Kinetics and dynamics of methanol steam reforming on CuO/ZnO/alumina catalyst

Article abstract of DOI:10.1016/j.apcata.2017.04.012

The kinetic and dynamic behavior of the methanol steam reforming (MSR) over CuO/ZnO/Al₂O₃ catalyst were followed at various steam to carbon (S/C) ratio using temperature ramping and steady state input conditions. The data is required for designing a methanol reformer-internal combustion engine (ICE) system, utilizing the reforming products as fuel for ICE while using the hot exhaust gases to heat the reformer. The catalyst exhibited high activity above 200?°C for MSR and above 250?°C for methanol decomposition reactions, making methanol a good hydrogen vector. The production of undesired byproducts, methyl formate and dimethyl ether, declined with increasing S/C and vanish at S/C?=?1. Rate oscillations were observed during MSR on CuO/ZnO/Al₂O₃ under isothermal conditions and S/C ～1 for the first time. Oscillations with a period of 10?min in order of magnitude, were observed in all MS signals with some phase shift between them. A simple kinetic model was developed for methanol decomposition (MD) using S/C?=?0 data and the rate and activation energies were determined. Calculations using S/C?=?1 data assuming MD followed by WGS, show that at high temperatures the WGS is close to equilibrium.

Full text of DOI:10.1016/j.apcata.2017.04.012

Accepted Manuscript
Title: Kinetics and Dynamics of Methanol Steam Reforming
on CuO/ZnO/alumina Catalyst
Authors: Rajesh Thattarathody, Moshe Sheintuch
PII:
S0926-860X(17)30169-2
DOI:
Reference:
http://dx.doi.org/doi:10.1016/j.apcata.2017.04.012
APCATA 16211
To appear in:
Applied Catalysis A: General
Received date:
Revised date:
Accepted date:
27-2-2017
19-4-2017
21-4-2017
Please cite this article as: Rajesh Thattarathody, Moshe Sheintuch, Kinetics and
Dynamics of Methanol Steam Reforming on CuO/ZnO/alumina Catalyst, Applied
Catalysis A, Generalhttp://dx.doi.org/10.1016/j.apcata.2017.04.012
This is a PDF file of an unedited manuscript that has been accepted for publication.
As a service to our customers we are providing this early version of the manuscript.
The manuscript will undergo copyediting, typesetting, and review of the resulting proof
before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that
apply to the journal pertain.

1
Graphical Abstract
Highlights
 Kinetics and dynamics of MSR over CuO/ZnO/Al2O3 were studied at 0<S/C<1 with the
application of feeding reforming products to combustion engines.
 Isothermal oscillations observed for the first time for this system.
 Increasing S/C led to decline in methyl formate and DME.
 Calculations using a simple model assuming MD followed by WGS for S/C =1 data show
that Ea for MD is 50 kJ/mol and that at high temperatures the WGS is close to equilibrium.
Kinetics and Dynamics of Methanol Steam Reforming
on CuO/ZnO/alumina Catalyst
Rajesh Thattarathody, Moshe Sheintuch*
Department of Chemical Engineering,
Technion- Israel Institute of Technology, Haifa, Israel 32000
* Corresponding author, e-mail: cermsll@technion.ac.il
Abstract
The kinetic and dynamic behavior of the methanol steam reforming (MSR) over
CuO/ZnO/Al2O3 catalyst were followed at various steam to carbon (S/C) ratio using temperature
ramping and steady state input conditions. The data is required for designing a methanol reformer-
internal combustion engine (ICE) system, utilizing the reforming products as fuel for ICE while
using the hot exhaust gases to heat the reformer. The catalyst exhibited high activity above 200 °C
for MSR and above 250 °C for methanol decomposition reactions, making methanol a good
hydrogen vector. The production of undesired byproducts, methyl formate and dimethyl ether,
declined with increasing S/C and vanish at S/C=1.

2
Rate oscillations were observed during MSR on CuO/ZnO/Al2O3 under isothermal
conditions and S/C~1 for the first time. Oscillations with a period of 10 min in order of magnitude,
were observed in all MS signals with some phase shift between them.
A simple kinetic model was developed for methanol decomposition (MD) using S/C=0 data
and the rate and activation energies were determined. Calculations using S/C=1 data assuming MD
followed by WGS, show that at high temperatures the WGS is close to equilibrium.
Keywords: Methanol steam reforming; methanol decomposition; water gas shift reaction; rate
oscillations
1. Introduction
Methanol is a good hydrogen vector that enables distribution in the liquid phase, using
available infrastructure while avoiding the safety and capacity problems associated with pure
hydrogen distribution [1,2]. It can be easily converted to hydrogen by decomposition or by steam
reforming. Methanol can be produced in large quantities from syngas, from natural gas or from
biomass either directly by fermentation or indirectly by gasification; thus, it can be considered as a
renewable energy source. Using the methanol steam reforming (MSR) products as a fuel for
Internal Combustion Engine (ICE) makes it possible to benefit from the hydrogen-enriched gaseous
fuel, to eliminate the known problems of on-board hydrogen storage and to reduce substantially
energy production expenses compared to those obtained with fuel cells. This work, focusing on
methanol SR, is conducted as part of a project aimed to develop a spark-ignition engine fueled by
methanol reforming products for hybrid propulsion system with significantly improved fuel
economy and ultra-low emissions. The efficiency improvement results from the use of ICE
effluents to heat the reformer [3-6].
Methanol steam reforming has been thoroughly studied in recent years [7]. Complete
combustion ensures a great reduction in CO, NOx and volatile hydrocarbons emissions. Prototypes
of on board reforming based on methanol have been exhibited by Toyota. Methanol as a fuel
enables reduction of benzene, butadiene and OMHCE (organic matter hydrocarbon equivalent)
emissions together with a lower photochemical reactivity of emitted pollutants [8]. Some of the
main drawbacks of liquid methanol as an automotive fuel are its relatively low heating value, poor
engine startability at low ambient temperatures, and higher aldehyde emissions.
The overall steam reforming reaction (Eq. 1-3), incorporates methanol decomposition (MD)
and water gas shift (WGS) described by
CH3OH + H2O ⇌ CO2 +3H2
CH3OH ⇌ CO + 2H2
ΔH = 49.7 kJ mol−1
ΔH = 90.2 kJ mol−1
ΔH= −41.2 kJ mol−1
(1)
(2)
CO + H2O ⇌ CO2 + H2
(3)
In fact, for ICE-feed purposes the MD reaction is superior, but it may lead to coking problems.

3
Some side reactions also occur during methanol dissociation, as outlined in the following
equations, leading to the formation of small amounts of dimethyl ether (eq. 4 and 7), methyl formate
(eq. 8), formaldehyde (eq. 9), etc.
2CH3OH ⇌ (CH3)2O +H2O
(dehydration)
(methanation)
(4)
2CH3OH ⇌ CH4 +2H2 + CO2
2CH3OH ⇌ CH4 + CO + H2 + H2O
3CH3OH ⇌ (CH3)2O +CO2+ 3H2
2CH3OH ⇌ HCO2CH3 + 2H2
(5)
(dehydration)
(6)
(7)
(8)
(decomposition)
(dehydrogenation)
CH3OH ⇌ CH2O+H2
CH3OH ⇌ C+H2+H2O
(formaldehyde synthesis)
(carbonation)
(9)
(10)
The catalyst selected must ensure high activity and stability of SR and WGS reactions, at
temperatures lower than those of the exhaust gas. Compared with the strict requirement for the
ultra-pure hydrogen for fuel cells, ICE is much more flexible and can efficiently burn different
mixtures of hydrogen, carbon monoxide and other gases. This feature greatly reduces the cost of
energy obtained from renewable fuels. As explained above methanol decomposition is desired from
energy considerations: Methanol conversion to CO2 and H2 is rather undesirable, because CO2 is a
diluent gas and not an energy carrier. On the other hand, MSR is expected to suffer less from coking
than methanol decomposition.
Two groups of catalysts were employed for this application: copper-based catalysts, which
are inexpensive and active, and group VIII-X catalysts which show better thermal stability and
long-term stability. For economic reasons, copper-based catalysts are preferred [9]. However, these
catalysts have some disadvantages like pyrophoricity and deactivation by thermal sintering, coke
deposition or change in oxidation state [10-12]. The copper crystallites have a tendency to undergo
sintering readily at temperatures >300 °C [13]. Indeed it is suggested that for conventional Cu-
based catalysts, the temperature of operation should not exceed 260 °C. Sintering of Cu is reported
also to be a function of steam concentration. As stated above several byproducts, such as methyl
formate, are formed; these in turn promote deactivation routes via pyrolysis. The deactivation by
CO disproportionation (Boudouard reaction) is considered less favorable in the case of MSR owing
to the higher CO2/CO ratio which minimizes the thermodynamic driving force. The changes in
oxidation state Cu(0)/Cu(I) can also results in decreased activity or changes in selectivity.
A detailed review of MSR studies was published by Sa et al. [9]. Table 1 summarizes the
main finding on Cu-based catalysts and on reaction kinetics. Note that most studies did not employ
the S/C=0 case to derive MD kinetics.
Several studies focused on the effect of catalyst structure: Zirconia was shown to have a
promotional effect on Cu-Zn catalysts by decreasing the CO selectivity [19,20]. In addition, ZrO2
can increase the Cu dispersion [19]. Al2O3 can also increase both the total surface area and Cu
dispersion [20]. Cr2O3 is also reported to stabilize the Cu against sintering [21,22]. The promotional
effects of CeO2 on Cu were reported to be better than that of ZnO or Al2O3 or ZnO-Al2O3 system
under the same reaction conditions [23]. This was attributed to the high dispersion of Cu metal
particles and strong Cu-CeO2 interaction. In the case of CuZnGaOx mixed oxide catalyst [24], the

4
interface between Cu metal-defective oxides plays an important role in determining the activity.
For the recently developed PdZn/ZnO catalyst [25] a strong synergism between active sites on the
PdZn alloy and those on the ZnO support is necessary for excellent catalytic activity and the activity
and selectivity are strongly influenced by the calcination atmosphere of ZnO precursor with
increase in activity from oxidizing to reducing. However the catalyst activity dropped by 24%
during 48 h stability test under MSR conditions. The activity and selectivity of Pd/ZrO2-m
(monoclinic) catalyst was shown further enhanced by adding Cu as a promoter particularly at low
temperature (<220 °C) [26].
1.1. Oscillations on CuO
Reaction rate oscillations were reported for a large class of reactions, since the 1970’s [27-
30]. Since methanol decompose into CO and hydrogen we should note that oxidation reactions of
either CO and H2, are known to oscillate when their mixture with oxygen is catalyzed by noble
metals (Pt, Pd, Rh) as well as by certain base metals. This was shown for supported catalysts or
wires at atmospheric conditions as well as on certain facets under vacuum. It is generally agreed
that the phenomenon in catalytic systems is not due to thermal effects (which may be significant at
atmospheric pressure, but negligible under vacuum). The high heat capacity of catalytic systems
improves their stability; also, the reactions of methanol decomposition and SR are endothermic.
There is no general agreement on the mechanism of oscillations under atmospheric
conditions. The period is usually of minutes in order of magnitude, much longer than the main
reaction time scales (TOF, usually larger than 1 s-1) indicating slow surface modifications, like
oxidation and reduction. The report here of oscillations during the endothermic reactions of
methanol decomposition or steam reforming (followed by WGS) is interesting and to the best of
our knowledge no such reports were published. We review some of the relevant information.
Reports on oscillatory CO oxidation on noble metal are quite common [31,32]; reports on
metal oxide are rare. Eckert et al. [33] observed oscillations of temperature and concentration
during CO oxidation on CuO in a certain region of feed concentrations at temperature of 430 K.
There are quite a few reports on oscillatory behavior of methanol oxidation on Cu catalysts.
Werner et al. [34] studied methanol oxidation on polycrystalline copper predosed with oxygen
(methanol to oxygen molar ratio of 1: 1). Oscillatory behavior was recorded in the temperature
range between two peaks of formaldehyde production. The oscillatory behavior was characterized
as thermochemical. The oscillations were attributed to copper oxidation and reduction.
Chen et al. [35] studied autothermal (i.e., without external heating) methanol reforming on
CuO/ZnO, using a mixture with water and oxygen and observed periodic temperature oscillations
at the four O/C ratios studied (O/C = 0, 0.125, 0.25 and 0.5) with S/C ratios 1, 1.5 and 2, at 250 and
300 °C. Also, the period of temperature oscillations increased as the O/C ratio increases and it is
particularly pronounced at O/C = 0.5. The observation of oscillations during oxidation reactions is
not surprising, but oscillations with O/C=0 (as in our case) require an explanation: one does not
expect the system to sustain itself without the exothermic oxidation reaction, and this is probably a
transient effect, that will disappear after a sufficiently long time.
Koga et al. [36] used copper-zinc oxide catalyst for autothermal methanol reforming with a
molar ratio for methanol/water/oxygen = 1.00/1.50/0.125 at a constant space velocity of 2260 h-1
in the temperature range of 200–295 °C. In the case of the catalyst powder, a large spike of the gas
flow rate with large variations in the temperature was observed once. Both the gas flow rate and

5
the internal temperature oscillated vigorously over time for catalyst pellets, resulting in unstable
hydrogen production.
2. Experimental
All experiments were conducted in a Micromeritics Autochem II 2920 system, having a
quartz U-tube shaped reactor of 1 cm inner diameter where a commercial Cu/ZnO/Al2O3 (Sud-
Chemie, now Clariant) catalyst sample is placed between plugs of quartz wool. A 1/16” k-type
thermocouple is placed inside the tube and is used to monitor and control the sample temperature
by means of a PID controller connected to a ceramic furnace. The gaseous effluent is then analyzed
by a means of a thermal conductivity detector and subsequently introduced into a mass
spectrometer (MKS instrument, Cirrus 2) through an electrically heated fiberglass capillary.
CuO/ZnO/Al2O3 catalyst pellets (3x3 mm) were crushed and sieved to particle sizes of 1-2
mm. About 300 mg catalyst was loaded and the catalytic bed length is about 5 mm. Pretreatment
procedure was as follows: temperature programmed reduction has been carried out prior to each
experiment at 350 °C (heating rate 5 °C) for 1 h under 10% H2 in Ar (total flow 100 ml/min). After
in situ reduction, the catalyst was allowed to cool down to room temperature and was subjected to
a reactive flow of saturated methanol vapors in Ar or of a mixture of methanol and water vapors in
Ar (with10 ml/min or 50 ml/min Ar flow). After observing the oscillations, several runs were
conducted at fixed temperatures.
The Autochem system consists of a built-in vapor generator which contains a flask (capacity
of 50 ml) and a reflux zone which can be heated independently. The carrier gas Ar enters the
Autochem II instrument and is saturated with methanol in the flask to a degree that depends on the
temperature of the reflux zone. Ar carrying particular amount of methanol is then passed through
the catalyst and the effluent gas is then analysed with the help of mass spectrometer. Concentrations
of the main products H2, CO2 and CH4 were determined using the MS signal by calibrating at
various compositions (10, 15 and 20%) of single component in Ar, while CO was calibrated by
passing 10% CO in He.
For methanol decomposition, the flask contains only methanol (40 ml) and the reflux
temperature was fixed at 48 °C to get 50% vapor composition according to the Antoine equation.
The flask was kept at 55 °C. For methanol steam reforming, the liquid composition of methanol
and water were determined using the literature data [37] to get 1:1 vapor composition. According
to the literature, for methanol-water system, a liquid phase methanol mole fraction of 0.14
corresponds to a vapor phase methanol mole fraction 0.508 at a temperature of 357.7 K. So the
amount of water and methanol charged were 30 and 11 ml respectively to restrict the total volume
to ~40 ml. The temperature of the flask was set to 83 °C and the reflux zone was at 79 °C. Similarly
for S/C =0.5, the amount of water and methanol taken were 20 ml each with flask and reflux
temperatures of 75 and 69 °C respectively.
In the case of MSR and Ar flow of 10 ml/min, with S/C=1, the change in feed composition
over the 4hrs of experiment was acceptable. In the case of 50 ml/min, since the feed composition
variation with time could not be ignored, the procedure was modified, as explained below. For both
isothermal and ramp up experiments, Ar was allowed to flow through the solution only after
reaching the initial temperature (250 °C in the isothermal and 130 °C in the ramp up). The liquid
methanol mole fraction was estimated to reduce from 0.14 to 0.127 after 30 min of reaction. We
decided, therefore, for both experiments, to increase the flask and reflux temperatures by 2 °C after
each 30 min till the end of the experiment to compensate for the change in composition.

6
After temperature programmed reaction, temperature programmed desorption (TPD) was
carried out under helium (flow rate 50 ml/min) up to 400 °C (heating rate 5 °C/min) and kept for
half an hour to desorb all the species adsorbed on the catalyst surface.
Temperature-programmed oxidation (TPO) experiments were conducted in order to estimate
the amount of carbon deposited on the catalytic surface. Following reaction and desorption, the
sample is cooled in He atmosphere to room temperature and then TPO with 10% O2 in He was
initiated with a heating rate of 5 °C/min to 700 °C and held for 40 minutes for methanol
decomposition and to 800 °C and held for 30 min in the case of reforming. The MS signal for CO2
was calibrated by injecting 0.5, 1 and 1.5 ml of CO2 from which the amount of carbon deposited on
the catalyst surface was determined.
3. Experimental results
We have studied the effect of temperature and of the steam to carbon (S/C) ratio on kinetics
and dynamics. We found and describe below slow activity changes as well as relatively fast
oscillatory behavior.
3(a). Product distribution and deactivation of CuO/ZnO/Al2O3 catalyst
After hydrogen pretreatment and cooling, the catalyst was subject to a reactive stream made
of Ar (10 ml/min) saturated with methanol to 50 mol% or with methanol and water vapors to 25
% of each reactant , by sweeping the temperature up and down at a rate of 1 °C/min in the range
130 to 350 °C. Three separate experiments were carried out with varying steam to methanol ratio
as 0:1, 0.5:1 and 1:1 (Fig. 1).
From the figure it can be noticed that:
(i)
Reaction starts at around 140 °C and complete conversion of methanol occurs at around
300 °C for S/C= 0.0, at 275 °C at 0.5 while it occurs at around 250 °C for 1:1 ratio. Note
that in S/C=0.5, CO2 builds up before CO does. Obviously more H2 and CO2 and less CO
is obtained in the case of reforming. In all the cases, methane concentrations were less than
1%.
(ii) The MS signal seems to be oscillatory and this is especially strong in the case of S/C=1, as we
elaborate below. Note that beyond 260 °C full methanol conversion is achieved, while the
signals of H2, CO, CO2 oscillate, suggesting that the oscillations are due to WGS. This also
explains the increasing amplitude with increasing steam contents.
(iii) Other than H2, CO, CO2 and CH4, small concentrations of methyl formate and dimethyl ether
were also produced, the amounts of which decline with increasing steam content (Fig. 2).
Since the use of MS analysis cannot be considered accurate, and the compositions are quite
close to those at equilibrium, we study the effect of flow rate by conducting methanol
decomposition experiment with a higher flow rate (Ar flow 50 ml/min) using a reactive streams of
Ar saturated with 50% methanol. The comparison of results with that of Ar flow rate of 10 ml/min
(Fig. 3, both at 50% in feed) showing a decline in methanol, hydrogen and CO composition, as
expected from a system far from equilibrium. The concentrations of methyl formate and dimethyl
ether also decreased, but note also that increasing flow rate increased the peak temperatures of these
components, since these products are highly sensitive to methanol concentration (Fig. 4).

7
Similarly the flow rate effect on methanol steam reforming with S/C =1 was also studied
(fig. 5) and a substantial reduction in the formation of products can be seen.
3(b). Surface composition after temperature scanning
In all the cases, after the reaction temperature scanning experiment (up and down), TPD
was carried out in helium (50 ml/min) up to 400 °C and kept for 30 minutes to remove all the
adsorbed materials from the catalyst surface. After cooling in helium, TPO was carried out with
10% O2/He (50 ml/min) to quantify the amount of coke produced on the catalyst surface during
reaction. In the case of methanol decomposition, the maximum temperature used was 700 °C and
kept at this temperature for 40 minutes while in the other cases, maximum temperature was 800 °C
and kept for 30 min. The results of the TPO experiments for the three different steam to methanol
ratios are shown in Fig. 6.
In all cases, only one peak is observed in the MS CO2 signal, at ~680 °C which can be
attributed to graphite formation [38]. The amounts of coke produced with the three different S/C
ratios (0, 0.5 and 1.0) are quite similar (0.0187, 0.017 and 0.018 g of carbon per g of catalyst,
respectively). This indicates coking of the catalyst, but surprisingly the S/C ratio does not affect
the coking rate.
3(c). Oscillatory behavior
As discussed in Introduction oscillations in methanol oxidation on Cu using air or oxygen
were reported before. Oscillations were reported also in autothermal methanol oxidation/reforming
using a mixture of air and steam and are probably due to the oxidation reaction. To the best of our
knowledge, this phenomenon has not been reported for steam reforming of methanol.
To ascertain that the transient results in Fig. 1 are truly kinetic we studied the behavior under
fixed input conditions (Fig. 7, S/C=1, 250 °C) with a pre-reduced catalyst. It is apparent that the
reaction is oscillatory. Large amplitude oscillation showing upward spikes (of H2, CO2) or
downward spikes of CO are evident. The total conversion is almost complete as evident from the
methanol spikes. It seems that the spike in methanol precedes that of CO2. To show that we plotted
(Fig. 8) the phase plane behavior of methanol (mass 32) vs CO2 (44); if the oscillations were in
phase we expect motion along a line. The phase plane behavior showed a drop in methanol before
a spike of CO2 followed by slow buildup of CO2. The phase diagram of the oscillation is shown in
fig. 9.
The pre-reduced catalyst was subjected to the methanol steam reforming reaction conditions
at various temperatures, 150, 200, 250 and 300 °C for 2h with steam to methanol ratio 1:1 (Fig. 10,
Ar flow 10 ml/min). The signal includes now many frequencies especially at 150 °C (shown on
expanded scale), and the large amplitude oscillations at this temperature are those of methanol and
water. At higher temperatures, the large amplitudes are those of hydrogen and CO2. Repeating the
experiments for S/C=0.5 showed a relatively stable behavior, with phase spikes at 250, 300 and
high frequencies especially at 150 °C (Fig. 11).
4. Discussion
Before analyzing our results we review kinetics modeling results: Several studies
correlated the data with a reaction rate expression, usually presented as a single reaction. Initial
attempts described the kinetics by power law rate expressions (Table 1). Jiang et al. [39] studied
MSR on Cu/ZnO/Al2O3, analyzed the reaction mechanism, and suggested and calibrated the
following Langmuir-Hinshelwood rate equation, assuming that the r-WGS reaction is negligible,

8
K
pM
M
[k
K
(
)(
)]
CH O
3
K
√
pH
√
H
T
2
2
2
)
(11)
(
r =
Cs
R
pM
M
(1+(
)(
)+√K √p )2
H H
2
2
K
√
pH
√
H
2
2
where 퐶푠푇 represents total surface concentration of site (mol m-2). Here only the adsorptions of
methanol and hydrogen have significant effects on the reaction rate. Peppley et al. [40] developed
of a comprehensive model for MSR on Cu/ZnO/Al2O3 catalysts based on an analysis of surface
mechanism, which account for all the three possible overall reactions (Eqns. 1-3). The rate
expressions, derived upon assuming 1) active sites for the decomposition reaction is distinct from
those for the MSR and the WGS reaction, 2) the rate-determining step for both the MSR and the
MD is the dehydrogenation of adsorbed methoxy groups, are
pH3
p
2
p p
p
∗
k
MSR
KCH O(1)
)CsT1CsT1a
2
CO
M
H
(
)
(1 −
3
k
p
√
MSR M W
2
r
=
(12)
MSR
1/2
Den1(1 + K1H/2
p
)
(
)
1a
H
2
p
p
M
W
∗
∗
∗
(
)
2
(
)
)
Den1 = (1 + KCH O(1)
+ KHCOO(1)p
p
√
+ KOH(1)
CO
H
2
3
2
p
√
p
√
H
H
2
p
p
p
p
k
KOH(1)
)CsT12
2
∗
CO W
H
CO
2
(
)
(1 −
WGS
k
p
p
p
WGS CO W
√
H
2
r
=
(13)
(14)
WGS
2
(Den1)
p
pH2
p
2
D M
∗
T
T
M
H
CO
)
)Cs2Cs2a
(
)
(
k
KCH O(2)
(1 −
MD
3
k p
p
√
2
r
=
MD
Den3
p
p
1/2
1/2
∗
∗
M
W
( )
2
( )
2
(
)
(
)
(
)
)(1 + KH
Den3 = (1 + KCH O
+ KOH
p
)
H
2
2a
3
p
p
√
H
√
H
2
2
1/2
/
1
KH(1a) (Ki
( )
1
∗( )
( )
1
where Ki
= Ki
stands for the adsorption coefficient for surface species, i
adsorbed on type 1 site), 퐶푠푇1 represents total surface concentration of type 1 active sites, which
catalyze the elementary surface processes involved in the mechanisms for the MSR and WGS, and
퐶푠푇1 is that of H2 adsorbing type. Similarly CsT2, are that of type 2 sites which catalyse surface
푎
processes for MD and CsT2a that of H2 adsorbing type for MD.
Ilinich et al [41] correlated the MSR rate (rate of methanol consumption) and the overall rate
(MSR followed by r-WGS), on a catalyst containing Pd and Zn on CeO2 as
−0.40
−0.13
−0.075
RMSR = 34.4e−54270/(RT) g(p);
Roverall = 53e−55860/(RT) g(p)
g(p)=pM0.54pW0.10pH2
p
p
(15)
CO
CO
2
(16)
(R is in mol s-1 gPd-1 and the activation energies are 54270 and 55860 J mol-1).
The apparent activation energy for methanol steam reforming was reported in the range 75-100
kJ/mol [40,42-44].
Our kinetic studies have shown that high conversions can be achieved at temperatures above
250 °C for methanol decomposition and 225 °C for methanol steam reforming and the main
products are H2, CO2 and CO. This makes methanol a good energy vector for fueling ICEs while
using the combustion effluents to heat the feed reformer.

9
Results indicate that the model should account for the rates of the two reactions (MD and
WGS), contrary to the single-reaction rate expressions in the literature (Eqn. 11).
To build a simple kinetic model and elucidate its parameters we suggest the following
analysis based on two reactions, MD (decomposition) and CO WGS. We will find first the
parameters of MD and use it then to calculate the WGS parameter. We use a PFR model to analyze
the data for the 5 mm long reactor; the balance on methanol (CM) when S/C=0 (i.e., MD), while
assuming first-order kinetics, is
df
(
F C
)
= k
C = k
C
(1 − f)/(1 + δf)
(17)
M in
,
MD M MD M in
,
dW
yields,
(
)
f
0
k
W
1+f df
MD
F
(
)
=
= −2 ln 1 − f − f
(18)
∫
1−f
where F is feed flow rate (20 or 100 cc/min in this study), CM,in is feed M conc, f is conversion and
the term (1+δf) is due to dilution by products (MCO+2H2), which for this case where the
methanol is diluted in Ar in equal parts is δ=1 (in the absence of such dilution, δ=2). We treat it as
practically irreversible. Since CM is not measured (its MS signal was not calibrated) we find f at
various T from the following procedure. At any steady state CTyH=2CM,inf/(1+) where
CM,in/CT=1/2. In complete conversion CTyHm=2CM,in2/2 where yH2m is the maximal (high T) value.
Thus we find
y
y
H
H
2
2
(
)
)
f = (yH m)/(2 − δ
(19)
y
H m
2
2
Now, kMD is the only unknown and can be estimated from data in Fig 3 (10 and 50 cc min-
1). While the results at the two flow rates differ, there seems to be a compensating effect. The
Arrhenius plot, i.e. plot of lnkMD vs 1/T (Fig. 12), at both flow rates yields similar lines with
activation energies for methanol decomposition of 62 and 39 kJ/mol at 10 and 50 ml/min
respectively. The difference may be due to several effects including: (i) external film diffusion
resistance; (2) consumption of methanol to DME; (3) kinetics that is not linear, or (4) shifting MS
calibration. The first effect would have led to larger rates at the larger flow rates, contrary to results.
There is no independent evidence for different reaction kinetics and the consumption due to DME
is small. Since the change in rates with flow rates is small we may define the kinetics by an
intermediate activation energies of about 50 kJ/mol, compared with literature values reported to be
76-77 kJ/mol [45,46].
Now a similar balance applies for feed of methanol and water where both reactions take
place, and we have determined already the kinetics of MD and need to determine the WGS kinetics.
The balances take the form
(
) = −α k
i MD M
C
− β k
i WGS CO W
(C
C
−
)
F dC
i
C
C
H
CO
2
2
(20)
dW
K
eqWGS
where αi and βi are stoichiometric coefficients of the various components (i=M, CO, W, H2, CO2).
Now,
C
= C
/D − C (y + y
M in T CO CO
,
)
(21)
M
2

10
where the dilution effect, 퐷 = 1 + (2푦
the system is close to equilibrium of WGS by checking the term (C
+ 3푦
)/4 for S/C=1. Now we will check whether
퐶푂
2
퐶푂
C
)/(C
C
). All
CO2 H2
CO W
concentrations were determined by calibrating the MS except water concentration which is
calculated from,
C
= C
/D − C y
W in
,
(22)
W
T CO
2
The rate constant kWGS could be determined for the data with Ar feed rate 50 ml/min
(concentrations at small flow rates almost vanished). We can use the hydrogen data (Eqn. 20, α=2,
β=1) using the equilibrium constants of the WGS reaction given in the literature [47]. Fig. 13
compares the calculated C
C
/C term with Keq,WGS showing that the system is close to
C
CO
H
2
CO W
2
equilibrium at high temperatures, but far from it, at the lowest 3 temperature points; the Arrhenius
plot of kWGS, (not shown), using a CSTR model with the three low temperature data points yields
an activation energy of 38 kJ/mol for WGS reaction (the difference between the PFR and CSTR
model at the low conversion end is small).
We
note that one argument leading the MSR research is whether one should consider the kinetics as
SR to CO2 followed by r-WGS [15,16,39,40,48,49] or as MD followed by WGS as was considered
here and in [46,50-52]. At high temperatures the two schemes are indistinguishable as WGS is at
equilibrium. But at lower temperatures (Fig. 13), the observation of CCO lower than its WGS
2
equilibrium values support the MD-WGS approach. Inspection of the hydrogen yield (moles of
hydrogen produced/moles of CO and CO2 formed) in Fig. 5 (right) reveals that it is ~2.25, 2.25, 2.0
at 250, 200 and 150 °C (Fig. 14), respectively, which support a MD-WGS mechanism at an Ar flow
of 50 ml/min. At low temperature the yield corresponds to that of MD, while at high temperature,
where the system is close to WGS equilibrium the yield somewhat higher than 2. If MS is the first
step one would expect yield of almost 3 at low temperature.
Oscillatory behavior was observed for the first time for this reaction. As discussed in the
Introduction at atmospheric pressure this behavior is common for oxidation reactions on noble
metal catalysts (wires or supported). There is no acceptable mechanism to account for it at high
pressures, as opposed to studies at low pressures on well defined planes. In spite of the many studies
of this intricate behavior, its effect on reactor design is not clear and we have chosen not to pursue
this line of research.
5. Conclusions
The system of a methanol reformer feeding an ICE with a hydrogen enriched stream while
the otherwise wasted hot engine exhaust gases are used to heat the endothermic reforming reaction,
can eliminate the known problems of on-board hydrogen storage and the economic problems
associated with fuel cells. We carried out a kinetic study of methanol steam reforming over a
commercial catalyst CuO/ZnO/Al2O3 with three different S/C ratio 0 (MD), 0.5 and 1, at
atmospheric pressure and at temperatures in the range from 140 to 350 °C, sweeping the
temperature up with a very slow rate. High methanol conversions are observed above 250 °C for
MD, while the MSR activities were high above 200 °C, making methanol a good hydrogen vector.
It was observed that increasing water/methanol ratio increased the production of hydrogen and
decreased the production of methyl formate and dimethyl ether. The concentrations of methane
obtained were less than 1% in all cases.

11
Rate oscillations were observed during MSR on CuO/ZnO/Al2O3 under isothermal
conditions and S/C~1 for the first time. Oscillations with a period of 10 min in order of magnitude
were observed in all MS signals with some phase shift between them. While rate oscillations were
previously reported for methanol oxidation and for autothermal reforming (i.e., in presence of
oxygen), conclusive evidence for the same in the case of MSR were not reported.
Upon increase in the feed flow rate all the products concentrations decline including methyl
formate and dimethyl ether during MD. The amount of coke deposits on the surface of the catalyst
was found to be similar in all cases of S/C ratio 0, 0.5 and 1 and is graphitic.
A simple kinetic model was developed for methanol decomposition (MD) using S/C=0 data
and the rate and activation energies were determined. This calibration step was missing in most
previous studies. This step is important for our reformer-ICE application, since a 2H2+CO may be
superior to that of 3H2+CO2. Calculations using S/C=1 data assuming MD followed by WGS, show
that at high temperatures the WGS is close to equilibrium. Most previous studies either calibrated
a single rate expression, or calibrated a scheme of MSR followed by r-WGS. We show evidence
that supports our suggested MD-WGS, which mechanistically seems to be simpler.
Acknowledgement
This research was supported by the I-CORE Program of the Planning and Budgeting Committee
and The Israel Science Foundation (Grant No. 152/11). The Micromeritics instrument was
supported by the Wolfson Foundation and by the I-Core project of the ISF through The Nancy and
Stephen Grand Technion Energy Program (GTEP).
References
[1] W. Cheng, Acc. Chem. Res. 32 (1999) 685-691.
[2] G. A. Olah, Angew. Chem. Int. Ed. 44 (2005) 2636 –2639.
[3] Z. Ivanic, F. Ayala, J. Goldwitz, J. Heywood, SAE Tech. Pap. 2005-01-0253, 2005.
[4] A. Poran, M. Artoul, M. Sheintuch, L.M. Tartakovsky, SAE Int. J. Engines, 7 (2014) 234-
242.
[5] V. K. Chakravarthy, C. S. Daw, J. A. Pihl, J. C. Conklin, Energy Fuels 24 (2010) 1529–
1537.
[6] A. Poran, L. Tartakovsky, Energy 88 (2015) 506-514.
[7] F. Joensen, J.R. Rostrup-Nielsen, J. Power Sources 105 (2002) 195-201.
[8] K. Cheekatamarla, C.M. Finnerty, J. Power Sources 160 (2006) 490-499.
[9] S. Sa, H. Silva, L. Brandaoa, J.M. Sousa, A. Mendes, Appl. Catal. B: Environ. 99 (2010)
43–57.
[10] T. Valdes-Solis, G. Marbán, A.B. Fuertes, Catal. Today 116 (2006) 354–360.
[11] W. Cao, G. Chen, S. Li, Q. Yuan, Chem. Eng. J. 119 (2006) 93–98.
[12] Y. Choi, H.G. Stenger, Appl. Catal. B 38 (2002) 259.

12
[13] M.V. Twigg, M.S. Spencer, Top. Catal. 22 (2003) 191–203.
[14] W. Gu, J.P. Shen, C.H. Song, Prepr. Pap.-Am. Chem. Soc., Div. Fuel Chem. 48 (2003)
804-807.
[15] H. Purnama, T. Ressler , R.E. Jentoft, H. Soerijanto, R. Schlögl, R. Schomäcker, Appl.
Catal. A: Gen. 259 (2004) 83-94.
[16] J.K. Lee, J.B. Ko, D.H. Kim, Appl. Catal. A: Gen. 278 (2004) 25-35.
[17] B. Lindstrom, L. J. Pettersson, Int. J. Hydrogen Energy 26 (2001) 923–933.
[18] J. Agrell, H. Birgersson, M. Boutonnet, J. Power Sources 106 (2002) 249–257.
[19] H. Jeong, K.I. Kimb, T.H. Kimb, C.H. Ko, H.C. Park, I.K. Song, J. Power Sources 159
(2006) 1296–1299.
[20] J. Agrell, H. Birgersson, M. Boutonnet, I. Melián-Cabrera, R.M. Navarro, J.L.G. Fierro, J.
Catal. 219 (2003) 389–403.
[21] L. Ma, B. Gong, T. Tran, M.S. Wainwright, Catal. Today 63 (2000) 499–505.
[22] X. Huang, L. Ma, M.S. Wainwright, Appl. Catal. A 257 (2004) 235–243.
[23] Y. Liu, T. Hayakawa, T. Tsunoda, K. Suzuki, S. Hamakawa, K. Murata, R. Shiozaki, T.
Ishii, M. Kumagai, Top. Catal. 22 (2003) 205–213.
[24] W. Tong, K. Cheung, A. West, K. Yu, S. C. E. Tsang, Phys. Chem. Chem. Phys. 15 (2013)
7240-7248.
[25] K. M. Eblagon, P. H. Concepcion, H. Silva, A. Mendes, Appl. Catal. B: Environ. 154–155
(2014) 316–328.
[26] C.S.R. Azenha, C. Mateos-Pedrero, S. Queiros, P. Concepcion, A. Mendes, Appl. Catal.
B: Environ. 203 (2017) 400–407.
[27] M. Sheintuch, R.A. Schmitz, Catal. Rev-Sci. Eng. 15 (1977)107-172.
[28] M.G. Slinko, M.M. Slinko, Catal. Rev-Sci. Eng. 17(1978) 119-163.
[29] R. lmbihl, G. Ertl Chem. Rev. 95 (1995) 697-733.
[30] M.M. Slinko, N.I. Jaeger, Oscillating Heterogeneous Catalytic Systems, in: Studies
in Surface Science and Catalysis, Elsevier Science, 1994, vol. 86, pp. 1-387.
[31] F. Schuth, B.E. Henry, L.D. Schmidt, Adv. Catal. 39 (1993) 51-127.
[32] J. Singh, M. Nachtegaal, E.M.C. Alayon, J. Stotzel, J.A. van Bokhoven, ChemCatChem
2010, 2, 653–657.
[33] E. Eckert, V. Hlavacek, M. Marek, Chem. Eng. Commun. 1(1973) 95-102.
[34] H. Werner, D. Herein, G. Schulz, U. Wild, R. Schlogl, Catal. Lett. 49 (1997) 109-119.
[35] W. Chen, Y.Syu, Int. J. Hydrogen Energy 36 (2011) 3397 -3408.
[36] H. Koga, S. Fukahori, T. Kitaoka, A. Tomoda, R. Suzuki, H. Wariishi, Appl. Catal. A: Gen.
309 (2006) 263–269.
[37] K. Kurihara, M. Nakamichi, K. Kojima, J. Chem. Eng. Data 38 (1993)446-449.
[38] C.H. Bartholomew, Catal. Rev.-Sci. Eng. 24 (1982) 67.
[39] C.J. Jiang, D.L. Trimm, M.S. Wainwright, N.W. Cant, Appl. Catal. A 97 (1993) 145-158.
[40] B.A. Peppley, J.C. Amphlett, L.M. Kearns, R.F. Mann, Appl. Catal. A 179 (1999) 31-49.
[41] O. M. Ilinich, Y. Liu, E. M. Waterman R. J. Farrauto, Ind. Eng. Chem. Res. 52 (2013)
638−644.
[42] T. Takeguchi, Y. Kani, M. Inoue, K. Eguchi, Catal. Lett. 83 (2002) 49-53.
[43] B. Frank, F.C. Jentoft, H. Soerijanto, J. Krohnert, R. Schlogl, R. Schomacker, J. Catal. 246
(2007) 177–192.
[44] Geissler, E. Newson, F. Vogel, T.-B. Truong, P. Hottinger, A. Wokaun, Phys. Chem.
Chem. Phys. 3 (2001) 289-293.
[45] P. Mizsey, E. Newson, T. Truong, P. Hottinger, Appl. Catal. A: Gen. 213 (2001) 233–237.
[46] E. Santacesaria, S. Carra, Appl. Catal. 5 (1983) 345-358.

13
[47] J. M. Moe, Chem. Eng. Prog. 58 (1962) 33-36.
[48] K. Takahashi, N. Takezawa, H. Kobayashi, Appl. Catal. 2 (1982) 363–366.
[49] C.J. Jiang, D.L. Trimm, M.S. Wainwright, N.W. Cant, Appl. Catal. A, 97 (1993) 245–255.
[50] V. Pour, J. Barton, A. Benda, Coll. Czech. Chem. Commun. 40 (1975) 2923.
[51] J.C. Amphlett, M.J. Evans, R.F. Mann, R.D. Weir, Can. J. Chem. Eng. 63 (1985) 605–611.
[52] V. Pour, J. Barton, A. Benda, Collect. Czech. Chem. Commun. 40 (1975) 2923–2934.

14
Figure 1. Methanol product composition during decomposition (S/C=0, feed 50% in Ar, left column) or SR
(S/C=0.5, middle (33% methanol and 17% steam) and 1, right, (25% methanol and 25% steam)); methanol
(mass 32, red) and water (18, grey) MS signals are shown on upper row while the lower row represents the
volume percentages of the main products obtained after calibration of H2(2, black), CO(28, green) and
CO2(44, blue). (Ar flow 10ml/min). The MS signals of H2O and CH3OH in data of S/C=1 were smoothed for
clarity; smoothing was not applied in any other case of this work unless specified.

15
Figure 2. MS spectra corresponding to methyl formate and dimethyl ether during steam reforming of
methanol with different steam to methanol ratios.

16
Figure 3. Flow rate effect: Comparison of product composition during methanol decomposition (50%
methanol in Ar, S/C = 0) with Ar flow of 10 ml/min (left), and 50 ml/min (right): note that H2, CO, and CO2
are denoted as vol% on left scale while other signals by a.u. on right.

17
Figure 4. MS spectra corresponding to methyl formate (Ar flow, black: 10 ml/min and grey: 50 ml/min) and
dimethyl ether (Ar flow, blue: 10 ml/min and light blue: 50 ml/min) during methanol decomposition.

18
Figure 5. Flow rate effect: Comparison of product composition during methanol steam reforming (S/C = 1)
with Ar flow of 10 ml/min (left), and 50 ml/min (right): H2, CO and CO2 are denoted as vol% on left scale
while other signals by a.u. on right. The MS signals corresponding to H2O (left and right) and CH3OH (left)
were smoothed for clarity.

19
Figure 6. MS spectra corresponding to CO2 during TPO under 10% O2/He (flow rate 50 ml/min) after
methanol steam reforming with three different steam to methanol ratios.

20
Figure 7. MS results obtained during methanol steam reforming with steam to methanol ratio 1:1 (300 mg
catalyst, Ar flow through methanol-water mixed solution was 10 ml/min, left, or 50 ml/min, right).

21
Figure 8. Phase plane plot showing counter-clockwise oscillations in plane of CO2 MS signal vs methanol
signal observed during methanol steam reforming over Cu/ZnO/Al2O3 with steam to methanol ratio 1:1 at
250 °C (using data of 60-120 mins of the plot shown on the left of Fig. 7).

22
Figure 9. Phase diagram of oscillation for methanol steam reforming.

23
Figure 10. MS results during methanol steam reforming on Cu/ZnO/Al2O3 with steam to methanol ratio 1:1
at various temperatures (Ar flow was 10 ml/min).Note that data for 150 °C is shown on expanded scale.

24
Figure 11. MS results for methanol steam reforming on Cu/ZnO/Al2O3 with steam to methanol ratio 1:0.5 at
various temperatures. (Ar flow was 10 ml/min). Data for 150 °C is shown on expanded scale.

25
Figure 12. Arrhenius plot of methanol decomposition rate constant over CuO/ZnO/Al2O3 at different feed
rates (kMD in cc g-1 s-1).

26
Figure 13. Calibration of WGS parameter: (The plot of ln[(C
C
)/(C
C
)] vs 1/T, calculated with an
CO2 H2
CO W
Ar feed rate of 50 ml/min, in comparison with the equilibrium coefficient of WGS reaction.

27
Figure 14. Plot of hydrogen yield vs temperature using data in figure 5 (right).

28
Table 1. Catalyst performance for MSR
Catalyst T, P S/C W/F
Main results
kg(cat.) s (M=CH3OH; W=H2O)
mmol-1
Cu/ZnO/Al2O3 110-360 °C, 0-
0.018-
Reduction of methyl formate,
DME and methane upon addition
of H2O
(Sud-Chemie)
[12]
1 atm
1.75 0.29
Cu/ZnO/Al2O3 180-270 °C, 1.4
(3 Sud-Chemie 1 atm
and 1 Synetix)
0.023
Above 230 °C, all catalysts
exhibited similar activities with
methanol conversion >90%
[14]
Cu/ZnO/Al2O3 230-300 °C,
1
Full conversion achieved at
W/F>0.01 and at T>250 °C.
0-0.025
(Sud-Chemie)
[15]
1 atm
rSR = k1PmMPnW m = 0.6; n = 0.4
Reaction rate depends on PM and
Cu/ZnO/Al2O3 160-260 °C, 1-2
(synetix 33-5) 1 atm
[16]
0.0067
PH2, and not on PCO, PCO PW.
2
Cu/Zn/Al2O3
(with Cr or Zr 1 atm
promoter) [17]
180-320 °C, 1; in 0.0081
Higher Cu content leads higher
conversion and selectivity. Use of
ternary components increases CO2
selectivity
N2
Cu/ZnO/Al2O3 175-350 °C, 1.3
0-0.025
Production
of
appreciable
(Sud-Chemie)
[18]
1 atm
amounts of CO at low contact
times and high conversions
(S/C=1), suppressed by increased
S/C ratio.
Cu/ZnO/Ga2O3 150 and 195
2
0.0062
0.083
Lower T and shorter contact time
reduces the CO level, trend
opposite to CuZnOx based
catalyst
Enhanced activity and selectivity
particularly at low temperature
(<220 °C)
[24]
°C, 1 atm
CuPd/ZrO2-m
[26]
<220 °C, 1
atm
1.5