SIMPLIFIED KINETIC MODEL FOR STEAM REFORMING OF ETHANOL ON A Ni/Al2 O3

CATALYST
Y.-J. Wu,1 J.C. Santos,1 Ping Li,2 J.-G. Yu,2 A.F. Cunha1 and A.E. Rodrigues1 *
1. Laboratory of Separation and Reaction Engineering (LSRE), Associated Laboratory LSRE/LCM, Faculty of Engineering, Department of
Chemical Engineering, University of Porto, Rua Dr. Roberto Frias s/n, 4200-465 Porto, Portugal
2. State Key Laboratory of Chemical Engineering, College of Chemical Engineering, East China University of Science and Technology (ECUST),
Shangai 200237, China

Ethanol steam reforming is a promising method for hydrogen production. Kinetic studies are carried out over a nickel-based catalyst from 200
to 600 C. A simplified LangmuirHinshelwoodHougenWatson (LHHW) kinetic model was proposed with nine parameters, where the surface
decomposition of methane is assumed as the rate determining step (RDS), and while all other reaction steps are set as reversible. It is postulated
that the velocity of the backward reaction step of the dissociative methane adsorption CH4 + 2 )  ) CH3 + ) H is similar to all other equilibrium
reactions proposed, while the forward reaction step is set as the RDS. In addition, the formation of adsorbed carbon species is excluded due to the
fact that SRE experiments were performed with excess water, suggesting that OH species, excluding O species formation, stick better on active
Ni surface than methane. In addition, a power law kinetic model was used for analysis; the activation energy was 31.8 kJ/mol, and the reaction
order of ethanol pressure was 1.52. Experimental results were successfully demonstrated by both kinetic models. The purpose of this work was to
develop a simplified mechanistic model to replace empirical models (power rate model) and other proposed kinetic models that are too complex
for an applied kinetic process. It was found that the simplified LHHW kinetic model has a better fit than that of the power law model. The proposed
kinetic model works well over a wide temperature range (200600 C).
Keywords: hydrogen, ethanol steam reforming, kinetic studies, kinetic model

INTRODUCTION

[1,2]

ydrogen is the energy carrier of the future.

It is used
in fuel cells for energy generation.[3,4] Currently, steam
reforming (SR) of natural gas and light oil fractions is
widely used to produce hydrogen.[5,6]
Renewable resources are an alternative for the production
of hydrogen via SR. Ethanol is among the candidates for a
good choice, because it can be produced in large amounts by
the fermentation of biomass (C12 H22 O11 (sacarose) + H2 O
4CH3 CH2 OH + 4CO2 ).[7,8] In addition, ethanol can be directly converted into hydrogen by steam reforming (SR), the so-called steam
reforming of ethanol (SRE):
CH3 CH2 OH + 3H2 O  2CO2 + 6H2

(1)
KINETIC MODELS
[911]

it is
Since research efforts have primarily focused on SRE,
clear that several reaction pathways are possible in SRE (Figure 1),
depending on the catalysts used.[10] The efforts are mainly focused
on Co, Cu, Pt and Ni based catalysts.[912] As typical transition
metals, the active outer layer electrons and associated valence
states make it possible for these metals to be candidates for SRE. It
is worth noting that the conversion of the reactants used, ethanol
and water and the product distribution are strongly dependent
on the nature of catalysts employed (active phase, promoters and
support material). Therefore, it is necessary to carry out a comprehensive kinetic study on SRE to understand the nature of the
reaction.
In this work, SRE reaction is studied in a large range of temperature (200600 C) and space time (0.0086.731 s) in which
the ethanol conversion goes from 0% to 100%. The goal is to

116

THE CANADIAN JOURNAL OF CHEMICAL ENGINEERING

prove that the established simplified kinetic model is convenient

to describe the kinetic behaviour of the catalyst used in SRE.
The novelty of this work is that a simplified Langmuir
HinshelwoodHougenWatson (LHHW) kinetic model is proposed with nine parameters, where the surface decomposition of
methane is assumed as the rate determining step (RDS) while all
other reaction steps are set as reversible. To the best of our knowledge, there is no literature report assuming that the dissociative
methane adsorption is the RDS in SRE. The purpose of this work
was to develop a simplified mechanistic model to move away from
empirical models (power rate model), but also to move away from
other proposed models that are too complex for an applied kinetic
process.

The kinetic behaviour of the SRE reaction has been studied over
various catalysts in recent years to describe the reaction rate using
a power law rate expression[11,1322] or an equation based on certain mechanistic assumptions.[13,2329] However, these reports on
kinetics show significant differences. This might be related with

Additional supporting information may be found in the online version of

this article.
Author to whom correspondence may be addressed.
E-mail address: arodrig@fe.up.pt
Can. J. Chem. Eng. 92:116130, 2014
2013 Canadian Society for Chemical Engineering
DOI 10.1002/cjce.21773
Published online 18 April 2013 in Wiley Online Library
(wileyonlinelibrary.com).

VOLUME 92, JANUARY 2014

Figure 1. Scheme of reactions involved in SRE.

the different catalytic materials and reaction conditions used as

well as the time scale of elementary step procedures (1012 s).
It is difficult to detect surface species with the available spectroscopic methods. Kinetic analysis can be classified into two groups,
according to the type of kinetic expressions used for analysis: the
Power rate law expressions derived from data fitting or the use of
a LangmuirHinshelwood and/or EleyRideal kinetic model.
Power Rate Law Expressions Derived From Data Fitting
In order to have a model independent description, many of the
measured kinetics are given by power law expressions without
considering the mechanism of elementary steps. Data from kinetic
measurements of the SRE reaction are normally fitted into the
following power law expression:


rSRE = k0 exp

EA
RT

pEtOH
pH2 O

(2)

A survey of reported power rate law expressions can be found

in Table 1. The rate constant k0 , activation energy EA , and
for the reaction order vary between different studies. The reasons
for the disagreement among previous research results on the values of the kinetic parameters for the investigated reaction can
be explained with the use of different catalysts (active phases,
supports, particle sizes, morphologies, catalyst activation/pretreatments) and/or operating conditions (temperature ranges,
feed ratios, different reactors such as differential type).
LangmuirHinshelwood (or EleyRideal) Mechanism
Power rate laws provide little information about the underlying
processes. Moreover, the power rate law kinetic model used in one
study might not be suitable for other types of catalysts under different reaction conditions. A better understanding of the involved
chemical processes can be used for the kinetic models, which can
be more generic. The establishment of a kinetic model with elementary steps requires a deep knowledge of the reactions involved

VOLUME 92, JANUARY 2014

in the overall SRE process. Kinetic models normally use the

LangmuirHinshelwood (or EleyRideal) mechanism. It assumes
that all reaction species are adsorbed on the catalyst surface before
they take part in network reactions. However, the mechanisms
proposed are based on assumptions and simplifications,[30] such
as rate determining steps (RDS), dominating species on surface,
quasi-equilibrium, steady state, etc. A survey of mechanisms can
be found in Tables 2a2c. In Table 2a, Ni and Co based catalysts used in SRE can be found with the corresponding kinetic
models. Table 2b summarises the kinetic models in SRE for Rh
and Ru based catalysts. Finally, in Table 2c the kinetic models used in SMR with Ni based catalysts are shown for a better
comparison.
It could be found that these published kinetic studies in the
field of SRE processes are mainly performed in a small temperature range. In order to have an intensive investigation of
SRE behaviour, a systematic kinetic study of SRE was performed from 200 to 600 C based on a LangmuirHinshelwood
HougenWatson mechanism model with nickel as the active metal
phase. The kinetic model is based on the reaction pathway (green
colour) shown in Figure 1, involving four reactions. In addition,
the power rate law kinetic expression was also examined for comparison purpose.
Kinetic Models Derived From Possible Mechanisms
SRE and the reaction pathways
Based on results reported in literature, a schematic representation
of the reactions involved in SRE was given in the introduction section (Figure 1). In a first step, ethanol can go through
dehydrogenation or dehydration reaction. Dehydration reaction
is favoured when the catalytic system shows acidic features
(e.g. Al2 O3 -based support material) resulting in the formation
of ethane,[31] which is a precursor for coke formation through
polymerisation. When the catalyst has basic features (e.g. MgObased support material), dehydrogenation is preferred, leading
to the formation of acetaldehyde.[3235] The acetaldehyde formed

THE CANADIAN JOURNAL OF CHEMICAL ENGINEERING

117

118

THE CANADIAN JOURNAL OF CHEMICAL ENGINEERING

VOLUME 92, JANUARY 2014

96

51

85

18.4

r = k pEtOH

r = kN3.64
EtOH

r = kCEtOH a

r = k pEtOH
pH

NiAl-LDH

Y 2 O3
After heat treated at 500 C
Al2 O3
La2 O3
Al2 O3

Support/precursor

Pt (15 wt.%)

RhPt (4 wt.%)

Ni (20.6 wt.%)

Ru (5 wt.%)

Ni (15 wt.%)

Ni (25 wt.%)

CeO2

Monolith (BASF:SR10D)

n/a

Al2 O3

Al2 O3

MgOAl2 O3

Raney-type 68.9 wt.%, Ni (28.2 wt.%), Cu (2.9 wt.%), Al2 O3

Ni (43.6 wt.%)

Ni (16.1 wt.%)
Ni (15.3 wt.%)
Ni

Ni (20.6 wt.%)

Active metal

300450

564700

400590

600700

320520

400600

250300

550650

400

130

T ( C)

1.56

310

4.3

10

4.3

318

5.5

4.3

RH2 O/EtOH
(mol/mol)

RDS
) C2 H5 OH CO + ) CH4 + H2
) C2 H5 OH + ) H2 O ) CH4 + CO2 + 2H2 + )
) CH4 + ) H2 O  CO + 3H2 + 2 )
) CH4 + 2 ) H2 O  CO2 + 4H2 + 3 )
) C2 H5 OH + ) H2 O 2CO + 4H2 + 2 )
) C2 H5 OH + 3 ) H2 O 2CO + 6H2 + 4 )
) CHO + ) ) CO + ) H
) CHO + ) OH ) CO2 + ) H2
) CHO + ) OH  ) CO2 + ) H2
) CO2 + ) H2  ) CO + ) OH
) C2 H5 OH + )  ) CH4 O + ) CH2
C2 H5 OH + )  ) C2 H5 OH
) C2 H5 OH + )  ) CH4 O + ) CH2
) CH3 CH2 O + )  ) CH3 CHO + ) H
) CH3 CHO + )  ) CH4 + ) CO
) HCOO + )  ) CO + ) OH

Reaction scheme

C2 H5 OH CH4 + CO + H2
C2 H5 OH + H2 O CH4 + CO2 + 2H2
CH4 + H2 O  CO + 3H2
CH4 + 2H2 O  CO2 + 4H2
C2 H5 OH + H2 O 2CO + 4H2
C2 H5 OH + 3H2 O 2CO2 + 6H2
C2 H5 OH CH4 + CO + H2
C2 H5 OH + H2 O CH4 + CO2 + 2H2
CH4 + 2H2 O  CO2 + 4H2
CO2 + H2  CO + H2 O
C2 H5 OH + 3H2 O 2CO2 + 6H2
C2 H5 OH + 3H2 O 2CO2 + 6H2
C2 H5 OH + 3H2 O 2CO2 + 6H2
C2 H5 OH + 3H2 O  2CO2 + 6H2
C2 H5 OH  CH4 + CO + H2
CO + H2 O  CO2 + H2

279
235
124
214
146
113
123
196
174
166
21.0
59.7
4.4
82.7
43.6
71.3

Active energy
(kJ/mol)

Ni (25 wt.%)
Ni
Ni (15 wt.%)
Co (1020 wt.%)

Ni

Active metal

MgOAl2 O3
n/a
Al2 O3
Al2 O3

NiAl-LDH precursor

Support

12
4.3
4.3
38

5.5

600650

400600
400590
320520
400700

3.510

RH2 O/EtOH
(mol/mol)

Mathure et al.[17]
Akpan et al.[14]
Akande et al.[13]
Sahoo et al.[29]

Llera et al.[25]

Mas et al.[16,27]

Refs.

Ciambelli et al.[15]

Simson et al.[19]

Akpan et al.[14]

Vaidya and Rodrigues[11]

Akande et al.[13]

Mathure et al.[17]

Morgenstern and Fornango[18]

Mas et al.[16]

Therdthianwong et al.[22]

Sun et al.[20,21]

Refs.

550650

T ( C)

Table 2a. Reported LangmuirHinshelwood and EleyRideal kinetics with rate determining steps (RDS) and activation energies for Ni and Co based catalysts

NEtOH represents the molar flow rate of ethanol.

= 0.2/1.2.
= 0.5 0.05, = 0 0.015.

4.4

r = kN0.43
EtOH

2O

23

0.711 2.71
r = k pEtOH
pH O

149

r = kCEtOH

144

7.0
5.9
16.9
1.9
n/a

EA
(kJ mol1 )

0.8
r = k pEtOH

2.52 7.0
r = k pEtOH
pH O

r = kNEtOH

Expression

Table 1. Power rate law kinetics and activation energies reported

VOLUME 92, JANUARY 2014

THE CANADIAN JOURNAL OF CHEMICAL ENGINEERING

119

110
77.5
162
90.9
85.9
418
151
107

C2 H5 OH+ ) ) C2 H5 OH
) CH4 + H2 O  CO + 3H2 + )
) CO4  CO2 + )
No RDS

) C2 H5 OH + ) H2 O  ) C2 H5 OOH + ) H2
) C2 H5 OH ) C2 H4 O + H2
) CH4 + ) H2 O  ) CH4 O + ) H2
) CO + ) H2 O  ) CO2 + ) H2
) CHO + ) ) CO + ) H
) CHO + ) OH  ) CO2 + H2 + )
) CH3 + ) OH  ) CO + 2H2 + )
) CO + ) OH  ) CO2 + ) H

C2 H5 OH  CH4 + CO + H2
CH4 + H2 O  CO + 3H2
CO + H2 O  CO2 + H2
C2 H5 OH + H2 O  CH4 + CO2 + 2H2
CH4 + H2 O  CO + 3H2
CO + H2  CO2 + H2
C2 H5 OH + H2 O 2CO + 4H2
C2 H5 OH CH4 + CO + H2
CH4 + H2 O  CO + 3H2
CO + H2 O  CO2 + H2
C2 H5 OH  CH4 + CO + H2
C2 H5 OH + H2 O  CH4 + CO2 + 2H2
CH4 + H2 O  CO + 3H2
CO + H2  CO2 + H2

RDS
) CH2 O+ ) ) CHO + ) H
) CHO+ ) O ) CO2 + ) H
) CO+ ) O ) CO2 + )
CH4 +2 ) ) CH3 + ) H
CH4 + 2 ) ) CH3 + ) H

Reaction

CH4 + H2 O  CO + 3H2
CO + H2 O  CO2 + H2
CH4 + 2H2 O  CO2 + 4H2
CH4 + 2H2 O  CO2 + 4H2

CH4 + 2H2 O  CO2 + 4H2

240.1
67.1
243.9
53.1
86.7
102

Active energy
(kJ/mol)

MgAl2 O4
K/MgAl2 O4
MgO

Ni
Ni

MgAl2 O4

Ni

Active metal

550600

275550

500575

T ( C)

MgAl2 O4 Al2 O3

Rh (1 wt.%)

Support

20 wt.% CeO2 /Al2 O3

Al2 O3

CeO2

Support

Ru (2 wt.%)

Ru (5 wt.%)

Rh (5 wt.%)

Active metal

Table 2c. SMR kinetics with rate determining steps (RDS) and activation energies for comparison purposes

558
400
496
110

RDS

Reaction scheme

Active energy
(kJ/mol)

11.25

48

10

48

Graschinsky et al.[24]

Peela and Kunzru[28]

Lu et al.[26]

Gorke et al.[23]

Refs.

Wei and Iglesia[68]

Rostrup-Nielsen et al.[69]

Xu and Froment[63]

Refs.

RH2 O/EtOH
(mol/mol)

RH2 O/CH4
(mol/mol)

500600

450550

600700

350660

T ( C)

Table 2b. Reported LangmuirHinshelwood and EleyRideal kinetics with rate determining steps (RDS) and activation energies for Ru and Rh based catalysts

is then decomposed via decarboxylation into methane and

carbon monoxide. These two reactions occur in low temperature ranges,[33] but it is also possible that ethanol directly cracks
into hydrogen, carbon monoxide and methane. This reaction step
can be considered as a sum of the first two reactions. Hereafter, the watergas shift reaction (WGS) takes place in a middle
temperature range. However, there is a possibility that a latent
WGS occurs, which means that acetaldehyde reacts directly with
water.[32] Some catalysts, such as copper based systems, which
have been reported in literature,[3638] can also force acetaldehyde
to react with water so that the acetic acid formed decomposes into
methane and carbon dioxide. Indeed, these two reactions can be
considered to be a latent WGS, because the sum of the individual
reactions corresponds to WGS.
The remaining methane, formed upon ethanol decarboxylation,
has a critical aspect in SRE; it limits the production of hydrogen
at low and intermediate temperatures. For methane conversion,
the SMR is required. Note that SMR is the reaction which generates the highest molar amounts of hydrogen. SMR occurs when
sufficiently high temperatures and excess of water is used, avoiding coke formation by methane decomposition,[39] Boudouard
reaction and ethylene polymerisation. Furthermore, at SMR reaction conditions, the WGS equilibrium becomes the limiting factor
for hydrogen yield.[40] Therefore, a second low temperature WGS
reaction step is necessary to complete hydrogen production. The
possible reactions that can occur during SRE are summarised in
Table 3.
Overall, it is possible to consider the following four important reaction steps for a kinetic modelling on a nickel catalyst
(Table 4).
Ethanol decomposition (ETD)
Several research groups reported the production of hydrogen
via ethanol decomposition under different conditions. Most of
the authors suggest that the first elementary step of SRE is the
molecular adsorption of ethanol.[13,14,41] This assumption can be
described by the Langmuir adsorption (associative adsorption).
The adsorbed ethanol dissociates into adsorbed ethoxy species
which further decomposes into acetaldehyde and hydrogen.[42]

Table 3. Possible reactions during SRE

0
H298
K
(kJ/mol)

Reactions

Chemical description

Dehydration
Dehydrogenation
Acetaldehyde
decomposition
Direct ethanol
decomposition
Watergas-shift
Latent
watergas-shift
Acetic acid
formation
Acetic acid
decomposition
Steam methane
reforming
Overall steam
methane
reforming

CH3 CH2 OH  C2 H4 + H2 O
CH3 CH2 OH  CH3 CHO + H2
CH3 CHO  CH4 + CO

+45.4
+68.4
18.8

CH3 CH2 OH  H2 + CH4 + CO

+49.7

CO + H2 O  CO2 + H2
CH3 CHO + H2 O  CO2 + H2 + CH4

41.1
59.9

CH3 CHO + H2 O  H2 + CH3 OOH

26.7

CH3 COOH  CH4 + CO2

33.3

CH4 + H2 O  CO + 3H2

+205.9

CH4 + 2H2 O  CO + 4H2

+164.7

120

THE CANADIAN JOURNAL OF CHEMICAL ENGINEERING

Table 4. Overall reaction steps for a kinetic modelling on a nickel

catalyst, and elementary reaction steps based on LHHW model
Overall reactions

Chemical description

Number

CH3 CH2 OH  CH3 CHO + H2

CH3 CHO  CH4 + CO
CO + H2 O  CO2 + H2
CH4 + H2 O  CO + 3H2
CH4 + 2H2 O  CO2 + 4H2
CH3 CH2 OH + H2 O  2CO2 + 6H2
CH3 CH2 OH + )  ) CH3 CH2 OH
) CH3 CH2 OH + )  ) CH3 CH2 O + ) H
) CH3 CH2 O + )  ) CH3 CHO + ) H
) CH3 CHO  ) + CH3 CHO
2 ) H  2 ) + H2

(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
(9)
(10)

ACD

CH3 CHO + )  ) CH3 CHO

) CH3 CHO + )  ) CH3 + ) CHO
) CHO + )  ) CO + ) H
) CH3 + ) H  2 ) + ) CH4
) CO  ) + CO

( 9)
(11)
(12)
(13)
(14)

WGS

) + CO  ) CO
H2 O + )  ) H2 O
) H2 O + )  ) H + ) OH
) CO + ) OH  ) COOH + )
) COOH + )  ) CO2 + ) H
) CO2  ) + CO2

(14)
(15)
(16)
(17)
(18)
(19)

2 ) H  2 ) + H2
) CH3 + ) H  2 ) + ) CH4
) CH3 + )  ) CH2 + ) H
) CH2 + )  ) CH + ) H
) CH + )  ) C + ) H

(10)
(13)
(20)
(21)
(22)

H2 O + )  ) H2 O
) H2 O + )  ) H + ) OH
) C + ) OH  ) CHO + )
) CHO + )  ) CO + ) H

(14)
(16)
(23)
(12)

2 ) H  2 ) + H2
) CO  ) + CO

(10)
(14)

ETD
ACD
WGS
SMR
Overall SMR
Overall SRE
ETD

SMR

Acetaldehyde may then be released in molecular form together

with hydrogen. Infrared studies made on different catalysts
have shown that ethanol adsorbs dissociative to form ethoxy
species.[4345] In these studies, the increase in temperature led to
the complete disappearance of the bonds corresponding to ethoxy
species at 200 C.[46]
Acetaldehyde decomposition (ACD)
In the next elementary step, the adsorbed acetaldehyde molecule
may go again through CC bond dissociation, forming an
adsorbed methyl and aldehyde species. The adsorbed aldehyde
species further decomposes into adsorbed CO and hydrogen
species. NiH may then react with the adsorbed NiCH3 species
intermolecular to release methane and two free nickel sites. At
this point the backward reaction can be set as the RDS. Finally,
adsorbed carbon monoxide species may be released as well.
Several works have reported the appearance of carbon monoxide, methane and hydrogen at the low-temperature region
during temperature programmed desorption analysis of adsorbed
ethanol. Such an observation can be fairly attributed to the decomposition of ethoxy species.[4749]
A pure Ni (111) surface, as Gates et al.[50] have shown, causes
bond breaking of ethanol in the following order: first, the

VOLUME 92, JANUARY 2014

scission of the OH bond in the ethanol molecule to form adsorbed

ethoxy species; second, the scission of the CH bond in the CH2
group in the ethoxy species to form an adsorbed intermediate of
acetaldehyde; and finally, the scission of the CC bond of intermediate adsorbed acetaldehyde to form adsorbed methyl species,
CH3 and carbon monoxide species CO. Therefore, it is possible to assume that the key reaction for the nickel catalyst tested
at temperatures up to 300 C is the dehydrogenation of ethanol
to surface adsorbed CH3 CHOads . The following sequence seems
to describe the transformations of CH3 CHOads at this temperature
range:
CH3 CHO(g)  CH3 CHO(ads)  CO(ads) + H(ads)
+ CH3(ads)  CO(g) + CH4(g)

(3)

Watergas shift (WGS)

The WGS reaction is an exothermic reaction without volume
change, and equilibrium conversion of CO (and H2 O) is favoured
with the application of low temperatures. The WGS reaction is
very fast at reforming conditions. As a result, the WGS equilibrium
is always established during steam reforming.
A comparison of metal-based catalysts shows that WGS
depends also on the nature of the active phase. Tests made
at 300 C under forward WGS feed conditions have shown
the following experimental trend for catalysts investigated on
alumina oxide support,[51] which has weak support effects:
Cu > Co > Rh > Ni > Pt > Au > Fe > Pd > Rh > Ir, which means Cu
shows higher WGS activity than Ni. Indeed, Ni catalysts have
lower WGS activities[2124] than Cu catalysts.[2527]
It must be pointed out that in WGS, O or OH species tend to
stick better on Fe, Mo or W, whereas Ni, Rh, Ru, Ir, Pd and Pt
are covered with CO species and Cu, Au and Ag present only
small coverage of both of these species. This is the reason why
in practice the WGS process is applied in two stages, consisting
of a high temperature shift (HTS) at 350 C with metaloxide catalysts such as Fe2 O3 /Cr2 O3 for a better H2 O activation and a low
temperature shift (LTS) between 190 and 210 C with, for example
Cu/ZnO/Al2 O3 catalysts.
The activation of water plays an important role and can
follow different routes: associative adsorption of H2 O, dissociative adsorption into OH and H[29,5254] or the water reacts in
the gas phase with adsorbed species following an EleyRideal
mechanism.[29,53] It is claimed that the latter mechanism has
lower probability of occurrence.[55] In this work, a monomolecular water adsorption (Langmuir adsorption) is proposed. However,
the manner in which water is activated is, of course, temperature
dependent.
Adsorbed carbon monoxide species reacts with adsorbed
hydroxyl species, forming intermediate carboxyl groups which
then decompose into adsorbed hydrogen and carbon dioxide.
Hydrogen and carbon dioxide are then released.
Steam methane reforming (SMR)
The first known kinetic studies on SMR were published in the
1950s.[56,57] Later, Bodrov et al.[5860] investigated SMR kinetics on
a nickel foil. The following expression was proposed to describe
SMR reaction rate:
rSMR =

kSMR pCH4
1 + apH2 O pH12 + bpCO

where a and b are temperature dependent constants.

VOLUME 92, JANUARY 2014

(4)

Allen et al.[61] also carried out SMR experiments using a commercial nickel catalyst and concluded that the RDS was the
desorption of products. Later, De Deken et al.[62] determined the
SMR kinetics on a Ni/Al2 O3 catalyst. Hydrogen was added to the
feed stream. The kinetic model was composed of two expressions
that accounted for the formation of CO and CO2 .
The proposed SMR kinetics[5860,62,63] were compared by
Elnashaie et al.[64] The most important conclusion was the different dependences for the partial pressure of steam: negative,[5860]
positive De Deken et al.[62] and positive or negative.[63] The conclusion was that the work of Xu and Froment[63] is the most generic
kinetic model. They postulated that carbon intermediates react
with adsorbed oxygen, which should be the RDS.
Another mechanism for SMR was proposed later by RostrupNielsen.[65] Hou and Hughes[66] later developed a similar model
to fit the kinetics on a Ni/-Al2 O3 catalyst with more parameters. Reports in the literature for SMR kinetics indicate that
the key step for a successful hydrogen production must be the
sufficient supply of activation energy to crack the CH bonds
in the methane molecule. It has been also reported that the
theoretical methane activation in SMR may occur at temperatures higher than 300 C.[67] It is well-known that the XH bond
strength in covalent hydrides depends on the electronegativity
and element size (X = H, C, N, O). When increasing the atomic
number, the bond force decreases in a group. Bonds between
hydrogen, carbon, oxygen or nitrogen atoms are the most stable covalent bonds. The dissociation energy of methane is very
high (diss H = 435.3 kJ/mol). Also, the electronic configuration
of methane is 1s2 2s2 2p2 , with two unpaired electrons present
in two 2p orbitals. The transfer of one electron from the 2s to
the 2p orbitals results in four single-filled orbitals (1s2 2s1 2p3 )
for bonding. Thus, the arrangement of the four CH bonds in the
methane molecule is the result of a tetrahedral symmetry, which
minimises electron repulsion.
Wei and Iglesia[68] found that the reaction rates are only proportional to the partial pressure of CH4 . This observation leads
to the conclusion that the dissociation of methane is the RDS
for SMR. The methane dissociation is limited kinetically by the
initial activation of the CH bonds and unaffected by the H2 O
concentration.
The current understanding of methane dissociation[69] leads to
the conclusion that the reaction does not proceed via an adsorbed
precursor state. It has been found that the adsorbed methyl
species may decompose so far that an adsorbed carbon species
is obtained.[70]
The first reaction step is the dissociative adsorption of CH4 :
CH4 + 2 )  ) CH3 + ) H

(5)

The ) CH3 species are subsequently dissociated by many possible surface reactions until chemisorbed carbon atoms are formed
on the metal surface, as shown in Figure 2a.
DFT calculations of these steps on Ni surface, which are reproduced in an energy diagram given in Figure 2b, were performed by
Bengaard et al.[71] They proved that the energies of the CHx species
are higher than those for methane, while the adsorbed methyl
species CH3 show a similar energy to CH2 and CH species.
In addition, the energies of the chemisorbed carbon C* species
are lower than for CHx species and CH4 , with higher activation
energy for decomposition of CHx species to C* then to the hydrogenation of CHx species to methane. The dissociative adsorption
of methane Equation (5) seems to be the RDS.

THE CANADIAN JOURNAL OF CHEMICAL ENGINEERING

121

The adsorbed acetaldehyde (2s) is released from the free active

nickel site, or it further reacts (3s) to form methane and adsorbed
carbon monoxide simultaneously. Hereafter CO can be released
(4s), while the produced methane and two free active sites can go
through the reversible dissociative adsorption reaction, which is
set as the RDS.
In the next step adsorbed CO, OH and H react to form three
active nickel sites plus CO2 and H2 in the gas phase (6s). Adsorbed
hydrogen species may also react with intermolecular (2NiH) to
form two free active nickel sites and molecular hydrogen (7s).
Finally, the adsorbed methyl species, obtained in (3s), can react
with adsorbed hydroxyl groups (5s) to generate the adsorbed CO,
2 molecular hydrogen products and one active nickel site.
EXPERIMENTAL AND MODELLING
Reaction Studies

Figure 2. The mechanism of methane decomposing (a) and potential

energy diagram (b).

If water is introduced in excess, the adsorbed CH3 species may

react with the adsorbed hydroxyl species or even oxygen species
intermolecular, thus forming adsorbed CO and hydrogen:
) H3 C + ) OH  ) H3 COH + )

(6)

) H3 COH + )  ) H2 COH + ) H

(7)

) H2 COH + )  ) H2 CO + ) H

(8)

) H2 CO + )  ) CHO + ) H

(9)

and
) CHO + )  ) CO + ) H

(10)

A commercial 15 wt.% Ni/-Al2 O3 catalyst (Octolyst 1001), supplied by Degussa AG in form of extrudates, was used. The
experiments were performed from 200 to 600 C in 100 C intervals
and atmosphere pressure with a water-to-ethanol molar ratio of
10 in the feed. Details on the characterisation of the catalyst and
reaction units can be found in a previous work.[33] In addition,
a Dycor ProLine Process Mass Spectrometer (AMETEK, Berwyn)
was used to measure all the partial pressures of products in real
time. The schematic representation of the experimental units can
be found in the Supplementary Information.
The conversion of the reactants ethanol and water were calculated as follow:
n EtOH,0 n EtOH
100,
n EtOH,0
n H O,0 n H2 O
XH2 O (%) = 2
100
n H2 O,0

XEtOH (%) =

where n EtOH,0 and n H2 O,0 represents the molar flow rate of ethanol
and water in the feed, respectively. n EtOH and n H2 O is the outlet
molar flow rate of unreacted ethanol and water. The partial pressures were obtained from all species (molar distribution) present
in the outlet goes stream:
n

i
yi = 

n i

As a result, the steps in this model could be simplified due to the

excess water used. The adsorbed CO species will undergo again
a WGS reaction mechanism, as already proposed above, or be
released.

122

THE CANADIAN JOURNAL OF CHEMICAL ENGINEERING

pi
ptotal

ptotal =

n


(12)

pi

n i
where n i represents the molar flow rate of one species and
represents the total molar flow rate of species in the outlet stream.
The yield of species i was calculated from:

Proposed Mechanism
Based on findings in the literature, Table 4 depicts the possible
elementary reaction steps based on a LangmuirHinshelwood
HougenWatson (LHHW) model. This model is associated with
Figure 3, where a SRE mechanism with excess water is proposed.
It is clear that the proposed SRE mechanism with excess water is
too complex; therefore, the establishment of a new simple kinetic
model is described in the following section (Table 5), associated
with the corresponding LH expressions and using the HW methodology where the dissociative adsorption of methane (Equation 5)
is set as the RDS.
In the first step (1s), ethanol reacts with a nickel site directly
into adsorbed acetaldehyde and releases molecular hydrogen
simultaneously.

(11)

Yi (%) =

n i n i,0 vEtOH
100
n EtOH,0 vi

(13)

and the stoichiometric coefficient  is defined as: EtOH =

acetaldehyde = 1, CH4 = CO = CO2 = 2 and H2 = 6. The selectivity
of species i can be obtained by:
Si =

Yi

(14)

XEtOH

The catalytic results are discussed in terms of space time :

=

mcat 1
VR
=
b V
V

(15)

VOLUME 92, JANUARY 2014

Figure 3. Proposed mechanism for SRE with excess water.

where VR is the reactor volume, V is the volume flow rate of the

reactants and inert under the standard condition, and mcat and
b represent the mass of the catalyst and the bulk density of the
catalyst, respectively. Table 6 collects the experimental conditions
used. The reproducibility of experimental results was checked,
and the experimental error was around 5%. For all the measurements made in this work, an average error of around 5% should
be taken into account.
Kinetic Modelling
Thermodynamic analysis only provides the limit of the reaction
system.[40] However, kinetics must be taken into account. It is
known that carbon formation can occur during SRE, but at a

Table 5. Simplified mechanism and calculation of coverage

expressions

dXi
d(mcat /n i,0 )

Variable

CH3 CH2 OH + )  ) CH3 CHO + H2

) CH3 CHO  ) + CH3 CHO
) CH3 CHO  ) CO + CH4
) CO  CO + )
CH4 + 2 )  ) CH3 + ) H
H2 O + 2 )  ) H + ) OH
) CO + ) OH + ) H  3 ) + CO2 + H2
2 ) H  2 ) + H2
) CH3 + ) OH  ) + ) CO + 2H2

VOLUME 92, JANUARY 2014

r=

(16)

Table 6. Experimental conditions

Simplified model

certain RS/E (steam/ethanol molar ratio) the formation can be

suppressed. It was found that carbon deposits are formed below
RS/E = 3, associated also with a low hydrogen yield in this region.
In this work, the simplified kinetic model proposed does not consider the formation of adsorbed carbon due to the fact that we
operate SRE with excess water.
The SRE reaction was investigated in an integral reactor; the
reaction rates were obtained using the differential method by
analysing the conversions versus mcat /n i,0 . The initial concentration of the ethanolwater mixture is used to find the reaction rate,
using the differential method of analysis:

(1s)
(2s)
(3s)
(4s)
(RDS)
(5s)
(6s)
(7s)
(8s)

Operative conditions

T ( C)
p (kPa)
mcat (kg)
V inert (N cm3 /min)
 (s)
yEtOH,0 (kPa/kPa)
yH2 O,0 (kPa/kPa)
RS/E (mol/mol)
Rinert/EtOH,0 (mol/mol)

200600
100
0.00010.05
0200
0.0086.731
0.0560.0901
0.5600.901
10:1
06.9

THE CANADIAN JOURNAL OF CHEMICAL ENGINEERING

123

The rate of reaction is determined by the slope of dXi versus

d(mcat /n i,0 ), where n i,0 represents the molar flow rate of reactant
i in the feed.The power rate law model can be expressed as:
(17)

200 C
300 C
400 C
500 C
600 C

40

20

0
0

[s]

30

24

m exp
(ri ricalc )2
i
= 
m
exp
2

(18)

where m is the total number of experimental measurements and r

is the average reaction rate. The kinetic parameters were estimated
by minimising this objective function using LevenbergMarquardt
method with the nonlinear regression software (MATLAB 2010b).
The values of the LevenbergMarquardt parameters used in
the optimisation were function tolerance = 1 1010 , max iterations = 1 103 and max function evaluations = 1 104 .

r )

XH O [mol%]

(ri

60

XEtOH [mol%]

where k0 is the pre-exponential factor, EA is the energy of activation, pi is the partial pressure of the reactant ethanol and n in the
superscript is the reaction order with regard to the reactant. These
parameters can be determined with a nonlinear regression of the
experimental results.
In a following step, the establishment of a simplified LHHW
kinetic model for SRE was prepared; details can be found from
the supplementary information.
In order to estimate kinetic parameters, experimental data
points obtained were fitted with the use of the power rate model
and the LHHW model. The sums of squared residuals were used
to obtain the objective function Fobj :

100

80

n
rSRE = k0 e(EA /(RT)) pEtOH

Fobj

18

200 C
300 C
400 C
500 C
600 C

12

0
0

[s]

Figure 4. Ethanol (a) and water (b) conversions.

RESULTS AND DISCUSSION

Catalyst Performance
Ethanol and water conversion versus space time, at different temperatures, are shown in Figure 4. Dashed lines used in Figures 46
were introduced to follow the experimental points. As can be seen,
ethanol conversion increases with temperature and space time.
Particularly at temperatures higher than 300 C, the total ethanol
conversion can be reached with a space time of 0.78 s. The water
conversion also increases with temperature and space time for
space times lower than 0.2 s. At 600 C and space times greater
than 0.2, the water conversion remains constant in correspondence with the total conversion of ethanol or decreases slightly
due to the water formation through the reverse WGS reaction
favoured at the higher temperature.
The intra-particle diffusion was determined with the Weisz
Prater criterion, and the calculated value was less than 1 (0.25
at 600 C and 0.0085 at 200 C), which means the concentration
on the catalyst surface is very close to the concentration within
its pores. Therefore, the effect of the intra-particle diffusion was
eliminated. In order to eliminate the external mass transfer limitation, preliminary tests at constant space time but with varying
feed rates were carried out.[33] It was found that the conversion of
ethanol at constant space time became independent from feed rate
when the inert flow was 200 N cm3 /min. In addition, the external
mass transfer effect was determined with the Mears criterion, and
the calculated values are less than 0.15 (0.028 at 600 C and 0.001
at 200 C) with an inert flow of 200 N cm3 /min in the feed. As
a result, only experimental results obtained with 200 N cm3 /min
inert in the feed were used for LHHW kinetic model analysis.

124

THE CANADIAN JOURNAL OF CHEMICAL ENGINEERING

During SRE reaction on a nickel catalyst, acetaldehyde, carbon

monoxide and methane are hereby intermediate products, while
hydrogen and carbon dioxide are the desired products.
The variation of intermediate product distributions with space
time and temperature is shown in Figure 5. As a result of the
decomposition reaction, yields of acetaldehyde were decreased
with space time in all temperatures. With increasing temperature,
the CH4 yields first increased because of the ETD then decreased
due to the existence of the SMR at high temperatures, whereas
the CO yields initially decreased and then increased marginally
by the temperature owing to the equilibrium of WGS reaction.
On the other hand, with increasing space time, the yields of CO
decreased while CH4 increased slightly, which is probably due to
the methanation reaction (4H2 + CO2  CH4 + 2H2 O).[28] Therefore, low space times are recommended for hydrogen production
from SRE.
The distributions of the desired products versus space time at
different temperatures are shown in Figure 6. Below 500 C, the
selectivity towards H2 and CO2 increased with the temperature
and space time, whereas above 500 C, the yields of CO2 decreased
due to the existence of the reverse WGS at high temperatures. At
600 C, the selectivity of H2 remains constant with different space
times, which is in agreement with the complete conversion of
ethanol at these conditions.
Figure 4 shows results of Xi (mol%) versus  (s), while Figure 6
depicts results of yi (kPa/kPa) versus  (s). As the conversions of
the reactants increase, it is expected that the partial pressures

VOLUME 92, JANUARY 2014

1.2x10-2

6.0x10-1

200 C
300 C
400 C
500 C
600 C

1.0x10-2

4.5x10-1

yH [kPa/kPa]

yAc [kPa/kPa]

8.0x10-3
6.0x10-3

3.0x10-1

4.0x10-3

200 C
300 C
400 C
500 C
600 C

1.5x10-1

2.0x10-3
0.0

0.0

[s]

5.0x10-2

200 C
300 C
400 C
500 C
600 C

200 C
300 C
400 C
500 C
600 C

1.2x10-1

yCO [kPa/kPa]

3.0x10-2

8.0x10-2

2.0x10-2

4.0x10-2

0.0

[s]
1.5x10-1

200 C
300 C
400 C
500 C
600 C

1.2x10

9.0x10

6.0x10-2

3.0x10-2

0.0
1

[s]

Figure 5. Partial pressures of the intermediate products: acetaldehyde

(a), carbon monoxide (b) and methane (c).

of the desired products also increase, suggesting the conversions

of reactants are proportional to the partial pressures of the desired
products.
Kinetic Analysis
By considering the reactions and products involved mentioned
above in Kinetic Models Derived From Possible Mechanisms Section, the following reaction scheme is presented in Table 4,
according to findings in literature reported. Unfortunately, this
model is too complex, so it must be simplified to a satisfactory
description of the kinetic behaviour, as shown in Table 5. The

VOLUME 92, JANUARY 2014

Figure 6. Partial pressures of hydrogen (a) and carbon dioxide (b).

-2

[s]

-1

yCH [kPa/kPa]

[s]

0.0

1.6x10-1

1.0x10-2

4.0x10-2

yCO [kPa/kPa]

RDS of this model is the dissociative adsorption of mechanism,

which is similar to the reports in the literature for SMR kinetics.
Research efforts on SMR kinetics[68,69] agree that methane is
adsorbed on nickel surfaces as carbon atoms. Water is also
adsorbed either as OH species or O species, and hydrogen formation is a fast reaction compared to the RDS.
However, there is no agreement in the adsorption mechanism for methane. Xu and Froment[63] say that only one Ni
atom is required for the adsorption of methane. Methane is
subsequently dehydrogenated, but the model predicts that three
RDS are involved together with 13 reaction steps, as shown in
Table 2c. They found also an unlikely negative heat of steam
adsorption, which is not a strong argument in favour of their
mechanism; it is also known that the WGS reaction is very fast at
reforming conditions. As a result, the WGS equilibrium is always
established during steam reforming.
In contradiction to Xu and Froment,[63] Rostrup-Nielsen et al.[69]
postulates that methane is not dissociative adsorbed on Ni. The
RDS is assumed to be the irreversible adsorption of methane associated with substantial energy and entropy barriers and requires
at least a certain amount of active sites for carbon atoms formation, since methane is a stable and highly symmetrical molecule
with low sticking coefficients for adsorption.
Wei and Iglesia[68] postulate that the reactivity of the active
metal phase towards CH bond breaking governs the overall
reaction kinetics. It was found that reaction rates are proportional
to methane partial pressure but independent of carbon dioxide and

THE CANADIAN JOURNAL OF CHEMICAL ENGINEERING

125

126

THE CANADIAN JOURNAL OF CHEMICAL ENGINEERING

a
40

200 C
300 C
400 C
500 C
600 C

rSRE[molkg-1h-1]

30

20

10

0
0

pEtOH [kPa]

40

200 C
300 C
400 C
500 C
600 C

30

rSRE [molkg-1h-1]

water pressures, which led them to the conclusion of sole kinetic

relevance of CH bond activation steps. The catalyst surface may
be uncovered by reactive intermediates, due to the fast steps of
the activation of the co-reactant and methane activation through
the foraging of chemisorbed carbon intermediates. Further, it is
postulated that recombinative desorption steps of H atoms with
OH species form hydrogen and water.
Finally, Gates et al.[50] describes the mechanism of ethanol
decomposition on the Ni (111). The sequence of bond scission
steps occurs as ethanol undergoes dissociative reactions. Bond
activation occurs in the order OH, CH2 , CC and CH3 . The
products observed are CH3 CHO(g) , CH4(g) , CO(g) , H2(g) and surface
carbon. The latter species dissolves into the Ni bulk. Acetaldehyde
and methane desorb, and the formation of both of these products
is controlled by scission of the methylene (CH2 group). The CH3
group is cleaved from the intermediate surface CH3 CHO species
to form CH3(ads) . H2 exhibits a broad, the carbonoxygen bond in
ethanol remains intact and CO ultimately desorbs in a single desorption. A small fraction of CO(a) species undergo exchange with
the carbidic surface carbon in a minor process observed.
Mas et al.[16] reported kinetic experiments on SRE with the
formation of H2 , CO, CO2 and traces of CH4 . Experiments with
different methane concentrations showed that ethanol conversion
decreases as methane concentrations increase, revealing the existence of competitiveness between both ethanol and methane for
adsorption on the same active site. A power rate law was used
first to estimate kinetic parameters then to complete the kinetic
study with a LH model.
Llera et al.[25] analysed the kinetic results on SRE, and they
found that both CO2 and CH4 are intermediate products, while
CO is a final product. They proposed a kinetic model involving
two reaction steps where ethanol reacts irreversible, while the
CH4 and CO2 are reversible in the reaction. They postulate that
the surface reactions are the RDS steps and that the dissociative
adsorption of methane is a reversible step in quasi-equilibrium as
well as in all remaining reactions. Eighteen reaction steps with
four RDS steps were used, making the model quite complex. The
model parameters have shown acceptable fit with a goodness of
R2 = 0.95218.
Sahoo et al.[29] proposed a mechanistic kinetic model using LH
approach, considering the surface reactions of SRE, WGS and ETD
as the RDS in the mechanism. The fitted data indicate that the
formation of acetaldehyde from ethoxy is the RDS for reforming,
and the correlation coefficient R2 value for all cases was above
0.95.

Gorke
et al.[23] reported that the RDS steps in the model were
CO2 desorption, dissociative ethanol adsorption and reaction of
adsorbed methane with steam from the gas phase for WGS, ETD
and SMR, respectively. Since a micro reactor has been used, no
heat or mass transfer limitations were detected.
Mathure et al.[17] used acquired experimental data from SRE
to fit them in the power-law kinetic model. A deviation of 10.2%
was calculated, and considerable amounts of coke formation were
observed during the process as well, reducing catalytic activity. Hereafter an EleyRideal mechanism postulated by Akande
et al.[13] was proposed for a better fitting and description of reactions involved in SRE. Four RDS are proposed: adsorption of
ethanol on active sites, dissociative adsorption of ethanol, surface
reaction of the adsorbed oxygenated hydrocarbon fraction with no
adsorbed steam, and surface reaction of the adsorbed hydrocarbon
fraction with non-adsorbed steam.
The values of kinetic parameters and the fitness value of objective function for power rate law model and LHHW model obtained

20

10

0
0

pEtOH [kPa]
Figure 7. Comparison between the mathematical model (lines) and
experimental results (filled spots): power rate model (a) and LHHW
model (b).

from fitting are summarised in Table 7. The coefficients of determination R2 is obtained from:
R2 = 1 Fobj

(19)

As it is shown in Table 7, R2 calculated by the power rate law

model and LHHW for ethanol reaction rates are 0.918 and 0.981,
respectively. The comparison of reaction rates from experimental
and calculated results at different temperature conditions is shown
in Figures 7 and 8 with a power rate law model (a) and a LHHW
model (b). As it can be seen from these figures, the predicted
values from both models are in agreement with the experimental
results under different temperature conditions, and the values of
LHHW model are closer to the experimental values than the power
rate law model, due to the higher coefficient of determination of
LHHW model (R2 = 0.981) compared with that of power rate law
model (R2 = 0.918). These results confirmed the LHHW model
with assumption of the surface decomposition of methane, since
the RDS has a better fit than the power rate law model.
The order values of activation energies reported for SRE has
a deviation from 100 to 102 kJ/mol. The deviation of activation
energies could be due to the different active phases and reaction
conditions employed.

VOLUME 92, JANUARY 2014

Table 7. Calculated results for rSRE with 200 N cm3 /min inert for the kinetic models, pre-exponential factor, energies of activation, reaction
enthalpies, heat of adsorptions, reaction order and coefficients of determination
Model

ri (mol/h)

Relation

k0,j (mol/g h)

Power law
LHHW

rSRE
rSRE

k(T ) = k0 e(EA /(RT ))

k0
kRDS (T ) = k0,RDS e(EA /(RT ))
kRDS
Equilibrium of elementary steps

e((Hj )/R)((1/T )(1/673.15)) K1

Kj (T ) = Kj(673.15)
K2
K3
K4
K5
K6
K7
K8
Equilibrium of adsorbed species

e(Hi /R)((1/T )(1/673.15)) KAc

Ki (T ) = Ki(673.15)

KCO

KOH
KH

KCH

(2.22 0.95) 102

(1.67 0.54) 109
Kj(673.15) (mol/g h)
(3.75 0.10) 105
(1.66 0.29) 105
(3.95 0.65) 1014
(4.03 0.89) 101
(1.01 0.21) 100
(4.82 1.18) 102
(1.11 0.23) 103
(3.42 0.54) 1012
Ki(673.15) (mol/g h)
(6.02 0.34) 106
(4.03 0.89) 101
(3.02 0.79) 102
(3.00 0.37) 102
(1.78 0.42) 107

R2

31.8 2.1
1.52 0.11
0.918
31.8 1.6
n/a
Hj (kJ/mol)
Sj,673.15 a (J/mol K) 0.981
1
(7.20 0.58) 10
22.2 1.0
(6.55 0.41) 101
100.9 10.5
(2.33 0.28) 101
244.8 34.4
(5.53 0.38) 100
15.8 1.9
(4.43 0.32) 100
6.7 0.8
(4.82 0.35) 101
20.2 2.6
(1.73 0.24) 101
84.0 14.2
(1.23 0.18) 102
36.8 5.6
Hi (kJ/mol)
Si,673.15 a (J/mol K)
1
(6.55 0.41) 10
100.9 6.0
(5.53 0.38) 100
15.8 1.9
(4.22 0.88) 100
35.4 8.2
(8.65 0.12) 100
42.0 1.7
(1.01 0.21) 102
11.2 2.5

When S is positive and H is negative, a process is always spontaneous. When S is positive and H is positive, a process is spontaneous at high
temperatures. When S is negative and H is negative, a process is spontaneous at low temperatures. When S is negative and H is positive, the reverse
process is spontaneous.

While the reaction order with respect to ethanol has a deviation

from 0.43 to 3.64 over nickel based catalysts according to Akande
et al.[13] and Akpan et al.,[14] the large difference could be due to
the presence or absence of heat and mass transfer resistances as
well as a different mechanism related to the catalyst type.
The obtained energy of activation for the SRE reaction in this
study is 31.8 kJ/mol, which is between the values obtained by
Mathure et al.[17] (23 kJ/mol) and Akpan et al.[14] (51 kJ/mol)
where nickel-based catalysts were employed, while the reaction
order (1.52) of the ethanol partial pressure is also in the range of
the values reported by Mathure et al.[17] (0.71) and Therdthianwong et al.[22] (2.52).
Parity plots for rates of reaction were calculated using the
power rate model and LHHW-model versus the experimental reaction rates. R2 values were obtained from the linear fitting of the
points. It might be observed that the LHHW-model has a quite
good fitting, while the power rate model has stronger deviations. The relatively low activation energy obtained for the LHHW
model is very likely attributed to the reaction conditions because
a high water-to-ethanol molar ratio (10) was used. Mathure
et al.[17] investigated a commercial Ni-based catalyst at similar
reaction conditions (water-to-ethanol molar ratio of 12) where
the gas-phase and intra-particle mass transfer resistance was eliminated, and the activation energy was determined with a value of
21 kJ/mol.
The obtained values of Kj are listed in Table 7. It can be found
that K3  K1 , which indicates that the acetaldehyde decomposition reaction is much faster than the ethanol decomposition
reaction. On the other hand, K2 > K7 > K4 suggests the unreacted acetaldehyde could be released from the active phase faster
than hydrogen and carbon monoxide. As a result, a low yield
of acetaldehyde was obtained especially at high temperatures.
It can be found that K8 with a relatively large enthalpy value
(123 kJ/mol) increases as the temperature is raised, indicating
that the elementary reaction between CH3 and  OH is also a
strong endothermic reaction similar to SMR, which has an average enthalpy of 180 kJ/mol among the reaction conditions used

EA (kJ/mol)

VOLUME 92, JANUARY 2014

in this study (200600 C). In addition, the equilibrium constant

(K6 ) of the elementary reaction (Equation 6s in Table 5), which
is related with the reversible WGS reaction, decreases with the
temperature rise. As a result, the reaction temperature becomes
the deterministic function of SRE process in order to obtain a high
yield of hydrogen with a low yield of carbon monoxide from the
product gas.[72]

CONCLUSIONS
In this work, the kinetic behaviour of SRE over a commercial
nickel-based catalyst has been investigated from 200 to 600 C
with a water-to-ethanol mole ratio of 10 in the feed. The purpose
of this work was to develop a simplified mechanistic model to
replace empirical models (power rate model) and other proposed
kinetic models that are too complex for an applied kinetic process. The following conclusions can be drawn from the analysis
and discussion above.
(1) Both the power rate law and LHHW kinetic models employed
in this work can successfully demonstrate the catalytic SRE
process in a fixed-bed reactor among a large temperature
range.
(2) The energy of activation for SRE reaction was 31.8 kJ/mol,
and the reaction order of the ethanol pressure obtained from
the power rate law model was 1.52, which are in good agreement with other reported results.
(3) According to the comparison of experimental data between
the results from two kinetic models, the LHHW model with
assumption of the surface decomposition of methane as the
RDS has a better fitting than that from the power rate law
model.
(4) The values of the parameters for LHHW kinetic model indicate that the acetaldehyde decomposition reaction is much
faster than the ethanol decomposition reaction, while the
elementary reaction between CH3 and  OH is also a strong
endothermic reaction as the SMR.

THE CANADIAN JOURNAL OF CHEMICAL ENGINEERING

127

Figure 8. Parity plot for rates of reaction calculated using the power rate model (a) and LHHW-model (b) versus the experimental reaction rates.

NOMENCLATURE
ACD
EA
ETD
EtOH
LHHW
m
mcat

128

n
n i
n i,0
rSRE
r
RS/E
RDS
Si
SMR

acetaldehyde decomposition
activation energy (kJ/mol)
ethanol decomposition
ethanol
LangmuirHinshelwoodHougenWatson
total number of experimental measurements
catalyst mass (kg)

THE CANADIAN JOURNAL OF CHEMICAL ENGINEERING

reaction order of the reactant pressure

molar flow rate of species i at the outlet (mol/h)
molar flow rate of species i in the feed (mol/h)
reaction rate of SRE (mol kg/h)
average reaction rate (mol/kg h)
steam/ethanol molar ratio in the feed
rate determining step
selectivity of product i
steam methane reforming

VOLUME 92, JANUARY 2014

SR
SRE
T
p
pi
VR
WGS
XEtOH
XH2 O
yi
Yi

steam reforming
steam reforming of ethanol
temperature ( C)
total pressure (kPa)
partial pressure for component i (kPa)
reactor volume (m3 )
watergas shift
ethanol conversion (%)
water conversion (%)
mole fraction for component i in the gas phase
(kPa/kPa)
yield of product i (%)

Greek Symbols
b

i

bulk density of the catalyst

space time (s)
stoichiometric coefficient of component i

Subscripts
0
t

standard states
total

ACKNOWLEDGEMENTS
Y.J.W. gratefully acknowledges doctoral grant from China Scholarship Council (2010674011).A.F.C. is gratefully to Fundaca o
para a Ciencia e Tecnologia (FCT), Portugal, for providing
financial support for this research program and postdoctoral
grant (SFRH/BPD/62968/2009).The authors also thank for LSRE
financing by FEDER/POCI/2010.
REFERENCES
[1]
[2]
[3]
[4]
[5]
[6]
[7]

[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]

C. Hulteberg, Int. J. Hydrogen Energy 2012, 37, 3978.

C. Parmesan, G. Yohe, Nature 2003, 421, 37.
F. de Bruijn, Green Chem. 2005, 7, 132.
K. V. Kordesch, G. R. Simader, Chem. Rev. 1995, 95, 191.
N. Z. Muradov, T. N. Veziroglu, Int. J. Hydrogen Energy
2008, 33, 6804.
J. A. Turner, Science 2004, 305, 972.
A. Vert`es, N. Qureshi, H. Yukawa, H. Blaschek, Biomass to
Biofuels: Strategies for Global Industries, Wiley, Hoboken,
New Jersey 2009.
C. Wyman, Handbook on Bioethanol: Production and
Utilization, Taylor & Francis, Washington, DC 1996.
A. Haryanto, S. Fernando, N. Murali, S. Adhikari, Energy
Fuels 2005, 19, 2098.
M. Ni, D. Y. C. Leung. M. K. H. Leung, Int. J. Hydrogen
Energy 2007, 32, 3238.
P. Vaidya, A. E. Rodrigues, Chem. Eng. J. 2006, 117, 39.
F. Frusteri, S. Freni, L. Spadaro, V. Chiodo, G. Bonura, S.
Donato, S. Cavallaro, Catal. Commun. 2004, 5, 611.
A. Akande, A. Aboudheir, R. Idem, A. Dalai, Int. J.
Hydrogen Energy 2006, 31, 1707.
E. Akpan, A. Akande, A. Aboudheir, H. Ibrahim, R. Idem,
Chem. Eng. Sci. 2007, 62, 3112.
P. Ciambelli, V. Palma, A. Ruggiero, Appl. Catal. B 2010, 96,
190.
V. Mas, M. L. Dieuzeide, M. Jobbagy, G. Baronetti, N.
Amadeo, M. Laborde, Catal. Today 2008, 133135, 319.

VOLUME 92, JANUARY 2014

[17] P. V. Mathure, S. Ganguly, A. V. Patwardhan, R. K. Saha,

Ind. Eng. Chem. Res. 2007, 46, 8471.
[18] D. A. Morgenstern, J. P. Fornango, Energy Fuels 2005, 19,
1708.
[19] A. Simson, E. Waterman, R. Farrauto, M. Castaldi, Appl.
Catal. B 2009, 89, 58.
[20] J. Sun, D. Luo, P. Xiao, L. Jigang, S. Yu, J. Power Sources
2008, 184, 385.
[21] J. Sun, X. P. Qiu, F. Wu, W. T. Zhu, Int. J. Hydrogen Energy
2005, 30, 437.
[22] F. Aupretre, C. Descorme, D. Duprez, Catal. Commun.
2002, 3, 263.
M. Laborde, N. Amadeo, Chem. Eng. J.
[23] J. Comas, F. Marino,
2004, 98, 61.
[24] A. N. Fatsikostas, X. E. Verykios, J. Catal. 2004, 225, 439.
[25] J. R. Rostrup-Nielsen, Catal. Sci. Technol., 1984, 5, 1.

[26] P. Haussinger, R. Lohmuller,

A. M. Watson, Ullmanns
Encyclopedia of Industrial Chemistry, Wiley-VCH Verlag
GmbH & Co. KGaA, Weinheim 2000.
[27] J. D. Holladay, J. Hu, D. L. King, Y. Wang, Catal. Today
2009, 139, 244.
[28] N. R. Peela, D. Kunzru, Ind. Eng. Chem. Res. 2011, 50,
12881.
[29] D. R. Sahoo, S. Vajpai, S. Patel, K. K. Pant, Chem. Eng. J.
2007, 125, 139.
[30] I. Chorkendorff, J. W. Niemantsverdriet, Concepts of
Modern Catalysis and Kinetics Concepts, Wiley VCH,
Weinheim, Germany 2003.
[31] M. C. Sanchez-Sanchez, R. M. Navarro, J. L. G. Fierro, Int.
J. Hydrogen Energy 2007, 32, 1462.
[32] M. Akiyama, Y. Oki, M. Nagai, Catal. Today 2012, 181, 4.
[33] A. F. Cunha, Y. J. Wu, F. A. D. Alvarado, J. C. Santos, P. D.
Vaidya, A. E. Rodrigues, Can. J. Chem. Eng. 2012, 90, 1514.
[34] S. Freni, S. Cavallaro, N. Mondello, L. Spadaro, F. Frusteri,
Catal. Commun. 2003, 4, 259.
G. Baronetti, M. Jobbagy, M. Laborde, Appl.
[35] F. Marino,
Catal. A 2003, 238, 41.
[36] A. F. Cunha, Y. J. Wu, J. C. Santos, A. E. Rodrigues, Chem.
Eng. Res. Des. 2012, DOI: 10.1016/j.cherd.2012.09.015.
[37] A. F. Cunha, Y. J. Wu, J. C. Santos, A. E. Rodrigues, Ind.
Eng. Chem. Res. 2012, 51, 13132.
[38] N. Iwasa, N. Takezawa, Bull. Chem. Soc. Jpn. 1991, 64,
2619.
ao, J. L. Figueiredo, Appl. Catal. A
[39] A. F. Cunha, J. J. M. Orf
2008, 348, 103.
[40] Y. J. Wu, F. D. Alvarado, J. C. Santos, F. Gracia, A. F.
Cunha, A. E. Rodrigues, Chem. Eng. Technol. 2012, 35, 847.
[41] F. Wang, Y. Li, W. Cai, E. Zhan, X. Mu, W. Shen, Catal.
Today 2009, 146, 31.
[42] J.-H. Wang, C. S. Lee, M. C. Lin, J. Phys. Chem. C 2009, 113,
6681.
[43] H. Idriss, C. Diagne, J. P. Hindermann, A. Kiennemann, M.
A. Barteau, J. Catal. 1995, 155, 219.
[44] A. Yee, S. J. Morrison, H. Idriss, J. Catal. 1999, 186, 279.
[45] A. Yee, S. J. Morrison, H. Idriss, J. Catal. 2000, 191, 30.
[46] L. V. Mattos, F. B. Noronha, J. Catal. 2005, 233, 453.
[47] M. A. S. Baldanza, L. F. de Mello, A. Vannice, F. B.
Noronha, M. Schmal, J. Catal. 2000, 192, 64.

THE CANADIAN JOURNAL OF CHEMICAL ENGINEERING

129

[48] E. M. Cordi, J. L. Falconer, J. Catal. 1996, 162, 104.

[49] L. F. de Mello, F. B. Noronha, M. Schmal, J. Catal. 2003,
220, 358.
[50] S. M. Gates, J. N. Russell, Jr. J. T. Yates, Jr. Surf. Sci. 1986,
171, 111.
[51] N. Schumacher, A. Boisen, S. Dahl, A. A. Gokhale, S.
Kandoi, L. C. Grabow, J. A. Dumesic, M. Mavrikakis, I.
Chorkendorff, J. Catal. 2005, 229, 265.
[52] N. Amadeo, M. Laborde, Trends Chem. Eng. 1996, 3, 159.
M. Boveri, G. Baronetti, M. Laborde, Int. J.
[53] F. Marino,
Hydrogen Energy 2004, 29, 67.
A. Hancz, A. Erdohelyi, Appl. Catal. A 2004, 269,
[54] J. Rasko,
13.
[55] M. A. Vannice, Kinetics of Catalytic Reactions, Springer
Science Business Media Inc., New York 2005.
[56] W. W. Akers, D. P. Camp, AIChE J. 1955, 1, 471.
[57] H. A. Dirksen, C. H. Riesz, Ind. Eng. Chem. Res. 1953, 45,
1562.
[58] N. M. Bodrov, L. O. Apelbaum, M. I. Temkin, Kinet. Catal.
1964, 5, 696.
[59] N. M. Bodrov, L. O. Apelbaum, M. I. Temkin, Kinet. Catal.
1967, 8, 821.
[60] N. M. Bodrov, L. O. Apelbaum, M. I. Temkin, Kinet. Catal.
1968, 9, 1065.
[61] D. W. Allen, E. R. Gerhard, M. R. Likins, Ind. Eng. Chem.
Res. 1975, 14, 256.
[62] J. C. De Deken, E. F. Devos, G. F. Froment, Steam
Reforming of Natural Gas: Intrinsic Kinetics, Diffusional
Influences, and Reactor Design, in Chemical Reaction
Engineering-Boston, J. Wei, C. Georgakis, Eds., American
Chemical Society, Boston 1982, pp. 181197.
[63] J. Xu, G. F. Froment, AIChE J. 1989, 35, 88.
[64] S. S. E. H. Elnashaie, A. M. Adris, A. S. Al-Ubaid, M. A.
Soliman, Chem. Eng. Sci. 1990, 45, 491.
[65] J. R. Rostrup-Nielsen, Catal. Today 1993, 18, 305.
[66] K. Hou, R. Hughes, Chem. Eng. J. 2001, 82, 311.
[67] A. Gritsch, G. Kolios, U. Nieken, G. Eigenberger, Chem. Ing.
Tech. 2007, 79, 821.
[68] J. Wei, E. Iglesia, J. Catal. 2004, 224, 370.
[69] J. R. Rostrup-Nielsen, J. Sehested, J. K. Nrskov, Adv.
Catal. 2002, 47, 65.
[70] A. F. Cunha, A. F. Carrapatoso, Hydrogen Production by
Catalytic Decomposition of Methane, Thesis, Department of
Chemical Engineering. Universidade do Porto, 2009.
[71] H. S. Bengaard, J. K. Norskov, J. Sehested, B. S. Clausen, L.
P. Nielsen, A. M. Molenbroek, J. R. Rostrup-Nielsen, J.
Catal. 2002, 209, 365.
[72] F. Daz Alvarado, F. Gracia, Chem. Eng. J. 2010, 165, 649.
Manuscript received July 5, 2012; revised manuscript
received October 5, 2012; accepted for publication October 5,
2012.

130

THE CANADIAN JOURNAL OF CHEMICAL ENGINEERING

VOLUME 92, JANUARY 2014