Methane steam reforming over a Ni/Nial2o4 model catalyst - Kinetics

Article abstract of DOI:10.1002/cctc.201402017

The steam reforming reaction of methane to synthesis gas (CO and H ₂) was investigated in a very broad steam-to-carbon (S/C) ratio range (0.2-7.1) over a Ni/NiAl₂O₄ catalyst at 843, 858, and 873 K. The dataset includes 179

Full text of DOI:10.1002/cctc.201402017

CHEMCATCHEM
FULL PAPERS
DOI: 10.1002/cctc.201402017
Methane Steam Reforming over a Ni/NiAl O Model
2
4
Catalyst—Kinetics
Christoph Sprung, Bjørnar Arstad, and Unni Olsbye*
[a]
[b]
[a]
The steam reforming reaction of methane to synthesis gas (CO
methane at high S/C ratios. The extent of the water gas shift
13
and H ) was investigated in a very broad steam-to-carbon (S/C)
reaction was assessed by CO co-feed experiments and ther-
2
2
ratio range (0.2–7.1) over a Ni/NiAl O catalyst at 843, 858, and
modynamic modelling and found to be not in equilibrium. A
change in the kinetic isotope effect (i.e., inverse and normal),
as well as the apparent activation energy, was observed de-
pending on the applied S/C ratios. The experimental data were
satisfactorily described by a Langmuir–Hinshelwood expres-
sion.
2
4
873 K. The dataset includes 179 data points at each tempera-
ture giving a detailed insight in the gradual change of kinetic
orders in both methane and steam with changing S/C ratios as
well as with total reactant partial pressures at differential con-
version levels. At low S/C ratios, steam dominates the kinetics
with high kinetic orders (low for methane) and vice versa for
view of such studies is given in Table 1 (see also Ref. [11]). As
observed from Table 1, vastly different apparent activation en-
Introduction
ꢀ
1
Ni-catalysed methane steam reforming (MSR) to synthesis gas
is a key step in industrial natural gas conversion processes,
either alone or as part of autothermal reformer units.
ergies (ranging from 29 to 364 kJmol ) have been reported
for Ni catalysts during methane steam reforming, and rate
equations range from simple first-order dependencies in
p(CH ), to complex equations in which steam and/or the reac-
4
ꢀ
1
CH þ H O Ð CO þ 3 H
D Hð873 KÞ ¼ 223:1 kJ mol
tion products slow down the reaction rate. The values in
Table 1 nicely illustrate that the validity of kinetic investigations
of catalytic reactions is restricted to the catalytic system under
study, including the main active site, support and promoters,
partial pressures, temperatures and conversion range. Typically,
a temperature increase will lead to enhanced reaction rates be-
cause a higher fraction of reactant molecules will possess suffi-
cient energy to react, but also because (in this case) the NiO–
Ni equilibrium is shifted towards the active species/reduced Ni
4
2
2
R
ðR1Þ
ðR2Þ
ðR3Þ
ꢀ
1
CH þ 2 H O Ð CO þ 4 H
D Hð873 KÞ ¼ 186:9 kJ mol
4
2
2
2
R
ꢀ
1
CO þ H O Ð CO þ H
D Hð873 KÞ ¼ ꢀ36:2 kJ mol
2
2
2
R
for CH activation at higher temperatures. Another parameter
4
Steam reforming may globally be described by Reac-
tions (R1)–(R3), which represent the direct conversion of meth-
that influences the apparent activation energy is the active-site
coverage of reactants, intermediates, and products. The cover-
age of intermediates might be so scarce at a given tempera-
ture–partial pressure that adsorbate formation becomes a rate-
limiting step. Looking again at the literature reports cited in
Table 1, it is observed that the investigations performed at the
lower temperatures include adsorption terms in the denomina-
ane to CO (R1), CO (R2) and the water gas shift (WGS) reac-
2
tion (R3), respectively. During the past five decades, huge ef-
forts have been allocated to catalyst and process optimisation
for MSR, by varying the metal elements shown to be active for
the title reaction (see Ref. [1] and citations therein), the sup-
[
2]
[3]
[4]
[5]
[6,7]
port materials (MgO, MgAl O , Al O , SiO2, NiAl O
4
and
tor implying a competition of reactants (and products) for the
2
4
2
3
2
[
8]
[9] [10]
[3,12–17]
using bare Ni foil ) or adding various promoters (Sn,
S
).
active site on the catalyst.
At higher temperatures (above
[4,8,18]
A number of kinetic investigations of the MSR reaction over
Ni-based catalysts have been published, and a selective over-
873 K), adsorbate coverage decreases significantly
until
[2]
the kinetics depend solely on the methane partial pressure.
This trend has later been described also by DFT methods.
[1]
[a] Dr. C. Sprung, Prof. U. Olsbye
Furthermore, the support material reveals its influence: Steam
and reaction products) plays a significant role for investiga-
University of Oslo
(
Department of Chemistry
P.O. Box 1033, NO-0315 Oslo (Norway)
Fax: (+47)22855441
tions of non-active supports (e.g., spinel phase of Ni, Mg or
Ca), whereas supports that efficiently activate steam (e.g.,
MgO, g-Al O ) usually lead to kinetic independence in steam,
E-mail: unni.olsbye@kjemi.uio.no
2
3
[b] Dr. B. Arstad
probably because steam is activated on the support and sup-
plied to the metal phase by a spillover process.
SINTEF Materials and Chemistry
Pb. 124, NO-0314 Oslo (Norway)
A well-known strategy to possibly circumvent artificial ef-
fects on kinetic parameters is to study the kinetics of a selected
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/cctc.201402017.
ꢀ
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 0000, 00, 1 – 15
&
1
&
ÞÞ
These are not the final page numbers!

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
Table 1. Selected overview on kinetic models for methane steam reforming.
ꢀ
1
Catalyst
Ni foil
T [K]
S/C
E
A
[kJmol
]
r=
Reference
[4]
kpðCH4Þ
1073–1173
>1.4
129.8
pðH OÞ
2
1
þa
þbpðCOÞ
pðH2 Þ
Ni/a-Al
Ni/a-Al
Ni/Kieselgur
2
2
O
O
3
973–1173
673–773
882–1183
823–948
76.6
151.6
108.7
364ꢁ19
kp(CH
4
)
[8]
[18]
[5]
pðCH4 Þ
3
0.95–7.4
2–17
3–5
k
þ
pðH2 Þ
kp(CH
4
)
ꢀ
ꢁ
Ni/a-Al
2
O
3
[14]*
pðCH4 ÞpðH2 OÞ
pðH2 Þ
kCO
3
ꢀpðCOÞ=KP1
2
ð1þKCO kCO Þ
1:2
k0 expðꢀ46:9=TÞpðH2 OÞ
1:22
Ni/NaY
773–838
673–773
2.8–5.7
2.5–7.8
46.9
69
[19]
1
þkpðH2 Þ
0
:24
0:28
4
pðCH4 Þ pðH2 OÞ
Ni/NiAl
2
O
4
7:3 ꢂ 10 expðꢀ69=RTÞ
1:86
[7]
ð1þKpðH2 Þ
Þ
Ni/MgO
773–923
773–953
913
4
99.77
29 (67)
k
1
p(CH
kp(CH )p(H
ðpðCOÞ ꢀ Kp1
4
)
[16]
[20]
[15]*
ꢀ0.5
Ni/g-Al
2
O
3
0.2–5
3
4
2
O)
pðCH4 ÞpðH2 OÞ
G-56B
ꢀk
1
3
Þ=DEN
1
pðH2 Þ
Ni wire
Ni/CaAl
873–1023
748–823
2.5
2.7–5.8
213.5
185
[21]
[22]*
k1
3
n
2
O
4
pðH2 Þ
m
ðpðCH
4
ÞpðH
2
OÞ ꢀ pðH
2
Þ pðCOÞ=K
1
Þ=ðDEN
2
Þ
3
k1 pðCH4 ÞpðH2 OÞ½1ꢀðpðCOÞpðH2Þ Þ=KpðCH4 ÞpðH2 OÞꢃ
Ni
743–973
748–823
2
2
3
[12]
[13]
pðH2 OÞþl2 pðH2 Þ þl3 pðH2 Þ
0:5
1:25
3
k1 ðpðCH4ÞpðH2 OÞ =pðH2 Þ ½1ꢀðpðCOÞpðH2 Þ Þ=Kp1 ðCH4 ÞpðH2 OÞꢃ
Ni/Al
2
O
3
4–7
209.2
2
ðDEN3 Þ
Ni/MgO
Ni/MgAl
Ni/MgAl
823–1023
773–848
773
0.8–2.8
3–5
4
102
240.1
231.6
kp(CH
4
)
[2]
[3]
[1]*
2
2
O
O
4
k1
3
2
4
pðH2 Þ
2:5 ðpðCH
4
ÞpðH
2
OÞ ꢀ pðH
2
Þ pðCOÞ=K
1
Þ=ðDEN
4
Þ
2
pðCH4 ÞpðH2 OÞ
pðCH4ÞpðH2 OÞ
4
DEN
DEN
DEN
DEN
1
2
3
4
¼ 1 þ K
a
pðH
2
OÞ þ K
a
K
S1
3
þ K
a
K
S1
:5
K
S2
pðH2 Þ
pðH2 Þ
0
¼ 1 þ KCOpðCOÞ þ KH2 pðH
2
Þ þ K
H
pðH
2
Þ
þ KCH4 pðCH
4
Þ þ KH2
O
pðH
2
OÞ=pðH Þ
2
0:5
¼ 1 þ KCOpðCOÞ þ KH2 pðH
¼ 1 þ KCOpðCOÞ þ KH2 pðH
2
Þ
þ K
H
O
pðH
2
OÞ=pðH
2
Þ
2
2
Þ þ KCH pðCH
4
4
Þ þ K pðH
H
2
O
2
OÞ=pðH
2
Þ
*
Additional equations in the model, see reference for more detail.
reaction–catalyst system under a sufficiently wide range of re-
action conditions that an elementary step reaction mechanism
may be deduced and used as basis for a global rate equation.
Several elementary step mechanisms have been proposed
CH* þ * Ð C* þ H*
C* þ O* Ð CO* þ *
ð6Þ
ð7Þ
ð8Þ
ð9Þ
ðrdsÞ
*
Ð 1=2 H ðgÞ þ *
[
23]
H
2
for the Ni-catalysed MSR reaction. Aparicio presented a 13-
step micro-kinetic model in which a surface reaction is rate-de-
termining. Further developments incorporating DFT modelling
showed the structure sensitivity of the title reaction and a pref-
erence for steps on the surface owing to the significantly
CO* Ð COðgÞ þ *
As observed from Table 1, the methane dissociation, step (1),
has been found to be rate-limiting in several experimental
[10,24]
lower activation barriers than in the closed-packed surface.
[
2,4,5,8]
[2,5]
studies,
by partial-pressure variation
and/or isotopic la-
The Nørskov group pioneered the use of DFT methods to find
scaling relationships for key steps in, for example, MSR and to
employ these relationships to predict reactivity trends among
[
2]
belling experiments. Other studies, using different catalysts
and/or reaction conditions, concluded that oxidation with sur-
[3,6,13,26,27]
[
1,25]
face oxygen was rate-limiting,
as suggested by posi-
various metal elements.
The models are based on experi-
[3,13]
tive kinetic orders in steam during space time
and partial
mental studies, and led to the set of elementary steps present-
ed in Equations (1)–(9). rds: rate determining step
[6]
[26]
pressure variation, as well as from isotopic labelling experi-
ments. However, discrepancies exist with respect to the surface
species reacting with surface oxygen.
CH ðgÞ þ 2* Ð CH * þ H*
ðrdsÞ
ð1Þ
ð2Þ
ð3Þ
ð4Þ
ð5Þ
4
3
Xu and Froment suggested that methane was only partially
dehydrogenated before reacting with surface oxygen [Eqs.
H OðgÞ þ 2* Ð OH* þ H*
2
[3]
(
10)–(15), with the other steps similar to steps (1)–(9)]. Three
OH* þ * Ð O* þ H*
independent oxidation channels were considered rate-deter-
mining for the overall reactions (R1)–(R3).
CH * þ * Ð CH * þ H*
3
2
CH * þ * Ð CH* þ H*
CH * þ O* Ð CH O* þ *
ð10Þ
2
2
2
ꢀ
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 0000, 00, 1 – 15
&
2
&
ÞÞ
These are not the final page numbers!

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
CH O* þ * Ð CHO* þ H*
ð11Þ
ð12Þ
ð13Þ
ð14Þ
ð15Þ
abundance of Ni/NiO than of NiAl O , which supports the former
2 4
2
statement. TEM investigations pictured the spinel phase as large
single crystallites and NiO had varying ranges of 5–30 nm. Such
small particle sizes for NiO were additionally confirmed by the spe-
cific intensity distribution of certain Raman bands. XPS spectra
were also recorded after treating the catalyst in different reaction
conditions and handling the sample in vacuum; the NiO/Ni in-
creased with increasing S/C ratio (which was expected). XRD and
CHO* þ * Ð CO* þ H*
ðrds1Þ
ðrds2Þ
CO* þ O* Ð CO * þ *
2
CHO* þ O* Ð CO * þ H*
ðrds3Þ
2
CO * Ð CO ðgÞ þ *
2
2
O chemisorption data are reported in this manuscript. XRD data
2
[
1]
were collected on a Bruker D5000 in Bragg–Brentano geometry.
The CuKa1 radiation was selected by a Ge (111) monochromator. In
Figure 1, the diffractogram is shown with the characteristic reflec-
Equation (1) and (7) were predicted by Jones et al. as rate-
determining. Surface reactions on nickel have been found to
[
1,23]
[3,28]
be the rds and might involve a C*
or a CH * species.
x
tions for the two materials NiAl O4 and NiO according to their
2
Recent modelling investigations (both DFT and micro-kinetics)
elaborate possible variations in the oxidation mecha-
cubic face-centred space groups.
[
16,24,29,30]
nism.
The present work was performed in the context of a broader
project on sorbent-enhanced reforming of methane, in which
methane and steam are converted to a mixture of hydrogen
and sorbed CO . Owing to the presence of a sorbent material
2
during the reforming reaction, the equilibrium is shifted to the
product side, and the reaction temperature may be lowered
[
31–33]
significantly (down to ꢄ823–873 K
) in comparison to con-
[34]
ventional steam reforming of methane (up to 1223 K ). The
aim of the present study was to provide a kinetic investigation
of the title reaction in a relevant range of sorbent-enhanced
reforming reaction temperatures, and covering a wider range
of reaction conditions than any open literature study known to
date, to provide a consistent data set bridging the gap be-
tween methane-lean and methane-rich conditions. A second
aim was to elucidate further mechanistic details about the title
reaction over a Ni-based catalyst. For this purpose, a 2 wt%
Figure 1. Characterisation of the 2 wt% NiO/NiAl
onstrating the typical reflections for the space groups of NiO and NiAl
2 4
O catalyst by XRD, dem-
2
4
O .
NiO/NiAl O catalyst was chosen because of the exceptional
2
4
Dynamic O2 chemisorption was performed to determine the
amount of nickel surface atoms. A catalyst portion of 1 g was
placed inside a quartz reactor under a He 5N (99.999 %) flow of
stability under a wide range of steam-to-carbon (S/C) ratios. In
a first part of this work, kinetic measurements were performed
ꢀ1
[
6]
1
00 SmLmin (Sml, volume at standard conditions) (5.0). The
in a limited range of S/C ratios at 873 K. Herein, the kinetic
sample was reduced in 10 vol% H /He at 873 K for 2–3 min before
2
study has been extended to cover 843–873 K and S/C=0.2–
cooling down to RT (i.e., ꢄ300 K) under helium flow. A pneumatic
7
.1. Furthermore, kinetic isotope effect (KIE, CH /CD transient)
4 4
6
-port valve equipped with a 250 mL sample loop sent pulses of
13
and CO co-feed experiments were performed to elucidate
2
a 1:1 O /Ar mixture to the reactor, still under He flow. The reactor
2
further mechanistic details.
effluent was monitored by a Pfeiffer quadrupolar mass spectrome-
ter. The recorded mass numbers (m/z=4 and 32) were evaluated
as follows: The m/z=32 signal was normalised with respect to
m/z=4 and after baseline subtraction each peak area was deter-
mined. A sufficient number of pulses were sent to the reactor to
have at least ten peaks with no adsorption (saturated surface),
which were taken as a reference to determine the amount of
Experimental Section
Catalyst preparation and characterisation
The catalyst was a NiO/NiAl O4 material. Respective amounts of
2
nickel and aluminium nitrate were dissolved in a 1:1 water/isopro-
pyl alcohol solution, which was evaporated under vigorous stirring
to obtain a solid precursor residue. The residue was dried in
a drying cabinet for >48 h at 343 K and subsequently calcined in
static air for 8 h at 1473 K, thereby forming the spinel phase. Sur-
plus amounts of nickel were present as NiO, in this study 2 wt% of
the NiAl O .
oxygen adsorbed. Converting the number of adsorbed moles O
2
to
the number of nickel surface atoms is depending on the support
[
36–38]
and size of the Ni particles.
In this characterisation, a O
/Ni
2
[39]
ꢀ6
ratio of 3:4 was chosen. Thus the 7.24ꢁ10 moles O
give 9.654ꢁ10 moles of surface nickel per gram catalyst (N
adsorbed
(Ni)).
2
ꢀ
6
s
The dispersion of the Ni(O) particles was determined (ratio of
moles of surface Ni to total moles of Ni(O) on the catalyst) to D=
2
4
0
.036. Under the assumption of spherical Ni particles, this number
The catalyst material was characterised by SEM, TEM, and Raman
spectroscopy and were presented in detail in Ref. [35]. NiO and
NiAl O were clearly distinguished by SEM. NiO appeared in small
corresponds to an average Ni particle diameter of 26 nm, in good
agreement with the TEM measurements cited above.
2
4
rounded particles of <500 nm in size, which densely covered (and
which were also in contact with each other) the surface of large
sharp-edged NiAl O particles. X-ray photoelectron spectroscopy
2
4
(
XPS) and Raman microscopy investigations revealed a higher
ꢀ
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 0000, 00, 1 – 15
&
3
&
ÞÞ
These are not the final page numbers!

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
Reactor setup
Table 2. Reaction conditions: S/C ratios in the reactor feed, determined by the inlet mGC, used for the investi-
gations at 843–873 K.
The reactor experiments were per-
formed in a fixed-bed reactor unit
heated by an electrical furnace.
The catalyst bed consisted of the
catalyst (10 mg, 2 wt%NiO/NiAl O )
!
p(CH
4
) variation!
1.29
0.91
1.27
1.49
1.78
2.03
2.34
2.66
3.06
3.29
3.69
3.97
4.31
4.58
5.02
5.34
0.79
0.96
1.26
1.41
1.66
1.9
2.13
2.36
2.6
2.87
3.17
3.37
3.67
3.87
4.13
0.61
0.84
1.01
1.18
1.41
1.63
1.84
2.0
0.53
0.67
0.85
1.01
1.16
1.36
1.5
1.71
1.9
2.05
2.2
0.39
0.58
0.75
0.91
1.08
1.18
1.34
1.47
1.62
1.77
1.96
2.07
2.26
2.4
0.45
0.57
0.69
0.81
0.92
1.03
1.19
1.32
1.46
1.36
1.76
1.86
2.02
2.13
2.33
0.33
0.44
0.55
0.68
0.76
0.84
0.96
1.09
1.21
1.3
1.46
1.54
1.67
1.77
1.9
0.31
0.4
0.25
0.33
0.39
0.48
0.58
0.66
0.72
0.81
0.89
0.97
1.04
1.14
1.21
1.31
1.37
0.23
0.29
0.35
0.43
0.5
0.57
0.64
0.7
0.79
0.86
0.92
1.0
1.08
1.14
1.22
0.21
0.27
0.31
0.4
0.45
0.51
0.57
0.64
0.7
0.76
0.83
0.9
0.96
1.08
*
1
2
2
3
.7
2
4
.17
.56
.0
0.48
0.56
0.66
0.74
0.83
0.94
1.02
1.12
1.22
1.31
1.39
1.5
and a-alumina (490 mg, both 250–
20 mm) and was held in place by
4
alumina wool inside an alumina re-
actor (inner diameter 8 mm). The
temperature was measured by
a thermocouple placed in an alu-
mina thermocouple well (outer di-
ameter 3 mm) in the catalyst bed.
Ar, H , N , and CH gas flows were
regulated by mass flow controllers,
which were fed from a central gas-
distribution system. All gas transfer
lines were heated to temperatures
›
O) variation
fl
3.35
3.81
4.13
4.59
4.98
5.4
p(H
2
2.17
2.4
2.58
2.79
3.02
3.23
3.43
5
6
6
7
.8
.3
.690
.09
2.38
2.56
2.76
2.96
2
2
4
2.55
1.59
>
393 K. The desired amount of
steam was generated by a two-
[40]
stage steam generator, built in-house. First, steam was generat-
ed by bubbling Ar through boiling water. Attached to this genera-
tor part was a vigreux column (the second stage), through which
the generated steam was guided. This column was surrounded by
a water jacked kept at a specific temperature (e.g., 343 K), which
determined the H O/Ar ratio in the outlet of the steam generator
2
setup.
N was added to the feed as an internal standard (1.5 vol%) to cal-
2
culate mass balances (see below). Two micro-gas chromatographs
(mGC) (Agilent3000), each equipped with two columns (molsieve
and U-Plot) and thermal conductivity detectors, simultaneously an-
alysed the feed and effluent gas, respectively. The analysis time
was 4 min.
Test procedure
The catalyst was activated (for all investigations) at 873 K in
1
0 vol% H /Ar for 10 min and finally stabilised under reference
2
conditions for steam methane reforming (i.e., 6.3% CH , 12.3%
4
ꢀ1
ꢀ1
H O, 1.5% N , 79.9% Ar, W/F(CH )=0.34 g
h mol(CH ) ) for
2
2
4
cat
4
1
20 h (8 h for isotopic labelling experiments, see below). The stabi-
lisation period was necessary to reach the linear deactivation state
of the catalyst, which began approximately 6–8 h after start-up of
the reaction. This initial period was characterised by an exponential
decrease in activity. The extended 120 h stabilisation period was
meant to emphasise the durability of the catalyst. The total feed
ꢀ
1
rate (for all experiments) was 200 Smlmin . The whole range of
partial-pressure variation experiments (segments) are listed in
Table 2. Its S/C ratio values were determined by mGC feed analysis.
For each temperature, the experiment started with conditions in
the upper left corner of Table 2, proceeding in a column to the
lower right corner. Each S/C segment was followed by a segment
under reference conditions. Both lasted for 30 min to get a suffi-
cient number of GC measurements. A constant conversion level
was rapidly observed after a switch of feed gas compositions,
shown in Figure 2 for the reactants and two products. Table 2 con-
tains several repetitions of similar S/C values which have a lower/
higher total reactant partial pressure. Such data points are there-
fore part of a different partial pressure variation sequence and are
thus not the same.
Figure 2. Experimental raw data at 873 K, 10:490 mg (catalyst/inert),
4
ꢀ1
2
7.91ꢁ10 h for one set of p(H O) variation: outlet concentrations of
a) reactants and b) products; taken from Ref. [6].
C, H and O mass balances were determined and for high methane
and steam partial pressure levels, all balances were within a range
of ꢁ5%. The H balance was between 0.95 and 1.0 in all experi-
ments. However, the lowest methane partial pressure led to sys-
tematically lower C balances and the lowest steam partial pres-
sures to higher O balances. Measurement uncertainties of the de-
termined concentrations by GC analysis were quantified as stan-
ꢀ
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 0000, 00, 1 – 15
&
4
&
ÞÞ
These are not the final page numbers!

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
dard deviations of the repeated analysis runs for each experimental
condition. The uncertainty of the catalyst mass was assumed to be
Calculations
Concentrations determined by mGC analysis were converted into
ꢁ
0.1 mg. In Figure 6, the propagated error on the TOF is illustrated
ꢀ
1
molar flows (for example molmin ) based on the analysis of the
for the lowest four sets of partial pressure variation. The absence
of inner and outer diffusion limitations were confirmed by particle
size and bed height variation, respectively. The latter is presented
in Figure 3a (reprint from Ref. [6]). To ensure that diffusion limita-
internal standard (N ), which was always fed with the same flow
2
rate. The gas flows were assigned as reactor feed (a) and effluent
(w), determined by the mGC at the reactor inlet and outlet, respec-
tively. The methane and steam conversion (x) were determined by
Equation (16) and (17), respectively. The turnover frequency (TOF)
ꢀ1 [41]
is defined by Equation (18), in s . m is the catalyst weight in
gram and N (Ni) the specific number of nickel surface atoms per
s
gram catalyst. Finally, the CO selectivity was calculated by Equa-
2
tion (19).
The approach to equilibrium h was determined according to Equa-
[
2]
tion (20). The equilibrium constants K and K2 for reactions (R1)
1
[42]
and (R2) were determined from the HSC database at 873 K. Max-
imum values for h and h were 3ꢁ10 and 2ꢁ10 , respectively,
ꢀ6
ꢀ8
1
2
hence insignificant contribution to the determination of the con-
version rate. For the highest reactant partial pressures, a methane
conversion of 12.7% was determined, whereas the thermodynamic
[42]
equilibrium predicts a methane conversion of 95.1%.
The as-
sumption of negligible influence of the reverse reaction was con-
firmed by CD /H O experiments, in which no H/D exchange was
4
2
observed in methane (see Results). The rates reported in this work
therefore represent the forward methane reforming reaction.
wCO þ w
CO2
xCH4 ¼ w
ꢅ 100%
ꢅ 100%
ð16Þ
þ w
CO2
þ wCH4
CO
wCO þ 2w
CO2
xH2O ¼ w
ð17Þ
ð18Þ
ð19Þ
þ 2w
þ wH2O
CO
CO2
xCH4 ꢅ a
CH4
TOF ¼
m ꢅ N ðNiÞ
S
wCO2
SCO2 ¼ w
CO
þ wCO2
3
4
1
pðCOÞ ꢅ pðH Þ
1 pðCO Þ ꢅ pðH Þ
2
2
2
h1 ¼
h2 ¼
ð20Þ
2
K pðCH Þ ꢅ pðH OÞ
K pðCH Þ ꢅ pðH OÞ
1
4
2
2
4
2
Figure 3. a) Variation of catalyst bed height at 873 K and constant gas
hourly space velocity (CH /H O/N /Ar=6.4:4.8:0.8:87) for catalyst/inert
4 2 2
Unless stated explicitly, all data were corrected for deactivation.
The methane conversion at the end of the stabilisation period was
weights: (*) 10:490 mg; (&) 15:735 mg; (~) 20:980 mg; activation ener-
gies are determined by the Arrhenius equation; figure taken from Ref. [6];
b) Activation energy according to the Arrhenius equation for S/C=0.3, 1.2,
and 3.4 and temperatures between 843 and 873 K, 10:490 mg (catalyst/
taken as reference conversion x . A correction factor b was deter-
ref
mined for each reference methane conversion (following a partial-
pressure variation step, see Figure 2a) according to Equation (21).
This factor b was then multiplied with the conversion at partial-
pressure variation conditions to get the corrected value. All other
gas flows were as well corrected (Equations (22)–(26)). To correct
all flow values, the prior-determined balances for carbon (CB), hy-
4
ꢀ1
inert), 7.91ꢁ10 h
.
tions were not rate-determining at other S/C ratios, Arrhenius plots
were made for S/C=0.3, 1.2 and 3.4, respectively. The plots are
presented in Figure 3b, and show a constant slope in the tempera-
ture range tested, thus suggesting that the reaction is not diffu-
sion-limited at S/C ratios (0.3–3.4). The plots in Figure 3a and 3(b)
further show a gradual increase in the apparent activation energy
from low to high S/C ratios. This observed increase could be the
result of a change in rate-determining step from low to high S/C
or of a change in surface coverage from low to high S/C. To eluci-
date the effect of product formation on the observed reaction
rates, product gases were added to the reactor feed at 873 K. The
drogen (HB), oxygen (OB) and the CO selectivity (SCO2) were consid-
2
ered to remain constant under experimental and corrected condi-
tions. A capital letter (e.g., W, X) in the following refers to a correct-
ed value (lower case letters, for example, w, x are uncorrected).
ꢂ
ꢃ
ꢀ
1
xCH4
ref
b ¼
XCH4 ¼ xCH4 ꢅ b
ð21Þ
x
wCO2 ꢅ CB ꢅ a ꢅ b
CH4
WCO2
¼
w
ð22Þ
ð23Þ
ð24Þ
CO
2
þ w
SCO2
CH4
data are reported in Figure 4 for H and CO co-feeds, and for CO2
2
WCO2 ꢀ S ꢅ WCO2
CO2
co-feed in Table 4, respectively. The effect of each product on the
overall methane conversion rate was negligible, and suggests that
the product concentrations do not need to be taken into account
when modelling the rates of reaction (see Results).
WCO
¼
S
CO2
WCO þ WCO2 ꢀ X ꢅ ðWCO þ WCO2
Þ
CH4
WCH4
¼
X
CH4
ꢀ
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 0000, 00, 1 – 15
&
5
&
ÞÞ
These are not the final page numbers!

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
[43]
factor for CH , which is 0.763. Additionally there was
4
off-line analysis by a GC–MS (HP6890 Gas Chromato-
graph equipped with
a GSGASPRO column (60 m,
0
.32 mm) and an HP-5973 mass selective detector). Gas
samples were taken with a syringe through a septum at
the reactor exit and injected into the GC–MS. The kinetic
isotope effect (KIE) was studied by switching between
CH /H O and CD /H O mixtures at steady state. The KIE
4
2
4
2
was determined by the ratio of the conversion during
ꢀ
ꢁ
xðCH4Þ
CH4 and CD4 feed
. In a second set of experi-
xðCD Þ
4
13
ments, labelled CO was co-fed (the amount was ap-
2
proximately that of CO generated by the steam reform-
2
ing reaction itself) to the reactant mixture. The methane
conversion was determined by Equation (16), with a cor-
13
rected value for the co-fed CO2.
Results and Discussion
Kinetic tests
We will first discuss in detail the data collected at
873 K. The plot of TOFs versus S/C ratios is shown for
the 873 K experiments in Figure 5. Therein, the solid
lines connect experimental values at constant
a) p(CH ) and b) p(H O) in the reactor feed. The data
4
2
in Figure 5 appear (note the logarithmic scale of the
abscissa) as an equal-sided triangle. The logarithmic
illustration of the present data set in Figure 5 clearly
visualises a transition in kinetic regimes from low to
high S/C ratios, mirrored at a plane at S/C=1. How-
ever, it has to be pointed out that the appearing tri-
angular shape of the data set is not universal, but limited to
the TOF=f(S/C) range in which the reaction was studied. Ap-
plying reaction conditions outside the parameter space of par-
tial pressure variation conditions of this work might very well
expand the TOF=f(S/C) area.
Figure 4. Methane conversion as a function of time on stream, alternating feed between
4 2 2
reference conditions (i.e.: 6.3% CH , 12.3% H O, 1.5% N , 79.9% Ar) and addition of
4
ꢀ1
product at 873 K, S/C=2, 10:490 mg (catalyst/inert), 7.91ꢁ10 h : a) hydrogen and
b) CO.
WH2O ¼ OB ꢅ aH2O ꢀ WCO ꢀ 2WCO2
ð25Þ
ð26Þ
1
WH2 ¼ 2 ꢅ ðHB ꢅ 4 ꢅ aCH4 þ HB ꢅ 2 ꢅ aH2O ꢀ 2 ꢅ WH2O ꢀ 4 ꢅ WCH4
Þ
Thermodynamic equilibrium calculations were conducted with raw
[42]
(
uncorrected) data by using the HSC 5.1 database as above. In
High reaction rates are found at the maxima of the sum of
the applied methane and steam partial pressures, and for a
S/C ratio of 1. From Figure 5a, it is observed that the reaction
Table 3, input values for thermodynamic calculations are summar-
ized, assuming that all converted methane gave CO and H . The
2
methane concentration was used as a reference, because it does
not take part in the WGS reaction.
rate generally increases with p(H
p(CH ) at low partial pressures of CH . The rate limit increases
O), but is quickly limited by
2
4
4
with an increasing p(CH ), thus suggesting that the access to
4
Table 3. Input amounts of H
2
O, CO and H
2
for WGS equilibrium calcula-
CH and/or to adsorbate species derived from it may represent
4
tion, x is the moles of methane converted.
a limiting parameter for methane conversion. At very low
methane partial pressures, an increasing steam partial pressure
has a negative influence on the methane TOF. This last obser-
vation may suggest that methane and water compete for the
same active site(s) on the catalyst surface. Considering subse-
quently the data in Figure 5b, it is observed that the rate of
methane conversion generally increases with an increasing
p(CH ), except at low p(H O) (i.e., S/C<1 and p(H O)ꢆ
CH
4
H
2
O
CO
H
0
2
before reaction
m
n
nꢀx
0
x
after steam reforming
mꢀx
3x
Isotopic labelling
4
2
2
Isotopic labelling experiments were performed for three different
S/C feed compositions (0.5, 1.4 and 3.6) according to the proce-
dure described above (amount of catalyst, activation, space veloci-
0
.026 mbar), at which increasing p(CH ) has no impact on the
4
methane conversion rate.
The kinetic data are further plotted as TOF(CH ) versus
13
4
ty, etc.). Labelled gases (CD and CO ) were purchased from Cam-
4
2
p(CH ) and TOF(CH ) versus p(H O) in Figure 6a and b, respec-
4
4
2
bridge Isotope Laboratories, Inc., with an isotope purity of >99%
1
3
tively. The data in Figure 6 clearly show the gradual change of
kinetic orders for both methane and steam with changing S/C
ratio. This implies the incorporation of both reactants in the ki-
(
D and C, respectively). The gas conversion factor (mass-flow-con-
troller setting) for CD was determined to be 0.641 (relative to N2
4
at RT) by preliminary volumetric experiments, in contrast to the
ꢀ
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 0000, 00, 1 – 15
&
6
&
ÞÞ
These are not the final page numbers!

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
Figure 5. Methane TOF as function of S/C at 873 K, 10:490 mg (catalyst/
4
ꢀ1
inert), 7.91ꢁ10 h : Symbols are the experimental values with lines con-
necting experiments at constant a) p(CH ) and b) p(H O). For all data pre-
4
2
sented in this and the following figures, the constant reactant partial pres-
sures increase in the following way: partial pressures levels (in bar) were for
4
ꢀ1
CH
4
: 0.015 (*), 0.020 (!), 0.026 (&), 0.031 (^), 0.036 (~), 0.041 (ˆ), 0.046
Figure 6. Methane TOFs at 873 K, 10:490 mg (catalyst/inert), 7.91ꢁ10 h , as
a function of a) methane and b) steam partial pressure; the respective other
partial pressure is held constant, see Table 2. The measurement uncertainty
was added for the lowest four sets of partial pressure variation in both fig-
ures.
(
*), 0.055 (!), 0.065 (&), 0.075 (^), 0.084 (~), 0.093 (ˆ); and for H
2
O:
0
.020 (*), 0.026 (!), 0.032 (&), 0.037 (^), 0.043 (~), 0.049 (ˆ), 0.055 (*),
0
.061 (!), 0.066 (&), 0.072 (^), 0.078 (~), 0.084 (ˆ), 0.090 (*), 0.095 (!),
0.101 (&).
netic expression. The reported TOFs in Figures 5 and 6 are be-
tween 2.0 and 8.0 (depending on the S/C ratio and reactant
partial pressure). Our values are within the range of the report-
ed values in the literature for nickel catalysts supported on
methane and steam are in the range of 0 to 0.7 and ꢀ0.2 to
0.7, respectively. The gradual change in reaction orders was ac-
companied by the change in apparent activation energy re-
ported in Figure 3b, which increased with increasing S/C ratio.
Together these two observations suggest a change in rate-de-
termining step with changing gas composition (i.e., S/C).
Besides the variations in temperature, the support material
contributes significantly, which is reviewed in more detail by
[
44,45]
[46,47]
[2]
[48]
TiO (1.7
), g-Al O (13.6
), MgO (4.1; 2.6 ), Mg(Al)O
2
2
3
[
46,47]
[1]
[48]
[45]
[9]
(8.81
), MgAl O (1.0; 4.2 ), C (0.58 ) and YSZ (7.4 ).
2
4
A selection of data from each of the three datasets obtained
at 873, 858 and 843 K are shown in Figure 7, in which TOFs are
plotted versus partial-pressure variations. Except for an overall
lower reaction rate at the lower temperatures, the data follow
the same trends with very similar slopes for a given parameter
variation series.
[11]
Elnashaie et al. Some support materials are able to activate
steam (e.g., MgO, g-Al O ) and negative kinetic orders in
2
3
p(H O) are less likely to be observed because the buffering
2
Most of the previous investigations (see Table 1) were per-
formed at S/C ranges greater than unity, in which ranges the
results show high kinetic orders in methane and low or slightly
negative orders in steam. Empirical power rate laws with a sole
dependence on the p(CH4) satisfy these experimental
effect of the support can prevent this. In MSR, experiments
using a support with little or no steam activation (e.g.,
MgAl O , a-Al O , NiAl O , CaAl O ) or bare nickel (e.g., Ni foil,
2
4
2
3
2
4
2
4
Ni wire) do, in contrast, adsorb steam mainly on the nickel sur-
face, hence, both reactants compete for the active sites (pref-
[
2,5,8,16,18]
[10]
data.
Others, operating in a similar S/C range have in-
erably the step site ) and may therefore have kinetic rele-
[
1,3,4,7,12–15,22]
corporated steam in the rate expressions
and are
vance. The investigations presented here are consistent with
observations made at lower temperatures (873 K and lower)
and over nickel on a non-active support (in terms of steam ac-
more consistent with the observations in this study. We have
also described the presented experimental data by simple
power rate laws after dividing the data in three segments on
the S/C scale as shown in Figure S2, S3 and Table S1 in the
Supporting Information. The determined kinetic orders for
[1,3,13–15]
tivation).
Compared with those studies, the novelty of
the data presented here is the investigation of the S/C ratio
<1 regime, which clearly demonstrates the low surface cover-
ꢀ
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 0000, 00, 1 – 15
&
7
&
ÞÞ
These are not the final page numbers!

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
Figure 8. Comparison of experimental raw data and simulated thermody-
namic (TD) equilibrium values for the evolution of a) CO and b) H
2
(colour
4
ꢀ1
Figure 7. Methane TOFs, 10:490 mg (catalyst/inert), 7.91ꢁ10 h , at 873
symbols only), 858 (b) and 843 K (g) for constant-feed partial pres-
sures of (a) steam and (b) methane.
coding for constant p(CH )) at 873 K (see Figure 5), 10:490 mg (catalyst/
4
(
4
ꢀ1
inert), 7.91ꢁ10 h : raw experimental data (symbols), predicted outlet con-
centration at thermodynamic equilibrium (c), predicted outlet concentra-
tions for the WGS reaction in thermodynamic equilibrium (b).
age of water molecules in this range, leading to high kinetic
orders in steam.
[
3,5,6,13,20,22,49]
ies concluded that it was not in equilibrium.
By
13
One of the few systematic investigations in a very broad S/C
CO co-feed experiments Wei and Iglesia found the WGS re-
2
[
6]
[2]
range (0.2–5) similar to our previous work was performed by
action to be in equilibrium. Our investigations revealed, in
[
20]
13
Ross and Steel,
monotonic kinetic orders for methane (n=1) and steam (m=
0.5) throughout the S/C ratio range studied, which is in con-
over a Ni/g-Al O3 catalyst. They observed
contrast, huge differences in the C fractions in CO and CO2
2
species (Figure 10) and we therefore conclude the WGS reac-
tion not to be equilibrated. These differences may be caused
by the different catalysts studied (Ni/MgO versus Ni/NiAl O ) or
ꢀ
trast to our investigations.
2
4
different methane conversion levels under investigation. Wei
13
and Iglesia performed their experiments (for CO co-feed) at
2
WGS reaction
significantly higher methane conversions than those studied
here. Rennhack and Heinisch elaborated the influence of meth-
ane conversion and found that an equilibrium for the WGS re-
action could be expected for methane conversion levels above
Experimental outlet concentrations of CO and H at 873 K, to-
2
gether with corresponding concentrations predicted for ther-
modynamic equilibration of the WGS [see Equation (R3) and
Table 3], are shown in Figure 8. The equilibrium constant for
[21]
50% for a Ni-wire catalyst.
[
42]
WGS at 873 K is KWGS(873 K)=2.62, hence, product formation
is favoured. The thermodynamically predicted outlet concen-
trations for the products are clearly different from the ones ex-
perimentally determined (Figure 8). These data strongly sug-
gest that the WGS reaction is not at equilibrium under the
conditions studied here. Both H and CO (data not shown for
Isotopic labelling and kinetic isotope effect
Isotopic labelling experiments were performed at three differ-
ent S/C ratios under the same (differential) conditions as in the
experiments described above. The data are reported in Table 4.
Methane conversions (i.e., CH and CD , respectively) were de-
2
2
CO ) concentrations are expected higher for an equilibrated
2
4
4
WGS reaction with a correspondingly lower amount of CO.
It has been reported that the WGS reaction was in equilibri-
termined for CH /H O and CD /H O feed, respectively. The ratio
4 2 4 2
of conversions during labelled to unlabelled feed (x(CD )/
4
[
1,2,15,16,18]
um in some studies of Ni-based MSR,
but other stud-
x(CH )) is called the kinetic isotope effect (KIE). The observed
4
ꢀ
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 0000, 00, 1 – 15
&
8
&
ÞÞ
These are not the final page numbers!

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
shown observed for CD alone and for the CD /H O
4
4
2
reactant mixture fed during the KIE experiments, re-
spectively.
Normal KIEs are assumed to be caused by a
[
50,51]
CꢀH(D) bond cleavage
(primary effect). Thus, the
isotopic labelling data are consistent with the obser-
vations from the partial pressure variation data pre-
sented in Figure 5–7, in which a linear, positive de-
pendence of methane partial pressure was observed
at high S/C ratios. A normal KIE has been widely ob-
served for methane conversion in experimental work:
[52]
1
.08–1.02 (dry reforming ), 1.97–1.19 (dry reform-
[52]
[53]
ing ), 1.07–1.91 (dry reforming ), 1.45 (dry reform-
[26]
[54]
ing ), 1.09–1.48 (oxidative coupling ), 1.62:1.66
[2]
(
dry/steam reforming, 873 K ) and from calculations
[
55]
[56]
[57]
1
.98 (873 K ), 2.87 (1000 K ), 1.59 (800 K ) and
[
58]
2
.44 (800 K ). Our observed normal KIE (at high S/C)
is similar to those reported in Ref. [2,26] and consis-
tent with the absence of H/D exchange in the meth-
[20]
ane species. Thus, a kinetic relevance of the reac-
[1]
tion between adsorbed carbon and oxygen is ex-
cluded.
An inverse KIE (secondary effect), on the other
hand, may have several origins: for example, close/
[
59]
[60]
tight transition state (TS), agostic interaction, re-
[61]
action mechanism (S 2 versus E2), or hybridization
N
[62]
change in the transition state. All but the last pos-
sibility are excluded from further discussion because
of the small size of the molecules involved and limit-
ed selection of nucleophiles (O*/OH*). Similarities to
our reported inverse KIE are found, possibly owing to
the close correlation in mechanism, with results in
4
ꢀ1
Figure 9. Isotopic labelling at 873 K, 10:490 mg (catalyst/inert), 7.91ꢁ10 h : a) CH
CD conversions at CH /CD +H O switching experiments, b) GC–MS analysis, H/D mixing
in methane during CD /H O feed at 873 K.
4
and
4
4
4
2
4
2
[63–66]
the literature on methanation
and Fischer–
[62]
Tropsch synthesis. Loc et al. observed both inverse
and normal KIEs for the methanation reaction de-
[67]
pending on the support applied.
The range of KIEs from inverse to normal at low
and high S/Cs, respectively, clearly indicates the in-
volvement of CꢀH moieties in the rate determining
step(s) (otherwise no KIE would have been observed).
13
The results from the CO co-feed experiments are
2
shown in Figure 10. For each S/C ratio the amount of
13
co-fed CO was similar to the amount of CO gener-
2
2
ated by the MSR reaction in non-isotopic labelling ex-
13
periments. The C fraction in CO was >46%, which
2
is much higher than that of <7% in CO. This obser-
vation suggests that the WGR (R3), is not in equilibri-
um at our experimental conditions, and is, therefore,
in agreement with the conclusion drawn from the
comparison between experimental data and thermo-
dynamic calculations made in Figure 8.
1
3
Figure 10. CO
catalyst/inert), 7.91ꢁ10 h ; peak areas of GC–MS analysis are normalized with respect
to the strongest m/z peak of CO and CO , respectively.
2
co-feed experiment at three different S/C ratios at 873 K, 10:490 mg
4
ꢀ1
(
2
Notably, the fragmentation patterns of methane
were similar irrespective of the S/C ratio, strongly in-
dicating that possible back-formation of methane from inter-
KIEs are further illustrated in Figure 9a. For an S/C ratio=0.5,
a slightly inverse KIE of 0.9 is observed. With increasing S/C
ratio, the KIE increases through KIE=1.1 for S/C=1.4 to 1.3 for
S/C=3.6. In Figure 9b, the methane fragmentation pattern is
mediates and products, giving CH D , was negligible for all
x
4ꢀx
S/C ratios and conversions (up to 7%) investigated. The ab-
sence of H/D exchange during CD /H O feed was interpreted
4
2
ꢀ
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 0000, 00, 1 – 15
&
9
&
ÞÞ
These are not the final page numbers!

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
These observations suggest that the reverse reaction of
MSR, as well as surface coverage of product-derived species
may be neglected during reaction modelling.
Table 4. Results of the isotopic labelling experiments at 873 K,
4
ꢀ1
1
0:490 mg (catalyst/inert), 7.91ꢁ10 h and different S/C ratios: mean
1
3
4 2
CH conversions for KIE and CO co-feed.
*
The KIE shifted from slightly inverse at low S/C ratios to
1
3
S/C
feed)
x(CH
4
) [%]
x(CD
4
) [%]
KIE
x(CH
CO
2
co-feed
positive values at high S/C ratios (Table 4) via approximately
zero at S/C ratios close to 1. Furthermore, the apparent activa-
tion energy increased with increasing S/C ratios (Figure 3).
These effects suggest that the rate-limiting step changes
with changing S/C ratios and that CꢀH bonds were involved in
the rate-limiting step under all conditions. For the normal KIE,
the CꢀH bond is cleaved in the rate-limiting step. This might
occur during a simple CꢀH cleavage (CH *!CH *+H*; x=
(
4
)/x(CD
4
)
x(CH ) [%]
4
0.5
1.4
3.6
3.5ꢁ0.36
6.5ꢁ0.37
8.6ꢁ0.20
4.0ꢁ0.21
6.2ꢁ0.29
6.8ꢁ0.47
0.9
1.1
1.3
3.6
6.6
8.2
to be likely a result of the low methane (i.e., CH and CD ) con-
4
4
4ꢀx
3ꢀx
version and, therefore, reverse reactions are relatively scarce
0, 1, 2, 3) or an oxidation step (e.g., CH4 *+O*!CH O*+
ꢀx
3ꢀx
[
6]
compared to forward reactions, which was similarly conclud-
ed in Ref. [53]. This observation adds to several other observa-
tions reported above, that is, that the extent of reaction [h1
H*; x=0, 1, 2, 3). For the case of an inverse KIE, the CꢀH bond
has to stay intact during the rate-limiting step, which could be
an oxidation step (e.g., CH *+O*!CH O*; y=0, 1, 2)
3ꢀy
3ꢀy
and h for reactions (R1) and (R2), respectively, see Experimen-
*
The WGS reaction was not in equilibrium under the
2
tal Section] were found insignificant, and that addition of prod-
ucts to the feed did not influence the rate of methane conver-
sion (see Experimental Section). Together, these results suggest
that the rates reported in this study are differential in nature,
and that product concentrations do not need to be taken into
account when developing a rate expression for the reaction.
Theoretical investigations provide similar activation energies
conditions tested, as verified by thermodynamic calculations as
well as isotopic labelling experiments (Figure 8 and 10, respec-
tively), which suggests that the reactions leading to methane-
and steam-derived adsorbate species had rate constants within
the same order of magnitude allowing for a change in rds with
changing S/C ratio.
*
All the above observations are consistent with the gen-
for the elementary steps for CH * oxidation and methana-
eral reaction scheme presented in Equations (1)–(15). The oxi-
dation mechanism in Equations (10)–(14), supports an inverse
KIE at low S/C ratios (see Figure S4).
x
[
1,30]
tion,
hence, the large excess of reactants (compared to
products) will drive the reaction to the product side.
Thus, more than a single oxidation pathway needs to be
[3]
considered. The established Scheme by Xu and Froment
offers several oxidation possibilities (two of which are kinetical-
ly relevant) that can explain a normal KIE. However, their inves-
tigations were performed at high S/C ratios, and thus, their
model fails to explain the inverse KIE at low S/C ratios (which
Mechanism
The main findings from this experimental data may be sum-
marised as follows:
[45]
was observed in Figure 9a). Both Bradford and Vannice and
*
The MSR reaction may be divided into two kinetic re-
Xu and Froment studies suggest CH O* species is most abun-
x
gimes; one at S/C ratios below 1, and another one at S/C
dant species, however, also the CH dissociation was assessed
4
ratios above 1. For S/C<1, the methane conversion rate had
to have a high reaction barrier, leading to a normal KIE. Blay-
[24,30]
a positive reaction order in p(H O). For S/C>1, the reaction
lock et al.
suggested a pathway that would be able to de-
2
order in p(H O) was either zero or slightly negative. Conversely,
scribe an inverse KIE. The key steps from the investiga-
2
[3,24,30]
at S/C>1, the reaction order in methane was positive, but at
S/C<1 it approached zero (Figure 5–7).
tions
are the oxidation of a CH * species by O* (also OH*
x
is possible), which is illustrated in Figure S4. During such an ox-
*
With total reactant pressure in the range of 0.037 to
idation, the hybridization of the CH * (x=1 or 2) changes from
x
2
2
3
0
.224 bar, the maximum methane conversion rate was ob-
sp to sp for x=1 (sp to sp , for x=2) and an inverse KIE of
2
3 [51]
served at an S/C ratio of 1 (Figure 5).
0.8–0.9 would be expected (sp to sp ), which matches the
observed KIE of 0.9 at low S/C ratios in this study. At high S/C
ratios, on the other hand, Scheme (B) in Figure S4 exemplifies
a reaction pathway for the surface reaction during an attack of
a surface oxygen (or hydroxyl) where a hydrogen atom is liber-
ated, causing a normal KIE (of 1.3), in agreement with the liter-
ature reports cited above.
Together, these observations suggest that the coverage in
surface species was low in the range of conditions studied (i.e.,
S/C=0.2–7.2), and that the concentration of surface species
derived from methane and steam became rate-limiting for S/C
ratios higher or lower than 1, respectively.
*
The extent of reaction of methane and steam [(R1) and
R2)] was low (see above), and back formation of methane as
observed by CD /H O co-feed experiments was negligible
(
The observation of both an inverse and a normal KIE in the
same reaction system by varying conditions requires a model
with a shift in the rate-determining step. Similar observation of
4
2
under the conditions tested (Figure 9), suggesting that these
reactions were far from thermodynamic equilibrium and con-
sidered irreversible. Addition of products (H , CO or CO , re-
[
54]
a KIE change with S/C were reported earlier by Shi et al.
[
53]
during oxidative coupling of methane and Cant et al. during
dry reforming. We thus suggest two reaction pathways, one
2
2
ported in Figure 4a, b and Table 4, respectively) to the feed
did not alter the rate of methane conversion.
[24,30]
[3]
for low
and one for high S/C and interpret the transition
ꢀ
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 0000, 00, 1 – 15
&10
&
ÞÞ
These are not the final page numbers!

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
kinetic region as a crossover from one region to the other
without a distinct rate-controlling kinetic step. The different ki-
netic regions are presented in Table S1.
Kinetic modelling
As a first approach for developing a model describing the ob-
servations and analyses presented above, a steady-state model
was derived from a simplified reaction scheme for MSR in-
[
12]
spired by Khomenko et al.:
k1
CH þ S G H CꢀS þ 2 H2
ð27Þ
ð28Þ
ð29Þ
ð30Þ
4
k2
CꢀS þ H O G H CO þ S þ H2
2
k3
H O þ S G H OꢀS þ H2
2
k4
CO þ OꢀS G H CO þ S
2
in which S is the active site and: S+OS+CS=1.
Re-formation of methane from carbon containing surface
species was negligible. Furthermore, no influence of the meth-
ane conversion rate was observed during addition of CO and
H , respectively. This suggests, that reactions (27) and (28) are
2
irreversible, and that the surface coverage in H and CO is negli-
gible. In addition, it was observed that the WGS reaction
[Eqs. (29) and (30)], was not at equilibrium under the condi-
tions tested in this study and points to the steady-state ap-
proach as the most suitable for the description of the current
reaction system, giving the following rate expressions:
Figure 11. Steady-state model inspired by the reaction scheme by Khomen-
ko et al. The TOFs are plotted as a function of a) methane and b) steam
[12]
partial pressure, respectively: experimental data points (symbols) and simu-
ꢀ
d½CH ꢃ=dt ¼ k ꢅ ½CH ꢃ ꢅ ½Sꢃ
ð31Þ
ð32Þ
lated data (c).
4
1
4
d½CSꢃ=dt ¼ k ꢅ ½CH ꢃ ꢅ ½Sꢃꢀk ꢅ ½CSꢃ ꢅ ½H Oꢃ ¼ 0
1
4
2
2
d½OSꢃ=dt¼ k ꢅ ½H Oꢃ ꢅ ½Sꢃꢀk ꢅ ½H ꢃ½OSꢃꢀk ꢅ ½COꢃ
3
2
ꢀ3
2
4
ment with the actual values. The sudden change in the trend
of the TOF as a function of (one) reactant partial pressure, that
is, increase followed by a plateau with increasing reactant par-
tial pressure, is not observed for the simulated values. A
reason for that might be found in the applied rate expression.
ꢅ½OSꢃ þ kꢀ4 ꢅ ½CO ꢃ ꢅ ½Sꢃ
ð33Þ
2
¼
0
Combination of Equations (31)–(33) yields the final rate
expression for methane conversion:
Note that the value of k has no influence on the simulation
3
ꢀ
1
result and was only applied with 9.88 s because of the preset
ꢀ
d[CH ]/dt=k [CH ]/(1+A+B), in which:
condition (k >k ).
4
1
4
3
1
A=(k /k )·[CH ]/[H O], and
In a second approach, a more complex model was devel-
oped and investigated, incorporating dissociative adsorption
1
2
4
2
B=(k [H O]+k [CO ])/(k [CO]+k [H2])
3
2
ꢀ4
2
4
ꢀ3
The model parameters were fitted with experimental data
using an iterative procedure (generalized reduction gradient
nonlinear algorithm). Literature reports suggest WGS reaction
of steam prior to CH oxidation. Such features in a model have
x
previously been suggested by Xu and Froment, based on
Equations (1)–(4) and (10)–(15) above, and lead to the rate ex-
pressions presented in Equation (34) and (35). Methane conver-
sion was represented by the sum of the methane conversions
[
Eqs. (29) and (30)] to be much faster than methane activation
[
21]
step over Ni-based catalysts [Eq. (27)]. Hence, rate constants
k >k constraint is applied to the model. Furthermore, k was
of reactions (R1) and (R2) to CO and CO , respectively.
3
1
ꢀ4
2
fixed to: k 4 =k k /[k K(WGS)], where K(WGS) is the equilibri-
ꢀ
3
4
ꢀ3
[42]
!
um constant for the WGS reaction, K(WGS)873 K =2.62.
The
3
H2
k
p pCO
ꢀ
1
1
:5
2
2
rate constants fitted to the model are: k =9.88 s , k =
r1 ¼ p2 pCH4
p
4
H O
ꢀ
ꢀ
=ðDENÞ
=ðDENÞ
ð34Þ
ð35Þ
1
2
K
ꢀ
1
ꢀ1
5
ꢀ1
6
ꢀ1
H2
1
1
4
4.1 s , k =9.88 s , k =7.14·10 s , k =2.51·10 s , k
=
3
ꢀ3
4
ꢀ4
ꢀ
1
.58 s .
!
Simulated data are presented together with experimental
3
k1
p pCO
H2
data points in Figure 11. Clearly, the model describes the
trends of the TOF qualitatively correctly and in good agree-
r3 ¼ p
pCH4 pH4O
3:5
H2
K
3
ꢀ
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 0000, 00, 1 – 15
&
11
&
ÞÞ
These are not the final page numbers!

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
DEN ¼ 1 þ KCOpCO þ K p þ KCH4 pCH4 þ KH2OpH2O=pH2
ð36Þ
second-order rate expression failed to describe the experiment,
whereas the first-order expression at least showed correlations
at higher methane partial pressures. In conclusion, both first-
and second-order rate expression are required simultaneously
to represent the experiments as shown in Figure 12b. An aver-
H2 H2
A large negative kinetic influence of hydrogen was part of
their model with kinetic orders of ꢀ2.5 and ꢀ3.5 for the path
along Reaction (R1) and (R2), respectively. Products of elemen-
tary steps appear with a negative exponent in Xu and Fro-
ment’s model. For example, hydrogen is a product of methane
ꢀ1
age deviation of the TOF of approximately 0.3 s per data
point was achieved, which is considered to be a reasonable
result. All but the lowest methane partial pressure sequence
(black colour) were represented well.
and water dehydrogenation steps to reach CH * and O*, re-
x
spectively. The exponents for hydrogen partial pressure range
therefore between ꢀ0.5 and ꢀ3 (this was explained in the
An issue related to our derivation of a Langmuir–Hinshel-
wood rate expression from elementary steps mechanism is Xu
[
3]
model by Xu and Froment. ) Our experimental data yielded
[3]
a zero-order influence of H partial pressure (see Figure 4a),
and Froment’s negative kinetic contribution of hydrogen par-
tial pressure, which also was derived from an elementary step
mechanism. We believe the key to that answer may be found
in the co-fed hydrogen to the reaction mixture. In Figure 4a,
the independence of the hydrogen partial pressure on the
methane conversion was confirmed under differential condi-
tions. The co-fed hydrogen, however, never exceeded that of
methane, as in the case of Xu and Froment (H /CH =1.25).
2
and Xu and Froment’s model is, therefore, not applicable to
the full range of our data.
We investigated several simplified reaction pathways (see
rate expressions S1–S5 and Figure S5), all of which have a CHx
intermediate in common, thus leaving the option to interpret
the observed normal KIE originating from the CꢀH bond cleav-
age as a result of CH dissociation or CH * oxidation, respec-
4
x
2
4
tively. Furthermore, the oxidation of CH * to a CH O* species
Furthermore, during the activation (i.e., reduction) of the cata-
lyst, they employed pure hydrogen over a period of 12 h,
x
x
could satisfy the inverse KIE. Eventually, the reaction scheme
involving a partial equilibrium between reduced and oxidised
surface species [see pathway in Figure S5e] was chosen, from
which the final Langmuir–Hinshelwood rate expression was de-
rived.
whereas in this work 10 vol% H was applied for 10 min. High
2
concentrations and exposure times enhance the dissolution of
hydrogen in nickel. The presence of hydrogen in the nickel
particles and additionally in the gas phase could be reasons
for the negative contributions of hydrogen found by Xu and
Froment.
We assumed that adsorbed methane and steam on the sur-
face is in equilibrium with their gas-phase species:
As a concluding remark, the models derived in this study
represent the experimental data fairly well. However, both
models represent simplifications of the complete reaction
scheme. In particular, the first model does not take into ac-
count the dissociative adsorption of water and the second
model does not include the WGS reaction. For a broader view
and a unifying kinetic description of the MSR reaction for
a larger range of experimental conditions, a microkinetic ap-
proach would be required. This task is currently in progress
and intended to be a follow-up of this manuscript.
CH þ * Ð CH *
ð37Þ
ð38Þ
4
4
H O þ * Ð H O*
2
2
Irreversible dissociation of both reactants in one step:
CH * ! CH * þ x=2 H
ð39Þ
ð40Þ
4
x
2
H O* ! O* þ H2
2
Reversible oxidation to CH O* and irreversible oxidation to
x
CO* and CO *:
2
Conclusion
CH * þ O* Ð CH O* þ *
ð41Þ
ð42Þ
ð43Þ
x
x
k1
An extensive dataset was presented for the steam reforming
reaction at 843, 585 and 873 K over a Ni/NiAl O catalyst. De-
CH * þ O* ꢄ! CO* þ x=2 H þ *
x
2
2
4
k2
CH O* þ O* ꢄ! CO * þ x=2 H þ *
x
2
2
pending on the steam-to-carbon (S/C) ratio in the feed and
total reactant partial pressure, both reactants were kinetically
relevant and did show changes in the kinetic order (0–0.7 for
methane and ꢀ0.2–0.7 for steam). From the dataset, three dif-
ferent kinetic regimes were proposed; 1) at low, 2) at inter-
The final rate expression including the denominator, which
incorporates solely the adsorption of the reactants (no prod-
ucts), is then given by Equation (44) below in which k and K
i
j
13
are the kinetic and adsorption constants, respectively:
mediate, and 3) at high S/C ratios. From CO co-feed experi-
2
ments and thermodynamic modelling, the water gas shift reac-
tion was found to be not equilibrated under the conditions ap-
plied. Kinetic isotope effects (KIEs), determined by CH /H O
2
2
H O
k1KCH4 KH2OpCH4 pH2O þ k K
K
pCH4
p
2
CH4 H O
2
2
TOF ¼
ð44Þ
2
ð1 þ KCH4 pCH4 þ KH2OpH2O
Þ
4
2
and CD /H O feed switching at steady state, revealed three dif-
4
2
Comparison of the model with the experimental data is
shown in Figure 12. For analyses and evaluation purposes the
rate expressions containing only the first and second order de-
pendencies are presented separately in Figure 12a and c, re-
ferent KIEs: inverse, no, and normal at low, intermediate, and
high S/C ratios, respectively. These observations indicate a tran-
sition between two rate-determining steps depending on the
conditions. At high S/C ratios, a normal KIE was observed and
the rate-determining step should, therefore, involve a CꢀH
spectively, and significant differences were observed.
A
ꢀ
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 0000, 00, 1 – 15
&12
&
ÞÞ
These are not the final page numbers!

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
Figure 12. Langmuir–Hinshelwood model from Equation (44) is compared to the experimental data represented as TOF and methane conversion, respectively.
ꢀ
1
ꢀ1
ꢀ1
ꢀ1
The determined kinetic and equilibrium constants k
16.65, 1734.12, respectively.
1 2 4 2
[s ], k [s ], K(CH ) [bar ] and K(H O) [bar ] are a) 46.0, –, 18.58, 19.98; b) 11, 0.56, 230, 426; c) –, 0.29,
5
Nielsen, I. Chorkendorff, J. Sehested, J. K. Nørskov, J. Catal. 2008, 259,
47–160.
bond cleavage, whereas at low S/C ratios an inverse KIE was
observed suggesting that the rate-determining step involves
a hybridization change. The nature of the involved CH species
might be the same, but they might react through different
routes (as a result of, among others, different surface coverag-
es, restructuring etc.). Two models were developed: 1) a
steady-state model that represents observed trends qualitative-
ly correctly and 2) a Langmuir–Hinshelwood model that de-
scribes the dataset satisfactorily, including the observed
changes in kinetic orders of the reactants.
1
[
2] J. Wei, E. Iglesia, J. Catal. 2004, 224, 370–383.
3] J. Xu, G. F. Froment, AIChE J. 1989, 35, 88–96.
[
[4] I. M. Bodrov, L. O. Apelbaum, M. I. Temkin, Kinet. Catal. 1964, 5, 696–
05.
7
[
[
[
[
5] W. W. Akers, D. P. Camp, AIChE J. 1955, 1, 471–475.
6] C. Sprung, B. Arstad, U. Olsbye, Top. Catal. 2011, 54, 1063–1069.
7] A. Al-Ubaid, E. E. Wolf, Appl. Catal. 1988, 40, 73–85.
8] I. M. Bodrov, L. O. Apelbaum, M. I. Temkin, Kinet. Catal. 1967, 8, 821–
828.
[
9] E. Nikolla, J. Schwank, S. Linic, J. Catal. 2009, 263, 220–227.
[
[
[
10] H. S. Bengaard, J. K. Nørskov, J. Sehested, B. S. Clausen, L. P. Nielsen,
A. M. Molenbroek, J. R. Rostrup-Nielsen, J. Catal. 2002, 209, 365–384.
11] S. S. E. H. Elnashaie, A. S. Al-Ubaid, M. A. Soliman, A. M. Adris, J. Eng. Sci.
Acknowledgements
1
988, 14, 247–273.
12] A. A. Khomenko, L. O. Apelbaum, F. S. Schub, Y. S. Snagovskii, M. I.
Temkin, Kinet. Catal. 1971, 12, 423–430.
The Norwegian Research Council, through the RENERGI project
[
[
13] K. Hou, R. Hughes, Chem. Eng. J. 2001, 82, 311–328.
14] J. C. De Deken, E. F. Devos, G. F. Froment in Chemical Reaction Engineer-
ing Boston, chap. 17, pp. 181–197.
“
Hydrogen Production by Sorbent Enhanced Reforming”,
P 178129/S30, is acknowledged for financial support.
[
[
[
15] D. W. Allen, E. R. Gerhard, M. R. Likins, Jr., Ind. Eng. Chem. Process Des.
Dev. 1975, 14, 256–259.
Keywords: carbon · hydrogen · isotopic labeling · kinetics ·
nickel
16] A. K. Avetisov, J. R. Rostrup-Nielsen, V. L. Kuchaev, J. Bak Hansen, A. G.
Zyskin, E. N. Shapatina, J. Mol. Catal. A 2010, 315, 155–162.
17] S. S. E. H. Elnashaie, A. M. Adris, A. S. Al-Ubaid, M. A. Soliman, Chem. Eng.
Sci. 1990, 45, 491–501.
[1] G. Jones, J. G. Jakobsen, S. S. Shim, J. Kleis, M. P. Andersson, J. Rossmeisl,
F. Abild-Pedersen, T. Bligaard, S. Helveg, B. Hinnemann, J. R. Rostrup-
ꢀ
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 0000, 00, 1 – 15
&13
&
ÞÞ
These are not the final page numbers!

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
[
18] I. M. Bodrov, L. O. Apelbaum, M. I. Temkin, Kinet. Catal. 1968, 9, 1065–
[46] D. Li, T. Shishido, Y. Oumi, T. Sano, K. Takehira, Appl. Catal. A 2007, 332,
98–109.
1071.
[
[
19] A. Al-Ubaid, E. E. Wolf, Appl. Catal. 1987, 34, 119–134.
20] J. R. H. Ross, M. C. F. Steel, J. Chem. Soc. Faraday Trans. 1 1973, 69, 10–
[47] D. Li, I. Atake, T. Shishido, Y. Oumi, T. Sano, K. Takehira, J. Catal. 2007,
250, 299–312.
2
0.
21] R. Rennhack, R. Heinisch, Erdoel Kohle Erdgas Petrochem. 1972, 25, 22–
9.
[48] J. R. Rostrup-Nielsen, J. H. B. Hansen, J. Catal. 1993, 144, 38–49.
[49] L. Maier, B. Schꢂdel, K. Herrera Delgado, S. Tischer, O. Deutschmann,
Top. Catal. 2011, 54, 845.
[
[
2
22] M. A. Soliman, A. M. Adris, A. S. Al-Ubaid, S. S. E. El-Nashaie, J. Chem.
Technol. Biotechnol. 1992, 55, 131–138.
[50] M. Wolfsberg, W. A. Van Hook, P. Paneth, L. P. N. Rebelo in Isotope Ef-
fects, Springer 2010, chap. 10, pp. 313–342.
[
[
[
23] L. M. Aparicio, J. Catal. 1997, 165, 262–274.
[51] E. V. Anslyn, D. A. Dougherty in Modern Physical Organic Chamistry, Uni-
versicy Science Books, Sausalito, CA 2006, chap. 8, pp. 421–431.
[52] Z. Zhang, X. E. Verykios, Catal. Lett. 1996, 38, 175–179.
[53] N. W. Cant, R. Dꢃmpelmann, A. M. Maitra, Stud. Surf. Sci. Catal. 1997,
107, 491–496.
24] D. W. Blaylock, Y.-A. Zhu, W. H. Green, Top. Catal. 2011, 54, 828.
25] R. M. Watwe, H. S. Bengaard, J. R. Rostrup-Nielsen, J. A. Dumesic, J. K.
Nørskov, J. Catal. 2000, 189, 16.
[
26] T. Osaki, T. Horiuchi, K. Suzuki, T. Mori, Catal. Lett. 1997, 44, 19–21.
27] Y. S. Snagovskii, A. A. Khomenko, R. A. Ivanova, F. S. Schub, M. I. Temkin,
Kinet. Catal. 1978, 19, 942–949.
[
[54] C. Shi, M. Xu, M. P. Rosynek, J. H. Lunsford, J. Phys. Chem. 1993, 97,
216–222.
[
28] T. Osaki, H. Fukaya, T. Horiuchi, K. Suzuki, T. Mori, J. Catal. 1998, 180,
[55] O. Roberto-Neto, E. L. CoitiÇo, D. G. Truhlar, J. Phys. Chem. A 1998, 102,
4568–4578.
106–109.
[
[
29] Y. Zhu, D. Chen, X. Zhou, W. Yuan, Catal. Today 2009, 148, 260–267.
30] D. W. Blaylock, T. Ogura, W. H. Green, G. J. O. Beran, J. Phys. Chem. C
[56] H. Burghgraef, A. P. J. Jansen, R. A. Van Santen, Faraday Discuss. 1993,
96, 337–347.
2
009, 113, 4898–4908.
31] A. L. A. Lopez Ortiz, D. P. Harrison, Ind. Eng. Chem. Res. 2001, 40, 5102–
109.
[57] O. Roberto-Neto, F. B. C. MacHado, F. R. Ornellas, Chem. Phys. Lett. 2005,
409, 38–42.
[58] S. T. Banks, D. C. Clary, Phys. Chem. Chem. Phys. 2007, 9, 933–943.
[59] R. A. Poirier, Y. Wang, K. C. Westaway, J. Am. Chem. Soc. 1994, 116,
2526–2533.
[
5
[
[
32] D. P. Harrison, Ind. Eng. Chem. Res. 2008, 47, 6486–6501.
33] N. Hildenbrand, J. Readman, I. M. Dahl, R. R. Blom, Appl. Catal. A 2006,
3
03, 131–137.
[60] M. J. Tanner, M. Brookhart, J. M. DeSimone, J. Am. Chem. Soc. 1997, 119,
7617–7618.
[61] S. M. Villano, S. Kato, V. M. Bierbaum, J. Am. Chem. Soc. 2006, 128, 736–
737.
[62] B. Shi, C. Jin, Appl. Catal. A 2011, 393, 178–183.
[63] T. Mori, H. Masuda, H. Imai, A. Miyamoto, S. Baba, Y. Murakami, Chem.
Lett. 1981, 10, 831–834.
[64] M. Agnelli, H. M. Swaan, C. Marquez-Alvarez, G. A. Martin, C. Mirodatos,
J. Catal. 1998, 175, 117–128.
[
[
34] J. Rostrup-Nielsen, L. J. Christiansen, Concepts of syngas manufactore,
Vol. 10, Imperial College Press, catalytic science series ed., 2011.
35] C. Sprung, Kinetic investigation in methane steam reforming over a nickel
catalyst—Thesis, University of Oslo, 2012.
36] J. Smith, P. Thrower, M. Vannice, J. Catal. 1981, 68, 270–285.
37] C. Crisafulli, S. Scira, S. Minica, L. Solarino, Appl. Catal. A 2002, 225, 1–9.
38] K. Ione, V. Romannikov, A. Davydov, L. Orlova, J. Catal. 1979, 57, 126–
[
[
[
1
35.
[
[
39] V. D. Stytsenko, Appl. Catal. A 1995, 126, 1–26.
40] G. Ertl, Knçzinger, J. Weitkamp, in Handbook of heterogeneous catalysis,
Vol. III, VCH, Weinheim, 1997, pp. 1376–1386.
41] M. A. Vannice, Kinetics of Catalytic Reactions, Springer, 2005.
42] A. Roine, Outokumpu hsc chemistry for windows Outokumpu Research
Oy, FIN - 28101 PORI.
[65] T. Mori, H. Masuda, H. Imai, A. Miyamoto, S. Baba, Y. Murakami, J. Phys.
Chem. 1982, 86, 2753–2760.
[66] Z. A. Ibraeva, N. V. Nekrasov, V. I. Yakerson, B. S. Gudkov, E. Z. Golosman,
Z. T. Beisembaeva, S. L. Kiperman, Kinet. Catal. 1987, 28, 386–391.
[67] L. Loc, N. Huan, N. Gaidai, H. Thoang, N. Nekrasov, Y. Agafonov, A. Lapi-
dus, Kinet. Catal. 2011, 52, 749–755.
[
[
[43] Brooks instruments, gas conversion list, source j-836-d-508 rev. 14, In-
stallation and operating manual.
[
[
44] M. C. J. Bradford, M. A. Vannice, Appl. Catal. A 1996, 142, 73–96.
45] M. C. J. Bradford, M. A. Vannice, Appl. Catal. A 1996, 142, 97–122.
Received: February 6, 2014
Published online on && &&, 0000
ꢀ
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 0000, 00, 1 – 15
&14
&
ÞÞ
These are not the final page numbers!

FULL PAPERS
Kinetics change: A detailed investiga-
tion of methane steam reforming over
a nickel catalyst is performed at steam-
to-carbon (S/C) ratios of 0.2–7.1 at 843,
C. Sprung, B. Arstad, U. Olsbye*
& – &&
&
Methane Steam Reforming over a Ni/
NiAl O Model Catalyst—Kinetics
858, and 873 K. Kinetic orders (red for
2
4
high and blue for low values) of the re-
actants change as a function of S/C. The
background color represents steam and
the color of the symbols methane kinet-
ic order. The actual position of the sym-
bols represents the steam conversion
rate as a function of S/C.
ꢀ
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 0000, 00, 1 – 15
&15
&
ÞÞ
These are not the final page numbers!