoukides and Pavlou [2]
Al-Saleh et al. [31]
Hall and Neubauer [7]
Park and Gau 1161
Tan et al. [24]
Gleaves et al. [21]
Liu et al. [35]
Bulushev and Khasin [29]
Eliyas and Petrov [33]
Borman and Westerterp [4]
co2
W
-I-
=I=
-I=
=I=
-I-
- -/=/- -I=
-/--
-I--
-I-
-I--
-I-I-
=/-
=I-
-1-k
--I-
-I=
- -I-l-l-=/- -I- =I-I-I--
-1-f
-I--
110
k fK&P,P,
ri = 1 + K&P, + K&P, + K&PI+, + K;,P,,
ri =
k;KkfiP,&
(1 + KkP, + dm
+ KLP, + K&P,
+ Ki,P,,)2
UC>
ri = (1 + KLP, + KLP, + K&P, + K&P,
+ KQ,P,J2
(Id)
kf = ki,,exp
Kj = K),, exp
details
Fig. 2. Experimental
set-upfor the additionof watervapourto the
feed:A, supplyline; B, supplyvessel;C, electricalheater;D, saturation vessel;E, dewpoint measuring
device;F, flow controller.
111
Table 2
Investigated ,range of parameters in the kinetic experiments
Parameter
Berty-type reactor
Bobo reactor
T W
P (bar)
P, (mbar)
PO (bar)
P,o War)
P, (mbar)
P, (mbar)
455-528
1.8-10.0
3-177
0.37-2.08
5-58
l-79
l-63
455-512
1.8-9.0
4-151
0.37-I .86
0.2-22
0.4-21
0.4-21
434-496
2.0-10.0
4-101
0.42-2.09
0
O-25
o-47
_1
Ni=
IRexp,i
1
Rexp,i
Rxlc,il
(4)
- Rcalc,i
si = Rexp,i
Rexp,i
(5)
112
Table 3
Kinetic constants obtained by fitting: (a) Berty reactor data set; (b) Bobo reactor data set; (c) tubular reactor data set; (d) all kinetic data
(a) Berty data
Parameter
Model 1
k r,m
Tact
K E.-Z
Model 2
Model 3
ri
r2
I1
12
r1
r2
rl
12
35 x 10
5.9 x lo3
1.49 x 103
8.7 x 10
16 x lo3
5.9 x 10
0.55 x 106
8.7 x 103
61.7
1.9 x 10
1.2 x 103
1.0 x 103
25.8
35.5
5.2 x lo3
78
78
32.4
20.6
2.2 x 103
1.4 x 103
1.2,x lo3
29.6
24.0
5.4 x 10
77
78
33.8
0.97 x 106
9.5 x 10
0.0122
3.5 x 10
35.2
15
98
55
13.0
30.4 x 106
11.2 x 103
0.144
2.4 x lo3
5.9
77
64
8
10.8
0.50 x 106
9.2 x IO3
0.0130
3.5 x 10
5.8
10
101
55
12.8
7.34 x 106
10.7 x 103
0.222
2.5 x lo3
4.9
93
55
14
11.0
TUd%E
K Q-2
K EO,cc
K C.02
k r,m
Tact
K E,m
T ads,E
K 0.m
K EO,w
K CCC
Model 4
Model 4
Parameter
rl
12
rl
r2
0.49 x lo6
9.1 x lo3
0.026
3.2 x lo3
20.9
9
0.15 x 103
41
13.3
1.24 x lo9
1.27 x IO3
0.018
3.5 x 10
5.0
58
0.20 x 103
0.13 x IO
12.0
0.48 x lo6
9.1 x IO3
0.026
3.2 x lo3
4.6
9
0.13 x 103
40
11.8
1.32 x lo9
12.7 x lo3
0.018
3.5 x lo3
6.0
62
0.29 x 10
0.16 x lo3
9.4
k r,m
Tact
K Em
Tads.E
K
ZTCC
K C.3)
Model 3
Model 4
rl
r2
rl
r2
0.76 x lo6
8.8 x 103
0.010
3.0 x 103
2.2
10.5 x 109
12.8 x 103
0.0021
3.5 x 103
10.4
0.92 x 10
8.8 x 103
9.6 x lO-3
3.0 x 103
1.5
16.1 x lo9
12.6 x IO3
1.6 x lO-3
3.4 x 103
3.2
24
50
27.0
89
40
24.2
27
50
27.1
96
43
25.1
k r.33
Tact
K 6-2
T ads,E
K O,K E0.m
K C.02
Model 3
Model 4
11
12
0.695 x lo6
9.0 x 10
0.012
3.1 x 103
8.8
88
75
63
25.4
0.15 x lo9
11.4 x 10
0.14
2.3 x lo3
14.9
0.38 x lo3
0.13 x lo3
28
24.4
r2
36.5 x lo6
8.8 x 103
0.022
3.0 x 103
1.5
0.10 x 10
60
40
25.2
0.316 x lo9
12.8 x 10
0.0065
3.6 x lo3
3.0
0.25 x lo3
0.13 x 103
55
23.9
E.P.S.
Schouten
et al. I Chemical
Engineering
and Processing
35 (1996)
107-120
113
R /moYs.kg
expf
II0
To
oml 7
o.oooo1 O.ooool
0.0001
0.001
O.OOcml
0.0001
R /moYs.kg
=PJ
JTb
0.001
0.00001
0.0001
0.001
o.ooo1
~xp;moVs.kg
IId
I-LY*s*kg
0.00001
o.woo1
Ie
o.ooo1
0.001
R moJ/s.kg
expJ
Fig. 3. Comparison of the experimental and calculated reaction rates of the epoxidation reaction (I) and the total combustion reaction (II) for the
Berty data set fitted to model 2 (a) and model 3 (b), the Bobo data set fitted to model 4 (c), the tubular reactor data set fitted to model 3 (d) and
all kinetic data fitted to model 4 (e); 1, Berty reactor data; 2, Bobo reactor data; 3, tubular reactor data.
114
0.05
0.1
0.15
0.2
P, /bar
Ethene (%)
Oxygen (x)
Ethene oxide (%)
Carbon dioxide (%)
Water (%)
Empty (%)
Model 4
ri
r2
rl
r2
2-8
41-74
3-11
38-75
O-32
4-13
31-63
o-17
38-77
O-23
O-18
o-15
O-12
11-29
O-16
o-3
8-24
o-15
O-10
15-48
2-9
O-17
o-7
8-34
Parameter
Case A
Case B
Case C
T K)
P, (mbar)
PO (bar)
Pa, (mbar)
P, (mbar)
P, (mbar)
rl a
r2 a
465.5
35.8
1.05
501.0
30.3
1.03
3.3
4.6
4.6
0.201
0.334
0.275
0.456x
0.452x
1.145
0.277
476.2
23.2
rl
r2
b
b
i-1 c
r2 c
i-1 d
r2
0
0
0
0.104x
0.110 x
0.145 x
0.185 x
0.264 x
0.423 x
0.190 x
0.278 x
IO-3
10-a
10-a
10-a
1O-3
1O-3
10-S
lO-3
x10-3
x 10-s
x 1O-3
1O-3
10-s
x 10-s
x lO-3
0.561 x 1O-3
1.01
6.5
12.5
12.5
0.064x
0.089 x
0.080 x
0.070 x
0.131 x
0.227 x
0.067 x
0.086 x
lO-3
1O-3
lo-)
10-s
10-I
lo-)
lO-3
10-s
E.P.S.
Schouten
et al. / Chetnical
Engineering
1
1.5
n R--
0.5
1
1.5
P /bar
0
115
107-120
2.5
-0
0.5
2.5
II b
1.5
2.5
nmr
1P ,uan
0
11 a
-0
35 (1996)
0.5
Ia
Ib
and Processing
0.5
1 1.5
P /bar
0
2.5
Fig. 5. Relative distribution of errors of the epoxidation (I) and total combustion (II) reactions vs. the oxygen partial pressure for all kinetic data
fitted to model 3 (a) and to model 4 (b); 1, Berty reactor data; 2, Bobo reactor data; 3, tubular reactor data.
116
-1
420
,,,I
440
460
480
SW
0
520
-11
420
540
T/-K
440
34u
480
520
T/K
II a
Ia
02
-zd
460
0.5
,s:
0
-0.5
-1
-0
0.05
Ib
0.02
0.1
PElbar
0.15
0.04
0.06
0.2
0.05
0.02
II b
0.08
P /bar
0.15
0.2
0.04
0.06
0.08
P /bar
EO
IC
0.1
P, /bar
II c
EO
0.5
5; O
-0.5
-1 -.
-1
0
0.02
0.04
0.06
0.08 0.1
P /bar
C
Id
I02
II d
Y-----l
Y-----l
0.5
5;
0
-0.5
-1
0
Ie
0.02
0.04
0.06
P /bar
H
0.08 0.1
II e
0.02
0.04
0.06
0.08
0.1
P /bar
Ii
Fig. 6. Relative distribution of errors of the epoxidation (I) and total combustion (II) reactions vs. the temperature (a) and partial pressures of
ethene (b), EO (c), CO, (d) and H,O (e) based on all data and model 4.
EO (%)
co* (%)
H,O (%)
r,: case A
r2: case A
r,: case B
r2: case B
Y,: case C
rz: case C
0
0
11.6
25.1
26.4
35.3
0
0
14.8
19.0
29.5
35.3
0
0
10.1
8.3
20.2
15.7
117
520
Y
c
4.4.2. JVater
In Fig. 9(a), the influence of water on the selectivity
and conversion is shown. Water reduces both the selectivity and conversion only slightly. The trends in model
calculations and experiments are the same: the conversion decreases at high water concentrations and the
selectivity remains more or less the same. In Fig. 9(b),
the axial temperature profiles are shown. The results
are similar to those found for the addition of carbon
TabIe 7
Influence of the addition of carbon dioxide and water to the feed on
the performance of the cooled tubular reactor. Conditions:
T,,,=465.8 K, &,=0.82 kg m--2s-, P=O.50 MPa and ~,,~,=0.91
mol.%
Experiment
Addition
1
2
3
4
0.2% H,O
0.1% co,
0.1% CO, i- 0.2% H,O
0.444
0.381
0.338
0.283
0.575
0.575
0.606
0.545
exp.
exp.
exp.
. . . . TW
460
exp.
500
480
100
axial
200
300
400
position/mm
5. Conclusions
A Langmuir-Hinshelwood
mechanism was chosen
for both the epoxidation and total combustion reactions, because it was found necessary to include an
adsorption term for ethene in the reaction rate expressions to obtain an order in ethene smaller than unity at
high ethene partial pressures. It is difficult to discrimi-
118
*O /I
0.80
0.60
o.40
0.60
talc.
+
0.40
-.-.
0.00
(a)
0.25
water
500
Y
;2
0.50
conc./mol
talc.
460
lb)
0.25
COP
(a)
0.50
0.75
conc./mol
480
+Q
470
P
460
(iI)
m 0.39
0.50
i
I
axial
100200300400
position/mm
0.75
1.00
*00/o
+ 0.09
0 0.23
. 0.38
v 0.56
%
%
%
%
96
0.94
axial
100200300400
position/mm
1.00
A
Y
c
480
0.00
,ss;;Eflvity
470
selectivity
0.00!---A
0.20
0.00
cab
0.20
1.00
conversion
__
_
- zc
490
A
conversion
E.P.S.
Schouten
et al. / Chemical
Engineering
and Processing
-25%
107-120
119
T
temperature (K)
T,,, ( = E/R), activation temperature (K)
Tads ( = AH/R), adsorption temperature (K)
ew
0%
3.5 (1996)
mole fraction
axial position (m)
Greek symbols
(J mol- )
axial
position/mm
Acknowledgements
We gratefully acknowledge the technical assistance
of K. v. Bree, F. t. Borg, A.H. Pleiter and O.D.
Veehof and the experimental work of J.H. Hendriks
and W.F. Hendriksen.
Appendix: Nomenclature
4
4
k:,
Table 8
Influence of a change in frequency on tubular reactor performance
(see Fig. 10 for conditions)
Experimental
Calculated
Frequency factor x 0.75
Frequency factor x 1.25
act
ads
C
talc
E
EO
exp
0.392
0.488
0.397
0.566
0.588
0.536
0.548
0.522
in
j
0
W
W
1
2
3
activation
adsorption
carbon dioxide
calculated
ethene
ethene oxide
experimental
of reaction i
inlet
of component j
oxygen
wall
water
epoxidation reaction
total combustion reaction
ethene oxide combustion reaction
References
[l] E.P.S. Schouten, P.C. Borman and K.R. Westerterp, Determination of the kinetics of the ethene epoxidation, Cizem. Eng. Proc.,
submitted for publication.
[2] M. Stoukides and S. Pavlou, Ethylene oxidation on silver catalysts: effect of ethylene oxide and of external transfer limitations,
Gem. Eng. Commun., 44 (1986) 53-74.
[3] R.A. v. Santen and H.P.C.E. Kuipers, The mechanism of the
ethylene epoxidation, Adu. Catal., 35 (1987) 265-321.
[4] P.C. Barman and K.R. Westerterp, An experimental study of the
kinetics of the selective oxidation of ethene over a silver on
e-alumina catalyst, Irzd. Eng. Gem. Ref., 34 (1995) 49-58.
[5] M.I. Temkin, The kinetics of some industrial heterogeneous
catalytic reactions, Adu. Catal., 28 (1979) 173-290.
[6] C. Backx, J. Moolhuysen, P. Geenen and R.A. v. Santen,
Reactivity of oxygen adsorbed on silver powder in the epoxidation of ethylene, J. Catal., 72 (1981) 364-368.
[7] R. Hall and G. Neubauer, Kinetic studies on the mechanism of
the catalytic ethene oxidation, J. Catal., 105 (1987) 39-54.
[8] H. Worbs, Ph.D. Thesis, Technische Hochschule, Breslau, 1942.
[9] R.E. Kenson and M. Lapkin, Kinetics and mechanism of
ethylene oxidation. Reaction of ethylene and ethylene oxide on a
silver catalyst, J. Phys. Gem., 74 (1970) 1493-1502.
[lo] P.A. Kilty and W.M.H. Sachtler, The mechanism of the selective
oxidation of ethylene to ethylene oxide, Cat. Rev. Sci. Eng., IO
(1974) 1-16.
[l l] N.W. Cant and W.K. Hall, Catalytic oxidation VI. Oxidation of
labelled olefins over silver, J. Catal., 52 (1978) 82-94.
[12] S. Kagawa, M. Iwamoto, H. Mori and T. Seiyama, Formation
120
[26] G.R. v. Schultze and H. Theile, Uber die Oxydation von Athylen
in der Gasphase an Silberkontakten, Erc(iil mi Kohle, 5 (1952)
552-560.
[27] K.E. Hayes, The role of reaction products in the silver catalyzed
oxidation of ethylene, Cm. J. Chern, 38 (1960) 2256-2268.
[28] J. Mikami, S. Satoh and H. Kobayashi, Studies on the catalytic
oxidation of ethylene by means of the pulse technique, J. Catal.,
18 (1970) 265-270.
1291 D.A. Bulushev and A.V. Khasin, Effect of carbon dioxide on the
catalytic oxidation of ethylene over silver. Reaction rates and
composition of adsorbed layer, React. Kinet. Catal. Lett., 44
(1991) 421-425.
[30] L.G. Nault, D.W. Bolme and L.N. Johanson, Reaction rate
studies of catalytic oxidation of ethylene, hf. Wg, Cknr. Proc.
Des. Deo., 1 (1962) 285-292.
1311 M.A. Al-Saleh, M.S. Al-Ahmadi and M.A. Shalabi, Kinetic
study of ethylene oxidation in a Berty reactor, Chew. Big. J., 73
(1988) 35-41.
1321 P.L. Metcalf and P. Harriot, Kinetics of silver-catalysed ethylene
oxidation, Id Eng. Chew Proc. Des. Dev., 11 (1972) 478-484.
[33] A. Eliyas and L. Petrov, Modelling of the inhibiting effect of
carbon dioxide on the selective oxidation of ethene over silver
catalyst, Appl. Catal., 62 (1990) 1l-21.
[34] M. Bowker, M.A. Barteau and R.J. Madix, Oxygen induced
adsorption and reaction of PI,, H,O, CO and CO2 on single
crystal Ag(llO), Surf. Sci., 92 (1980) 528-548.
1351 W. Liu, A. Eliyas and L. Petrov, Influence of feed water vapour
on the selective oxidation of ethylene over silver catalyst, Appl.
Catal., 61 (1990) 265-274.
[36] K.E. Murray, A study of the oxidation of ethylene on a silver
catalyst, Austr. J. Sci., Ser. A, 3 (1950) 433-449.
[37] P.D. Klugherz and P. Harriot, Kinetics of ethylene oxidation on
a supported silver catalyst, AIChE J., 17 (1971) 856-866.
[38] J.M. Berty, Reactors for vapour-phase catalytic studies, Chern.
Eng. Proc., 70 (1974) 78-84.
[39] P.C. Borman, A.N.R. Bos and K.R. Westerterp, A novel reactor
for determination of kinetics for solid catalysed gas reactions,
AIChE J., 40 (1994) 862-869.
1401 E.P.S. Schouten, P.C. Borman and K.R. Westerterp, Oxidation
of ethene in a wall cooled tubular reactor, Chem. Eng. Sci., in
press.
[41] E.C. Steiner, T.D. Rey and P.S. McCroskey, SIMUSOLVReference Guide, Dow Chemical Company, 1990.
1421 R.K. Sharma, D.L. Cresswell and E.J. Newson, Kinetics and
fixed-bed reactor modelling of butane oxidatidn to maleic anhydride, AIChE J., 37 (1991) 39-47.