Ethanol steam reforming over Rh and Pt catalysts: Effect of temperature and catalyst deactivation

Article abstract of DOI:10.1039/c2cy20703f

In this study 0.2% Rh/alumina and 0.2% Pt/alumina catalysts were tested for ethanol steam reforming (ESR) over a range of temperatures (773-873 K) at 20 barg with a 5:1 steam to ethanol ratio. Hydrogen was always the main product over Rh/Al₂O₃ although liquid products such as acetaldehyde, diethyl ether and acetone were also produced (<12%). Pt/Al ₂O₃ also produced hydrogen as the main product but at 773 K significant levels of ethene were formed. Less liquid product was formed over the Pt/Al₂O₃ (<5%). Hydrogenation of ethene to ethane was inhibited and an activation energy comparable with low temperature studies was obtained for rhodium (～42 kJ mol^-1) however the platinum catalyst gave a low activation energy more typical of a diffusion controlled system (～20 kJ mol^-1). At 873 K the platinum catalyst is more active (90% conversion cf. ～60%) and gives a higher hydrogen selectivity (55% cf. 49%) than the rhodium catalyst. Temperature-programmed oxidation (TPO), Raman spectroscopy, BET and SEM analysis were used to characterise the nature of the coke species for reactions at 773 K, 823 K and 873 K. The TPO results indicated that different types of coke were deposited on both catalysts during ESR. Carbon nanotubes and filamentous coke were observed on the catalysts after ESR at 873 K. Raman analysis revealed that the coke deposited on the catalysts was graphitic in nature and the disorder in the graphitic type coke generally increased with an increase in the reaction temperature.

Full text of DOI:10.1039/c2cy20703f

Catalysis
Science & Technology
View Article Online
View Journal
PAPER
Ethanol steam reforming over Rh and Pt catalysts:
effect of temperature and catalyst deactivation
Cite this: DOI: 10.1039/c2cy20703f
Muhammad Bilal and S. David Jackson*
In this study 0.2% Rh/alumina and 0.2% Pt/alumina catalysts were tested for ethanol steam reforming
(ESR) over a range of temperatures (773–873 K) at 20 barg with a 5 : 1 steam to ethanol ratio. Hydrogen
was always the main product over Rh/Al2O3 although liquid products such as acetaldehyde, diethyl ether
and acetone were also produced (o12%). Pt/Al2O3 also produced hydrogen as the main product but at
773 K significant levels of ethene were formed. Less liquid product was formed over the Pt/Al2O3 (o5%).
Hydrogenation of ethene to ethane was inhibited and an activation energy comparable with low
temperature studies was obtained for rhodium (B42 kJ molÀ1) however the platinum catalyst gave a low
activation energy more typical of a diffusion controlled system (B20 kJ molÀ1). At 873 K the platinum
catalyst is more active (90% conversion cf. B60%) and gives a higher hydrogen selectivity (55% cf. 49%)
than the rhodium catalyst. Temperature-programmed oxidation (TPO), Raman spectroscopy, BET and
SEM analysis were used to characterise the nature of the coke species for reactions at 773 K, 823 K and
873 K. The TPO results indicated that different types of coke were deposited on both catalysts during
ESR. Carbon nanotubes and filamentous coke were observed on the catalysts after ESR at 873 K. Raman
analysis revealed that the coke deposited on the catalysts was graphitic in nature and the disorder in
the graphitic type coke generally increased with an increase in the reaction temperature.
Received 16th October 2012,
Accepted 23rd November 2012
DOI: 10.1039/c2cy20703f
www.rsc.org/catalysis
The main products for steam reforming of ethanol are
hydrogen, carbon monoxide and carbon dioxide, however the
Introduction
Steam reforming has been used for over fifty years as a means of
producing synthesis gas and today the steam reforming of methane
is the principal method used for hydrogen production.1 Over that
period of time the feedstock has changed from naphtha to
methane and in future we can envisage steam reforming of bio-
mass as a major process. To prepare for such a significant change
in feedstock there has been considerable research into different
catalyst systems2–5 utilising different bio-derived feedstocks.6–10
The formation of hydrogen/syn gas from biomass derived sources
is perceived as promising future technology because the raw
material used is renewable. Among the different renewable
sources, a keen interest has been taken in ethanol over the last
few years as an ethanol feedstock is considered renewable and
environmentally benign. Also a higher hydrogen yield per mole
can be obtained from steam reforming when ethanol is used as a
reactant (6 : 1) compared to methanol (3 : 1) or methane (4 : 1).
reaction pathway of steam reforming of ethanol is complex and
varies significantly with reaction parameters such as tempera-
ture, steam to ethanol ratio, catalyst, support and pressure,
equally other reactions such as dehydration, dehydrogenation,
cracking, water gas shift and methanation can also occur. Some
of the reactions that can occur during ethanol steam reforming
(ESR) are detailed below:
C2H5OH + H2O - 2CO + 4H2 DH298 = +256 kJ molÀ1
(2)
C2H5OH - CH3CHO + H2 DH298 = +69 kJ molÀ1
CH3CHO - CH4 + CO DH298 = À19 kJ molÀ1
(3)
(4)
CH3CHO + H2O - 3H2 + 2CO DH298 = +187 kJ molÀ1
(5)
2CO - CO2 + C DH298 = À172 kJ molÀ1
CO + H2O - CO2 + H2 DH298 = À41 kJ molÀ1
(6)
(7)
C2H5OH + 3H2O - 2CO2 + 6H2 DH298 = +174 kJ molÀ1
(1)
Ideally for hydrogen production, ethanol steam reforms
and forms CO2 and H2 as shown in eqn (1). However as can
be seen in eqn (2) ESR can also deliver an ideal syn-gas for
Centre for Catalysis Research, WestCHEM, School of Chemistry, University of
Glasgow, Glasgow G12 8QQ, Scotland, UK. E-mail: david.jackson@glasgow.ac.uk
c
This journal is The Royal Society of Chemistry 2012
Catal. Sci. Technol.

View Article Online
Paper
Catalysis Science & Technology
Fischer–Tropsch synthesis. If the reaction proceeds through
The molar selectivity for the dry gas was calculated as
acetaldehyde then subsequent decomposition can give rise to follows;
methane and carbon monoxide (eqn (4)).
selectivity of X = (mmoles of X/[S(mmoles of all gas phase
Previously we had examined ESR over 0.2% Ru/Al2O3,11 in products)]).
this work we wished to extend the study to include rhodium
Post-reaction catalyst samples were analysed using BET,
and platinum. Ethanol steam reforming over rhodium and Raman spectroscopy, SEM and TGA-DSC with evolved gas
platinum has been studied by other groups4,12–16 but there analysis. BET surface areas and pore volume of pre- and post-
have been no studies at moderate pressure and extended reaction catalysts were measured using
a Micromeritics
periods considering activity, selectivity and deactivation. In this Gemini III 2375 Surface Area Analyser. Prior to analysis,
paper we have examined the activity and selectivity of 0.2 wt% between 0.04–0.05g of catalyst was placed into a vial and purged
Rh/Al2O3 and 0.2 wt% Pt/Al2O3 catalysts at 20 barg and tem- under a flow of nitrogen (30 ml minÀ1) over night at 383 K to
peratures between 773 K and 873 K. The catalysts were run for remove moisture and any physisorbed gases from the catalyst
100 h and the deactivation and surface species were also sample. Raman spectra of post reaction catalysts were obtained
studied.
with a Horiba Jobin Yvon LabRAM High Resolution spectro-
meter. A 532 17 nm line of a coherent Kimmon IK series He–Cd
laser was used as the excitation source for the laser. Laser light
was focused for 10 seconds using a 50Â objective lens and
Experimental
The catalysts used in this project were low metal loading (0.2%) grating of 600. The scattered light was collected in a back-
Rh/Al2O3 and Pt/Al2O3 prepared via incipient wetness impreg- scattering configuration and was detected using nitrogen
nation. The alumina was determined by XRD analysis to be a cooled charge-coupled detector. A scanning range of between
mixture of g- and d-alumina phases. The metal salts used were 100 and 4100 cmÀ1 was applied. SEM images of the post
Rh(NO3)3 and H2PtCl6. After drying, the catalysts were calcined reaction catalysts were obtained using a Philips XL30 Environ-
at 723 K for 4 h. The BET surface area of the calcined catalysts mental SEM. The sample was irradiated with a beam of
was 104 Æ 3 m2
g
À1, while the metal dispersions were deter- electrons, which was followed by changing magnification and
mined by hydrogen chemisorption as 4% for Rh/alumina and focusing for increasing resolution of the catalyst surface.
18% for Pt/alumina. The catalysts were crushed and particle Temperature Programmed Oxidation (TPO) was carried out
sizes between 600 to 425 mm were used for the steam reforming on post-reaction samples using a combined TGA/DSC SDT
of ethanol reactions.
Q600 thermal analyser connected to an ESS Mass Spectrometer
Ethanol steam reforming was performed in a continuous for evolved gas analysis. Each sample was heated from room
flow high pressure microreactor. Prior to the reaction, catalysts temperature to 1273 K using a heating ramp of 5 deg minÀ1
(0.25 g) were reduced in situ at 873 K for 2 hours using hydrogen under 2% O2/Ar at a flow rate of 100 ml minÀ1
.
gas at a flow rate of 50 ml minÀ1. The reactor was then purged
with argon and the temperature set to the desired reaction
temperature. The pressure in the reactor was increased to
Results
20 barg using argon. The water–ethanol solution was mixed The alumina support was tested in the absence of metal and
in order to obtain steam to ethanol ratio of 5 : 1 in the gas was found to be active. The main gaseous products at 873 K
phase. This ratio was chosen as it is close that found with were hydrogen (44%), methane (28%), carbon dioxide (14%)
fermentation ethanol. The ethanol–water solution was intro- and carbon monoxide (13%): only trace amounts of ethene and
duced through a vaporizer set at a temperature of 773 K. The liquid products (B1% acetone) were detected. The conversion
gas flow rate of the steam/ethanol was set at 760 ml minÀ1
,
of ethanol was always >98% over the 100 h TOS.
which was generated by pumping the liquids through a Gilson
pump at a rate of 0.412 ml minÀ1. The argon gas flow rate was
Rh/Al2O3 catalyst
set at 10 ml minÀ1 giving an overall GHSV of 50 000 hÀ1. Once ESR was carried out over Rh/Al2O3 at 773 K, 823 K and 873 K.
all the reaction parameters had been fixed, analysis was begun The conversion of ethanol at different temperatures was com-
by flowing reactants from vaporizer to reactor. The eluant from pared and is shown in Fig. 1. The conversion of ethanol was
the reactor tube in gaseous form entered a knockout pot where Z 99% at all temperatures in the initial 1.5 hours TOS, however
high boiling point products were collected and analysed by a after B24 h TOS, at 773 and 823 K, the conversion of ethanol
Trace GC-2000 Series using a Zebron column and FID detector. decreased to B83% before stabilising, whereas at 873 K the
The temperature of knockout pot was kept at 273 K. The conversion of ethanol did not stabilise, gradually decreasing
gaseous products were analysed by an on-line Varian GC 3400 until, at the end of the reaction (100 h TOS), it had reduced
using a TCD detector and a carboxenTm1010 plot column. The to 60%.
reaction was performed at 773 K, 823 K and 873 K for 100 hours
The liquid and gaseous products formed during the steam
time on stream. Ethanol conversion was calculated as follows; reforming of ethanol over Rh/Al2O3 at 773 K, 823 K and 873 K
conversion = [(ethanol in – ethanol out)/ethanol in]
The yields for liquid products were calculated as follows;
yield = (mmoles of product X/mmoles of ethanol in)
are shown in Fig. 2–4 and Table 1.
Fig. 2 illustrates that during the ethanol steam reforming reaction
at 773 K for the first 24 h ethene was a significant product.
c
Catal. Sci. Technol.
This journal is The Royal Society of Chemistry 2012

View Article Online
Catalysis Science & Technology
Paper
Fig. 1 Conversion (%) of ethanol over Rh/Al2O3 at 773 K, 823 K and 873 K.
Fig. 2 Molar selectivity of dry gas over Rh/Al2O3 at 773 K.
However the selectivity to ethene decreased throughout the
TPO analysis was performed on the Rh/Al2O3 catalyst to
reaction. All other gaseous products, except hydrogen, had identify the type and amount of coke deposition during ESR at
selectivities below 8%. At 823 K and 873 K, as at 773 K, 773 K, 823 K and 873 K. The TGA profile of the post-reaction
hydrogen had the highest selectivity but by 873 K hydrogen catalyst at each temperature is shown in Fig. 5. The carbon
selectivity had decreased slightly while the minor gaseous dioxide evolution profiles match the derivative curve. Three
components had all increased with methane having a selectiv- evolutions of carbon dioxide can be observed at all tempera-
ity B15%. The ethene : ethane ratio at each temperature tures with the peak maxima shifting to higher temperature with
suggests that the hydrogenation is not under equilibrium higher reaction temperature. The peak maxima for carbon
control.
dioxide evolution, for the catalyst used at 873 K, were 717 K,
Acetaldehyde was always the major liquid product but yields 960 K and 1051 K. The amount of carbon removed is reported
were low (o11%) and for other liquid products the yields were in Table 2. The formation of coke significantly decreased the
below 3% (Table 1).
BET surface area and pore volume. The Raman spectra for the
c
This journal is The Royal Society of Chemistry 2012
Catal. Sci. Technol.

View Article Online
Paper
Catalysis Science & Technology
Fig. 3 Molar selectivity of dry gas over Rh/Al2O3 at 823 K.
Fig. 4 Molar selectivity of dry gas over Rh/Al2O3 at 873 K.
spent Rh/Al2O3 catalyst after ESR at 773 K, 823 K and 873 K observed. However, at 873 K only large diameter filamentous
reveals characteristic bands for graphitic-type carbon as shown carbon deposits were observed, which were scattered over the
in Fig. 6. The ratio of the band intensities is reported in Table 2. whole catalyst surface.
The SEM images show that over the Rh/Al2O3 catalyst, with
an increase in the reaction temperature, a remarkable change
Pt/Al2O3 catalyst
took place in the coke morphology. Fig. 7 illustrates that the The conversion of ethanol over the Pt/Al2O3 catalyst at 773 K,
post-reaction sample, ex-773 K, produced a different size of 823 K and 873 K is shown in Fig. 8. Initially the conversion of
filamentous type carbon. The increase in the reaction tempera- ethanol was >97% at all temperatures, over the next 30 h, at
ture from 773 K to 823 K produced two types of filamentous each temperature, conversion decreased before stabilising for
carbon: a long small diameter filamentous carbon and a large the rest of the reaction.
diameter helicoid type carbon. For further clarification, other
Fig. 9–11 show the dry gas selectivities, while Table 3 reports
particles of the sample were scanned and similar results were the yields for the liquid products over Pt/Al2O3 at 773 K,
c
Catal. Sci. Technol.
This journal is The Royal Society of Chemistry 2012

View Article Online
Catalysis Science & Technology
Paper
Table 1 Yield of liquid products over Rh/Al2O3 as a function of temperature and
Table 2 Post-reaction BET analysis and weight loss of Rh/Al2O3 at 773 K, 823 K
time on stream
and 873 K
Yield of liquid products (%)
Temp. (K) TOS (h) Acetald.a DEEa Acetone AAcida 1,1-DEEa
BET surface
Pore volume Weight loss
Temperature (K) area (m2 gÀ1
)
(cm2 gÀ1
)
in TPO (%) (ID/IG)
Reduced (873)
106
47
50
0.46
0.04
0.04
0.03
—
41
41
41
—
773
823
873
24
48
72
100
24
48
72
100
24
48
72
10.2
7.2
6.3
4.6
7.2
4.7
3.9
3.4
9.6
9.7
9.8
10.6
2.8
2.2
2.1
1.8
2.4
2.5
2.6
2.5
1.9
2.4
2.0
1.9
1.4
0.8
0.7
0.6
1.5
1.0
1.0
0.8
1.9
1.6
1.5
1.5
0.1
0.1
0.1
0.1
0.2
trb
tr
tr
tr
tr
tr
0.3
0.4
0.5
0.5
0.6
0.6
0.5
0.5
1.0
1.2
1.1
1.3
773
823
873
0.91
1.04
0.93
45
with the peak maxima shifting to higher temperature with
higher reaction temperature. From the catalyst used at 873 K
for ESR, the peak maxima for carbon dioxide evolution were
890 K and 994 K. The amount of carbon removed is reported in
Table 4. The Raman spectra for the spent Pt/Al2O3 catalysts,
after ESR at 773 K, 823 K and 873 K, reveal characteristic bands
for graphitic-type carbon at 1337 cmÀ1 and 1578 cmÀ1 as shown
in Fig. 13. Table 4 reports the extent of carbon removed during
the TPO, the BET analysis and the ID/IG ratios from the Raman
100
tr
a
Acetald. = acetaldehyde, DEE = diethyl ether, AAcid = acetic acid,
b
1,1-DEE = 1,1-diethoxy ethane. tr = trace.
823 K and 873 K. As expected hydrogen exhibited the highest spectra. Clearly coke formation significantly decreased the BET
selectivity at all temperatures however it was lower than that surface area and pore volume.
found with Rh. Whereas the selectivity to carbon dioxide
The SEM images in Fig. 14 show the spent Pt/Al2O3 catalyst
noticeably increased from that found with Rh. The selectivity after use at 773 K and 873 K. From the images, it can be seen
to ethane was low and suggests that the hydrogenation of that no filamentous coke was present on the catalyst surface of
ethene was under kinetic control as was observed with Rh.
the 773 K reaction sample. This indicates that amorphous type
Like the Rh/Al2O3 catalyst, small amounts of liquid products coke is deposited. However, increasing the reaction tempera-
were observed at all temperatures (Table 3). However the yield ture from 773 K to 873 K produced both single and multi-wall
of acetaldehyde (o5%) and indeed the overall liquid product carbon nanotubes (CNTs). These results indicate that the coke
yield was much less over Pt/Al2O3 compared to Rh/Al2O3.
TPO analysis was performed on the Pt/Al2O3 catalyst to
identify the type and amount of coke deposition during ESR
at 773 K, 823 K and 873 K. The TGA profile of the post-reaction
catalyst at each temperature is shown in Fig. 12. The carbon
formation is significantly influenced by temperature.
Discussion
Alumina
dioxide evolution profiles match the derivative curve. Two An analysis of the alumina reactivity has been performed
evolutions of carbon dioxide were observed at all temperatures previously11 and it was shown that the main reaction over the
Fig. 5 Post reaction TPO for Rh/Al2O3 at 773 K, 823 K and 873 K.
c
This journal is The Royal Society of Chemistry 2012
Catal. Sci. Technol.

View Article Online
Paper
Catalysis Science & Technology
Fig. 6 Raman spectra of post-reaction Rh/Al2O3 catalyst at 773 K, 823 K and 873 K.
2C2H5OH - 2CO + 2CH4 + 2H2 DH298 = +99 kJ molÀ1
(8)
CO + H2O - CO2 + H2 DH298 = À41 kJ molÀ1
(7)
overall, 2C2H5OH + H2O - 3H2 + CO + 2CH4 + CO2
(9)
Even though there was no significant deactivation of the
alumina (o2% over 100 h TOS) and no extensive carbon was
detected by SEM, graphitic species were detected by TPO and
Raman spectroscopy.11 Clearly although these graphitic type
species are present they do not significantly affect the activity of
the support.
Rh/Al2O3
At 773 K the conversion of ethanol over Rh/Al2O3 was initially
high Z 99% but rapidly decreased to give a steady state B83%.
High conversion of ethanol over rhodium catalysts has been
reported in the literature over a range of conditions and
rhodium has usually been found to be the most active metal
for steam reforming.17–19 Over the support, in the absence of
metal, direct decomposition was observed but in the presence of
metal this route is partially inhibited with ethanol dehydration
to ethene, ethanol dehydrogenation to acetaldehyde, ethanol
decomposition and ethanol steam reforming to carbon monoxide
and hydrogen all observed. In the first 24 h the ratio of ethanol
decomposition to ethanol steam reforming is 4 : 1, while the
ratio of ethanol dehydration to ethanol steam reforming is
15 : 1. Nevertheless throughout the reaction the main product
is hydrogen while ethene systematically decreases before
Fig. 7 Post-reaction SEM images of Rh/Al2O3 at (a) 773 K and (b) 823 K.
alumina support in the absence of metal was ethanol decom- reaching a steady state at B72 h. This is in agreement with
position, followed by the water gas shift reaction.
the literature13,20 although the yield of ethene is higher in our
c
Catal. Sci. Technol.
This journal is The Royal Society of Chemistry 2012

View Article Online
Catalysis Science & Technology
Paper
Fig. 8 Conversion (%) of ethanol over Pt/Al2O3 at 773 K, 823 K and 873 K.
Fig. 9 Molar selectivity of dry gas over Pt/Al2O3 catalyst at 773 K.
study. By the end of the reaction the ratio between ethanol not being converted to gas phase products. Hence much of the
dehydration and ethanol steam reforming is B1 : 1. No deactivation seen relates to the loss of dehydrogenation activity
significant hydrogenation of ethene is observed, even after of the catalyst.
100 h TOS when hydrogen is in a considerable excess the ratio
At 823 K hydrogen was the principal gaseous product at all
C2H4 : C2H6 > 5, which is significantly removed from thermo- times on stream with steam reforming the principal reaction.
dynamic equilibrium (C2H4 : C2H6 B 0.05). Rather surprisingly However ethene selectivity slowly increased with time, while at
the yield of liquid products decreases with time. It would be the same time the selectivity to methane, carbon monoxide and
expected that as the catalyst deactivates that the liquid product carbon dioxide decreased. This would be consistent with sites
yield would increase, especially if acetaldehyde were a reaction that convert ethene, through steam reforming or hydrogenolysis,
intermediate, nevertheless there is a clear trend in the opposite to carbon monoxide, carbon dioxide or methane, slowly being
direction with acetaldehyde decreasing from B10% to o5% deactivated. It is also consistent with the slight increase in
(Table 1). There is no corresponding increase in all gas phase hydrogen selectivity. As at 773 K the acetaldehyde yield
products except ethane confirming that the liquid products are decreased with time on stream as did the yield of acetone. This is
c
This journal is The Royal Society of Chemistry 2012
Catal. Sci. Technol.

View Article Online
Paper
Catalysis Science & Technology
Fig. 10 Molar selectivity of dry gas over Pt/Al2O3 catalyst at 823 K.
Fig. 11 Molar selectivity of dry gas over Pt/Al2O3 catalyst at 873 K.
to be expected as acetone is thought to be a secondary product the CO2 : CH4 ratio is B1 indicating an overall reaction
of acetaldehyde formed via aldol condensation, oxidation and described by,
decarboxylation.21
C2H5OH - CO + CH4 + H2 DH298 = +49 kJ molÀ1 (10)
At 873
K there is continuous catalyst deactivation
with conversion decreasing throughout the experiment. This
is in contrast to 773 K and 823 K where, although there is
carbon deposition, activity stabilises after B40 h on-stream.
Hence the catalyst is constantly changing throughout the
reaction. Noticeably the selectivity to methane has increased
with it becoming the most selective of all the carbon containing
gases. The reaction chemistry can be split into three distinct
units, ethanol decomposition, ethanol steam reforming
and ethanol dehydration. In the early stages of the reaction
CO + H2O - CO2 + H2 DH298 = À41 kJ molÀ1
(7)
overall, CH3CH2OH + H2O - CH4 + CO2 + 2H2 (11)
Taking account of the hydrogen formed in reaction (11) the
carbon monoxide and residual hydrogen have a H2 : CO ratio of
B2 suggesting a steam reforming reaction of,
CH3CH2OH + H2O - 2CO + 4H2 DH298 = +256 kJ molÀ1
(2)
c
Catal. Sci. Technol.
This journal is The Royal Society of Chemistry 2012

View Article Online
Catalysis Science & Technology
Paper
Table 3 Yield of liquid products over Pt/Al2O3 as a function of temperature
not achieving equilibrium we maintain a higher hydrogen
selectivity than predicted as it is not consumed by the metha-
nation reaction.
and TOS
Yield of liquid products(%)
TGA with evolved gas analysis of the spent catalysts confirms
the presence of carbon deposits on the catalyst. From the
weight loss it can be calculated that the amount of carbon
deposited is o0.1% of the carbon passed over the catalyst
during its time on stream. The profiles reveal that different
types of carbon are deposited on the catalyst surface, which is
consistent with results obtained by Can et al.20 over Rh cata-
lysts. In that study of ESR over Rh/Al2O3 at 673 K, they found
that after 8 h four different types of coke could be identified
from combustion events in the TPO. Type I (583 K) was
assigned to coke deposited on the metal, type II (653 K) to coke
deposited around the metal–support interface, type III (793 K)
to carbon deposited on the support, while type IV (968 K) was
assigned to a graphite phase generated by the thermal ethanol
decomposition of ethanol. Our previous results11 confirmed the
assignment of the type IV material. After use in ESR, alumina
gave a TPO peak at 985 K, while the Raman spectrum con-
firmed the presence of graphitic species, while analysis of the
reaction products confirmed that only ethanol decomposition
was taking place. However if we examine the carbon dioxide
evolution at 773 K and 873 K (Fig. 15) we see no evidence of
combustion at 583 K or 653 K. It is possible that these species
are only present when low temperature ESR conditions are used
(673 K cf. 773–873 K) or that the TOS makes a difference (8 h
cf. 100h) and indeed if we examine the shift between the 773 K
and 873 K TPO profiles in Fig. 15 it can be seen that there is a
shift of B65 K. Hence it is likely that the type I and type II
species detected at 583 K and 653 K respectively by Can et al.20
are contained within the shoulder centred at 717 K in our TPO.
The shift to higher temperatures of the combustion maxima
Temp. (K) TOS (h) Acetald.a DEEa Acetone AAcida 1,1-DEEa
773
823
873
24
48
72
100
24
48
72
100
24
48
72
3.8
2.3
2.0
2.0
4.5
3.0
3.6
2.9
4.5
4.5
3.5
3.7
2.1
2.3
2.4
2.4
2.2
2.4
2.2
2.3
1.1
1.3
1.5
1.6
1.5
0.6
0.5
0.5
0.9
0.7
0.7
0.6
0.9
0.8
0.8
0.8
0.6
0.5
0.4
0.3
0.4
0.1
0.5
0.4
trb
tr
0.2
0.2
0.2
0.2
0.3
0.3
0.3
0.3
0.3
0.3
0.4
0.4
tr
tr
100
a
Acetald. = acetaldehyde, DEE = diethyl ether, AAcid = acetic acid,
b
1,1-DEE = 1,1-diethoxy ethane. tr = trace.
The ratio of ethanol being converted by each reaction is
B4 : 1 in favour of eqn (11). The rest of the ethanol, reacting to
give gas phase products, is converted to ethene by dehydration.
The C2H4 : C2H6 ratio has decreased to B3 : 1 compared to
5 : 1 at 773 K. Nevertheless this ratio is still well removed from
equilibrium at 873 K (1 : 4 ethene : ethane). From the rate of
hydrogenation a pseudo-activation energy was calculated over
the 773–873 K range and the value obtained was B42 kJ molÀ1
.
11
This is close to the value obtained over Ru/Al2O3 and the
literature value obtained over Rh at 298 K of 35 kJ mol .
À1 22 This
confirms that ethene hydrogenation is significantly inhibited,
probably by the presence of carbon monoxide and a low surface
coverage of hydrogen.
Examining the non-C-2 gas phase products, it is clear
that they are also well removed from equilibrium. At high
pressure methanation is favoured and as the temperature is
increased methane becomes the principal C-1 gas. However by as the reaction temperature increases suggests that as the
Fig. 12 Post-reaction TPO for Pt/Al2O3 catalyst at 773 K, 823 K and 873 K.
c
This journal is The Royal Society of Chemistry 2012
Catal. Sci. Technol.

View Article Online
Paper
Catalysis Science & Technology
Fig. 13 Raman spectra of post-reaction Pt/Al2O3 at 773 K, 823 K and 873 K.
Table 4 Post-reaction BET analysis and weight loss of Pt/Al2O3 after TPO at 773 K,
823 K and 873 K
BET surface
Pore volume Weight loss
Conditions (K) area (m2 gÀ1
)
(cm3 gÀ1
)
in TPO (%)
(ID/IG)
Reduced (873)
100
41
27
0.46
0.09
0.05
0.05
—
37
40
40
—
773
823
873
0.92
1.03
1.38
71
reaction temperature is increased the carbonaceous deposit on
the surface changes from one with a high H : C ratio to
one with a lower H : C ratio. This is supported by the loss of
water evolution (not shown), which was only observed for the
773 K and 823 K ex-reaction samples. The presence of more
dehydrogenated coke at 873 K is also supported by the SEM
images (Fig. 7), which gave a larger number of multi wall
carbon nanotubes. The coke formation follows a similar pat-
tern to that seen over Ru/Al2O3,11 although no filamentous
carbon was detected at 773 K over Ru/Al2O3. The Raman ID/IG
ratio reveals that the intensity and disorder in graphitic-type
carbon is significantly increased as the temperature is
increased. The increase in disorder is not completely under-
stood but it may be due to the formation of coiled coke, which
may produce more disorder in the graphitic carbon. To the best
of our knowledge it is the first time such coke has been seen
over Rh/Al2O3 catalyst during a steam reforming reaction.
Recently it was reported that formation of coiled type coke took
place by catalytic chemical vapour deposition using different
hydrocarbons23 suggesting that different types of graphitic
carbon play a role in the formation of such carbon.
Fig. 14 Post reaction SEM images of Pt/Al2O3 catalyst at 773 K and 873 K.
Hydrogen is the main product although there is significant
formation of ethene. This is in agreement with other
studies16 that have shown that ethene is favoured at these
temperatures over Pt/alumina. The ethene : ethane ratio is
B5 similar to that found with Rh/Al2O3 and well removed
Pt/Al2O3
Over the first 48 h TOS at 773 K the Pt/Al2O3 catalyst deactivates from equilibrium. The carbon dioxide selectivity is higher
before settling to a steady state value of B75% conversion. than that observed for Rh/Al2O3 and suggests
a more
c
Catal. Sci. Technol.
This journal is The Royal Society of Chemistry 2012

View Article Online
Catalysis Science & Technology
Paper
Fig. 15 Carbon dioxide evolution during TPO of catalysts used at 773 K and 873 K in ESR.
effective water gas shift reaction. The principal reactions at methane and carbon dioxide show selectivities around 15%.
773 K are;
This is the highest selectivity observed for carbon dioxide over
the three catalysts (Rh/Al2O3 and Ru/Al2O311) and confirms the
high water gas shift (WGS) activity of the platinum catalyst. This
is in agreement with the literature28 and platinum catalysts are
now being offered commercially for WGS processes for fuel cell
applications by Sud-Chemie. Analysis of the dry gas ratios
suggests that the reactions to generate non-C2 gases are,
C2H5OH - C2H4 + H2O DH298 = +45 kJ molÀ1 (12)
1.5C2H5OH + 0.5H2O - CH4 + 2CO + 3H2
DH298 = +178 kJ molÀ1
(13)
(7)
CO + H2O - CO2 + H2 DH298 = À41 kJ molÀ1
4C2H5OH - 4CH4 + 4CO + 4H2 DH298 = +198 kJ molÀ1
The ratio of ethanol reacting via ethanol dehydration relative to
ESR, as represented by eqn (13), is, B3 : 1 and for every six CO
molecules formed by eqn (13), five undergo WGS to form CO2.
Hence the overall water : ethanol ratio in the reaction is 1.4 : 1,
which higher than obtained for Rh/alumina (B1 : 1).
The liquid product yield is lower for Pt/alumina than
Rh/alumina and is not a function of conversion. Therefore
the Pt/alumina is much more effective at converting secondary
liquid products, either by decomposition or reaction, to simple
gas phase components. This is surprising as rhodium is the
metal that has the highest propensity for steam reforming,
certainly for hydrocarbons, and it is also more effective at C–C
bond breaking by hydrogenolysis.24–26 However we note that
platinum is the metal of choice for reforming.27
(10)
C2H5OH + H2O - 2CO + 4H2 DH298 = +256 kJ molÀ1
(2)
4CO + 4H2O - 4CO2 + 4H2 DH298 = À164 kJ molÀ1
(7)
Giving an overall reaction of:
5C2H5OH + 5H2O - 4CH4 + 4CO2 + 2CO + 12H2 (14)
The selectivity to ethene is low but increases with TOS
however the selectivity to ethane stays relatively unchanged.
The increase in ethene selectivity is balanced by a decrease in
carbon dioxide and methane (hydrogen, carbon monoxide and
ethane are relatively constant). This suggests that there is a
minor underlying reaction sequence relating to ethene;
By 873 K after slight deactivation over the first 24 h the
catalyst achieves a steady state conversion B89%, revealing
the Pt/Al2O3 catalyst to be the more active than Rh/Al2O3 and
C2H5OH - C2H4 + H2O DH298 = +45 kJ molÀ1 (12)
C2H4 + H2O - CH4 + CO + H2 DH298 = +4 kJ molÀ1 (15)
11
Ru/Al2O3 at 873 K. This high activity is in keeping with the
study by Erdohelyi and co-workers,16 who obtained 100%
conversion at B773 K but, in contrast to our study, found
acetaldehyde as a major product. Verykios and co-workers15
CO + H2O - CO2 + H2 DH298 = À41 kJ molÀ1
(7)
found Pt very much less active than Rh with effectively no activity with ethanol being dehydrated over acid sites producing ethene
at 873 K (however the activity of their 1% Rh/Al2O3 was also low (eqn (12)), which is then partially steam reformed (eqn (15)) to
giving o20% conversion at 873 K). Over the first 48 h TOS both give methane, hydrogen and carbon monoxide. The carbon
c
This journal is The Royal Society of Chemistry 2012
Catal. Sci. Technol.

View Article Online
Paper
Catalysis Science & Technology
monoxide and steam react to give carbon dioxide and hydrogen Using this relationship with the ID/IG ratios reported in
(eqn (7)) (WGS). As the catalyst deactivates the dehydration Table 3 gives microcrystalline planar sizes of 4.8 nm from the
continues but the partial steam reforming decays. This results 773 K sample, 4.3 nm from the 823 K sample and 3.2 nm from
in the yield of both methane and carbon dioxide decreasing the 873 K sample. This shows that an increase of the reaction
while ethene increases. Calculation of an activation energy for temperature from 773 K to 873 K significantly reduces the
the hydrogenation of ethene to ethane, using the values after crystallinity of the carbon and increases its amorphous nature.
100 h TOS, over the 773–873 K temperature range gives a value of
The SEM analysis of the used catalysts confirms the presence
21 kJ molÀ1, which is approximately half the value that would be of whisker carbon/carbon nanotubes (CNTs) at 873 K. At lower
expected from room temperature determinations22 and signifi- temperatures no clear evidence of CNT formation was observed.
cantly lower than values obtained from the rhodium and ruthe- Although we have been unable to find any previous report of
nium catalysts.11 Activation energy values as low as this often are ethanol giving rise to CNTs in steam reforming, a recent paper31
indicative of a degree of diffusion control over the reaction and it has shown that CNTs can be formed over a platinum catalyst in
is possible that there are diffusional constraints in the system the absence of steam at temperatures Z 873 K. No filaments or
due to the extent of carbon deposition.
CNTs were observed at 773 K in agreement with our results. The
Once again the non-C-2 gases are well removed from equili- addition of ethanol to ethene feedstreams has been shown to
brium however the distribution is different from that seen with enhance the production of CNTs32 over iron catalysts. In these
rhodium. Although methanation is favoured, it is carbon systems it is proposed that ethanol (or indeed other oxygenates)
dioxide that is the principal C-1 gas, while the level of carbon acts to remove amorphous carbon and allow a higher yield of
monoxide is much lower than found with rhodium. This CNTs. Therefore given the presence of ethanol and ethene in our
suggests that over Pt/alumina, carbon monoxide is reacting reaction gases the formation of CNTs/whisker carbon is not
with steam via the water gas shift reaction faster than it does surprising. Note that the Pt/Al2O3 catalyst is the least deactivated
with hydrogen via methanation, hence reducing the carbon system under reaction conditions at 873 K and this may be due
monoxide level but maintaining a high hydrogen selectivity.
to the ethanol cleaning amorphous carbon from the surface
The weight loss profile obtained from the TGA of the used more efficiently than with rhodium or ruthenium.
Pt/Al2O3 shows a similar weight loss profile to that of Ru/Al2O3,11
Mechanistic implications
although with the catalyst tested at 773 K the weight loss was
complete by a relatively low temperature i.e. 1093 K. Also the total The mechanism of ethanol steam reforming has been the
weight loss from the catalyst tested at 773 K is lower than that subject of various investigations over a number of different
11
observed with Ru/Al2O3 and Rh/Al2O3 catalysts at the same metals and it is important to set the current work into the
temperature and represents only 0.05% of the carbon passed over mechanistic context that exists.33 In a recent study Davis and
the catalyst. The formation of different types of coke can be easily co-workers34 showed that at low temperatures (o773 K) ethanol
seen from the derivative profiles. The profile obtained is similar to could interact with the support and decompose in the presence
that observed by Shamsi et al.,29 who studied the steam reforming of steam to methane and carbon dioxide with carbon monoxide
of tetradecane and decalin over 0.61% Pt/alumina. This common- being produced by the reverse water gas shift reaction. In this
ality between TPO profiles further supports the idea that different case the support was a partially reduced ceria and it was
types of coke formation do not depend upon the feed.30 In the proposed that much of the reaction occurred on the support
study by Shamsi et al.29 the peak maxima occur at lower tempera- and the metal–support interface. However in a similar study by
tures compared to the maxima we observe (673 K and 793 K Erdohelyi et al.,16 where the support was alumina, no methane
cf. 893 K and 988 K) however their catalyst was on-stream for only was reported. In this case ethene was a significant product
30 min in comparison with 100 h for ours and the amount of formed by dehydration. Nevertheless the studies do agree that
carbon deposited is only B10% of that deposited during our ethanol does adsorb on the support surface to form an
reaction. Therefore it is likely that our coke has aged with time adsorbed ethoxy species and can subsequently form adsorbed
and become more recalcitrant in nature shifting the TPO profile acetate. In our system the alumina in the absence of metal is
to higher temperatures. As expected, and like the Ru/Al2O311 and active for ethanol decomposition and forms methane, carbon
Rh/Al2O3 catalysts, these weight losses are due to carbon dioxide oxides and hydrogen but, in contrast to the Erdohelyi et al.16
and water evolutions, interestingly, and unlike the Ru/Al2O311 and study, no ethene. The ratio of CH4 : (CO + CO2) was 1 : 1 as was
Rh/Al2O3 catalysts, water evolution was also observed from the found by Davis34 but we interpreted this as ethanol decomposi-
873 K reaction sample (not shown), which suggests that relatively tion followed by water gas shift. It is perfectly possible given the
more hydrogenated coke was present on the Pt/Al2O3 catalyst after difference in the supports that each interpretation is correct for
reaction at 873 K. This would be in keeping with the higher the given system. What these comparisons reveal is a system
activity observed with the platinum catalyst.
that is highly sensitive to the support used, even with alumina
The ID/IG ratio, obtained from the Raman bands, reported in it is clear that differences exist between the reactivity of
Table 3 shows that the disorder in the graphitic carbon steadily g-alumina and d-alumina. The picture is further complicated
increases with an increase in the reaction temperature. by the fact that when metal is added to the support there is a
A measure of the microcrystalline planar size of graphite can significant change in product distribution.11,16 Over our
be obtained from the relationship La (nm) = 4.4[ID/IG]À1 29
.
alumina support the major change, in the presence of either
c
Catal. Sci. Technol.
This journal is The Royal Society of Chemistry 2012

View Article Online
Catalysis Science & Technology
Paper
rhodium or platinum, is the formation of ethene. The formation of ability of the ethanol to remove amorphous carbon. In both cases
ethene over a metal function is well known35 and the mechanism we believe this to be the first example of carbon filaments formed
of ethanol dehydration has also been examined.36 It is likely that over rhodium and platinum during ESR.
the ethene formed is the main precursor to the carbonaceous
species and CNTs detected on the catalyst surface. Studies have
References
shown that ethanol31 and ethene37 will form CNTs over platinum
1 J. R. Rostrup-Nielsen and T. Rostrup-Nielsen, CATTECH, 2002, 6,
150–159.
and ethene will form CNTs over rhodium, whereas carbon deposi-
tion from carbon monoxide, over these catalysts, is inhibited in the
presence an equal amount of carbon dioxide.38 As there is always a
CO2 : CO ratio >1 (the rhodium catalyst at 873 K after 100 h TOS is
the only exception) it is unlikely that carbon monoxide is signifi-
cantly adding to the deposited carbon. Similarly the extent of
carbon deposition from methane, over these catalysts, even in the
absence of hydrogen and steam, is minimal.38
At low temperatures there is convincing evidence that the
dehydrogenation of ethanol to acetaldehyde and the dehydra-
tion to ethene are the two initial reactions over the available
sites in ESR, with methane and carbon monoxide being formed
from acetaldehyde decomposition and then subsequent water
2 A. J. Akande, R. O. Idem and A. K. Dalai, Appl. Catal., A, 2005, 287,
159–175.
3 I. A. Carbajal Ramos, T. Montini, B. Lorenzut, H. Troiani, F. C. Gennari,
M. Graziani and P. Fornasiero, Catal. Today, 2012, 180, 96–104.
4 T. Montini, L. De Rogatis, V. Gombac, P. Fornasiero and
M. Graziani, Appl. Catal., B, 2007, 71, 125–134.
5 S. M. de Lima, A. M. da Silva, L. O. O. da Costa, U. M. Graham,
G. Jacobs, B. H. Davis, L. V. Mattos and F. B. Noronha, J. Catal.,
2009, 268, 268–281.
6 S. D. Jones and H. E. Hagelin-Weaver, Appl. Catal., B, 2009, 90, 195–204.
7 X. Hu and G. Lu, Appl. Catal., B, 2010, 99, 289–297.
8 R. Trane, S. Dahl, M. S. Skjøth-Rasmussen and A. D. Jensen, Int. J.
Hydrogen Energy, 2012, 37, 6447–6472.
9 F. Bimbela, M. Oliva, J. Ruiz, L. Garcia and J. Arauzo, J. Anal. Appl.
Pyrolysis, 2009, 85, 204–213.
10 Y. Matsumura and H. Ishibe, J. Catal., 2009, 268, 282–289.
gas shift to form carbon dioxide.33,34 If the temperature is high 11 M. Bilal and S. D. Jackson, Catal. Sci. Technol., 2012, 2, 2043–2051.
enough methane can subsequently be steam reformed to give
12 X. Wu and S. Kawi, Energy Environ. Sci., 2010, 3, 334–342.
13 S. Cavallaro, Energy Fuels, 2000, 14, 1195–1199.
14 F. Aupretre, C. Descorme, D. Duprez, D. Casanave and D. Uzio,
carbon monoxide and hydrogen. At higher temperatures the
mechanism is less clear. It is often assumed that the same
mechanistic steps will be followed, only faster, however reac-
tions that are not readily observed at low temperatures due to
high activation energies can rapidly become dominant at high
temperatures. Carbon laydown and carbon–carbon bond break-
ing are two typical examples of this. From our study we cannot
delve too deeply into the mechanism, we can confirm produc-
tion of acetaldehyde, ethene and C-1 components. It is clear
that the mechanism is complex with multiple breakdown and
J. Catal., 2005, 233, 464–477.
15 D. K. Liguras, D. I. Kondarides and X. E. Verykios, Appl. Catal., B,
2003, 43, 345–354.
16 M. Domok, M. Toth, J. Rasko and A. Erdohelyi, Appl. Catal., B, 2007,
69, 262–272.
17 P. O. Graf, B. L. Mojet, J. G. van Ommen and L. Lefferts, Appl. Catal.,
A, 2007, 332, 310–317.
18 D. Duprez, P. Pereira, A. Miloudi and R. Maurel, J. Catal., 1982,
75, 151.
19 E. Kikuchi, K. Ito, T. Ino and Y. Morita, J. Catal., 1977, 46, 382.
20 F. Can, A. Le Valant, N. Bion, F. Epron and D. Duprez, J. Phys. Chem.
C, 2008, 112, 14145–14153.
synthesis routes, e.g. the formation of acetone is thought to be a 21 T. Nishiguchi, T. Matsumoto, H. Kanai, K. Utani, Y. Matsumura,
secondary product of acetaldehyde21 formed via aldol conden-
W. J. Shen and S. Imamura, Appl. Catal., A, 2005, 279, 273–277.
22 G. C. Bond, Metal-catalysed reactions of hydrocarbons (fundamental
sation, oxidation and decarboxylation. Further work involving
and applied catalysis), Springer, New York, 2005.
´
´
´
higher space velocities and, ideally, selectively labelled species 23 A. Csato, A. Szabo, A. Fonseca, D. Vuono, Z. Konya, A. Volodin,
would be useful to delineate the mechanism further.
C. Van Haesendonck, L. P. Biro, G. Giordano and J. B. Nagy, Catal.
Today, 2012, 181, 33.
24 J. H. Sinfelt and D. J. C. Yates, J. Catal., 1967, 8, 82–90.
25 D. E. Resasco and G. L. Haller, in Stud. Surf. Sci. Catal.,
Conclusions
ed. B. Naccache, C. Coudurier, G. Praliaud, H. Meriaudeau,
P. Gallezot, G. A. Martin and J. C. Vedrine, Elsevier, Amsterdam,
In this paper the steam reforming of ethanol over 0.2%
1982, vol. 11, pp. 105–112.
Rh/Al2O3 and 0.2% Pt/Al2O3 catalysts has been investigated at 26 S. D. Jackson, G. J. Kelly and G. Webb, J. Catal., 1998, 176, 225–234.
27 Z. Francisco, Appl. Catal., A, 2002, 229, 75–91.
28 D. W. Flaherty, W.-Y. Yu, Z. D. Pozun, G. Henkelman and
773 K, 823 K and 873 K at 20 barg over 100 h on-stream.
Hydrogen is the principle product at all temperatures with both
C. B. Mullins, J. Catal., 2011, 282, 278–288.
catalysts. At 773 K and 823 K Rh/Al2O3 is more active than 29 A. Shamsi, J. P. Baltrus and J. J. Spivey, Appl. Catal. A, 2005, 293, 145–152.
Pt/Al2O3, which is in keeping the literature33 however this is due
more to ethanol decomposition and ethanol dehydration rather
30 S. D. Jackson, S. J. Thomson and G. Webb, J. Catal., 1981, 70, 249.
31 N. Fukuoka, Y. Mizutani, S. Naritsuka, T. Maruyama and S. Iijima,
Jpn. J. Appl. Phys., 2012, 51, 06FD23.
than ESR. In contrast by 873 K Pt/Al2O3 is more active and stable 32 D. N. Futaba, J. Goto, S. Yasuda, T. Yamada, M. Yumura and K. Hata,
Adv. Mater., 2009, 21, 4811–4815.
33 N. Bion, F. Epron and D. Duprez, in Catalysis, ed. J. J. Spivey and
than Rh/Al2O3, which systematically deactivates, and reveals a
high activity for WGS. There is significant carbon laydown on
K. M. Dooley, RSC, London, 2010, vol. 22, p. 1–55.
both catalysts however it has little effect on catalyst activity at 34 G. Jacobs, R. A. Keogh and B. H. Davis, J. Catal., 2007, 245, 326–337.
35 M. Benitoa, J. L. Sanz, R. Isabel, R. Padilla, R. Arjona and L. Daza,
773 K and 823 K. At 773 K only the Rh/Al2O3 catalyst produces
J. Power Sources, 2005, 151, 11–17.
carbon filaments. At 873 K the Rh/Al2O3 catalyst shows a slow
36 S. D. Jackson, J. R. Jones, A. P. Sharratt, L. F. Gladden and R. J.
but steady deactivation. The Pt/Al2O3 catalyst also shows evidence
of CNT/whisker growth but does not significantly deactivate over
the reaction period. This may be due to a higher H : C ratio of the
Cross, G. Webb, in Catalysis of Organic Reactions, ed. D. Morrell,
Marcel Dekker, New York, 2002, pp. 453–460.
37 S. Takenaka, T. Iguchi, E. Tanabe, H. Matsune and M. Kishida,
Carbon, 2009, 47, 1251–1257.
deposit on the Pt/Al2O3 catalyst, as indicated by TGA, and/or the 38 E. Opara, S. D. Jackson, unpublished results.
c
This journal is The Royal Society of Chemistry 2012
Catal. Sci. Technol.