Understanding the role of TiO2 crystal structure on the enhanced activity and stability of Ru/TiO2 catalysts for the conversion of lignin-derived oxygenates

Article abstract of DOI:10.1039/c3gc41377b

Although Ru catalysts supported on reducible oxides such as TiO₂ hold significant promise for the deoxygenation of biomass derived oxygenates, a significant drawback is their instability under oxidation conditions necessary for catalyst regeneration. In this contribution, the role of TiO₂ crystal structure on resistance to metal particle sintering during calcination treatments at 400 and 500 °C is investigated. The resulting impact of the calcination temperature and TiO₂ support phase for the conversion of guaiacol at 400 °C under atmospheric pressure of hydrogen over supported Ru catalysts is presented. Results suggest that the rutile TiO₂ phase plays an important role in stabilizing Ru particles during calcination pretreatment in comparison with anatase supported Ru catalysts. Furthermore, rates normalized to the area of the support and the Ru suggest that the high activity of Ru/TiO₂ systems for guaiacol conversion is attributed to defect sites created by hydrogen spillover from the Ru metal to the reducible TiO₂ as opposed to only the sites located at the Ru/TiO₂ interface.

Full text of DOI:10.1039/c3gc41377b

Green Chemistry
View Article Online
View Journal | View Issue
PAPER
Understanding the role of TiO crystal structure on
2
the enhanced activity and stability of Ru/TiO2
catalysts for the conversion of lignin-derived
oxygenates†
Cite this: Green Chem., 2014, 16,
45
6
Taiwo Omotoso, Sunya Boonyasuwat and Steven P. Crossley*
Although Ru catalysts supported on reducible oxides such as TiO
genation of biomass derived oxygenates, a significant drawback is their instability under oxidation con-
ditions necessary for catalyst regeneration. In this contribution, the role of TiO crystal structure on
resistance to metal particle sintering during calcination treatments at 400 and 500 °C is investigated. The
resulting impact of the calcination temperature and TiO support phase for the conversion of guaiacol at
00 °C under atmospheric pressure of hydrogen over supported Ru catalysts is presented. Results
2
hold significant promise for the deoxy-
2
2
4
2
suggest that the rutile TiO phase plays an important role in stabilizing Ru particles during calcination pre-
Received 11th July 2013,
Accepted 21st October 2013
treatment in comparison with anatase supported Ru catalysts. Furthermore, rates normalized to the area
of the support and the Ru suggest that the high activity of Ru/TiO systems for guaiacol conversion is
attributed to defect sites created by hydrogen spillover from the Ru metal to the reducible TiO as
opposed to only the sites located at the Ru/TiO interface.
2
DOI: 10.1039/c3gc41377b
2
www.rsc.org/greenchem
2
properties necessitate additional catalytic upgrading of the oil
to make it suitable for use as a fuel source with existing
Introduction
2
Finite petroleum reserves and environmental considerations infrastructure.
serve as motivation for developing alternative sources of
Bio-oil is also highly unstable, and many of the compounds
energy that are not only environmentally friendly but also present in the complex mixture react over time after the liquid
1
,2
economically viable.
Thermochemical conversion routes product is collected. It has been shown that while the mole-
such as fast pyrolysis, gasification and liquefaction of biomass cular weight and viscosity of untreated bio-oil changes signifi-
have been proposed as possible strategies for the production cantly upon aging, the stability is greatly improved after
1
7
of renewable fuels and chemicals from lignocellulosic catalytic upgrading of the oil vapors prior to condensation.
biomass.
1
,3–11
Fast pyrolysis is a popular technique for the con- This demonstrates the importance of tuning the properties of
version of biomass into liquid fuels, and when combined with the catalyst that would come in contact with the pyrolysis
the process of catalytic upgrading to produce stable deoxy- vapors before they condense into the liquid phase to convert
genated products in the liquid fuel boiling range, several the compounds that cause the instability of the oil while also
studies have shown this to be a less costly route compared to minimizing reactions which lead to the loss of valuable
1
,12,13
18
other processes such as gasification and fermentation.
Fast pyrolysis oil, which is the liquid phase product derived
carbon–carbon bonds in the product.
Ruthenium catalysts have shown promise for upgrading of
from fast pyrolysis, commonly referred to as bio-oil, contains bio-oil vapors and have been found to be active under
over 400 different compounds that are representative of the conditions in which oxygenated compounds and water are
1
7,18
thermal break-down of the hemicellulose, cellulose and lignin present.
It should also be noted that, while Ru is expensive
1
3
fractions of biomass. These compounds confer undesir- upon comparison with metals such as Ni or Fe, Ru is more active
1
9–21
able characteristics such as low energy density, high oxygen for HDO
and an order of magnitude cheaper than Pt and Pd.
1
4–16
content, high viscosity, corrosivity, among others.
These
In spite of these advantages, the use of ruthenium catalysts
for industrial applications has been limited primarily due to
the high mobility of the ruthenium oxides in high temperature
oxidation environments. These conditions employed either
during synthesis/pre-treatment, during reaction, or during the
regeneration by burning off coke deposits lead to sintering
School of Chemical, Biological and Materials Engineering, University of Oklahoma,
Norman, OK 73019, USA. E-mail: stevencrossley@ou.edu
†
Electronic supplementary information (ESI) available. See DOI:
0.1039/c3gc41377b
1
This journal is © The Royal Society of Chemistry 2014
Green Chem., 2014, 16, 645–652 | 645

View Article Online
Paper
Green Chemistry
of Ru species. It is pertinent therefore to develop stable ruthe- Temperature-programmed reduction (TPR) of the catalysts was
−
1
nium catalysts that can be useful for industrial applications. carried out by flowing 5% H2 in Ar at 30 mL min over
Critical to this development is the improved basic understand- 30–40 mg of sample. The temperature ramp was set at a 10 °C
−
1
ing of the role of support phase in the stabilization of Ru par- min linear rate from room temperature to 800 °C. An SRI
ticles under oxidizing conditions, and the resulting impact on 110 thermal conductivity detector (TCD) was the analyzer of
catalytic performance and the ability to create highly active the effluent gas that was dried with drierite before introduc-
sites for the conversion of oxygenates.
Recent results from our group demonstrate that the combi- at the same flow rate. The amount of ruthenium metal on each
nation of Ru with the reducible oxide support TiO results in a catalyst was determined by quantitative TPR using a standard
significant enhancement in both the activity and catalyst stabi- copper(II) oxide sample (Sigma-Aldrich) to determine the
tion into the TCD and then compared with the pure argon gas
2
1
8
lity for the conversion of lignin-derived model compounds.
This same catalyst prepared by our group was also tested in present as RuO
the presence of real bio-oil vapors under similar conditions, X-ray diffraction (XRD) determinations were made using a
demonstrating promise for improving the thermal stability of Rigaku Automatic diffractometer (Model D-MAX A) equipped
hydrogen uptake on each catalyst assuming all of the Ru is
2
.
1
7
the resulting bio-oil.
In this work, the effect of calcination temperature and TiO
with a curved crystal monochromator. It was operated at 40 kV
and 35 mA with Cu Kα radiation as a source using an angle
2
support phase on the activity and selectivity of Ru/TiO for the range of 10–75° with a count time of 1.0 s and a step size of
2
conversion of a model compound representing lignin-derived 15. The well ground catalyst sample was put on a plastic slide
species present in bio-oil, guaiacol, under atmospheric and spread to get a flat surface. In order to estimate Ru metal
pressure of hydrogen at 400 °C is presented. It is shown that particle size by XRD, samples were pre-reduced in H flow at
2
an increase in calcination temperature has little effect on the 400 °C for 1 h and then cooled down to room temperature in
properties of P25-supported Ru catalysts, consisting of a mix of nitrogen.
anatase and rutile phases, when compared to purely anatase-
High resolution transmission electron microscopy (TEM,
supported Ru. In addition, evidence is presented that indicates JEOL JEM-2100 model) was employed to determine the Ru par-
that the new active sites resulting from the combination of Ru ticle size distributions. Samples were pre-reduced as described
and TiO for the conversion of guaiacol are due to the spillover above. The samples were then suspended in ethanol and soni-
2
of hydrogen to the support to create defects, and not necess- cated to obtain a uniform suspension before depositing a few
arily the interface between the metal and the support.
drops on carbon-coated copper TEM grids for measurements.
Over 100 particles were analyzed for the particle size distri-
bution of each sample.
Experimental
Catalytic activity measurements
Catalyst preparation
Catalytic activity was tested in a 1/4 inch OD stainless steel
reactor tube operating at 400 °C and atmospheric pressure.
The catalyst (15–200 mg, 40–60 mesh) was fixed between two
layers of quartz wool in the reactor tube which was then filled
with inert glass beads (1 mm diameter) from the feed inlet to
the packed catalyst bed in order to obtain an appropriate
vaporization zone. The reactor outlet connects to a six-port
valve with the lines heated to 300 °C to prevent the conden-
sation of vapors. An ice–water system was used to collect
the liquid product in a condenser. An external thermocouple
was placed at the center of the catalyst bed with an online gas
chromatograph HP 5890 GC equipped with a flame ioniza-
tion detector (FID), and a HP-5 column (30 m, 0.25 μm) used
to analyze the reactant conversion and product distribution.
Catalyst pretreatment was carried out under flowing hydrogen
Ru catalysts were synthesized by incipient wetness impreg-
nation of the titanium(IV) oxide (Alfa Aesar, Catalyst support,
Anatase 1/8″ pellets or Aeroxide P25) support with an aqueous
solution of the ruthenium(III) chloride hydrate precursor
(Aldrich, 99.98% trace metals basis). The pore volume of the
3
−1
titanium oxide support was 0.38 cm
g
pure anatase and
3
−1
22–24
0
.25 cm g P25, with an anatase/rutile ratio of ∼4/1.
The
pure anatase titanium oxide pellets were reduced to a particle
size smaller than 500 μm, dried overnight for 12 h and then
cooled down to room temperature prior to impregnation
while the P25 support was already in powder form. After impreg-
nation, the catalyst was dried at 120 °C for 12 h, and then oxi-
dized in flowing air at either 400 or 500 °C for 4 h. Then the
catalyst was pelletized, crushed and sieved to yield particle sizes
ranging from 250 to 420 μm (Mesh no. 40–60). Catalysts sup-
ported on anatase calcined at 400 and 500 °C are hereafter
referred to as RuTi400 and RuTi500, respectively. Similarly, cata-
lysts prepared on P25 are denoted as RuTiP400 and RuTiP500.
−
1
at 100 mL min
1
with a linear temperature ramp rate of
0 °C min− from room temperature to 400 °C and then held
1
for 1 h before the feeding began.
Ethylene hydrogenation reactions were carried out at 40 °C
and at atmospheric pressure in the same flow reactor system
described above. The catalysts were reduced at 400 °C and
Catalyst characterization
The BET surface area was measured by nitrogen adsorption on then cooled down to 40 °C prior to the introduction of the
a Micromeritics ASAP 2010 instrument. Prior to each determi- ethylene and hydrogen mixture at a molar ratio of 1 : 50. An
nation, the samples were degassed for 3 h at 300 °C. online gas chromatograph system with a flame ionization
646 | Green Chem., 2014, 16, 645–652
This journal is © The Royal Society of Chemistry 2014

View Article Online
Green Chemistry
Paper
2 3
detector and a Varian CP-Al O PLOT column (50 m, 0.32 μm)
was used to analyze the product stream and reactant
conversion.
Results and discussion
Catalyst characterization and Ru mobility
The surface areas for the pure titania support and ruthenium
catalysts are presented in Table 1. All the catalysts have surface
2
−1
Fig. 1 Temperature programmed reduction profiles of (a) RuTi400, (b)
RuTi500, (c) RuTiP400, (d) RuTiP500.
areas within the range of 50–170 m g . It can be seen that
even though the surface area of the P25 support is almost
three times lower than that of pure anatase, the loss in surface
area upon introduction of the Ru metal on this support is little
compared to the Ru supported on pure anatase. The anatase
polymorph of TiO is thermodynamically less stable than the
2
rutile phase. Increases in calcination temperature lead to a
loss in the surface area of pure anatase, which has been
2
5,26
ascribed to pore collapse.
P25 is a mixture of both anatase
and rutile phases, with a lower surface area compared to the
pure anatase, resulting in improved stability upon increases in
calcination temperature.
Temperature Programmed Reduction (TPR) profile peaks
for the catalysts are shown in Fig. 1. The reduction of the RuO
species occurs around 160 °C. The small peak at about
2
2 2
Fig. 2 XRD profiles for unreduced samples of (a) TiO -anatase, (b) TiO
2
7
P25, (c) RuTi400, (d) RuTi500, (e) RuTiP400, (f) RuTiP500, A and R
1
00 °C for the RuTi400 catalyst has been identified as occur- denote the peaks corresponding to anatase and rutile crystal phases,
2
8
respectively.
3 3
ring due to the reduction of RuCl species. The RuCl
species is as a result of the residual chloride on the sample
even after calcination at 400 °C. A higher calcination tempera-
ture of 500 °C is sufficient to eliminate these species hence the
absence of this peak in the TPR profile of RuTi500. The metal
loading of each catalyst as determined by quantitative TPR is
shown in Table 1. It can be seen that the Ru metal was not lost
as a result of increasing calcination temperature for all the
ruthenium catalysts. Therefore the difference in catalytic
activity cannot be attributed to this cause.
XRD profiles for the catalysts are presented in Fig. 2 and 3.
No bulk phase change from anatase to rutile due to the calci- Fig. 3 XRD profiles for reduced samples of (a) RuTi400, (b) RuTi500, (c)
RuTiP400, (d) RuTiP500.
nation temperature for RuTi400 and RuTi500 is observed
so the only bulk phase present for this titania support is
anatase as seen from the absence of any rutile peaks. The dis- RuTi500 at 2θ = 28° and 34.95° but this peak overlaps with the
tinct anatase and rutile phases are observed on the RuTiP400 rutile phase of the P25 samples, RuTiP400 and RuTiP500
and RuTiP500 as expected. In Fig. 2, the RuO peaks on the within the range of 2θ = 26.95° to 28.65° and 34.75° to 36.75°.
2
unreduced catalyst samples can be seen for RuTi400 and The metallic Ru peak can be seen at 2θ = 44° for all the
Table 1 Measured physical properties and characteristics of the catalysts used in this study
Diameter (nm)
XRD
Dispersion (%)
XRD
BET (m g−1)
2
Ru wt%
TEM
TEM
Catalyst
T
calcination (°C)
TiO
TiO
RuTi400
RuTi500
RuTiP400
RuTiP500
2
anatase
P25
—
—
400
500
400
500
165
60
150
102
55
—
—
3.66
3.86
3.82
3.96
—
—
5.5
7.7
5.3
5.7
—
—
3.3
—
—
16.3
11.7
17.1
15.8
—
—
27.3
2
2.4
2.7
37.5
33.4
52
This journal is © The Royal Society of Chemistry 2014
Green Chem., 2014, 16, 645–652 | 647

View Article Online
Paper
Green Chemistry
reduced catalyst samples in Fig. 3. It is important to note that the poor dispersion on this support due to the difference in
3
6
under the calcination conditions reported here, no measurable crystal structures between anatase TiO and rutile RuO .
2
2
shift of the bulk crystal structure from anatase to rutile
The results presented here are in agreement with the find-
TiO is observed in any of the cases. For the P25 supported ings by Carballo et al., therefore it is reasonable to propose
2
samples, no shift is observed in the ratio of intensities of that the high temperature calcination treatment on the pure
peaks representing anatase (2θ = 25.5°, 38.1°, 48.2°, and 54°) anatase support could lead to an agglomeration of the Ru oxide
to rutile (2θ = 27.6° and 36.3°) upon calcination. This is in particles since they are not stabilized on this titania phase
3
6
agreement with the literature results that demonstrate the hence the very low dispersion measured on RuTi500.
anatase–rutile transition at much higher temperatures than In addition to the average metal particle size due to agglo-
meration of RuO crystals, Ru/TiO systems are known to exhibit
Ruthenium metal particle sizes were estimated using the strong metal support interactions under reducing conditions
2
9–34
those reported here.
2
2
3
9,40
Scherrer equation from XRD line broadening measurements. above ∼400 °C.
Because of this, it is necessary to measure
Table 1 shows that the Ru particle size increases with the calci- not only metal particle size, but also the exposed metal, as
nation temperature for the anatase supported Ru catalysts part of the metal may be covered by the support. When
while the change in particle size is not as significant for the supported on TiO , complications arise with most standard
2
P25 supported catalysts. This is a very interesting result consid- chemisorption techniques for a variety of reasons. Typical
ering the greater support surface area of anatase supports in hydrogen chemisorption can lead to overestimated metal
2
all cases. This trend in particle size was also confirmed with surface area over reducible supports such as TiO due to spill-
4
1
TEM as shown in Table 1. Representative TEM images are over to the reducible oxides. CO chemisorption can also
shown in Fig. S1–S3.† Metal dispersion was estimated by lead to exaggerated estimated conversions due to adsorption
assuming, in accordance with the previous literature, that a of CO on the reduced support.
blend of the 001, 100 and 110 planes are exposed for Ru with a
Because of these challenges, a structure-insensitive hydro-
2
metal atom surface area of 0.0909 nm and exposed metal genation reaction could serve as a probe for the exposed metal
3
35
atom volume of 0.01365 nm . The Ru metal particle sizes surface area. We have chosen low temperature ethylene hydro-
4
2–44
estimated for RuTiP400 and RuTiP500 were 2.4 nm and genation, a classical structure insensitive reaction,
.7 nm, respectively, which are similar despite the increase in probe of the exposed Ru. It should be noted that while ethy-
calcination temperature. This was not the case for the anatase lene hydrogenation is structure insensitive, the hydrogenolysis
as a
2
4
5
supported Ru catalysts, RuTi400 and RuTi500, where an of ethane is not. While higher temperatures can lead to a
4
6
increase in particle size was observed upon calcination at variety of surface species, we observe no hydrogenolysis
00 °C. It is important to note that the average particle sizes products under the conditions utilized here.
estimated via XRD are larger than those estimated via TEM. The rates of ethylene hydrogenation compared at 50% con-
5
This is likely due to the inability of XRD measurements to version of ethylene in all cases are reported in Table 2. These
detect the smaller particles due to line broadening, shifting rates were used to estimate the percentage of exposed surface
the particle size distribution to larger diameters. The mobility area, by first assuming that the exposed surface area of
of ruthenium oxides in the presence of air is well known, and, RuTi400 was equal to the surface area estimated by TEM. This
as will be discussed below, the crystal phase of the support has rate was then used to estimate the percentage of exposed Ru
shown to have an influence on the mobility of the oxides and for all of the other Ru/TiO
final Ru particle sizes obtained. that no measurable conversion of ethylene was observed over
alone.
The ratios of the exposed Ru surface areas estimated from
2
catalysts. It is important to note
3
6
Carballo et al. compared different ruthenium based cata- TiO
lysts supported on pure anatase, P25 and silica–alumina using
2
infrared spectroscopy and CO hydrogenation reactions to eluci- all techniques follow a similar trend, with one exception. The
date the influence of the support on the properties and activity exposed Ru surface area is slightly higher over RuTiP400 and
of the catalysts. They found, in agreement with our results, RuTiP500 than over RuTi400 as measured by TEM and XRD,
that Ru supported on P25 had a significantly better dispersion while the exposed area among the three are similar when
3
6
than that supported on pure anatase TiO
2
.
The authors pro-
posed that the rutile phase of P25 can serve as an anchor for
the RuO2 species since they have similar crystal structures.
Electron microscopy results from the study showed that the Ru
Table 2 Ethylene hydrogenation rates at T = 40 °C and atmospheric
pressure at conversion ∼50%
3
6
was preferably anchored on the rutile phase of P25 TiO
2
,
and
2
this is reasonable based on findings that the anatase and
Exposed Ru (m )
3
7
per g catalyst
rutile phases in P25 exist distinctly. The differentiation
Ethylene hydrogenation
between the anatase and rutile phases was achieved by either
Catalyst
rate (mmol per g Ru per s)
TEM
Ethylene Hyd.
3
6,38
identifying the lattice d-spacing of both phases
or by
RuTi400
RuTi500
RuTiP400
6.51
1.81
6.51
6.12
5.4
—
7.6
7.2
5.4
1.6
5.5
5.4
indexing the diffraction patterns obtained from the high
3
6
resolution images of the particles. It was claimed that
agglomeration of the Ru species on pure anatase resulted in RuTiP500
648 | Green Chem., 2014, 16, 645–652
This journal is © The Royal Society of Chemistry 2014

View Article Online
Green Chemistry
Paper
estimated via ethylene hydrogenation rates. The discrepancy including anisole, deoxygenated aromatics, and alkylated
may be due to the presence of SMSI preferentially occurring species with greater than three methyl groups account for less
over the P25 samples, decreasing the relative amount of than ∼5% of the product selectivity below 100% guaiacol con-
exposed surface area for hydrogenation to a greater degree version in all cases.
upon reduction at 400 °C. This result is in agreement with
The very similar product distribution as a function of con-
observations that smaller Ru particles undergo encapsulation version implies that a similar type of active site is responsible
4
0
more readily than larger ones when supported on TiO
worth noting that if a fraction of the RuTi400 sample is encap- This is even more convincing considering the very different
sulated with amorphous TiO due to SMSI, the exposed ratios of TiO to exposed Ru surface area available for guaiacol
2
.
It is for guaiacol conversion across the various catalysts employed.
2
2
surface areas of Ru for all catalysts reported in Table 2 will be conversion across the various catalysts. We have previously
slightly smaller than reported, but the trends and conclusions shown that the rate of guaiacol conversion over Ru/TiO2 is
will be the same.
greatly enhanced when compared with Ru supported on
1
8
other non-reducible supports, as well as over TiO
2
alone.
2
Hydrodeoxygenation of guaiacol over Ru/TiO catalysts
This synergy was proposed to be due to the creation of new
1
8
Our group has recently shown that a synergistic effect exists highly active sites either at the interface between the metal
when combining Ru with the reducible oxide TiO
for the con- and the support, or through spillover to create oxygen
2
version of guaiacol. Ru supported on TiO was found to be sig- vacancies on the support. Because the product selectivity is so
2
nificantly more active than the other non-reducible supports similar across the various catalysts, the rate of guaiacol conver-
1
8
such as C, SiO
ingly, it was observed that while Ru supported on non-reduci- sites.
ble supports containing Lewis acidity such as Al catalyzed
Table 3 shows the rates of guaiacol conversion estimated at
transalkylation to methylcatechols, no dioxygenated species a constant guaiacol conversion of 50% normalized to a surface
were observed over Ru/TiO2 catalysts. It was proposed that area of Ru and a surface area of TiO . While the rates reported
defects created on TiO
due to the presence of Ru play an here are lower than the initial rates, the near identical selec-
important role in guaiacol deoxygenation.
tivity of the various catalysts implies that the conclusions drawn
2 2 3
and Al O
for guaiacol conversion. Interest- sion can help to better understand the nature of these active
2 3
O
2
2
The results for the conversion of guaiacol to the major on this basis will hold true in both cases. Rates per Ru surface
products consisting of phenol, cresols, di- and tri-alkylated area are expressed both in terms of the surface areas estimated
phenols, and deoxygenated aromatics as a function of W/F are via TEM, as well as those estimated via ethylene hydrogenation
shown in Fig. 4. While it is clear that the rates of conversion to reported in Table 2. It is immediately obvious from Table 3
each product are substantially different, what may not be that the rate per Ru atom is greatly enhanced when compared
immediately evident is the remarkably similar product selec- to the rate per surface area of Ru supported on SiO
tivities obtained over all of the catalysts. This trend can be difference in rates is conservative, as the rate over Ru/SiO
2
. This
was
2
observed in Fig. 5. The sum of all other minor products, obtained at a lower conversion of 23% with an exposed
Fig. 4 Conversion/yield (mol%) vs. W/F for guaiacol conversion on (a) RuTi400, (b) RuTi500, (c) RuTiP400, (d) RuTiP500, T = 400 °C, P = 1 atm,
(
2
mol H per mol guaiacol) = 60, TOS = 50 min.
This journal is © The Royal Society of Chemistry 2014
Green Chem., 2014, 16, 645–652 | 649

View Article Online
Paper
Green Chemistry
Fig. 5 Selectivity vs. guaiacol conversion to phenol, cresols, and di-and tri-alkylated phenols over RuTi400 (hollow triangles), RuTi500 (solid tri-
angles), RuTiP400 (hollow circles), and RuTiP500 (filled circles), T = 400 °C, P = 1 atm, (mol H per mol guaiacol) = 60, TOS = 50 min.
2
Table 3 Rates of guaiacol conversion per metal and TiO
for all catalysts, T = 400 °C, P = 1 atm, H
/guaiacol = 60 mol mol−
TOS = 90 min, Conversion ∼50% unless otherwise noted
2
surface areas area estimated from ethylene hydrogenation. This is an inter-
1
2
,
esting result since it could be expected that the smaller Ru par-
ticle size of RuTiP400 when compared with RuTi400 would
result in higher guaiacol conversion rates. Several reactions are
known to have an enhancement at the perimeter between
the active metal and the TiO2 due to the ability to activate
specific bonds. For example, Resasco and Haller demonstrated
that the normalized rate of hydrogenolysis of ethane and
2
Rate (mol per m Ru per h)
4
a
×
10
2
Rate (mol per m
TiO per h) × 10
2
4
Catalyst
TEM
Ethylene Hyd.
RuTi400
RuTi500
RuTiP400
142
—
52
53
2
141
248
73
70
—
5.3
3.9
8.6
8.5
—
butane over Rh/TiO catalysts could be explained not by the
2
45
exposed Rh surface area, but by the normalized perimeter.
RuTiP500
b
Bell et al. demonstrated the importance of the Ru/TiO inter-
Ru/SiO
2
2
c
39
TiO
2
—
—
0.07
face for the activation of C–O in Fischer–Tropsch synthesis.
a
It is important to note that in our results, the exposed surface
ethylene hydrogenation rates. Rates obtained at a guaiacol conversion area of Ru is similar for both RuTi400 and RuTiP400, although
Rates were estimated using particle sizes estimated from TEM and
b
c
of 23%. Rates obtained at a guaiacol conversion of 14%.
the rate per exposed Ru surface area is roughly double or more
than double over the higher surface area anatase supported
catalysts. In terms of the perimeter around the Ru/TiO inter-
2
2
−1
18
Ru surface area estimate at 6.6 m g as previously reported,
which is within the range of particle sizes obtained for the Ru/ face, the TEM results indicate that the particles supported on
TiO catalysts employed in this study. This suggests that the P25 are smaller than those on anatase, while the exposed
2
Ru alone is not the active species responsible for guaiacol con- surface area is the same. Based on these results, the perimeter
version. Similarly, the rate of conversion per TiO
surface area surface area would be greater over RuTiP400 and RuTiP500, so
2
is greatly enhanced over each Ru/TiO catalyst when compared the perimeter sites are likely not solely responsible for the
2
to pure anatase TiO
results confirm the significant synergy between the Ru and
TiO to create active sites that are nearly two orders of magni- exposed surface area of TiO
2
at a guaiacol conversion of 14%. These enhanced activity.
The rates of guaiacol conversion shown in Table 3 per
demonstrate the opposite trend,
alone. The important with the higher rates observed on the P25 supported samples
question to address is whether these active sites are around indicating that the rate of hydrogen spillover over the TiO
2
2
tude more active than either Ru or TiO
2
2
the perimeter of the metal particles at the Ru/TiO interface, surface may be greater over the rutile polymorph. This result is
2
or whether they are active sites due to oxygen vacancies created in agreement with the results of reduction kinetics from
on the TiO
From Table 3, the rate of guaiacol conversion per TiO2 rutile TiO
surface area is higher for RuTiP400 when compared to of defects and rates of migration of defects on anatase and
RuTi400 while the opposite trend is observed for the rate per rutile phases of TiO are different, with a greater amount of
exposed Ru metal surface area. This trend holds true regard- defects existing on the surface of rutile TiO than anatase
less of whether the rate per Ru surface area is expressed in TiO . They also found that the rate of reduction on rutile
terms of surface area estimated via TEM or exposed surface was 2.5 times faster than anatase, and they attributed this
2
support that result from hydrogen spillover.
Barteau et al., who compared Ru supported on anatase and
4
7
2
.
They demonstrated that both the surface density
2
2
2
650 | Green Chem., 2014, 16, 645–652
This journal is © The Royal Society of Chemistry 2014

View Article Online
Green Chemistry
Paper
difference to the varying cation densities on the low
index planes. In the absence of Ru, they demonstrated through
kinetic fitting that the rate limiting step for reduction was the
interaction between a H atom adsorbed on the surface and a
surface O atom. They proposed that a possible explanation
could be due to a limited quantity of adsorbed H atoms on the
surface. The addition of Ru shifted this step to a non rate-limit-
ing step, shifting the limiting step of the reduction to the rate
of defect healing from the bulk. It was proposed that a possi-
ble role of the Ru is to supply more H atoms to the surface via
spillover. In the case of guaiacol conversion, however, it is con-
ceivable that one of the oxygen atoms in guaiacol heals the
defects on the surface at a more rapid rate and shifts the rate-
3 A. V. Bridgwater and G. V. C. Peacocke, Renewable Sustain-
able Energy Rev., 2000, 4(1), 1–73.
4 D. C. Elliott, D. Beckman, A. V. Bridgwater, J. P. Diebold,
S. B. Gevert and Y. Solantausta, Energy Fuels, 1991, 5,
399.
5 A. V. Bridgwater, Chem. Eng. J., 2003, 91(2–3), 87–102.
6 A. V. Bridgwater, J. Anal. Appl. Pyrolysis, 1999, 51, 3–22.
7 A. V. Bridgwater, Therm. Sci., 2004, 8, 21–49.
8 D. L. Klass, Biomass for Renewable Energy, Fuels and Chemi-
cals, Academic Press, San Diego, 1998.
9 A. A. C. M. Beenackers and W. P. M. Swaaij, in Thermo-
chemical Processing of Biomass, ed. A. V. Bridgwater, Butter-
worths, London, U.K., 1984.
limiting step back to hydrogen spillover and migration across 10 J. J. Hos and M. J. Groeneveld, in Biomass, ed. D. O. Hall
the surface.
It is proposed that the difference in rates per TiO
and R. P. Overend, Wiley and Sons, Chichester, U.K., 1987.
surface 11 J. M. Moffatt and R. P. Overend, Biomass, 1985, 7, 99.
2
area and per exposed Ru surface area observed between 12 T. P. Vispute, H. Zhang, A. Sanna, R. Xiao and G. W. Huber,
RuTi400 and RuTiP400 can be explained by the roles played by Science, 2010, 330, 1222–1227.
defects in the TiO during guaiacol deoxygenation to form phe- 13 G. W. Huber, S. Iborra and A. Corma, Chem. Rev., 2006,
nolics. The rutile phase of TiO has been shown to have more 106, 4044–4098.
surface defects compared to the anatase phase as discussed 14 S. Czernik and A. V. Bridgwater, Energy Fuels, 2004, 18,
above. It can therefore be envisioned that more defects, 590–598.
2
2
created by a greater ability of Ru to spillover hydrogen on P25, 15 A. Oasmaa and S. Czernik, Energy Fuels, 1999, 13, 914–921.
could be generated on RuTiP400, which could explain the 16 G. V. C. Peacocke, P. A. Russel, J. D. Jenkins and
higher rates per total surface area.
A. V. Bridgwater, Biomass Bioenergy, 1994, 7, 169–178.
7 S. Wan, T. Pham, S. Zhang, L. Lobban, D. Resasco and
R. Mallinson, AIChE J., 2013, 59, 7.
8 S. Boonyasuwat, T. Omotoso, D. E. Resasco and
S. P. Crossley, Catal. Lett., 2013, 143, 783–791.
1
1
Conclusions
It has been shown that the titania support phase plays an 19 E. P. Maris and R. J. Davis, J. Catal., 2007, 328–337.
important role in preventing agglomeration of ruthenium 20 D. C. Elliott and T. R. Hart, Energy Fuels, 2009, 23, 631.
species under oxidizing conditions. While the size of the 21 J. Wildschut, F. H. Mahfud, R. H. Venderbosch and
Ru particles on pure anatase supported Ru/TiO
largely affected by increases in calcination temperature, it was 22 R. I. Bickley, T. Gonzalez-Carreno, J. S. Lees, L. Palmisano
proposed that these particles were stabilized by the rutile and R. J. D. Tilley, J. Solid State Chem., 1991, 92, 178–190.
phase of the P25 support during calcination and this led to 23 V. Puddu, H. Choi, D. D. Dionysiou and G. L. Puma, Appl.
similar particle sizes upon calcination at higher temperatures. Catal., B, 2010, 94, 211.
This results in the potential for enhanced stability after catalyst 24 G. Busca and G. Ramis, J. Chem. Soc., Faraday Trans., 1994,
regeneration. Furthermore, by comparing the guaiacol conver- 90(20), 3181–3190.
sion rates normalized to the surface area of the metal and the 25 N. Wetchakun, B. Incessungvorn, K. Wetchakun and
support, it was concluded that hydrogen spillover to produce S. Phanichphant, Mater. Lett., 2012, 82, 195.
oxygen vacancies is the primary mechanism for the enhanced 26 T. Sreethawong, Y. Suzuki and S. Yoshikawa, J. Solid State
2
was
H. J. Heeres, Ind. Eng. Chem. Res., 2009, 48, 10324–10334.
guaiacol conversion over Ru/TiO
2
catalysts.
Chem., 2005, 178, 329.
2
2
7 D. Li, N. Ichikuni, S. Shimazu and T. Uematsu, Appl. Catal.,
A: General, 1999, 180, 227–235.
8 A. Bossi, F. Barbassi, A. Orlandi, G. Petrini and
L. Zanderighi, Stud. Surf. Sci. Catal., 1979, 3, 405.
Acknowledgements
Financial support from the National Science Foundation 29 T. Sreethawong, Y. Suzuki and S. Yoshikawa, J. Solid State
EPSCOR (0814361) is gratefully acknowledged.
Chem., 2005, 178, 329–338.
3
3
0 C. N. R. Rao, Can. J. Chem., 1961, 39, 498.
1 A. W. Czanderna, C. N. R. Rao and J. M. Honig, Trans.
Faraday Soc., 1958, 54, 1069.
References
32 E. F. Huer and C. W. Werss, Am. Mineral., 1972, 57, 10–23.
1
D. Mohan, C. U. P. Pittman Jr. and P. H. Steele, Energy 33 R. I. Bickley, T. Gonzalez-Carreno, J. S. Lees, L. Palmisano
Fuels, 2006, 20, 848–889.
and R. J. D. Tilley, J. Solid State Chem., 1991, 92, 178–
2
M. Stocker, Angew. Chem., Int. Ed., 2008, 47, 9200–9211.
190.
This journal is © The Royal Society of Chemistry 2014
Green Chem., 2014, 16, 645–652 | 651

View Article Online
Paper
Green Chemistry
3
3
4 F. C. Gennari and D. M. Pasquevich, J. Mater. Sci., 1998, 33, 40 T. Komaya, A. T. Bell, Z. Weng-Sieh, R. Gronsky, F. Engelke,
571–1578. T. S. King and M. J. Pruski, J. Catal., 1994, 149, 142.
5 X. Shen, L. Garces, Y. Ding, K. Laubernds, R. P. Zerger, 41 R. Prins, Chem. Rev., 2012, 112, 2714–2738.
1
M. Aindowa, E. J. Neth and S. L. Suib, Appl. Catal., A: 42 C. H. Bartholomew and R. J. Farrauto, Fundamentals of
General, 2008, 335, 187.
Industrial Catalytic Processes, John Wiley & Sons Inc.,
3
6 J. M. G. Carballo, E. Finocchio, S. Garcia, S. Rojas,
Hoboken, NJ, 2nd edn, 2006.
M. Ojeda, G. Busca and J. L. G. Fierro, Catal. Sci. Technol., 43 G. A. Somorjai, Introduction to Surface Chemistry and Cata-
011, 1, 1013–1023. lysis, John Wiley & Sons Inc., New York, 1994.
7 T. Ohno, K. Sarukawa, K. Tokieda and M. Matsumura, 44 J. N. Kuhn, C. K. Tsung, W. Huang and G. A. Somorjai,
J. Catal., 2001, 203, 82. J. Catal., 2009, 265, 209–215.
8 V. Puddu, H. Choi, D. D. Dionysiou and G. L. Puma, Appl. 45 D. E. Resasco and G. Haller, J. Catal., 1983, 82, 279–288.
2
3
3
3
Catal., B, 2010, 94, 211.
46 D. S. Jordan and A. T. Bell, J. Phys. Chem., 1986, 90, 4797–
9 T. Komaya, A. T. Bell, Z. Weng-Sieh, R. Gronsky,
4805.
F. Engelke, T. S. King and M. J. Pruski, J. Catal., 1994, 47 J. E. Rekoske and M. A. Barteau, J. Phys. Chem. B, 1997,
50, 400–406. 101, 1113–1124.
1
652 | Green Chem., 2014, 16, 645–652
This journal is © The Royal Society of Chemistry 2014