i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 2 ( 2 0 1 7 ) 3 0 2 7 e3 0 3 5

Available online at www.sciencedirect.com

ScienceDirect

journal homepage: www.elsevier.com/locate/he

Description of reaction rate for steam methane

reforming throughout a channel in combination
with diffusion in a porous catalytic wall: Extension
to concentration variation along flow direction

Motohiro Saito*, Syogo Sakurai, Hiroshi Iwai, Hideo Yoshida

Department of Aeronautics and Astronautics, Kyoto University, Nishikyo-ku, Kyoto 615-8540, Japan

article info abstract

Article history: In a reactor with a porous catalyst on a channel wall, the local reaction rate and compo-
Received 31 July 2016 sition cannot be easily measured. This study built a model for evaluating the overall re-
Received in revised form action rate by predicting the distribution of the concentration of chemical species in a
21 September 2016 porous catalyst as well as in a gas stream. Prior to taking into account the streamwise
Accepted 22 September 2016 distribution, the validity of the function of the concentration profile in steam methane
Available online 14 October 2016 reforming, which is one-dimensional to the thickness direction of catalyst, was confirmed
by comparison with the results of an experiment. Next, the model was coupled with the
Keywords: streamwise variation in the concentration. Consequently, we derived a model in which the
Steam methane reforming local concentration can be specified at any position with the concentration at the inlet of
Catalytic wall reactor the channel as the only input variable. Finally, a criterion for the condition in which the
Porous catalyst maximum yield of products can be obtained with the minimum volume was proposed with
Rate-determining step nondimensional groups.
Reaction rate © 2016 Hydrogen Energy Publications LLC. Published by Elsevier Ltd. All rights reserved.
Diffusion equation

supports have been examined [6] [7]. Although many expres-

Introduction sions for the reaction rate, which is one of the most important
factors in designing a reactor, have been proposed [8e10],
Hydrogen has been utilized for many purposes such as a fuel universal expression have not been proposed because each
in internal combustion engines [1], gas turbines [2] and fuel material has its own kinetics. Therefore, prior to the applica-
cells [3], and as a material to form other products, e.g. the gas tion of a newly developed catalyst in a reactor, the reaction
to liquid (GTL) process [4]. Steam methane reforming (SMR), rate must be obtained through experiments.
expressed by the following equation, is the most popular way In addition to chemical aspects, the arrangement of the
to produce hydrogen [5]. catalyst has a strong influence on the performance of a
CH4 þ H2 O/CO þ 3H2 (1) reactor. Although catalyst pellets are packed into a reactor in
most large-scale plants, a packed catalyst has the disadvan-
A catalyst is generally used to enhance the reaction of SMR. tages of a high pressure drop and the risk of plugging or
To improve the performance of the catalyst in terms of ac- channeling, particularly in small systems [11]. Therefore, in
tivity and durability, various catalyst materials and their some reactors, the catalyst is coated on a channel wall

* Corresponding author. Fax: þ81 75 383 3652.

E-mail address: msaito@mech.kyoto-u.ac.jp (M. Saito).
http://dx.doi.org/10.1016/j.ijhydene.2016.09.162
0360-3199/© 2016 Hydrogen Energy Publications LLC. Published by Elsevier Ltd. All rights reserved.
3028 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 2 ( 2 0 1 7 ) 3 0 2 7 e3 0 3 5

[12e14]. In such systems, the transport phenomena at the where Dj eff , Cj and nj are the effective diffusion coefficient,
interface between the gas stream and the catalytic wall are molar concentration and stoichiometric coefficient, respec-
important. Furthermore, we must take into account mass tively. The flux at the interface between the gas phase and
transfer inside the catalyst layer, that is, internal diffusion, the catalyst layer, that is, Dj eff dCCH4 =dy at y ¼ 0, corresponds to
because catalysts generally have a porous structure. the converted methane rate, which is the change in con-
In our previous paper [15], we proposed a method for centration in the bulk gas stream. To fit the experimental
determining the reaction rate of SMR in a porous catalyst on a data, we assumed that rst could be expressed by a power-law-
channel wall. Generally, it is difficult to acquire local infor- type reaction rate with an upper limit in our previous paper
mation not only within a porous catalyst but also in a gas [15],
stream. Thus, in the proposed method, changes in the bulk (  a  b
concentrations of gas species were measured experimentally
rst ¼ min kst pCH4 pH2 O : (3)
and introduced into the one-dimensional diffusion equation rc ð¼ const:Þ
in the thickness direction in the catalyst layer. As a result, a
power-law-type reaction rate with an upper limit was found to This expression for the reaction rate indicates that when
be appropriate for expressing the local reaction rate. More- the partial pressures of methane and steam, pCH4 and pH2O
over, the dimensionless numbers derived from the reaction respectively, are sufficiently large, the reaction rate settles to a
rate expression and diffusion equation provide a quantitative constant value of rc. In many papers that represented the re-
criterion for the rate-limiting condition. action rate of SMR with a power-law-type equation, the
The model developed in our previous paper [15] was one- exponential index a has a value close to unity, whereas the
dimensional in the thickness direction of the catalyst layer, absolute value of b is smaller than that of a at the wide range
because catalysts had a short length. To design an actual of the concentration ratio of steam to methane, i.e., the
catalyst with a certain length, numerical simulation using the steam-to-carbon ratio [15,17e19]. Moreover, the change in
reaction rate estimated from an experiment with a short pH2O is generally not significant because the steam-to-carbon
catalyst is a possible method. For a wider application range, ratio is relatively large in actual conditions of reactors.
the aim of this study is to develop a model that can specify the Therefore, we assume that the reaction rate is proportional to
local phenomena at any position in a reactor even if the gas the concentration of methane and independent of the con-
composition varies along the flow direction. To this end, we centration of steam as follows:
attempt to describe the concentrations of species in a gas
stream and inside a porous catalyst as functions of position. 
kC
rst ¼ min ; (4)
In this paper, we first derive a function for the concentra- rc ð¼ const:Þ
tion in the thickness direction of the catalyst on the basis of
the experimental data in our previous paper [15]. Next, an where, for simple notation, C and k are the concentration of
experiment is performed with catalysts of different lengths methane and the reaction rate constant, respectively. Fig. 1(a)
and thicknesses. Finally, we derive equations for the con- shows a sketch of the reaction rate as a function of the con-
centration distribution in the case of variation along the flow centration. If the concentration is larger than the threshold
direction. In this model, the overall reaction rate throughout concentration Cc, which is calculated as rc/k, the reaction rate
the reactor can be evaluated from the gas composition at the becomes constant.
inlet. By integrating Eq. (2) and using Eq. (4), the local concen-
tration of methane can be expressed as a function of the po-
sition y. As shown in Fig. 1(b), we consider three different
Concentration profile in one dimension cases with different values of the concentration at the inter-
face between the gas flow and the catalyst layer. Hereafter,
Before developing a two-dimensional model, we derive the such a concentration is denoted as Csur.
one-dimensional concentration distribution perpendicular to When Csur is smaller than Cc, the local reaction rate de-
the flow direction and verify its validity by comparison with pends on the concentration at position y, as in mode [i] in
experimental results. Fig. 1(b). In mode [ii], a region in which the reaction rate is
constant appears in the upper layer. Upon further increasing
Basic equations Csur, the reaction rate of the entire field becomes constant. The
diffusion equation at each mode is summarized in Table 1.
As described in our previous paper [15], the overall reaction Here, the diffusion coefficient is assumed to be constant
can be considered on the basis of the diffusion and reaction because it is almost constant along the depth direction as
inside the porous catalyst. Although there are some diffusion described in our previous paper [15].
models, as investigated by Solsvik et al. [16], a simple equation The equations in Table 1 are integrated with boundary
is chosen to express the concentration profile in this study. conditions. For every case, C ¼ Csur at y ¼ 0 and dC/
If the diffusion is one-dimensional in the y-direction, the dy ¼ 0 at y ¼ H. In mode [ii], at the intermediate point of y ¼ yc,
conservation equation for species j can be expressed in terms an additional condition that connects the two different reac-
of the local reaction rate rst as follows: tion rates is necessary. The obtained solutions are also listed
in Table 1, accompanied with the flux from the catalyst to the


d dCj gas stream, which corresponds to the converted methane
Dj eff ¼ nj rst ; (2)
dy dy rate.
 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 2 ( 2 0 1 7 ) 3 0 2 7 e3 0 3 5 3029

fixed at 180 ml/min with a constant steam-to-carbon ratio of 5

and the catalyst temperature was maintained at 520  C. The
result that a thicker catalyst converted more methane revealed
the effect of mass diffusion within the catalyst. For each catalyst
thickness, with increasing amount of supplied methane, the
converted methane rate increased and then became saturated.
When the values of rc, k, Csur and D in Table 1 are given, DdC/
dy at y ¼ 0, that is, the converted methane rate, can be specified.
These variables are evaluated on the basis of experimental re-
sults. To determine representative values of rc and k, a catalyst
thickness of 34.6 mm is used in this calculation. Under the
saturated condition with a sufficient supply of reactant, the
converted methane rate is divided by the apparent catalyst
volume to obtain rc. The concentration at the interface between
the gas flow and the catalyst layer, Csur, is assumed to be the
same as the bulk concentration of the reformed gas measured
in the experiment. This assumption is based on the calculation
result that the concentration profile in the gas phase was
almost uniform over the cross section of the flow because of the
small channel height [20]. The diffusion coefficient D is calcu-
lated by method in the appendix of paper 15.
As shown in Fig. 2, the converted methane rates calculated
using the equations in Table 1 were in agreement with the
results of the experiment. Therefore, we conclude that the
one-dimensional functions of concentration coupled with the
power-law-type reaction rate with an upper limit are valid,
not only qualitatively but also quantitatively. In addition,
because the calculation result of the catalyst thickness of
55.6 mm, where rc and k obtained from different thickness were
Fig. 1 e Conceptual diagram of local reaction rate inside a used, was appropriate, an experiment on one catalyst is suf-
porous catalyst. ficient to estimate the variables in Table 1.

Comparison with experimental results

Experiment
As described in the previous section, the concentration of
We next extend the one-dimensional model developed in the
methane is expressed as a function of position y. Here, we
previous chapter so that it can be applied in the case where
verify the validity of the equations in Table 1 by comparison
the concentration varies along the flow direction. To obtain
with the experimental results for SMR from our previous
converted methane rates for comparison with the calculation
paper [15]. At the same time, we confirm that the values of rc
and to evaluate the variables used in the calculation, an
and k are consistent regardless of the thickness of the catalyst.
experiment is carried out with catalysts of different lengths.
Fig. 2 shows the converted methane rate obtained from the
experiment and calculation. Prior to the comparison between Experimental apparatus
them, the experimental results are reviewed briefly. The
experiment was performed with catalysts of 1 mm length and Fig. 3 shows a schematic diagram of the test section, which
different thicknesses. The total flow rate of the reactant was was described in detail in our previous paper [15]. The

Table 1 e Functions of concentration.

Mode Region Equation Concentration profile Flux at interface

[iii] 0yH 2
D ddyC2 ¼ rc CðyÞ ¼ rc 2
 rc H
þ Csur 
2Dy D y D dC  ¼ rc H
dy 
y¼0

 
[ii] 0  y  yc 2
D ddyC2 ¼ rc 
rc 2
CðyÞ ¼ 2D y þ Cc Csur
 r2D
c yc
y þ Csur D dC  ¼ D Cc C sur
 rc2yc
yc dy  yc
y¼0
yc  y  H 2
D ddyC2 ¼ kC expðuyÞþexpð2uHuyÞ
CðyÞ ¼ Cc expðuy c Þþexp½uð2Hyc Þ


[i] 0yH 2
D ddyC2 ¼ kC CðyÞ ¼ Csur expðuyÞþexpð2uHuyÞ 
1þexpð2uHÞ D dC  ¼ DuCsur 1expð2uHÞ
dy  1þexpð2uHÞ
y¼0
qffiffiffi
Cc ¼ rkc , u ¼ k
D.
3030 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 2 ( 2 0 1 7 ) 3 0 2 7 e3 0 3 5

differences from the previous experiment are the materials relationship k ¼ rc/Cc; in the following chapter, we use 600 s1
and the configuration of the catalyst. To improve the dura- as a representative value of k, which was evaluated from the
bility of the catalyst, a small amount of chromium was added experimental results for the L1T32 catalyst.
in this experiment, referring to the study of Bui et al. [21].
Therefore, the reaction rate was unknown and different from
that in the previous experiment. In this experiment, we pre- Modeling in two dimensions
pared six configurations of the catalyst embedded in grooves
with different lengths i.e., 1, 3 and 5 mm, as tabulated in Fig. 3. If the concentration of the chemical species is known at all
The total flow rate was fixed at 180 ml/min under the standard points, the overall reaction rate of a reactor can be evaluated.
conditions, and the steam-to-carbon ratio was maintained at In this chapter, we attempt to obtain a simple solution for the
3.5 by balancing with argon. The results for a catalyst tem- concentration distribution, with the gas composition at the
perature of 520  C will be shown. The total flow rate and the inlet of the channel as the only input condition.
temperature were determined to make an appearance of the
reaction saturation. Concentration profile in gas stream

In chapter 2, we derived the concentration profile in one

Experimental results
dimension, i.e., the depth direction of the porous catalyst. The
equations in Table 1 include Csur, which is the concentration at
Fig. 4 shows the converted methane rate for various concen-
the interface between the gas phase and the catalyst layer, as
trations of methane at the inlet of the channel with catalysts
a boundary condition. Here, we treat Csur as a function of the
of two different lengths and almost the same thickness. As
position along the flow direction x. To obtain a formula for
expected, a longer catalyst converted more methane. For both
Csur, some assumptions are made. First, in the gas stream, the
lengths of catalyst, saturation in the converted methane rate
concentration profile of methane is expressed one-
was observed. If the distribution of the concentration is one-
dimensionally along the flow direction x. In addition, simi-
dimensional, i.e., in the thickness direction of the catalyst,
larly to in section 2.2, Csur is the same as the bulk concentra-
the converted methane rate should be proportional to the
tion. In other words, Csur(x) represents the concentration
catalyst area, that is, the catalyst length. However, in the
profile in the gas stream. Another assumption is that there is
experimental results, the points where the gradient of the
no diffusion of chemical species in the x direction.
converted methane rate becomes zero are greatly different for
With these assumptions, the mass balance of methane in
the two lengths of catalyst. This fact implies that the phe-
the gas stream is expressed as follows by integrating the
nomena are not one-dimensional in the thickness direction
convection-diffusion equation of the channel height Hg:
and that the methane concentration varies along the flow
direction. dCsur ðxÞ
Fig. 5 shows the saturated converted methane rate of each Hg u ¼ DR; (5)
dx
catalyst, except for L3T69, in which saturation did not occur.
where u is the velocity of the gas. The right-hand side, DR, is
The saturated converted methane rate was proportional to the
the source term of the chemical species. In this case, because
volume of the catalyst. This is because the local reaction rate
the reaction does not occur in the gas phase, DR corresponds
within the entire field of the porous catalyst attained the
to the diffusion flux from the catalyst. If the methane diffuses
constant value rc in Eq. (4). Thus, from the slope in Fig. 5, we
one-dimensionally in the depth direction y in the porous
evaluate rc to be 1026 mol/(s$mm3), where the apparent vol-
catalyst layer, the fluxes in Table 1 can be substituted into DR.
ume of the catalyst is used as a unit volume. In addition, the
Thus, one can obtain the overall reaction rate by discretizing
reaction rate coefficient k can be quantified from the
Eq. (5) from the inlet to the exit of the channel.

Effectiveness factor

Here, we propose a method of expressing the concentration of

methane at any position as a function of x and y. The aim of
this model is that, apart from the numerical simulation, the
overall reaction rate throughout the reactor is evaluated from
only the gas composition at the inlet. To this end, we intro-
duce the effectiveness factor of a porous catalyst, which is
defined as follows for thickness H:
Z H
rst dy
h¼ 0
: (6)
H$rst jy¼0

The numerator on the right-hand side in Eq. (6) corre-

sponds to the actual amount of the reaction; the denominator
Fig. 2 e Converted methane rate calculated from equations is the amount of reaction if the reaction rate at every point in
in Table 1 in comparison with experimental results. the catalyst is identical to that at the interface. When the local
 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 2 ( 2 0 1 7 ) 3 0 2 7 e3 0 3 5 3031

Fig. 3 e Schematic view of test section.

reaction rate given by Eq. (4) is applied, h can be mathemati-

cally obtained. Similar to the local reaction rate shown in
Fig. 1, three situations are considered, in which the local
concentration is above or below the threshold concentration
Cc. The equations for h are summarized in Table 2. In all cases,
Eq. (6) can be rewritten as follows by using h to express DR:

dCsur ðxÞ
Hg u ¼ hHrst jy¼0 ; (7)
dx
where the negative sign on the right-hand side indicates the
consumption of methane.
Next, to obtain the function for Csur, we approximate h by a
simple formula. Fig. 6 shows h with the variation of the con-
centration at the interface, Csur, for catalyst thicknesses of
Fig. 4 e Converted methane rate plotted against methane 32 mm and 65 mm. Obviously, h is unity in mode [iii] because the
mole fraction for different lengths of the catalyst. reaction rate is constant, rc, everywhere. On the basis of the
shape of the plots, h is assumed to be given by:
8
< h1 ð ¼ const:Þ ðCsur  Cc Þ
h ¼ aCsur þ b ðCc  Csur  Cd Þ : (8)
:
1 ðCd  Csur Þ

Fig. 7 shows the graph of the newly adopted function for h,

with the meaning of the applied range of Csur. Here, we
introduce a new criterion Cd. If Csur is larger than Cd, the local
reaction rate is constant throughout the catalyst, i.e., the case
is categorized into mode [iii]. We can obtain Cd by integrating
Eq. (2) with the constant reaction rate, rc.

Table 2 e Effectiveness factors.

Mode [i] [ii] [iii]

Fig. 5 e Saturated converted methane rate plotted against Effectiveness h ¼ uH

1 expð2uHÞ1
h ¼ yHc þ uHr
kCc expð2uHÞexpð2uyc Þ h¼1
expðuHÞþ1 c expð2uHÞþexpð2uyc Þ
factor
catalyst volume.
3032 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 2 ( 2 0 1 7 ) 3 0 2 7 e3 0 3 5

rc H2
Cd ¼ Cc þ (9)
2D
Using the assumption given by Eq. (8), Csur can be expressed
as a function of x. The mode of the concentration profile shown
in Fig. 1 and Table 1 depends on Cin, which is Csur at the inlet of
the channel. Fig. 8 depicts the relationship between Csur and
the resulting mode. Similar to case [I], if Cin is smaller than Cc,
the entire field inside the porous catalyst is in mode [i]. With
increasing Cin, different modes appear. When Cin exceeds Cc, up
to three modes may appear. The number of modes depends on
the conditions; for example, the appearance of only mode [iii]
is possible, although it is not drawn in Fig. 8.
The function for Csur derived by integrating Eq. (7) and
Fig. 6 e Effectiveness factors for different thicknesses.
using Eq. (8) are summarized in Table 3. In this table, the re-
gions are segmented by the positions of x32 and x21 and the
catalyst length L as shown in Fig. 8.

Calculation results

We can obtain Csur(x), the profile of the concentration at the

interface between the gas phase and the catalyst layer, for any
case using the equations in Table 3. At each position, by
substituting the value of Csur into Table 1, which is the solution
of the concentration profile perpendicular to the flow direction,
the concentration distribution inside the catalyst is completely
determined. This corresponds to the situation where the
diffusion in the x direction is neglected in the catalyst. In the
following results, the value of h1 is given at the lowest con-
centration of the reactant in the experiment on each catalyst.
Prior to the comparison with the experimental results, the
distribution of the local reaction rate calculated by the pro-
posed model is shown in Fig. 9. The molar fraction of methane
at the inlet of the channel is 0.139 for the L5T35 catalyst. This is
a typical example in which three modes appear simulta-
neously. At a high concentration of methane, the local reaction
rate is constant regardless of the depth. In the downstream
direction, the local reaction rate tends to decrease from the
bottom of the catalyst; finally, the whole region enters mode [i].
Fig. 10 shows the calculation results in comparison with
the experimental data. Again, it is emphasized that the values Fig. 7 e Obtained function for effectiveness factor with the
of rc and k estimated for the L1T32 catalyst are consistently meaning of the threshold concentration.
applied for all catalysts and that the only input variable is Cin.
The results from the calculation and experiment are in good
was proposed, which consisted of nondimensional groups
agreement. Although the converted methane rate obtained
derived from the one-dimensional equation. Here, the crite-
from the calculation is underestimated at low concentrations,
rion is extended to the two-dimensional system.
the deviation is acceptable in spite of the simplified modeling.
The condition under which the catalyst sites are fully used
It is noteworthy that the concentration at the point where the
with the maximum reaction rate is obtained from expressed
reaction rate saturates is accurately evaluated. In the case of
by using Table 3 and Eq. (9) as
the L3T69 catalyst, such a saturation was not observed in the
experiment or the calculation. Hrc rc H2
Csur ðLÞ ¼  L þ Cin  Cd ¼ Cc þ : (10)
uHg 2D
Criteria for efficient use of catalyst in saturated reaction
This inequality can be rewritten as
In general, to increase the conversion of reactants to products,
the flow rate of the supplied gas should be small. From the
viewpoint of the yield of products, although a large flow rate Cc rc H2 rc HL
may be favorable, an excessive supply of reactants is not þ þ  1: (11)
Cin 2DCin uHg Cin
|{z} |fflfflffl{zfflfflffl} |fflfflfflffl{zfflfflfflffl}
efficient because of the saturation of the reaction. In our NC ND NL
previous paper [15], a criterion for the rate-limiting condition
 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 2 ( 2 0 1 7 ) 3 0 2 7 e3 0 3 5 3033

Fig. 8 e Appearance of modes for different concentrations at the inlet.

Table 3 e Concentration in the gas stream along the flow direction.

Position Function of surface concentration Mode
0  x  x32 Csur ðxÞ ¼ Hrc
uH x þ C3 [iii]
g

uHg
x32 ¼ ðC3  C2 Þ
Hrc 

aHrc
x32  x  x21 Csur ðxÞ ¼ a1 aC2 þ bÞexp uHg ðx32  xÞ  b [ii]


uH aC1 þb
x21 ¼ x32  aHrgc ln aC þb h1 Hk
2
Csur ðxÞ ¼ C1 exp uHg ðx21  xÞ [i]
x21  x  L

[I] x32 ¼ x21 ¼ 0; C1 ¼ Cin .

[II] x32 ¼ 0; C2 ¼ Cin ; C1 ¼ Cc .
[III] C3 ¼ Cin ; C2 ¼ Cd ; C1 ¼ Cc .

The first and second terms on the left-hand side, which are found that L1T32 and L5T35 have an option to increase the
similar to the terms of Eq. (15) in our previous paper [15], re- catalyst volume. The left-hand side of Eq. (11) is more sensi-
lates to the amount of supplied reactant and the strength of tive to the catalyst thickness H than the catalyst length L,
the diffusion in the depth direction, respectively. The third because H has an influence on ND and NL; particularly, ND is
term, NL, is a newly derived nondimensional group which proportional to the square of H. Indeed, L2T65, which is the
means the concentration decrease along the flow direction. shorter and thicker catalyst than L5T35, has a larger value of
The left-hand side of Eq. (11) is plotted in Fig. 11 for each the left-hand side of Eq. (11) in spite of smaller catalyst
size of catalyst. In this graph, the diffusion coefficient is esti- volume.
mated from the composition of the input gas. By comparison When equality holds in Eq. (11), the maximum yield of
with Fig. 10, it is obvious that the value of the left-hand side of products can be obtained with the minimum volume. Here-
Eq. (11) corresponds to the condition under which the overall after, we call this condition “optimal design point.” The
reaction is saturated. Fig. 12 shows the ratio of the nondi- optimal design point for different mole fractions of methane
mensional groups of Eq. (11), in which the molar fraction of at the inlet of the channel are plotted as solid lines in Fig. 13.
methane in the inlet gas is fixed at 0.2. This ratio can be useful The region above the lines belongs to the condition in which
for the design of configuration of catalysts. For example, it was the reaction saturated throughout the catalyst, that is, the

Fig. 9 e Distribution of local reaction rate inside a porous catalyst (not to scale).
3034 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 2 ( 2 0 1 7 ) 3 0 2 7 e3 0 3 5

Fig. 13 e Optimal design point for different mole fractions

of methane at the inlet of the channel.
Fig. 10 e Calculated converted methane rate in comparison
with experimental data for different catalysts.
and thin catalyst is more advantageous than a short and thick
one. Thus, to determine the catalyst size and to obtain the
maximum yield of products, Eq. (11) can be used as a criterion.

Conclusions

In a reactor with a porous catalyst on a channel wall, the local

reaction rate and composition cannot be easily measured. The
aim of this study was to build a model for evaluating the
overall reaction rate by predicting the distribution of the
concentration of chemical species in a porous catalyst as well
as in a.gas stream. As the target reaction, SMR was focused on.
Prior to taking into account the streamwise distribution,
the one-dimensional diffusion equation and reaction rate
Fig. 11 e Criteria for efficient use of each catalyst. with an upper limit were coupled to represent the concen-
tration of methane as a function of the depth in the catalyst.
The validity of the derived function was confirmed by com-
parison with the results of an experiment with catalysts
volume of catalyst can be reduced without decreasing the
having short lengths.
yield. On the other hand, in the region below the lines, the
Next, the model was coupled with the streamwise varia-
catalyst has some room to accept more reactants. In Fig. 13,
tion in the concentration. Consequently, we derived a model
the catalysts used in the experiment are also dotted. A broken
in which the local concentration can be specified at any po-
line depicts the same volume as L2T65. A short and thick
sition with the concentration at the inlet of the channel as the
catalyst tends to go to the region above the line of the optimal
only input variable. Using data on one catalyst, it was
design point. Therefore, if the catalyst volume is fixed, a long
confirmed that the overall reaction rate throughout the
reactor can be evaluated properly even for catalysts with a
different length or thickness. Finally, a criterion for the con-
dition in which the maximum yield of products can be ob-
tained with the minimum volume was proposed.
Nondimensional groups derived from this criterion can be
useful for determining the length and thickness of catalysts.

Nomenclature

a, b exponential indexes
C concentration, mol/m3
Cc threshold concentration for the limiting modes, mol/
m3
Cd surface concentration when the concentration at the
Fig. 12 e Ratio of nondimensional groups in Eq. (11) for bottom of the catalyst is the threshold value, mol/m3
different catalysts. Cin concentration at the inlet of the channel, mol/m3
 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 2 ( 2 0 1 7 ) 3 0 2 7 e3 0 3 5 3035

Cj concentration of chemical species j, mol/m3 [6] Shekhawat D, Spivey JJ, Berry D. Fuel cells: technologies for
Csur concentration at the interface between the gas phase fuel processing. Elsevier; 2011.
and the catalyst, mol/m3 [7] Angeli SD, Monteleone G, Giaconia A, Lemonidou AA. State-
of-the-art catalysts for CH4 steam reforming at low
D, Dj eff effective diffusion coefficients, m2/s
temperature. Int J Hydrogen Energy 2014;39:1979e97.
H catalyst thickness, m [8] Numaguchi T, Kikuchi K. Intrinsic kinetics and design
Hg channel height, m simulation in a complex reaction network; Steam-methane
k, kst reaction rate constants, 1/s reforming. Chem Eng Sci 1988;43:2295e301.
L catalyst length, m [9] Xu J, Froment GF. Methane steam reforming, methanation
NC nondimensional group associated with the supplied and water-gas shift: I. Intrinsic kinetics. AIChE J
reactant 1989;35:88e96.
[10] Hou K, Hughes R. The kinetics of methane steam reforming
ND nondimensional group associated with the
over a Ni/a-Al2O3 catalyst. Chem Eng J 2001;82:311e28.
concentration change in the depth direction [11] Holladay JD, Wang Y, Jones E. Review of developments in
NL nondimensional group associated with the portable hydrogen production using microreactor
concentration change along the flow direction technology. Chem Rev 2004;104:4767e90.
nj stoichiometric coefficient [12] Tonkovich ALY, Yang B, Perry ST, Fitzgerald SP, Wang Y.
p partial pressure Pa From seconds to milliseconds through tailored microchannel
reactor design of a steam methane reformer. Catal Today
DR concentration change in reaction, mol/(m2 s)
2007;120:21e9.
rc reaction rate in the saturated region, mol/(m3 s)
[13] Zhai X, Cheng Y, Zhang Z, Jin Y, Cheng Y. Steam reforming of
rst reaction rate, mol/(m3 s) methane over Ni catalyst in micro channel reactor. Int J
u velocity, m/s Hydrogen Energy 2011;36:7105e13.
x direction along the flow, m [14] Lin KK, Saito M, Niina Y, Ikemura K, Iwai H, Yoshida H.
y depth direction of the catalyst layer, m Deduction of proper reaction rate of steam methane
yc depth position where the reaction rate changes, m reforming over catalyst surface validated with a combination
of one- and two-dimensional simulations. J Therm Sci
Greek symbols Technol 2012;7:633e48.
a coefficient of surface concentration in effectiveness [15] Saito M, Kojima J, Iwai H, Yoshida H. The limiting process in
factor, m3/mol steam methane reforming with gas diffusion into a porous
catalytic wall in a flow reactor. Int J Hydrogen Energy
b intercept in the equation of effectiveness factor
2015;40:8844e55.
h effectiveness factor [16] Solsvik J, Tangen S, Jakobsen HA. On the consistent modeling
h1 effectiveness factor in the case of a low of porous catalyst pellets: mass and molar formulations. Ind
concentration Eng Chem Res 2012;51:8222e36.
u square root of the ratio of k to D, 1/m [17] Ahmed K, Foger K. Kinetics of internal steam reforming of
methane on Ni/YSZ-based anodes for solid oxide fuel cells.
Catal Today 2000;63:479e87.
references [18] Arcotumapathy V, Vo D-VN, Chesterfield D, Tin CT,
Siahvashi A, Lucien FP, et al. Catalyst design for methane
steam reforming. Appl Catal A General 2014;479:87e102.
[19] Sugihara S, Kawamura Y, Iwai H. Rate equation of steam
[1] Verhelst S. Recent progress in the use of hydrogen as a fuel
methane reforming reaction on Ni-YSZ cermet considering
for internal combustion engines. Int J Hydrogen Energy
its porous microstructure. In: Proceedings of 7th European
2014;39:1071e85.
Thermal-Sciences Conference; 2016. FCS-2.
[2] Chiesa P, Lozza G, Mazzocchi L. Using hydrogen as gas
[20] Lin KK, Saito M, Kojima J, Kuratsu S, Iwai H, Yoshida H.
turbine fuel. J Eng Gas Turbines Power 2005;127:73e80.
Comprehensive method for evaluating reaction rates of
[3] Buonomano A, Calise F, d’Accadia MD, Palombo A,
steam methane reforming on catalyst-coated surface with
Vicidomini M. Hybrid solid oxide fuel cellsegas turbine
consideration of limiting steps. J Therm Sci Technol
systems for combined heat and power: a review. Appl Energy
2013;8:165e77.
2015;156:32e85.
[21] Bui QTP, Kim Y, Yoon SP, Han J, Ham HC, Nam SW, et al.
[4] Wood DA, Nwaoha C, Towler BF. Gas to liquids (GTL): a
Steam reforming of simulated biogas over plate Ni Cr
review on an industry offering several routes for monetizing
catalysts: influence of pre-oxidation on catalytic activity.
natural gas. J Nat Gas Sci Eng 2012;9:196e208.
Appl Catal B Environ 2015;166:335e44.
[5] Mazloomi K, Gomes C. Hydrogen as an energy carrier:
prospects and challenges. Renew Sustain Energy Rev
2012;16:3024e33.