Hydrogen Production From The Low-Temperature
Hydrogen Production From The Low-Temperature
Hydrogen Production From The Low-Temperature
949-956, 1995
Copyright @ International Associationfor Hydrogen Energy
Elsevier Science Ltd
Printed in Great Britain. All rights reserved
036&3199(94)0013&8
036&3 199/95 $9.50 + 0.00
Pergamon
HYDROGEN
PRODUCTION
FROM THE LOW-TEMPERATURE
WATER-GAS
SHIFT REACTION:
KINETICS AND SIMULATION
OF THE
INDUSTRIAL
REACTOR
N. E. AMADEO
Departamento
and M. A. LABORDE
de Ingenieria Quimica(FI)/PINMATE(FCEyN)
Universidad de Buenos Aires/CONICET,
Pabell6n de Industrias, Ciudad Universitaria 1428 Buenos Aires, Argentina
Argentina
studied. The experiments were carried out at 453-503 K and atmospheric pressure. A reactive mixture of similar
composition to that employed in the industrial process was used. An integral reactor, an integral procedure and a
data treatment valid for near equilibrium conditions were employed. A number of representative models were
examined. It was found that only a Langmuir-Hinshelwood model, which considers the adsorption of four species
(CO, CO,, H, and H,O) and the surface reaction as the controlling step, adequately describesthe reaction behaviour
at the temperature and concentration ranges investigated. Values of adsorption constants and adsorption heats for
the four components involved in the WGSR are given.
An algorithm for the simulation of an adiabatic fixed-bed reactor was developed with the aim of checking the
kinetics expression. Both the industrial and simulated compositions agree. It is proved that the kinetic expression
proposed which is in harmony with a Langmuir-Hinshelwood mechanism is useful in designing industrial lowtemperature converters.
NOMENCLATURE
CP
E,
E aP
F
FCOO
k
k
Kj
&
m
;
Pj
b rco)
W
X
Greek letters
P
P~~PH
MPCOPH,OKJ
Vahance
AH
AH,
Y
Subscripts
0
e
t
Inlet
Experimental
Theoretical
INTRODUCTION
950
N. E. AMADEO
and M. A. LABORDE
EXPERIMENTAL
Equipment
Experiments were performed in a conventional flow
system shown in Fig. 1. Distilled water was fed to an
evaporator using a high precision liquid pump. The
evaporator was filled with a bed of catalyst in order to
retain traces of chlorine present in water.
A mixture of the water vapour produced and gases
(CO, CO,, H, and NJ was passed through a preheating
zone prior to entering the reactor. The stainless steel
reactor consisted of a cylindrical cup of 10 mm diameter
and 23 mm length, perforated on the bottom with orifices
of 0.5 mm diameter, which enabled the flow of gases. The
catalyst was retained by a plug of glass wool. A thermocouple tube passed through the catalyst bed so that
temperature could be measured at different positions
along the bed. The bed temperature was controlled within
0.5C. The cup was placed in a stainless steel pipe of
12.5 mm diameter which was heated by a tubular furnace
of three heating sections.
After leaving the reactor, the remaining steam was
condensed out. Before passing through the sampling
chamber and through a wet gas meter the gas stream
was dried by silica gel.
Feed and effluent compositions were analysed by gas
chromatography.
A Hewlett-Packard
chromatograph
model 5730, equipped with thermal conductivity detectors, was employed to analyse CO, H, and N, using
a Porapak Q column, whereas CO, was separated by a
molecular sieve column.
Chemicals
CO 99.9% from LAir Liquid and CO,, H, and N,
99.9% from La Oxigena were filtered through silica gel
and used without any further purification. Distilled water
was used.
Characteristics of the commercial catalyst used were:
composition (%p/p), CuO : 32.7, ZnO : 47, Al,O, : 11;
specific surface, 42 mz g-l; poral volume, 0.11 cm3 g-i;
and copper metallic area, 13 mz g-.
Procedure
Catalyst was loaded in the reactor and reduced in situ
for 4 h at 453 K and 12 h at 503 K using a gas mixture
of 1% H, in N,. At this stage a strict temperature control
was established in order to avoid copper oxide sintering
and/or zinc oxide reduction. Both phenomena produce
copper crystal growth and, as a consequence, catalyst
activity decay [27].
Preliminary experiments with different particle sizes
and feed rates at the highest operating temperature
HYDROGEN
PRODUCTION
-r
1. flow meter
2. sample
N2
3. water reservoir
4. pump
5. vaporiser
6. reactor
7. water trap
8. silicagel
drier
9. gas meter
H2
AND DISCUSSION
conditions
(1)
952
0.6
0.5
1011(1)
0.5
0.3 -
0.05
III
0.10
&O
.I
0.15
0.05
0.10
0.15
0.20
x 10m5Pa
0.25
0.30
III
0.35
0.40
0.45
0.50
pi o x 10e5 Pa
2
0.5 -
0.4
(cl
0.3 -
0.2 -
80
0.2 -
000
0.1 -
0.25
0.30
0.35
0.40
0.45
III
0.50
0.60
0.05
p& x 10m5Pa
III
0.10
p&
0.15
x 10m5Pa
Fig. 2. (a) CO conversion vs inlet CO partial pressure (at). Experimental data: operative conditions: T = 503 K; P: atmospheric
pilo = 0.4 x lo5 Pa, p& = 0.36 x 10 Pa, pgo, = 0 0 Pa. (b) CO conversion vs inlet H,O partial pressure (at). Experimental data:
operative conditions: T = 503 K; P: atmospheric pgo = 0.12 x lo5 Pa, paH, = 0.44 x 10s Pa, p& = 0.0 Pa. (c) CO conversion vs
inlet H, partial pressure (at). Experimental data: operative conditions: T = 503 K; P: atmospheric pg, = 0.11 x lo5 Pa,
p& = 0.26 x 10s Pa, pg, = 0.0 Pa. (d) CO conversion vs inlet CO, partial pressure (at). Experimental data: operative conditions:
T = 503 K; P: atmospheric p& = 0.11 x 10 Pa, pg,, = 0.26 x 10s Pa, pi, = 0.26 x lo5 Pa.
model V
r = kpcopH20(l
FO = i (X, - X,)2.
-1
FO
(n-m)
(3)
(2)
+ PCO)
model II
model III
r = kp~op&l
- B)Ml + KiPi)
model IV
r = kp,o(l
- PM1 + K~~co,~~,pco,pH,:,
+ KH2pH,
Model
Objective function
Variance
I
II
0.58
1.80
0.10
0.49
0.27
0.0050
0.0145
0.0008
0.0040
0.0020
III
IV
V
+ &,o~H,o
+ KCO~PCOJ
HYDROGEN
953
PRODUCTION
(-rco) =
0.92e~454~3~7p,opH20(1
- /I)
+ 0.0047e2737~9~Tpco2
(1 + 2.2e101.5/TpC0+ 0.4e158.~TpHzo
454.3%JOPH,0(1 +0.92e~,yje1596.1!TpH,)2
1855.5
>
Table 3. Adsorption
co
co*
AH
-0.91 +
0.02
-- 24.72 k 0.4
KO
2.21 *
0.11
0.0047 f
0.0002
H,O
energy of
H*
-- 14.4 k
0.50
0.40 i
0.02
0.052 +
0.002
4.08 + 0.06
k,:
0.92 + 0.08
P)
(5)
- 1.42 f
0.05
Ea, :
kOPH,Ofl
02,
(4)
Pco,
P)
(6)
954
N. E. AMADEO
1,/r
(l,mol;g/
and M. A. LABORDE
s)
6200
P
5200
!"
/
4200
3200
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
conversion
--Et
model
III
Moe (2)
Fig. 3. Verification of the kinetic model used in this work. (Reaction rate)- vs conversion, adiabatic path, for Moe [2] and our
r = Yl.85
x lo-
exp 12.88 -T
1855.5
!
PCOPH,OU- /3).
(10)
The solution of the state differential equation (7) taking
into account equations (6) and (9) [or (lo)], has been
found by the fourth-order Runge-Kutta method.
The computed results, using the reaction rates given
by equations (9) and (lo), were compared with those
available from industrial data, as shown in Table 4. The
operating conditions used in the simulation are typical
of the industrial low-temperature converter employed in
AX
= ( - rcok
(7)
(8)
(- rco),
= w - rco)p
=1
(9)
H2
co
co2
CH,
Exit temperature
(K)
Industrial
Simulation
using Moes
[2] expression
Simulation
using our
expression
77.42
0.43
20.36
1.79
505
17.40
0.42
20.50
17.23
0.48
19.79
1.74
502
1.80
505
HYDROGEN
PRODUCTION
955
~ll/&/TF
F
0
0.2
Model III
0.4
r/
MOE (2)
0.6
0.8
1.0
4.50.
0
0.2
MOE (:2)
0.4
Model III
0.6
0.8
1.0
Dimensionless position
Fig. 4. (a, b) Conversion and temperature profiles predicted using MOoe[2] and our kinetics expressions. T = 453 K, P = 22 x 10
Pa, $o = 2.5 x 10 Pa, pi.o = 8.14 x lo5 Pa, pi, = 6 x 10 Pa. Pco, = 1.65 x IO5 Pa, M = 2.043 x IO g, F& = 0.216 mol s-l.
CONCLUSIONS
In the present paper attention was centred on reproducing the industrial procedure, i.e. kinetic data were
obtained using a commercial catalyst (CuO/ZnO/Al,O,)
and feed compositions similar to those of an industrial
reactor.
WGSR kinetic was determined in an integral reactor,
using an integral procedure and a data treatment valid
for near equilibrium conditions. A Langmuir-Hinshelwood kinetic expression which considers adsorption of
the four components and the surface reaction as the
controlling step was found to fit accurately the experimental data.
Values of adsorption constants and adsorption heats
for the four components involved in the WGSR are given.
In addition, the kinetic was compared successfully, in the
complete range of conversions and considering an
adiabatic path, with the empirical kinetics proposed by
Moe and largely used in the design of this kind of reactor.
Finally, our kinetic expression obtained at atmospheric
pressure and modified with the pressure factor suggested
by Rase [36] was used in the simulation of an industrial
reactor. The computed results agree very well with the
data produced from an industrial low-temperature converter employed in the ammonia process.
REFERENCES
1. D. Newsome, Cutc~l.Rec. Sci. Engng 21, 275 (1980).
2. J. M. Moe, Chem. Enyng Progr. 58, 33 (1962).
3. H. Bohlbro, Acta Chem. Scund. 16, 431 (1962).
4. G. Shchibrya, N. Morozov and M. Temkin. Kinrt. Cutul..
USSR 6, 1010 (1965).
5. E. Cherednik, N. Morozov and M. Temkin, Kinet. Caful.,
USSR 10, 94 (1969).
6. T. Semenova. B. Lyudkovskaya. M. Markina, Ya. Volynkina, G. Cherkasov, V. Sharkina, N. Khitrova and G. Shpiro,
Kinet. Katd., USSR 18, 1014 (1977).
7. E. G. M. Kuijpers, R. B. Tjepkema and W. J. J. van der
Wal, Appl. Cutal. 25, 139 (1986).
8. G. C. Chinchen, M. S. Spencer, K. C. Waught and D. A.
Whan. J. Chem. Sot.. Furaduv Trans. 1 83. 2193 (1987).
9. J. Nakamura, J. M. Campbelland C. T. Campbell: J. Chem.
Sot. Faraduy Trans. 86, 2725 (1990).
956
N. E. AMADEO
and M. A. LABORDE