Ruthenium Dispersion: A Key Parameter for the Stability of Supported Ruthenium Catalysts during Catalytic Supercritical Water Gasification

Article abstract of DOI:10.1002/cctc.201500995

The catalytic supercritical water gasification of isopropanol (450 °C, 30 MPa) over Ru catalysts supported on carbon and metal oxides was performed in a fixed-bed plug flow reactor. The Ru loading was between 1.2 and 2 %. The catalyst stability over a period of 50 h was in the order: Ru/C>Ru/ZrO₂>Ru/Al₂O₃≈Ru/TiO₂. Considerable coke deposits were found on Ru/Al₂O₃ and Ru/TiO₂, which suggests that coke formation was responsible for the loss of activity, whereas the coke content was much lower on Ru/C and Ru/ZrO₂, which confirms their better coking resistance. Clearly, Ru/C was the most stable catalyst as a loss of only 3 % of its initial activity was measured. The high Ru dispersion of Ru/C and Ru/ZrO₂ was beneficial for the improvement of the catalyst stability because of the higher gasification rate versus the coke formation rate.

Full text of DOI:10.1002/cctc.201500995

DOI: 10.1002/cctc.201500995
Communications
Ruthenium Dispersion: A Key Parameter for the Stability
of Supported Ruthenium Catalysts during Catalytic
Supercritical Water Gasification
Gal Peng,[a] Christian Ludwig,[a, b] and FrØdØric Vogel*[a, c]
The catalytic supercritical water gasification of isopropanol
(4508C, 30 MPa) over Ru catalysts supported on carbon and
metal oxides was performed in a fixed-bed plug flow reactor.
The Ru loading was between 1.2 and 2%. The catalyst stability
over a period of 50 h was in the order: Ru/C>Ru/ZrO2 >Ru/
Al2O3 ꢀRu/TiO2. Considerable coke deposits were found on Ru/
Al2O3 and Ru/TiO2, which suggests that coke formation was re-
sponsible for the loss of activity, whereas the coke content was
much lower on Ru/C and Ru/ZrO2, which confirms their better
coking resistance. Clearly, Ru/C was the most stable catalyst as
a loss of only 3% of its initial activity was measured. The high
Ru dispersion of Ru/C and Ru/ZrO2 was beneficial for the im-
provement of the catalyst stability because of the higher gasifi-
cation rate versus the coke formation rate.
rutile-TiO2, monoclinic-ZrO2, and a-Al2O3 were reported to be
stable.[3–5] In a previous study,[6] we reported that the lifetime
of Ru/C catalysts was affected by the decomposition of the re-
actant (isopropanol) to coke over the carbon surface, whereas
Ru leaching and Ru sintering were found to have minor effects
on the catalyst deactivation. Zçhrer et al.[7] studied the stability
of Ru supported on metal oxides (Ru/TiO2 and Ru/ZrO2) during
the CSCWG of glycerol and observed some coke deposits on
the spent catalysts. Therefore, coking is a serious issue in the
gasification of organic compounds under supercritical condi-
tions. In this work, we compare the stability of Ru catalysts
supported on carbon and metal oxides (rutile-TiO2, monoclinic-
ZrO2, and a-Al2O3) during the CSCWG of isopropanol (IPA) to
evaluate their respective coking resistance. The use of alcohols
as model compounds is relevant as they have been reported
as intermediate products during the supercritical water gasifi-
cation of real biomass.[8] Fundamentally, IPA is a good model
compound because it is a simple molecule yet it contains CÀC,
CÀH, and CÀO bonds as well as primary and secondary carbon
atoms. The characteristics of the supported Ru catalysts are
listed in Table 1. It seems that a higher BET specific surface
The production of gaseous biofuels (such as H2 and CH4) from
wet biomass (e.g. microalgae, biomass residues) has attracted
a lot of attention during the last decade. For this purpose, cat-
alytic supercritical water gasification (CSCWG) is a promising
technology. Its main advantage is its capability to process wet
biomass (water content >60 wt%) without the need for
a drying step, which allows a high thermal efficiency to be
reached (70–77%).[1] At moderate temperatures (374–5008C),
a catalyst is required to reach a high biomass conversion and
a high CH4 or H2 yield. For the production of CH4, supported
Ru catalysts are known to be the most suitable catalysts be-
cause of the high activity of Ru to decompose the large organ-
ic molecules by CÀC bond cleavage as well as their high CH4
selectivity because of the ability of Ru to cleave CÀO bonds.[2]
Only a few catalyst supports are able to preserve their physical
structure in the harsh environment of supercritical water (T>
3748C, p>22.1 MPa). Carbon and some metal oxides such as
Table 1. Characteristics of the fresh supported Ru catalysts.
[a]
[b]
[c]
Catalyst
BET SSA
Vtotal
Ru loading
[wt%]
DCO
dp,CO
[m2 gÀ1
]
[cm3 gÀ1
]
[nm]
Ru/TiO2
Ru/ZrO2
Ru/Al2O3
Ru/C
4
23
5
0.03
0.19
0.04
0.70
1.8
1.2
1.2
2
8
52
6
16.3
1.9
23.3
1.6
653
61
[a] Measured at p/p0 =0.99. [b] Ruthenium dispersion determined by CO
pulse chemisorption. [c] Ruthenium particle size calculated according to
[9]
´
the formula developed by Borodzinski and Bonarowska.
[a] Dr. G. Peng, Prof. Dr. C. Ludwig, Prof. Dr. F. Vogel
General Energy Research
Paul Scherrer Institut (PSI)
5232 Villigen PSI (Switzerland)
E-mail: frederic.vogel@psi.ch
area (SSA) favors a higher Ru dispersion. Although the BET SSA
of Ru/ZrO2 was significantly smaller than that of Ru/C, its Ru
dispersion was still high. This result demonstrates that a rela-
tively small BET SSA of ~20 m2 gÀ1 is already large enough to
achieve highly dispersed Ru nanoparticles by a wet impregna-
tion method.
[b] Prof. Dr. C. Ludwig
ENAC-IIE
Ecole Polytechnique FØdØrale de Lausanne (EPFL)
1015 Lausanne (Switzerland)
[c] Prof. Dr. F. Vogel
The catalytic performance of the supported Ru catalysts
during the CSCWG of 10 wt% isopropanol at 4508C and
30 MPa over 50 h is shown in Figure 1. To operate at carbon
conversions below 100%, a weight hourly space velocity nor-
malized to one gram of Ru (WHSVgRu) was used. For the Ru/C
University of Applied Sciences Northwestern Switzerland (FHNW)
5210 Windisch (Switzerland)
Supporting Information and ORCID(s) from the author(s) for this article
are available on the WWW under http://dx.doi.org/10.1002/
cctc.201500995.
ChemCatChem 2016, 8, 139 – 141
139
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Communications
that the water gas shift reaction was still favored at a low total
carbon conversion (XC).
Interestingly, the carbon gasification efficiency (GEC) values
were lower than the XC values, which indicates carbon accumu-
lation in the reactor. Hence, coke deposition from the decom-
position of isopropanol on the catalyst support surface seems
to take place for the tested Ru catalysts. According to the GEC
and XC values for Ru/Al2O3, no coking occurred. As a result of
its small BET SSA, it is likely that the coke deposits were more
difficult to quantify. Notably, GEC was calculated from the gas
flow rate, and some inaccuracies caused by gas accumulation
in the setup might occur.
The temperature-programmed oxidation (TPO) results of the
fresh and spent Ru catalysts supported on metal oxides con-
firm that coke formation occurred during the CSCWG of iso-
propanol as CO2 peaks were observed with the spent catalysts
(Figure 2). The higher peak intensity for Ru/TiO2 and Ru/Al2O3
in comparison to Ru/ZrO2 confirms the better coking resistance
of the latter. The different CO2 peaks are likely related to the
oxidation of carbonaceous species located on the Ru surface
and the catalyst support.
Figure 1. CSCWG of 10 wt% isopropanol over Ru catalysts supported on
carbon and on metal oxides at 4508C and 30 MPa for 50 h with
WHSVgRu =5202 gOrg gRuÀ1 hÀ1
.
catalyst, this space velocity was still too low to reach incom-
plete conversion but it was at the limit of our setup. The stabil-
ity of the catalysts decreased in the order: Ru/C>Ru/ZrO2 >
Ru/Al2O3 ꢀRu/TiO2. Although the initial activity of all the Ru
catalysts supported on metal oxides was similar, which
depended on the type of metal oxide, the difference of the ac-
tivity loss was considerable. For instance, Ru/TiO2 and Ru/Al2O3
lost approximately 25 and 14% of their initial activity, respec-
tively, during the first 5 h, whereas Ru/ZrO2 lost only ꢀ7%. The
better stability of Ru/ZrO2 can be related to its higher Ru dis-
persion, which favored a higher gasification rate compared to
the coke formation rate. In contrast, after 50 h the stability of
Ru/C remained almost unaffected. Such a good stability can be
explained by the high capacity of the small Ru nanoparticles
to convert isopropanol fully to gaseous products at such
a high WHSVgRu. It is known that a high Ru dispersion is
needed to achieve a high activity during the CSCWG of
isopropanol.[6]
Figure 2. TPO of the fresh and spent Ru catalysts supported on metal
oxides.
The gas composition confirms the good catalytic per-
formance of Ru/C as the CH4 concentration was close to the
thermodynamic chemical equilibrium (Table 2 and Figure S4).
As the coke deposits cannot be characterized by TPO for the
Ru/C catalyst because of the combustion of the carbon sup-
port itself, Ru/C was characterized by N2 physisorption. The
physical structure of the Ru/C was well preserved after 50 h
on-stream (Table 3). Although a small fraction of the mesopore
volume was lost, likely caused by coke deposits, the total pore
volume remained high. This observation suggests that the
amount of coke inside the pores is low. In previous work,[6] we
Table 2. Summary of results after 50 h of CSCWG of 10 wt% isopropanol
at 4508C and 30 MPa with WHSVgRu =5202 gOrg gRuÀ1 hÀ1
.
[a]
[b]
Catalyst
XC
GEC
[%]
CH4
[vol%]
CO2
[vol%]
H2
[vol%]
CO
[vol%]
[%]
Ru/TiO2
Ru/ZrO2
Ru/Al2O3
Ru/C
21
38
22
97
13
25
20
81
46
48
44
65
19
21
18
21
35
31
39
14
0
0
0
0
[a] XC =(1À(mol CLiq,out/mol CFeed))”100%. [b] GEC =total mol CGas/total
mol CFeed ”100%.
Table 3. Physical data for fresh and spent Ru/C.
[a]
[b]
Catalyst
BET SSA
Vtotal
Vmesopore
Vmicropore
[cm3 gÀ1
]
[m2 gÀ1
]
[cm3 gÀ1
]
[cm3 gÀ1
]
The low CH4 and high H2 concentrations for the Ru catalysts
supported on metal oxides revealed that the methanation
reaction was less favored, probably because of a decreased
rate of CÀO bond cleavage. As no CO was detected, it seems
Ru/C
spent Ru/C
653
670
0.70
0.66
0.58
0.53
0.12
0.13
[a] Measured at p/p0 =0.99. [b] Calculated with the t-plot method.
ChemCatChem 2016, 8, 139 – 141
www.chemcatchem.org
140
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Communications
3 gminÀ1. The amount of Ru/C catalyst in the reactor was ꢀ0.17–
0.28 g. The WHSVgRu was 5202 gOrg gRuÀ1 hÀ1 for each experiment.
The gas phase was analyzed offline by GC (HP 6890) with a thermal
conductivity detector. Liquid samples were collected regularly
manually and the total organic carbon (TOC) was measured by
using a TOC analyzer (Vario TOC cube, Elementar). The observed
activity is defined as the total carbon conversion (Xc) from the feed
to the reactor effluent. The carbon gasification efficiency (GEC) is
the relationship between the total amount of carbon in the gas
phase and the total amount of carbon in the feed.
reported that if a 0.5% Ru/C catalyst was fully deactivated
(XC =10%) during the CSCWG of isopropanol (4508C, 30 MPa),
the pore volume was reduced drastically by the coke deposits.
In this study, we showed that a higher Ru dispersion led to
a better stability of supported Ru catalysts during the continu-
ous CSCWG of isopropanol. For the catalysts that have a high
Ru dispersion (Ru/C and Ru/ZrO2), a lower amount of coke de-
posits was found, which suggests a better coking resistance.
As a result of its higher Ru dispersion in comparison to the cat-
alysts supported on the metal oxides, Ru/C exhibited the best
overall catalytic performance (stability, activity, and CH4 selec-
tivity). Hence, Ru/C appears to be the most suitable catalyst for
CSCWG.
Acknowledgements
This work was financially supported within the scope of the Sun-
CHem project by the Competence Center Energy and Mobility
(CCEM-CH) and Swisselectric Research (Switzerland). The authors
would like to thank Alexander Wokaun for the fruitful discus-
sions.
Experimental Section
Rutile-TiO2 (pellets, Norpro Saint Gobain SA), monoclinic-ZrO2 (pel-
lets, Norpro Saint Gobain SA), a-Al2O3 (pellets, Alfa Aesar), and
carbon (granular, DESOTEC) were sieved to a size fraction of 0.3–
0.8 mm. The supported Ru catalysts were prepared by wet impreg-
nation with Ru(NO)(NO3)3 in a water solution for 6 h followed by
solvent evaporation in a rotary evaporator. The materials were
washed with pure water during filtration. The Ru loss during wash-
ing was quantified by inductively coupled plasma optical emission
spectroscopy (ICP-OES; Liberty 110, Varian). After drying at 608C
overnight in a vacuum oven, Ru/C was reduced under flowing
H2/Ar (10:90, 20 mLminÀ1) at 4508C for 4 h. The Ru catalysts sup-
ported on metal oxides were calcined at 4508C for 4 h. The BET
SSA and the pore volume were measured by N2 physisorption by
using an Autosorb-1 (Quantachrome Instruments). CO pulse chemi-
sorption was performed by using a fully automated instrument
(TPD/R/O 1100, Thermo Scientific) for the measurements of the Ru
dispersion (DCO) and the Ru nanoparticles size (dp,CO) according to
a procedure described elsewhere.[6] TPO was performed by using
a NETZSCH STA 449 C thermobalance coupled to an FTIR detector
(Bruker). The samples (50 mg) were measured from RT to 6508C
Keywords: alcohols · carbon · ruthenium · supercritical fluids ·
supported catalysts
[1] M. Gassner, F. Vogel, G. Heyen, F. MarØchal, Energy Environ. Sci. 2011, 4,
1726.
[2] F. Vogel in Handbook of Green Chemistry (Ed.: R. H. Crabtree), Wiley-VCH,
Weinheim, 2009, pp. 281–324.
[3] A. A. Peterson, F. Vogel, R. P. Lachance, M. Froeling, M. J. Antal Jr., J. W.
Tester, Energy Environ. Sci. 2008, 1, 32–65.
[4] P. Azadi, R. Farnood, Int. J. Hydrogen Energy 2011, 36, 9529–9541.
[5] D. C. Elliott, Biofuels Bioprod. Biorefining 2008, 2, 254–265.
[6] G. Peng, M. Steib, F. Gramm, C. Ludwig, F. Vogel, Catal. Sci. Technol. 2014,
4, 3329.
[7] H. Zçhrer, F. Mayr, F. Vogel, Energy Fuels 2013, 27, 4739–4747.
[8] A. G. Chakinala, S. Kumar, A. Kruse, S. R. A. Kersten, W. P. M. van Swaaij,
D. W. F. Brilman, J. Supercrit. Fluids 2013, 74, 8–21.
´
[9] A. Borodzinski, M. Bonarowska, Langmuir 1997, 13, 5613–5620.
with
a a flow of O2/Ar (20:80,
ramp of 108CminÀ1 under
20 mLminÀ1). The catalysts were tested in a fixed-bed plug flow re-
actor described in detail elsewhere.[6] Isopropanol was purchased
from VWR BDH Prolabo (99.8%). The feed rate was kept at
Received: September 8, 2015
Revised: October 27, 2015
Published online on December 3, 2015
ChemCatChem 2016, 8, 139 – 141
www.chemcatchem.org
141
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim