Selective production of 1,2-propanediol by hydrogenolysis of glycerol over bimetallic Ru-Cu nanoparticles supported on TiO2

Article abstract of DOI:10.1016/j.apcata.2014.06.002

A series of TiO₂ supported Ru-Cu bimetallic catalysts was investigated for the hydrogenolysis of glycerol. The catalysts were characterized by H₂ chemisorption, X-ray diffraction (XRD), scanning transmission electron microscopy (STEM/EDS) and X-ray photoelectron spectroscopy (XPS). The addition of copper led to an increase in the selectivity to 1,2-propanediol. The highest activity and selectivity was observed for a 1:1 Cu/Ru mass ratio. The results were explained in terms of the interaction between Ru and Cu. Copper diluted large Ru ensembles that are responsible for CC bond cleavage leading to ethylene glycol production instead of the desired 1,2-propanediol. The presence of Cu also inhibited the deactivation exhibited by monometallic Ru. Among all the catalysts, the 2.5Ru-2.5Cu/TiO₂ catalyst exhibited the best performance in the hydrogenolysis of glycerol.

Full text of DOI:10.1016/j.apcata.2014.06.002

Accepted Manuscript
Title: Selective Production of 1,2-Propanediol by
Hydrogenolysis of Glycerol over Bimetallic Ru-Cu
Nanoparticles Supported on TiO2
Author: Joyce B. Salazar Derek D. Falcone Hien N. Pham
Abhaya K. Datye Fabio B. Passos Robert J. Davis
PII:
S0926-860X(14)00376-7
DOI:
Reference:
http://dx.doi.org/doi:10.1016/j.apcata.2014.06.002
APCATA 14862
To appear in:
Applied Catalysis A: General
Received date:
Revised date:
Accepted date:
21-3-2014
27-5-2014
1-6-2014
Please cite this article as: J.B. Salazar, D.D. Falcone, H.N. Pham, A.K. Datye, F.B.
Passos, R.J. Davis, Selective Production of 1,2-Propanediol by Hydrogenolysis of
Glycerol over Bimetallic Ru-Cu Nanoparticles Supported on TiO , Applied Catalysis
2
A, General (2014), http://dx.doi.org/10.1016/j.apcata.2014.06.002
This is a PDF file of an unedited manuscript that has been accepted for publication.
As a service to our customers we are providing this early version of the manuscript.
The manuscript will undergo copyediting, typesetting, and review of the resulting proof
before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that
apply to the journal pertain.

Highlights




 Bimetallic Ru-Cu/TiO catalysts were investigated in the hydrogenolysis of glycerol.
2
 The addition of Cu to Ru/TiO enhanced the selectivity towards 1,2-PDO.
2
 Cu atoms presumably diluted large Ru ensembles that facilitated C-C bond cleavage.
 Among the bimetallic catalysts, 2.5Ru-2.5Cu/TiO exhibited the best performance.
2
1
Page 1 of 24

Selective Production of 1,2-Propanediol by Hydrogenolysis of Glycerol over Bimetallic Ru-
Cu Nanoparticles Supported on TiO2.
a
b
c
c
Joyce B. Salazar , Derek D. Falcone , Hien N. Pham , Abhaya K. Datye ,
a*
b*
Fabio B. Passos , and Robert J. Davis
a
c
Departamento de Engenharia Química e de Petróleo, Universidade Federal Fluminense, Rua Passo da Pátria, 156, Niterói,
RJ CEP 24210-240, Brazil
Department of Chemical Engineering, University of Virginia, 102 Engineers Way, Charlottesville, VA 22904-4741, USA
Department of Chemical & Nuclear Engineering, University of New Mexico, 203 Farris Engineering Center, MSC01 1120
Albuquerque, NM 87131, USA
b
*Corresponding authors E-mail addresses fbpassos@vm.uff.br (F.B. Passos), rjd4f@virginia.edu (R.J. Davis)
Abstract
A series of TiO
hydrogenolysis of glycerol. The catalysts were characterized by H
2
supported Ru-Cu bimetallic catalysts was investigated for the
chemisorption, X-ray
2
diffraction (XRD), scanning transmission electron microscopy (STEM/EDS) and X-ray
photoelectron spectroscopy (XPS). The addition of copper led to an increase in the selectivity
to 1,2-propanediol. The highest activity and selectivity was observed for a 1:1 Ru/Cu mass
ratio. The results were explained in terms of the interaction between Ru and Cu. Copper
diluted large Ru ensembles that are responsible for C-C bond cleavage leading to ethylene
glycol production instead of the desired 1,2-propanediol. The presence of Cu also inhibited
the deactivation exhibited by monometallic Ru. Among all the catalysts, the 2.5Ru-
2
.5Cu/TiO catalyst exhibited the best performance in the hydrogenolysis of glycerol.
2
Keywords: Ruthenium, Copper, Glycerol, 1,2-Propanediol, Deactivation.
1. Introduction
Biomass has been targeted as an alternative, renewable feedstock to fossil sources to
produce cleaner, more sustainable fuels as well as commodity chemicals. The integration of
biofuels and bio-chemicals within a so-called biorefinery can increase the viability of a bio-
based economy [1]. Glycerol, an important intermediate in the future bio-refinery, has been
singled out by US DOE and others as a candidate to produce a high number of valuable
commodity chemicals and intermediates [2].
The increasing demand for biodiesel will produce a large amount of glycerol, as a
byproduct, leading to a sharp decrease in the price of glycerol [1]. Several studies to convert
glycerol into higher value materials have been conducted in the past several years [3–10].
A viable application to utilize glycerol is the selective catalytic hydrogenolysis of
glycerol to propylene glycol (1,2-propanediol, 1,2-PDO). Propylene glycol is an industrially
important chemical and is extensively used in the production of polyester resins, functional
2
Page 2 of 24

fluids (antifreeze, deicing), paints and coatings, foods, pharmaceuticals and cosmetics. It is
currently produced at an industrial scale using fossil-fuel-derived propylene as the carbon
feedstock [9]. Producing 1,2-PDO from glycerol hydrogenolysis is an attractive option
because glycerol is a more sustainable feedstock compared to fossil sources.
Noble metal (Ru, Pt, Rh) [7, 11–18] and non-noble metal (Cu, Ni) [19–25] catalysts
have been tested for the hydrogenolysis of glycols. Among the transition metals, Ru has been
shown to be the most active catalyst for hydrogenolysis; however, it has been well known to
have high rates of C-C bond cleavage, which leads to excessive production of shorter chain
polyols [15, 26, 27].
In recent a study, Wang et al. [28] found that Ru exhibited a much higher activity in
glycerol hydrogenolysis compared to Rh, Pt and Pd, all supported on m-ZrO . They also
2
observed that the turnover rate increased three fold while there was a decrease in the
selectivity to 1,2-PDO as the Ru particle size increased from 1.8 to 4.5 nm. The dependence
of the rate of hydrogenolysis on the particle size and the surface catalyst structure has also
been studied previously elsewhere [29, 30].
In another study, Feng et al. [31] evaluated Ru catalysts supported on various
materials and they found that support and reduction temperature can influence the Ru particle
size as well as the catalytic performance. Among the supports tested, TiO
smallest and most stable Ru particles, which lead to the highest conversion of glycerol.
Copper-based catalysts supported on different metal oxides such as Al , Cr
ZnO, and SiO among others, have also been widely studied for glycerol hydrogenolysis [3,
9–24, 32, 33]. Copper is known to have low rates of C-C bond cleavage while still
2
created the
2
O
3
2 3
O ,
2
1
exhibiting high rates of C-O bond cleavage in hydrogenolysis reactions [34]. Therefore, when
Cu is used to catalyze the hydrogenolysis of glycerol it leads to a high selectivity of 1,2-PDO
and limits the excessive C-C bond cleavage that causes the formation of undesired products.
However, it has been observed that Cu catalysts exhibit low activity and high amounts of
particle agglomeration during reaction [28,29]. Recently, copper catalysts have been
modified by B O [22] and Ga O [36] in attempt to obtain better dispersion and stability.
2
3
2
3
Bimetallic Ru-Cu catalysts have been studied extensively in the past several decades
for a variety of applications [5, 37–41]. Bimetallic Ru-Cu nanoparticles supported on silica
were used to catalyze ethane hydrogenolysis and showed a significant decrease in activity as
the percentage of Cu increased compared to monometallic ruthenium catalyst. The
researchers suggested that Cu at the surface of the particles has a significantly lower rate of
C-C bond breaking then Ru [40, 42]. Moreover, Montassier and co-authors [43] suggested
3
Page 3 of 24

that Cu metal has an adsorption capacity for water and alcohols due to its electrophilicity and
this adsorption leads to a weakening of the H-O or C-O bond. This propensity for adsorbing
water and alcohols might be enhanced when copper is associated with noble metals or metals
with a lower d orbital electron occupancy, which can accept electrons from Cu leading to an
increase in the Cu atom’s electrophilicity.
Supported Ru-Cu bimetallic catalysts have also been applied to hydrogenolysis of
glycerol. Jiang et al. [44] employed Ru-Cu bimetallic nanoparticles supported on clay with
the aid of ionic liquids. A catalyst with a Ru:Cu molar ratio of 3:1 achieved 100% conversion
of glycerol and 85% yield of 1,2-PDO at 503 K with a H pressure of 8 MPa after 18h. More
2
recently, 100% conversion of glycerol and 78.5% yield of 1,2-PDO was found with a catalyst
that had a Cu:Ru molar ratio of 1:10 at 453 K and 8 MPa of H after 24 h of reaction [5].
2
Bimetallic Ru-Cu catalysts have been shown to exhibit good performance and
selectivity to 1,2-PDO for glycerol hydrogenolysis. Unfortunately, the reaction rates and
turnover frequencies for Ru-Cu catalyst have not been adequately described in the literature,
which prevents elucidation of fundamental structure-function relationships associated with
this class of material. Thus, in this work, a series of TiO supported Ru-Cu bimetallic
2
catalysts with different Cu/Ru molar ratios were prepared and their turnover frequencies were
evaluated for the hydrogenolysis of glycerol. Furthermore, H
2
chemisorption, X-ray
diffraction (XRD), scanning transmission electron microscopy (STEM/EDS) and X-ray
photoelectron spectroscopy (XPS) were used to characterize the metal components of the
catalysts.
2
2
. Experimental Methods
.1 Catalyst preparation
-1
The support, TiO (S : 50 m² g ), was obtained from Evonik-Degussa GmbH-P25.
2
BET
The precursors used for ruthenium and copper were Ru(NO)(NO ) in dilute nitric acid
3
3
(
Sigma-Aldrich) and Cu(NO ) ∙3H O (Alfa Aesar), respectively. The catalysts were prepared
3 2 2
by successive incipient wetness impregnations of Ru and Cu. The appropriate amount of
precursor was dissolved in distilled, deionized water and the solution was introduced to the
support using a ratio of 0.65 g of solution per gram of TiO . The solutions of Ru precursor
2
were added dropwise with thorough mixing between additions. After saturating the pores, the
material was dried at 393 K for 2 h, and then the next impregnation with Cu precursor
4
Page 4 of 24

solution was performed. Catalysts with different Cu/Ru weight ratios (1:2, 1:1, 2:1 and 4:1)
were prepared, as well as their monometallic counterparts, 2.5 wt% Ru and 2.5 wt% Cu.
3
-1
The catalysts were reduced in H
73 K for 4 h with a heating rate of 5 K min before reactions were performed. In addition,
the 2.5Ru/TiO and 2.5Ru-2.5Cu/TiO catalysts were reduced at two different temperatures
493 and 573 K) to evaluate the influence of reduction temperature on the dispersion of the
metal particles supported on TiO . The catalysts were identified by their weight percentages
of metal, such as: 2.5Ru/TiO ; 5Ru-2.5Cu/TiO2; 2.5Ru-2.5Cu/TiO ; 2.5Ru-5Cu/TiO ; 2.5Ru-
2
(GTS Welco, 5.0 UHP) flowing at 100 cm min at
-1
5
2
2
(
2
2
2
2
1
0Cu/TiO and 2.5Cu/TiO .
2 2
2.2 Catalyst characterization
Hydrogen chemisorption was performed in a Micromeritics ASAP 2020 automated
adsorption analyzer. The samples were reduced at 493 or 573 K under flowing H for 60 min,
2
and then outgassed under vacuum at 493 or 573 K for 60 min. After this, the catalysts were
cooled to 308 K and evacuated again for 60 min followed by analysis at 308 K. The total
2
amount of H adsorbed was calculated by extrapolating the hydrogen uptake to zero pressure,
assuming a stoichiometry H/Rusurf equal to unity. The mean particle size was calculated from
the metal dispersion D, assuming spherical particles, where the equation d (nm) = v/D was
applied. (where v =1.32) [45].
The X-Ray Diffraction (XRD) patterns were obtained using a PANalytical X'Pert Pro
MPD (Multi Purpose Diffractometer) instrument equipped with a Cu X-ray (λ Kα =1.54 Å)
tube source. The XRD patterns were collected with 2θ spanning from 20° to 90° with
-1
integration times of 30 s step and a step size of 0.05°. The spent catalysts were also analyzed
after being washed with distilled, deionized water and filtered with 0.45µm pore sized filter
paper, then dried overnight at 393 K.
The metal particle size was assessed by scanning transmission electron microscopy
(STEM) utilizing a JEOL 2010F transmission electron microscope operating at 200 kV, in
high-angle annular dark field (HAADF) mode, equipped with an energy dispersive X-ray
analysis spectrometer (EDS). The samples were dispersed in ethanol and mounted on either a
Cu mesh grid (monometallic Ru catalyst) or a gold mesh grid (rest of the catalysts).
The XPS analysis was performed with a Thermo ESCAlab 250S Spectrometer at the
Virginia Commonwealth University (VCU) in Richmond, Virginia, employing a focused
monochromatic Al Kα X-ray (1486.6 eV) source, running with a pass energy of 20 eV and a
5
Page 5 of 24

95 mm field emission gun. In this work, the peak of Ti (2p3/2) = 459.4 eV was taken as a
reference instead of the adventitious carbon peak because of the superposition of the C 1s and
Ru 3d peaks around 285 eV.
2.3 Hydrogenolysis of glycerol
Reactions were carried out with 50 mL of a 20% wt/wt aqueous glycerol (Acros,
9.6% purity) solution in a 100 mL stainless steel autoclave (Parr Instruments). The
9
previously reduced catalyst was loaded into the reactor with the solution. The reactor was
sealed and flushed with flowing H at atmospheric pressure and room temperature for 25 min
2
to remove any air in the headspace. Then, the reactor was pressurized to 2.5 MPa with H and
2
heated to the reaction temperature of 473 K while stirring at 500 rpm. Samples were taken
periodically (1, 2 and 3 h) through a sampling loop during the reaction. At the end of the
reaction (12 h), the reactor was allowed to cool down to room temperature and the liquid
sample was collected. All samples were filtered with a 0.2 µm pore size filter and analyzed
using a high performance liquid chromatograph (HPLC - Alliance) equipped with an Aminex
HPX-87H (Bio Rad) column and a refractive index detector. The HPLC column was
-1
2 4
maintained at 318 K, with a mobile phase of 5 mM H SO flowing at 0.5 mL min . The
liquid-phase products observed were 1,2-propanediol, ethylene glycol, 1,3-propanediol, lactic
acid, ethanol, 1-propanol, and 2-propanol.
The initial TOF and conversion were calculated using the following equations:
Amount of formed products (mol)
1
TOF (s ) =
Total surface metal atoms (mol)Reaction time (s)
Amount of formed products (mol)
Initial amount of glycerol (mol)
1
Liquid product based conversion (s ) =
100
Initial amount of glycerol (mol) - Final amount of glycerol (mol)
Initial amount of glycerol (mol)
1
Glycerol based conversion (s ) =
100
The initial TOF was based on the initial rate calculated at conversions lower than 20%. To
compare rates at higher levels of conversion, a site time yield (STY) based on hydrogen
chemisorption has also been used.
3
3
Results and discussion
.1 Catalyst characterization
6
Page 6 of 24

The molar Cu/Ru ratio measured by XPS and wide area EDS for the catalysts used in
this study are reported in Table 1. The nominal Cu/Ru atomic ratio of the bimetallic samples
was always between the atomic ratio determined by XPS and wide area EDS (Table 1),
suggesting that both methods provide self-consistent trends in metal composition. The 3d5/2
and 3d3/2 XPS peaks for the monometallic Ru catalyst were located at binding energies of
280.5 and 286.4 eV, respectively, whereas the 2p3/2 and 2p1/2 binding energies for the
monometallic Cu catalyst were measured at 933.1 and 953.1 eV, respectively. As revealed in
Table 1, the binding energies of both Ru and Cu in the bimetallic catalysts were similar to
those of the monometallic catalysts and revealed no clear trend with composition, suggesting
electron transfer between the two metals in the bimetallic catalysts was minimal.
Representative STEM images of the samples are provided in Figs. 1 and 2. Fig. 1
shows images for the monometallic Ru and Cu catalysts along with their corresponding
particle size distributions (PSDs) and Fig. 2 presents STEM images of the bimetallic catalysts
along with single particle EDS. Particle sizes from STEM as well as hydrogen uptake from
H chemisorption are shown in Table 2. The particle size for the monometallic Ru catalyst
2
determined by H chemisorption (4.2 nm) and the number average size from STEM (1.8 nm)
2
showed some difference, but good agreement was found when results from chemisorption
were compared with the surface average particle size (3.2 nm).
The hydrogen adsorption capacity per gram of catalyst (Hads/gcat) for each sample was
determined by extrapolation of the high pressure, linear portion of the adsorption isotherm to
zero pressure. The monometallic 2.5Ru/TiO
2 2
catalyst exhibited the highest H chemisorption
capacity followed by 5Ru-2.5Cu/TiO . As expected, the monometallic Cu sample had a
2
negligible hydrogen adsorption capacity, but when Cu was mixed with Ru there was a slight
decrease in H/Ru, likely the result of Cu deposition onto the Ru surface (Table 2).
Presumably, this decrease in adsorption capacity is evidence for the formation of bimetallic
ruthenium-copper particles with some portion of the Cu adsorbed on the surface of the Ru.
This conclusion is also corroborated by single particle EDS analysis of the bimetallic
catalysts which confirmed that small particles contain both Ru and Cu (Figs. 2a, 2b, 2c and
2d).
Sinfelt et al. [42] studied the Ru-Cu system and found that the addition of 5% Cu to
Ru was enough to decrease markedly the hydrogen chemisorption capacity by approximately
0% compared to monometallic Ru. Rouco et al. [40] found that as the nominal Cu/Ru molar
ratio of a Cu-Ru/SiO catalyst was increased from 0 to 0.667, with the Ru weight percent held
constant at 3%, the H/Ru ratio determined by H
6
2
2
chemisorption remained approximately
7
Page 7 of 24

constant. The authors suggested that either some of the hydrogen might be adsorbed on Cu,
which is facilitated by spillover from surface Ru atoms, or the particle size slightly decreased
as more Cu was incorporated with Ru. Interesting behavior was also found with a series of
Ru-Cu particles supported on high surface area graphite to which a low amount of Cu content
was added; a higher Cu content led to higher dispersion even compared to monometallic Ru
[46]. In the current work, we observed the dispersions (H/Ru) of the 2.5Ru-2.5Cu/TiO
2
,
2
.5Ru-5Cu/TiO , and 2.5Ru-10Cu/TiO were similar (Table 2), indicating that an increase in
2
2
the Cu content in the sample did not suppress the chemisorption capacity of the Ru beyond
that accomplished by adding a low level of Cu. We attribute these results to the formation of
separated small clusters of exclusively Cu on samples that have high loadings of Cu. Single
particle EDS and the number of particles with very light contrast suggest that samples with
higher Cu/Ru molar ratio (Fig. 2c and 2d) have more monometallic Cu particles than samples
with lower Cu/Ru molar ratio. Furthermore, XRD analysis (Fig. 3f and 3g) revealed that
increasing the Cu content of the catalysts resulted in the formation of a crystalline Cu phase.
The diffraction peaks at 2θ = 43.2°, 50.4° and 74.5° corresponding to (111), (200) and (220)
crystalline planes of copper, respectively, were observed for the fresh 2.5Ru-5Cu/TiO and
2
2
2.5Ru-10Cu/TiO catalyst (Fig. 3f and 3g) [47]. These peaks only appeared in catalysts that
had a copper content greater than 5%.
The XRD patterns of the bimetallic catalysts and of the blank support showed peaks
2
associated with anatase and rutile phases of the TiO crystalline domains (Fig. 3) [48]. The
XRD patterns of all the fresh and spent samples did not show any peaks ascribed to Ru,
presumably because of the high dispersion of the Ru particles. A high dispersion of Ru was
confirmed by STEM, which revealed number average particle sizes smaller than 3.1 nm for
all of the catalysts studied, as reported in Table 2.
3.2 Effect of reduction temperature
The reduction temperature can influence the chemical interaction between noble metal
and the support surface known as a strong metal-support interaction (SMSI). Tauster et
al.[49] found that at elevated reduction temperatures the amount of exposed Ru atoms
supported on TiO decreased, but not due to metal agglomeration. The reducibility of TiO2
2
oxide is favored at elevated temperature. The authors found that the high reduction
temperature of 773 K lead to the partial coverage of Ru metal particles by Ti
2 3
O species
decreasing the number of Ru active sites compared to a catalyst reduced at 473 K [49].
8
Page 8 of 24

In this work, 2.5Ru/TiO
different temperatures, 493 and 573 K, to determine if active surface sites were covered by
Ti species. The lower temperature (493 K) was chosen based on previous TPR results
2 2
and 2.5Ru-2.5Cu/TiO catalysts were reduced at two
2
O
3
which showed monometallic Ru can be reduced to Ru metal at around 490 K [11]. Copper
oxide also reduces to metallic Cu at temperatures around 450 K [46].
As shown in Table 3, changing the reduction temperature within this range did not
2
significantly alter either the H chemisorption capacity or the catalytic activity. Also, the
particle diameter measured by STEM was similar at both temperatures (493 and 573 K).
Nevertheless, these results contrast those reported by Feng et al.[31]. They observed a
decrease in the conversion of glycerol and an increase in particle size of Ru supported on
TiO with increasing catalyst reduction temperature from 473 to 573 K. The difference
2
between our results and those of Feng et al. might be attributed to the slightly broader range
in reduction temperature used by Feng et al. [31]. Additionally, Feng et al.[31] and Tauster
et al.[49] used RuCl as precursor, while the Ru precursor used in this study was
3
Ru(NO)(NO3)3.
3.3 Glycerol hydrogenolysis activity and selectivity
Table 4 summarizes the reaction results determined at low conversion for all of the
catalysts used in this work. The carbon balance was based on total carbon accounted for in
the liquid phase. No conversion was observed when a reaction was performed with blank
2
TiO . As shown in Table 4, the presence of a metal catalyst was required to catalyze
hydrogenolysis at this temperature. The addition of Cu metal to the Ru-based catalyst
markedly enhanced the selectivity towards 1,2-PDO, despite the slight decrease in activity
compared to monometallic Ru. The glycerol hydrogenolysis activities of the catalysts were
measured by calculating the TOF based on both H chemisorption uptake as well as the total
2
number of metal surface atoms. Surface metal atoms counted by H chemisorption are
2
assumed to be associated with Ru, since Table 2 indicates a negligible hydrogen uptake on
Cu. The number of total metal surface atoms (Ru+Cu) was estimated from the particle size
measured by STEM and the total metal loading on the sample. The rate of glycerol
hydrogenolysis, based on total metal surface atoms, decreased substantially as the Cu/Ru
ratio increased. This is attributed to the substantial fraction of low activity monometallic Cu
particles present with the bimetallic RuCu particles at higher Cu loadings. The monometallic
Cu catalyst was the least active catalyst, as expected. In contrast, the normalized rate of
9
Page 9 of 24

glycerol hydrogenolysis based on total H
ratio increased (Table 4), consistent with the conclusion that high levels of Cu did not
substantially block active Ru sites. The only exception in this trend was the 5Ru-2.5Cu/TiO
2
uptake did not change substantially as the Cu/Ru
2
catalyst. The results also show that when the Cu/Ru ratio increased, the selectivity to 1,2-
PDO increased from 48% in the case of monometallic Ru up to 97% for the 2.5Ru-
1
0Cu/TiO
the 2.5Ru-10Cu/TiO
monometallic Cu catalyst.
These results suggest that the role of Cu within these bimetallic particles is to dilute
2
catalyst, at the expense of ethylene glycol production. It must also be noted that
2
catalyst had approximately the same product selectivity as the
the large Ru ensembles that are responsible for C-C bond cleavage that lead to ethylene
glycol production. However, the addition of copper does not seem to substantially decrease
the ability of Ru to carry out the cleavage of C-O bonds, which leads to the production of 1,2-
PDO. As shown by Sinfelt et al.,[42] only small ensembles or perhaps single atoms of Ru are
needed to catalyze C-O cleavage. These results also suggest that Ru and Cu need to be
atomically mixed within the catalyst in order to increase the selectivity to 1,2-PDO. A low
selectivity to 1,2-PDO was observed when the monometallic Cu and Ru catalyst were
physically mixed in the batch reactor.
Figure 4 illustrates the conversion and 1,2-PDO selectivity after 12 h of reaction for
2
monometallic Ru and the bimetallic catalysts. The 2.5Ru-2.5Cu/TiO catalyst exhibited the
best performance in the hydrogenolysis of glycerol with 53.9 % conversion, 93.2% of 1,2-
PDO selectivity and with a carbon balance of 95% after 12 h. Furthermore, among the
2
bimetallic catalysts, 2.5Ru-2.5Cu/TiO exhibited the highest initial TOF as reported in Table
4.
3.4 Deactivation and catalyst recycle
The conversion based on liquid phase products, the conversion based on glycerol, and
the carbon balance for the monometallic Ru and bimetallic 2.5Ru-2.5Cu/TiO over the course
of the 12 h reaction is shown in Figure 5. After 1 h of reaction, the activity of the
2
-1
monometallic Ru catalyst was initially high (STY of 0.14 s , Table 5), but as the reaction
-1
proceeded, the STY decreased substantially to 0.019 s after 12 h. In an analogous
experiment with 2.5Ru-2.5Cu/TiO
2
, the initial STY of the bimetallic catalyst after 1 h was
-1
-1
0
.055 s and only slightly decreased 0.038 s after 12 h, indicating little deactivation with
time on stream. Part of the decrease in the rate of liquid phase product evolution for
10
Page 10 of 24

monometallic Ru arises from the production of gas phase species from the difference between
the conversion based on liquid phase products and the conversion based on glycerol. Gas
phase products were not measured in this work, but as shown in Table 4, the monometallic
Ru catalyst exhibited a high selectivity to ethylene glycol which is formed by C-C bond
cleavage. Subsequent hydrogenolysis reactions involving C-C bond cleavage to species with
high vapor pressures is likely and has been demonstrated by Maris and Davis [15].
Nevertheless, the reaction profiles in Figure 5 still suggest that substantial deactivation of the
monometallic Ru catalyst was occurring and could not be completely accounted for by
unmeasured gas phase products since the carbon balance was greater than 80% after 12 h of
reaction.
The reusability of monometallic Ru catalyst and 2.5Ru-2.5Cu/TiO was evaluated by
2
recycling the catalyst and re-testing it under the same reaction conditions. After a 12 h
reaction run, the spent catalyst was washed thoroughly with distilled, deionized water,
filtered, and then dried at 393 K overnight. The results for two cycles are reported in Table 5.
The mass of catalyst used in each reaction is listed in Table 5 since complete catalyst
recovery was not achieved. The observed deactivation of the monometallic Ru catalyst
during the reaction profile in Figure5 was likely caused by adsorption of some unidentified
carbon-containing species on the Ru sites because a simple washing step was sufficient to
regenerate most of the active catalytic sites as shown in the recycle experiments in Table 5.
The 2.5Ru-2.5Cu/TiO
2
catalyst also had a small decrease in the initial STY between
was
the two recycle runs (Table 5). Interestingly, the STY after 12 h with 2.5Ru-2.5Cu/TiO
2
the same as that after 1 h of reaction in cycle II. The bimetallic catalyst apparently does not
suffer from deactivation by weakly adsorbed side products in the same manner as the
monometallic Ru catalyst.
We do not attribute the deactivation during the reaction profile of the monometallic
Ru catalyst and 2.5Ru-2.5Cu/TiO , shown in Figure 5, to substantial particle growth because
2
the XRD pattern for the spent catalysts (Fig. 3c,e) did not show any Ru or Cu diffraction
peaks, and STEM of 2.5Ru/TiO revealed only a slight increase in the mean particle
2
diameter. The fresh monometallic Ru catalyst had a mean particle size of 1.8 ± 1.0 nm
whereas that of the catalyst recovered after the recycle experiments was 3.4 ± 1.8 nm. In fact,
the reduction of initial STY after recycling (Table 5) can be attributed to the increase in
particle size since the STY in Table 5 is based on the dispersion of the fresh catalyst
2
evaluated by H chemisorption.
11
Page 11 of 24

4. Conclusions
The addition of Cu metal to a Ru-based catalyst markedly enhanced the selectivity
towards 1,2-PDO. The hydrogenolysis rate on the Ru-Cu bimetallic catalysts, based on total
metal atoms, was at least an order of magnitude lower than that on a monometallic Ru
2
catalyst, whereas the rate based on H chemisorption was only slightly lower. The role of Cu
within these bimetallic particles most likely acts to dilute large Ru ensembles that are
responsible for C-C bond cleavage that leads to ethylene glycol production instead of the
desired 1,2-PDO. The presence of Cu also appears to inhibit the deactivation exhibited by
monometallic Ru as the reaction progressed. Among all the catalysts, the 2.5Ru-2.5Cu/TiO2
catalyst exhibited the best performance in the hydrogenolysis of glycerol.
Acknowledgment
J.B.Salazar is grateful to CAPES (Brazil) for the grant of a scholarship for a 9-month stay at
University of Virginia. The authors also thank Dmitry Pestov and Everett Carpenter at the
Virginia Commonwealth University Nanomaterials Core Characterization Facility for their
invalubale assistance with XPS. The material was based upon work supported by the US
National Science Foundation under award numbers CBET-1157829 and EEC-0813570.
12
Page 12 of 24

References
[1] J.J. Bozell, G.R. Petersen, Green Chem. 12 (2010) 539–554.
[2] T. Werpy, G. Petersen, Top Value Added Chemicals from Biomass, vol. 1, Results of
Screening for Potential Candidates from Sugars and Synthesis Gas; US DOE Report, 2004.
[3] S. Xia, R. Nie, X. Lu, L. Wang, P. Chen, Z. Hou, J. Catal. 296 (2012) 1–11.
[4] C.W. Chiu, M.A. Dasari, W.R. Sutterlin, G.J. Suppes, Ind. Eng. Chem. Res. 45 (2006)
791–795.
[5] H. Liu, S. Liang, T. Jiang, B. Han, Y. Zhou, CLEAN - Soil, Air, Water 40 (2012) 318–
324.
[6] E.P. Maris, W.C. Ketchie, V. Oleshko, R.J. Davis, J. Phys. Chem. B 110 (2006) 7869–
7876.
[7] E.S. Vasiliadou, E. Heracleous, I.A. Vasalos, A.A. Lemonidou, Appl. Catal. B-Environ.
92 (2009) 90–99.
[
8] D.M. Alonso, S.G. Wettstein, J.A. Dumesic, Chem. Soc. Rev. 41 (2012) 7965–8216.
9] A. Martin, U. Armbruster, I. Gandarias, P.L. Arias, Eur. J. Lipid Sci. Technol. 115 (2013)
[
9–27.
[10] C.H.C. Zhou, J.N. Beltramini, Y.X. Fan, G.Q.M. Lu, Chem. Soc. Rev 37 (2008) 527–
549.
[11] M. Balaraju, V. Rekha, B.L.A. Prabhavathi Devi, R.B.N. Prasad, P.S. Sai Prasad, N.
Lingaiah, Appl. Catal. A- Gen. 384 (2010) 107–114.
[12] M. Balaraju, V. Rekha, P.S. Sai Prasad, B.L.A. Prabhavathi Devi, R.B.N. Prasad, N.
Lingaiah, Appl. Catal. A- Gen. 354 (2009) 82–87.
[
[
[
[
13] D. Lahr, B. Shanks, J. Catal. 232 (2005) 386–394.
14] L. Ma, D. He, Z. Li, Catal. Commun. 9 (2008) 2489–2495.
15] E. Maris, R. Davis, J. Catal. 249 (2007) 328–337.
16] T. Miyazawa, S. Koso, K. Kunimori, K. Tomishige, Appl. Catal. A- Gen. 318 (2007)
244–251.
[17] T. Miyazawa, S. Koso, K. Kunimori, K. Tomishige, Appl. Catal. A- Gen. 329 (2007)
30–35.
[18] E. Ryneveld, A.S. Mahomed, P.S. Heerden, H.B. Friedrich, Catal. Lett. 141 (2011) 958–
967.
13
Page 13 of 24

[19] E.S. Vasiliadou, T.M. Eggenhuisen, P. Munnik, P.E. de Jongh, K.P. de Jong, A.A.
Lemonidou, Appl. Catal. B-Environ. (2013) 1–12.
[20] A. Bienholz, H. Hofmann, P. Claus, Appl. Catal. A- Gen. 391 (2011) 153–157.
[21] A. Bienholz, F. Schwab, P. Claus, Green Chem. 12 (2010) 290–295.
[22] S. Zhu, X. Gao, Y. Zhu, Y. Zhu, H. Zheng, Y. Li, J. Catal. 303 (2013) 70–79.
[
(
23] M. Balaraju, V. Rekha, P.S. Sai Prasad, R.B.N. Prasad, N. Lingaiah, Catal. Lett. 126
2008) 119–124.
[
[
[
[
[
[
24] S. Wang, H. Liu, Catal. Lett. 117 (2007) 62–67.
25] Z. Wu, Y. Mao, M. Song, X. Yin, M. Zhang, Catal. Commun. 32 (2013) 52–57.
26] D. Roy, B. Subramaniam, R. V. Chaudhari, Catal. Today 156 (2010) 31–37.
27] C. Montassier, D. Giraud, J. Barbier, Stud. Surf. Sci. Catal. 41 (1988) 165–170.
28] S. Wang, K. Yin, Y. Zhang, H. Liu, ACS Catal. 3 (2013) 2112−2121.
29] C. Montassier, J.C. Ménézo, L.C. Hoang, C. Renaud, J. Barbier, J. Mol. Catal. A-Chem.
70 (1991) 99–110.
[
30] E. Maris, W. Ketchie, M. Murayama, R. Davis, J. Catal. 251 (2007) 281–294.
31] J. Feng, H. Fu, J. Wang, R. Li, H. Chen, X. Li, Catal. Commun. 9 (2008) 1458–1464.
[
[
(
32] M.A. Dasari, P.P. Kiatsimkul, W.R. Sutterlin, G.J. Suppes, Appl. Catal. A- Gen. 281
2005) 225–231.
[33] L. Guo, J. Zhou, J. Mao, X. Guo, S. Zhang, Appl. Catal. A- Gen. 367 (2009) 93–98.
[34] A. Baiker, W. Richarz, Ind. Eng. Chem. Prod. Res. Dev. 16 (1977) 261–266.
[35] E.S. Vasiliadou, A.A. Lemonidou, Appl. Catal. A- Gen. 396 (2011) 177–185.
[36] A. Bienholz, R. Blume, A. Knop-Gericke, F. Girgsdies, M. Behrens, P. Claus, J. Phys.
Chem. C (2011) 999–1005.
[
[
[
[
[
37] D.E. Damiani, E.D.P. Millan, A.J. Rouco, J. Catal. 168 (1986) 162–168.
38] T. Jiang, Y. Zhou, S. Liang, H. Liu, B. Han, Green Chem. 11 (2009) 1000–1006.
39] Z. Wu, Y. Mao, X. Wang, M. Zhang, Green Chem. 13 (2011) 1311.
40] A.J. Rouco, G.L. Haller, J.A. Oliver, C. Kemball, J. Catal. 84 (1983) 297–307.
41] A.J. Hong, A.J. Rouco, D.E. Resasco, G.L. Haller, J. Phys. Chem. 91 (1987) 2665–2671.
14
Page 14 of 24

[
42] J.H. Sinfelt, Y.L. Lam, J.A. Cusumano, A.E. Barnett, J. |Catal. 42 (1976) 227–237.
43] C. Montassier, J.C. Ménézo, J. Moukolo, J. Naja, L.C. Hoang, J. Barbier, J. Mol. Catal.
[
70 (1991) 65–84.
[44] T. Jiang, Y. Zhou, S. Liang, H. Liu, B. Han, Green Chem. 11 (2009) 1000–1006.
[45] J.R. Anderson, Structure of Metallic Catalyst, Academic Press, New York, 1975.
[46] E. Asedegbega-Nieto, B. Bachiller-Baeza, A. Guerrero-Ruíz, I. Rodríguez-Ramos, Appl.
Catal. A- Gen. 300 (2006) 120–129.
[47] J.R. Shi, S.P. Lau, Z. Sun, X. Shi, B.K. Tay, H.S. Tan, Surf. Coat. Technol. 138 (2001)
250–255.
[48] H. Yin, Y. Wada, T. Kitamura, S. Kambe, S. Murasawa, H. Mori, T. Sakata, S.
Yanagida, J. Mater. Chem. 11 (2001) 1694–1703.
[49] S.J. Tauster, S.C. Fung, R.L. Garten, J. Am. Chem. Soc. (1978) 170–175.
15
Page 15 of 24

Tables and Figures
4
4
3
3
2
2
1
1
50
00
50
00
50
00
50
00
PSD for (a)
Mean: 1.8 ± 1.0 nm
Surf. Avg.: 3.2 nm
5
0
0
1
2
3
4
5
6
7
8
Particle diameter (nm)
4
3
3
2
2
1
1
00
PSD for (b)
50
00
50
00
50
00
Mean: 1.7 ± 0.6 nm
Surf. Avg.: 2.1 nm
50
0
1
2
3
4
5
6
7
8
Particle diameter (nm)
Fig. 1. HAADF-STEM images with corresponding particle size distributions (PSD) for: (a)
.5Ru/TiO , (b) 2.5Cu/TiO
2
2
2
.
16
Page 16 of 24

Fig. 2. HAADF-STEM images with corresponding single particle EDS results for: (a) 5Ru-
.5Cu/TiO , (b) 2.5Ru-2.5Cu/TiO , (c) 2.5Ru-5Cu/TiO , (d) 2.5Ru-10Cu/TiO
2
2
2
2
2
.
17
Page 17 of 24

*
# Cu
Rutile
Anatase
(g)
*
#
#
#
#
#
#
(f)
(
e)
d)
c)
(b)
a)
(
(
*
*
*
*
*
(
*
*
*
*
20
30
40
50
60
70
80
90
2
Fig. 3. X-ray diffraction patterns (a) TiO
2 2 2
, (b) 2.5Ru/TiO , (c) 2.5Ru/TiO (spent), (d) 2.5Ru-
2
.5Cu/TiO , (e) 2.5Ru-2.5Cu/TiO (spent) (f) 2.5Ru-5Cu/TiO , (g) 2.5Ru-10Cu/TiO . (All
2
2
2
2
diffractograms are associated with fresh catalysts unless otherwise noted).
18
Page 18 of 24

100
90
80
70
60
50
40
30
20
10
0
2
.5Ru 5Ru-2.5Cu 2.5Ru-2.5Cu 2.5Ru-5Cu 2.5Ru-10Cu 2.5Cu
Fig. 4. Comparison of conversion, based on liquid phase products, and carbon balance from
glycerol hydrogenolysis at 473 K over supported Ru, Cu, and Ru-Cu catalysts after 12 h of
reaction. Conditions: 20 wt% gly solution, T = 473 K; H Pressure = 2.5 MPa; 0.6 g of
2
catalyst.
19
Page 19 of 24

1
00
90
80
70
60
50
40
30
20
10
0
glycerol conv. Ru-Cu
liq. production conv. Ru-Cu
carbon balance Ru-Cu
glycerol conv. Ru
liq. production conv. Ru
carbon balance Ru
0
2
4
6
8
10
12
Reaction time (h)
Fig. 5. Reaction profiles for glycerol hydrogenolysis at 473 K over supported 2.5Ru/TiO
.5Ru-2.5Cu/TiO catalysts.
2
and
2
2
20
Page 20 of 24

Table 1
Characterization results from wide area EDS and XPS of catalysts reduced at 573 K
a
Catalyst
Cu/Ru
mol/mol)
Nominal
Cu/Ru
(mol/mol)
EDS
Cu/Ru
(mol/mol)
XPS
XPS Binding Energies (eV)
(
Ru 3d5/2 Ru 3d3/2 Cu 2p3/2 Cu 2p1/2
2
5
2
2
2
2
.5Ru/TiO2
-
-
-
280.5
281.0
280.9
281.1
281.0
-
286.4
286.1
286.4
286.2
286.4
-
-
-
Ru-2.5Cu/TiO
2
0.8
1.6
3.2
6.4
-
1.0
2.0
4.3
6.4
-
0.8
1.4
1.7
3.4
-
933.2
933.2
933.5
933.9
933.2
953.2
952.9
953.5
953.7
953.1
.5Ru-2.5Cu/TiO
.5Ru-5Cu/TiO2
2
.5Ru-10Cu/TiO2
.5Cu/TiO
2
Table 2
Results from H
microscopy
2
chemisorption and scanning transmission electron
a
b
Catalyst
H Uptake
µmol/gcat )
Ru dispersion STEM particle Surface average
c
(
(H/Rusurf
)
diameter (nm)
diameter (nm)
2
5
2
2
2
.5Ru/TiO
Ru-2.5Cu/TiO
.5Ru-2.5Cu/TiO
.5Ru-5Cu/TiO
2
69
118
43
47
27
-
0.31
0.28
0.21
0.25
0.21
-
1.8
3.1
2.1
2.4
2.7
1.7
3.2
4.3
3.2
3.5
3.7
2.1
2
2
2
.5Ru-10Cu/TiO
2
2
.5Cu/TiO2
a
All catalysts were reduced at 573 K
b
c
Determined by H
2
chemisorption
3
2
Determined by STEM (Σd /Σd )
21
Page 21 of 24

Table 3
Effect of reduction temperature on Ru dispersion and TOF for monometallic Ru and
2
.5Ru-2.5Cu/TiO
2
a
b
Catalyst
Ru dispersion
Conversion
(%)
Chemi.
TOF
Product selectivity (%)
(H/Rusurf)
-1 c
d
(s )
1,2-PDO
EG
40
41
23
25
1,3-PDO Others
e
f
2
2
2
.5Ru/TiO
.5Ru/TiO2
2
0.34
0.31
0.21
0.21
18.7
19.1
9.6
0.13
0.14
0.11
0.11
51
48
68
69
2
3
0
0
7
9
9
6
e
f
.5Ru-2.5Cu/TiO2
.5Ru-2.5Cu/TiO
2
*
2
9.7
Reaction conditions: 20 wt% gly solution, T = 473K; H Pressure = 2.5 MPa; 0.6 g of catalyst; reaction time: 1h.
2
All carbon balances were between 95 and 103%
a
2
Determined by H chemisorption
Based on total liquid phase products. Conversions were the average of 3 experiments, with an error < 2%.
Based on total liquid phase products normalized to the H chemisorption capacity
Others: 1-propanol, 2-propanol, ethanol, methanol, lactic acid
Reduced at 493K
Reduced at 573K
b
c
d
e
2
f
Table 4
Catalyst performance in glycerol hydrogenolysis at low conversion
Catalyst
Time
h)
Conversion
Chemi.
TOF
Surf. Met.
TOF
Product selectivity (%)
a
(
(%)
-
1 b
-1 c
d
(s )
(s )
1,2-PDO
EG
1,3-PDO Others
Blank
TiO2
12
12
1
1
1
1
2
12
1
0.00
< 1.0
19
11
10
8.6
8.4
13
-
-
-
-
-
-
-
-
-
-
-
-
2
5
2
2
2
2
.5Ru/TiO2
Ru-2.5Cu/TiO
.5Ru-2.5Cu/TiO
.5Ru-5Cu/TiO
.5Ru-10Cu/TiO2
.5Cu/TiO2
0.14
0.048
0.11
0.091
0.081
-
0.13
48
64
69
74
96
97
58
41
24
25
21
3
3
4
0
0
0
0
5
9
8
6
6
2
1
7
2
0.034
0.027
0.015
0.0050
0.0027
0.056
2
2
2
30
2
.5Ru/TiO + 2.5Cu/TiO2
15
0.055
2
*
Reaction conditions: 20 wt% gly solution, T = 473 K; H Pressure = 2.5 MPa; 0.6 g of catalyst. All catalyst were reduced
2
previously at 573K. All carbon balances were between 95 and 103%
a
Based on total liquid phase products. Conversions were the average of 3 experiments which had an error < 2%
b
Based on total liquid phase products normalized to the H
2
chemisorption capacity
c
Based on total liquid phase products normalized to surface metal atoms (Ru+Cu) evaluated from the inverse of the surface
average particle size obtained by STEM
d
Others: 1-propanol, 2-propanol, ethanol, methanol, lactic acid
22
Page 22 of 24

Table 5
2 2
Recycle experiments with 2.5Ru/TiO and 2.5Ru-Cu/TiO catalysts for glycerol hydrogenolysis
Catalyst
Cycle
Mass cat.
g)
Time Conversion Site-Time
Product selectivity (%)
a
(
(h)
(%)
Yield
-
1 b
c
(s )
1,2-PDO
EG
1,3-PDO Others
Cycle I
0.608
1
19
31
14
0.14
0.019
0.11
48
57
50
41
15
40
3
14
2
9
14
8
2
.5Ru/TiO
2
12
Cycle II
0.552
0.603
1
d
2
.5Ru-2.5Cu/TiO2 Cycle I
1
4.7
39
0.055
0.038
0.038
84
90
85
14
6
1
1
2
2
3
2
12
Cycle II
0.519
1
2.8
12
*
Reaction conditions: 20 wt% gly solution, T: 473 K; H
2
Pressure = 2.5 MPa. All carbon balances were between 95 and
1
05%. The spent catalyst was recovered by filtration with a 0.45µm pore membrane, washed with distilled, deionized water
and dried at 393 K before recycle
a
Based on total liquid phase products
Based on total liquid phase products normalized to the H chemisorption capacity of the fresh catalyst
2
Others: 1-propanol, 2-propanol, ethanol, ethanol, lactic acid
b
c
d
These experiments were performed in a 300 mL volume Parr reactor whereas all other experiments were performed in a 100
mL Parr reactor
23
Page 23 of 24

1
00
Glycerol
1,2-propanediol
9
8
7
6
5
4
3
2
1
0
0
0
0
0
0
0
0
0
0
Ru/TiO2
RuCu/TiO2
Page 24 of 24