Tandem Site- and Size-Controlled Pd Nanoparticles for the Directed Hydrogenation of Furfural

Article abstract of DOI:10.1021/acscatal.6b03190

The conversion of biomass to useful chemical products requires precise catalytic properties to achieve the required activity, selectivity, and durability. Here we show, through optimized colloidal synthesis, the tandem control of Pd size and site availabi

Full text of DOI:10.1021/acscatal.6b03190

Subscriber access provided by University of Newcastle, Australia
Article
Tandem Site and Size Controlled Pd Nanoparticles
for the Directed Hydrogenation of Furfural
Scott M. Rogers, C. Richard A. Catlow, Carine E. Chan-Thaw, Arunabhiram Chutia, Nan Jian, Richard
E. Palmer, Michal Perdjon, Adam Thetford, Nikolaos Dimitratos, Alberto Villa, and Peter P. Wells
ACS Catal., Just Accepted Manuscript • DOI: 10.1021/acscatal.6b03190 • Publication Date (Web): 17 Jan 2017
Downloaded from http://pubs.acs.org on January 17, 2017
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
ACS Catalysis is published by the American Chemical Society. 1155 Sixteenth Street
N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 10
ACS Catalysis
1
2
3
4
5
6
7
8
9
Tandem Site and Size Controlled Pd Nanoparticles for the Di-
rected Hydrogenation of Furfural
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Scott M. Rogers,1,2 C. Richard. A. Catlow,1,2,5 Carine E. Chan-Thaw,3 Arunabhiram Chutia,1,2 Nan
Jian,4 Richard E. Palmer,4 Michal Perdjon,1,2,5 Adam Thetford,1,2 Nikolaos Dimitratos,1,5 Alberto
Villa,*1,3 and Peter P. Wells*1,2,6,7
1. UK Catalysis Hub, Research Complex at Harwell, Rutherford Appleton Laboratory, Harwell Oxon, Didcot, OX11
0FA, UK.
2. Department of Chemistry, University College London, 20 Gordon Street, London, WC1H 0AJ, UK.
3. Dipartimento di Chimica, Università degli Studi di Milano, via Golgi 19, 20133 Milano, Italy.
4. Nanoscale Physics Research Laboratory, School of Physics and Astronomy, University of Birmingham, Edgbaston,
Birmingham, B15 2TT, UK.
5. Cardiff Catalysis Institute, School of Chemistry, Cardiff University, Cardiff, CF10 3AT, UK.
6. Diamond Light Source, Harwell Science and Innovation Campus, Chilton, Didcot, OX11 0DE, UK.
7. School of Chemistry, University of Southampton, Highfield, Southampton, SO17 1BJ, UK.
ABSTRACT: The conversion of biomass to useful chemical products requires precise catalytic properties to achieve the
required activity, selectivity and durability. Here we show, through optimized colloidal-synthesis, the tandem control of Pd
size and site availability for the directed hydrogenation of the bio-derived intermediate, furfural. Adjusting the temperature
of colloidal reduction dictates the size of Pd nanoparticles; in some instances ultra-small clusters <20 atoms are achieved.
Whereas, changing the solvent-system, affects the PVA-Pd interaction and relative proportion of available surface-sites
(corners, edges, planes), allowing us to control the selectivity to the valuable hydrogenation products of furfuryl alcohol
and tetrahydrofurfuryl alcohol. We demonstrate, through combined experimental and computational studies, that Pd na-
noparticle planes are more prone to deactivation through the formation of Pd carbide, and the resulting reduced efficacy
of furfural binding. This approach to nanoparticle optimization is an important strategy for producing long-lasting,
high-performance catalysts for emerging sustainable technologies.
Keywords: Pd nanoparticles, Colloids, Clusters, Furfural hydrogenation, Heterogeneous catalysis, Pd carbide
Introduction
Many heterogeneous catalysts have been investigated for
the hydrogenation of furfural, with the majority of studies
focusing on Cu, Ni,7 Ru,8 or Pd9-10 based systems. Cu, in the
form of Cu chromite,11 or supported forms,12-15 have been
commonly used, and show good selectivity to furfuryl al-
cohol as a consequence of the preference of Cu to bind
C=O, over C=C.15 Ni catalysts show different product distri-
bution to Cu, with the primary step favoring decarbonyla-
tion, with secondary paths including opening of the furan
ring.16 Pd is an ideal catalyst for hydrogenation processes
as it readily dissociates hydrogen under ambient condi-
tions. Recent studies have shown that Pd/TiO2 based sys-
tems are able to effectively hydrogenate furfural and furfu-
ryl alcohol at mild conditions (room temperature, 1-3 bar
hydrogen).17-19 The studies focused on Pd nanoparticles
prepared by incipient wetness impregnation and showed
that methyl furan and furfuryl alcohol were the major
products and that the product distribution could be altered
by changing the solvent of reaction and metal loading.17 In-
deed, the differences in furfural binding (η1 CO, η1 O, η2 C-
Upgrading biomass to useful products, be it energy or plat-
form intermediates, is a crucially important part of a sus-
tainable chemicals industry. The hydrolysis of hemicellu-
lose is one example, where one of the major products, xy-
lose, is acid catalyzed to the valuable intermediate furfu-
ral.1-2 Furfural is readily valorized, and is an important pre-
cursor in the generation of biofuels3-4 and chemical inter-
mediates.2 Furfuryl alcohol, used in the manufacture of
resins, adhesives, and synthetic fibres,5 is produced from
the selective hydrogenation of furfural. Subsequent hydro-
genation of furfuryl alcohol produces tetrahydrofurfuryl al-
cohol, a ‘green solvent’, often used in printer inks and ag-
ricultural applications.6 Ultimately, the hydrogenation of
furfural results in a complex network of products (Figure
1), where there is a need to control the relative distribution.
The challenge is to find a selective catalyst that has control
over C=C or C=O hydrogenation and disfavors alternative
pathways, e.g. decarbonylation.
ACS Paragon Plus Environment

ACS Catalysis
Page 2 of 10
O, and η2 C-C) and their preference on specific crystal fac-
NPs on to the support. The slurry was filtered, washed
thoroughly with distilled water, and dried overnight at
room temperature.
ets,20-21 have been linked to variations in product selectivity
for different particle sizes.22 As well as particle size effects,
the ability to manipulate the binding orientation of furfu-
ral on to specific metal sites, to selectively control the hy-
drogenation pathway, has been effectively achieved using
self-assembled monolayers (SAMs) as blocking agents.23-25
Medlin et al. used thiolates to selectively block facets, leav-
ing only particles edges/corners exposed.
1
2
3
4
5
6
7
8
Transmission Electron Microscopy (TEM) and High
Angle Annular Dark Field Scanning TEM (HAADF
STEM) Imaging. Samples for examination by TEM and ab-
erration-corrected HAADF STEM were prepared by first
dispersing the catalyst powder in high purity ethanol using
ultra-sonication for 30 minutes. 40 µL of the suspension
was dropped on to a holey carbon film supported by a 300
mesh copper TEM grid before the solvent was evaporated.
A JEOL JEM 2100 EM was used for HAADF STEM analysis
at the Nanoscale Physics, Chemistry and Engineering Re-
search Laboratory at the University of Birmingham by us-
ing a JEOL JEM2100F STEM equipped with a spherical ab-
erration corrector (CEOS). The HAADF detector was oper-
ated with an inner angle of 62 mrad and an outer angle of
164 mrad. The integrated HAADF STEM intensity analysis
was used to obtain the size of ultrasmall clusters, with large
Pd clusters functioning as mass balances.27-28 The 3D inten-
sity plot of the small Pd clusters was performed using the
imagej software.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Infra-red CO chemisorption studies. Fourier transform
infra-red (FTIR) transmission spectra were obtained with
is10 Nicolet spectrometer at a spectral resolution of 2 cm-1
and accumulating up to 64 scans. For each experiment, ~25
mg catalyst was pressed to form a very thin pellet. The cell
was purged with helium for 30 minutes to obtain a back-
ground spectrum before CO was introduced using a 10 %
CO/He mixture at 70 mL min-1 over a 30 second period.
Three CO doses of this nature were administered for each
experiment. The gas was switched to helium for 30 minutes
at 70 mL min-1, in order to remove gaseous and phy-
sisorbed CO before obtaining a spectrum.
Figure 1. Schematic representation of reaction pathways
during furfural hydrogenation.
In this study, we report for the first time, how Pd/TiO2 cat-
alysts prepared through controlled sol-immobilization
routes,26 affords series of nanoparticles of tailored particle
size distributions, including populations of metal clusters.
We show, systematically, how particle size and the solvent
system of preparation influences the activity, selectivity
and stability of the catalysts during furfural hydrogenation
at mild reaction conditions of 25 and 50°C. Furthermore,
we demonstrate changes to the nanoparticle structure,
through carbidization, which causes catalyst deactivation.
To our knowledge, this is the first time this deactivation
route has been reported.
X-ray Absorption Fine Structure (XAFS). XAFS studies
were performed to examine the Pd oxidation state
(XANES) as well as the average Pd particle size from the
primary shell coordination number (EXAFS). Pd K-edge
XAFS studies were carried out on the B18 beamline at the
Diamond Light Source, Didcot, U.K. Measurements were
performed in transmission mode using a QEXAFS set-up
with a fast-scanning Si(311) double crystal monochromator
and ion chamber detectors. The time resolution of the
spectra was 1 min/spectrum (kmax=18). On average, 15 scans
were acquired to improve the signal-to-noise level of the
data. XAS data processing was performed using the Deme-
ter IFEFFIT package. 29-30
Methods.
Catalyst Preparation. Supported Pd NPs were prepared
using a standard sol-immobilisation method with temper-
ature control during the reduction process. K2PdCl4 was
used to prepare solutions with various H2O/EtOH ratios of
the desired palladium concentration (1.26 x 10-4 M), to
which aqueous solutions of PVA (PVA/Pd (wt/wt) = 0.65)
were added. 0.1 M solutions of NaBH4 (NaBH4/Pd
(mol/mol) = 5) were freshly prepared in the respective sol-
vent and added drop-wise to each solution over a one mi-
nute period with stirring to form dark brown/black sols.
After the complete reduction of Pd species, 30 minutes, the
colloidal solution was individually immobilized on TiO2
(commercial P25) under vigorous stirring conditions. The
amount of support material required was calculated so as
to give a final metal loading of 1 wt. %. The mixture was
acidified to pH 1-2 by sulfuric acid before stirred for 60
minutes to accomplish full immobilization of the metal
Catalytic studies. Furfural hydrogenation was performed
at 25 or 50°C, using a stainless steel reactor (30 mL capac-
ity), equipped with heater, mechanical stirrer, gas supply
system and thermometer. Furfural solution (15 mL; 0.3 M
in 2-propanol) was added into the reactor and the desired
amount of catalyst (Furfural/metal ratio=500 mol/mol)
was suspended in the solution. The pressure of the hydro-
gen was 5 bar. The mixture was let at room temperature
(25°) or in alternative heated 50°C and mechanically stirred
(1250 rpm). At the end of the reaction, the autoclave was
cooled down to room temperature (when performed at
ACS Paragon Plus Environment

Page 3 of 10
ACS Catalysis
50°C), H2 flow stopped and the autoclave purged with flow- Results and Discussion
1
2
3
4
5
6
7
8
ing nitrogen. Samples were removed periodically (0.2 mL)
and HP 7820A gas chromatograph equipped with a capil-
lary column HP-5 30m x 0.32mm, 0.25 µm Film, by Agilent
Technologies. Authentic samples were analyzed to deter-
mine separation times. Quantitative analyses with external
standard method (n-octanol) was used.
Characterisation of As-Prepared Catalysts. The experi-
mental conditions at which each 1 wt. % Pd/TiO2 catalyst
was prepared, as well as notation, is presented in table 1,
with elemental analysis, by microwave plasma atomic ab-
sorption spectroscopy (MP-AES), shown in table S1. The
prepared Pd/TiO2 catalysts were characterized using TEM
to assess the particle size distribution (table 1), all images
and histograms detailed in the supporting information
(figure S1).
DFT studies. The Vienna Ab-initio Simulation Package
(VASP) was used to perform DFT based calculations with
Grimme’s D2 corrections.31-35 The projector augmented
wave (PAW) method was used and the cut-off energy for
the expansion of the plane-wave basis set was set to 550eV,
which gave bulk energies converged to within 10E− 5 eV. A
convergence criterion of 0.01 eV/Å for our structural opti-
mizations was adopted. For all the preliminary calcula-
tions, the most commonly used Perdew-Burke-Ernzerhof
(PBE) version of the generalized gradient approximation
(GGA) was used to carry out total energy calculations and
perform geometry optimizations.36 For the bulk calcula-
tions, the Brillouin zone was integrated using a Monk-
horst-Pack (MP) grid of 11×11×11 k-points. Since it is known
that the Pd(111) surface is the most stable surface among
the low index surfaces, all our calculations were performed
on Pd(111) surface.37 The ideal Pd(111) surfaces were mod-
elled by a 3 × 3 supercell with 5 atomic layers. A lattice con-
stant of 3.904 Å and a k-point grid of 3×3×1 was used. Dur-
ing the optimization process, we relaxed the upper two
atomic layers along with the furfural molecule. Bottom
three atomic layers were fixed to mimic the bulk of the sys-
tem. The adsorption energy was calculated using the equa-
tion:
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Within each solvent environment (A series:H2O or B se-
ries:H2O/EtOH), decreasing the temperature of prepara-
tion, decreases the average Pd particle diameter, in agree-
ment with our previous work with Au based materials.26
PdA1 (2.5 nm) and PdB1 (1.4 nm) produced the smallest Pd
nanoparticles within their respective series, with PdB1 ex-
hibiting the narrowest particle size distribution. It is un-
derstood this average particle size is smaller than any pre-
viously reported TiO2 supported Pd metal nanoparticles,
synthesized using a PVA/NaBH4 colloidal system.38 The ef-
fect of solvent system can be assessed by comparing sam-
ples PdA1 and PdB2, and PdA2 and PdB3, which denote
samples prepared at 1°C and 25°C, respectively. Interest-
ingly, the B series of Pd/TiO2 produces a larger spread of
Pd particle size as a function of preparation temperature,
with catalyst PdB2 being evidently smaller than the A se-
ries equivalent (PdA1). However, the distribution of sizes
for the A series is much narrower, and the catalyst pre-
pared at 25°C (PdA2), has a smaller average particle diam-
eter compared to the B series analogue (PdB3). We ration-
alize these differences by the competing influences during
sol-immobilization; the rate of metal precursor reduction,
solubility of PVA, and interaction of solvent with the metal
salt and formed colloid. PdA1 and PdB1 were examined
with HAADF STEM to investigate the presence of Pd clus-
ters that are too small to image clearly with standard TEM
(figure 2 and S2). The presence of ultra-small Pd clusters is
clearly evident for catalyst PdA1, and are identified by the
HAADF STEM images. The ability to achieve significant
populations of
Ead = EPd(111)+Furfural – (EPd(111) + Efurfural
)
Where, Ead is the adsorption energy, EPd(111)+furfural is the en-
ergy of the system with furfural molecule adsorbed and Efur-
fural is the energy of the furfural molecule.
Table 1. Solvent and temperature conditions applied for each Pd/TiO2 catalyst preparation, average Pd particle diameter
calculated by TEM analysis and XANES linear combination analysis data for the 1 wt. % Pd/TiO2 catalysts.
Tempera-
ture / °C
H2O/EtOH solvent Sample notation
ratio (vol:vol)
Reference standards (%)
Pd2+
TEM Average Pd
diameter / nm
Pd0
67
71
Rfactor
0.043
0.037
0.036
0.025
0.056
0.045
0.043
1
100
100
100
100
50
PdA1
PdA2
PdA3
PdA4
PdB1
PdB2
PdB3
2.5 ± 0.8
2.7 ± 0.9
2.9 ± 1.5*
5.2 ± 2.1*
1.4 ± 0.4
2.1 ± 0.6
3.4 ± 0.9
33
25
50
75
−30
1
29
20
12
80
88
29
64
67
71
50
36
25
25
50
* As a consequence of dense agglomeration only 100 particles were counted.
ACS Paragon Plus Environment

ACS Catalysis
Page 4 of 10
metal clusters through standard chemical means remains
is assigned to CO linearly adsorbed on corner sites of Pd
a challenge, and in accordance with our previous study,26
we demonstrate how adjusting the temperature of colloi-
dal reduction is able to influence this.
nanoparticles.40 There is a secondary linear adsorbed CO
band at 2063 cm-1, which can be ascribed to adsorption on
Pd nanoparticle edges. Bridge bonded CO can also be iden-
tified at ~1975 and 1945 cm-1, with the differences in posi-
tion attributed to adsorption of facets and edges, respec-
tively.40 Moreover, there are adsorption bands at 2140,
2120, and 1875 cm-1, assigned to CO adsorbed on Pd2+, Pd+
and threefold sites, respectively.41-42 The spectra obtained
for all four catalysts exhibit noticeable differences in the
distribution of available sites. The relative intensities of the
linearly adsorbed CO bands, compared to that from
bridge-bonded CO, is much higher for the catalysts pre-
pared in a mixed ethanol-water solvent system. Further-
more, there are differences in the ratio between different
types of linear and bridge-bonded sites. The series of NPs
prepared in the ethanol water-solvent have a greater pro-
portion of the 2086 cm-1 linear adsorbed CO, and 1975 cm-1
bridge-bonded CO. It is clear that changing the solvent sys-
tem of preparation affords Pd NPs with different surface
characteristics, most notably the increase in available Pd
corner and edge sites. We propose that the different sur-
face sites observed are a consequence of the interaction be-
tween the solvent and PVA, which affects the extent of PVA
binding on the metal surface.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 2. Processed HAADF STEM image of PdA1 showing
Pd11 and Pd17 clusters.
X-ray Absorption Near Edge Structure (XANES) is a valua-
ble tool in probing the speciation of metal nanoparticles,
in this instance for determining the Pd oxidation state in
the Pd/TiO2 catalysts under atmospheric conditions (dur-
ing hydrogenation conditions the oxidic content will re-
duce to Pd metal). The ratio between Pd2+ and Pd0 was per-
formed by linear combination analysis (LCA) of the 1st de-
rivative of the XANES profile, using PdO and Pd foil as ref-
erence standards (table 1 and figure S3). It is evident that
for each solvent system, lowering the temperature at which
colloidal Pd is prepared results in an increase in Pd2+. All
X-ray absorption fine structure (XAFS) measurements rep-
resent the whole sample, which means that for NP systems
the surface speciation can only influence the overall re-
sponse for very small NP size; the ratio of surface:core spe-
cies increases as the average NP diameter decreases. Small
Pd NPs form an oxidic surface layer at room temperature
when exposed to air,39 and indeed, we observe a clear cor-
relation between average Pd particle diameter (by TEM)
and the extent of Pd2+. The k2-weighted forward Fourier
transform of PdA1 (figure S3) confirms there are no large
PdO crystallites. The Pd-Pd scattering path that would
otherwise be present at 3Å, and can be observed in the PdO
reference, is not observed in the PdA1 data. With this
knowledge, it is understandable that the contribution in
the PdB1 catalyst is 71% Pd2+ and only 29% Pd0, due to the
very small average Pd size in this catalyst. This infor-
mation, consistent with TEM analysis, confirms that pre-
paring colloidal Pd NPs at a lower temperature results in a
decrease in particle size and the formation of clusters. Pd
colloidal preparation at −30°C in a mixed EtOH/H2O sol-
vent shows the most significant interest as the stabilization
of ultra-small metal nanoparticles is challenging, but
achievable.
Figure 3. FTIR spectra from CO-adsorption studies on
PdB1 (solid black line) and PdA1 (red dashed line).
Catalytic Testing. Both series of catalysts were evaluated
for the hydrogenation of furfural (Furfural = 0.3 M; F/metal
ratio=500 mol/mol, 5 bar H2, solvent 2-propanol) at 25°C
(table S2) and 50°C (table S3), with the data illustrated
graphically in figure 4. Comparing the data between the A
series (prepared in water) and B series (prepared in water-
ethanol), interesting relationships can be observed. For the
A series there is a clear correlation between Pd particle size
and activity; smaller Pd NPs are more active, regardless of
the reaction temperature. Indeed, the catalyst of the A se-
ries of smallest Pd particle size (A1, 2.5 nm), showed the
highest catalytic activity (440 and 620 mol (mol Pd)-1 h-1, at
25 and 50°C, respectively).
To elucidate the specific adsorption sites generated using
the different preparation routes, catalysts PdA1, PdA2,
PdB1 and PdB2 were evaluated using CO as a probe mole-
cule (figures 3 and S4). Here, the IR frequency of the CO
band was used to assess differences in the surface structure
of the Pd NPs. The adsorption band present at 2086 cm-1,
ACS Paragon Plus Environment

Page 5 of 10
ACS Catalysis
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 4. Catalytic performance plots for hydrogenation of furfural using tailored Pd/TiO2 catalysts (top left) Pd NPs
A1→A4 tested at 25°C, (top right) Pd NPs A1→A4 tested at50°C, (bottom left) Pd NPs B1→B3 tested at 25°C, and (bottom
right) Pd NPs B1→B3 tested at 50°C. Reaction conditions: Furfural = 0.3 M; F/metal ratio=500 mol/mol, 5 bar H2, solvent 2-
propranol, Converted mol (mol Pd)-1 h-1 calculated after 15 min of reaction, selectivity calculated at 50% conversion, except
catalysts PdA3 and PdA4 at 25°C, where selectivity is calculated at 10% conversion. Open black circles denote conversion.
Open blue diamonds denote activity. Red shapes indicate selectivity to the products denoted by the inset key.
For the A series other distinct trends can be seen; as the Pd
particle size increases (PdA1 → PdA4), selectivity to furfu-
ryl alcohol increases (48, 65, 76 and 78% for PdA1, PdA2,
PdA3 and PdA4, respectively, Table S3) with a decrease in
selectivity to tetrahydrofurfuryl alcohol (39, 25, 20 and 15%
for PdA1, PdA2, PdA3 and PdA4, respectively, Table S3). On
some occasions, the A series catalysts showed improved ac-
tivity, compared to the B series at comparable particle
sizes. Catalyst PdA1, with mean particle size of 2.5 nm has
a higher activity at 50 °C than PdB2 (620 and 470 converted
mol (mol Pd)-1 h-1, respectively, Table S3), despite the
smaller particle size of PdB2 (2.1 nm). It must be noted that
these Pd mass normalized activities are high when com-
pared against current Pd based catalysts for this reaction.10,
22 There are clearly other parameters, besides particle size,
that influence activity and selectivity, as can be observed
by the catalytic performance of the B series. In the case of
the B series tested at 25°C, the catalyst, B1 (1.4 nm) was less
active than B2 (2.1 nm) with activities of 491 and 568 mol
(mol Pd)-1 h-1, respectively. The selectivity of B1 is compara-
ble to that of the only previously reported Pd/TiO2 catalyst
(Pd particle also < 2 nm) used for this reaction, despite the
different reaction solvent.17 However, when tested at 50 °C,
there was a direct relationship between particle size and
activity. Another characteristic difference of the B series is
the variation in selectivity profile with increasing particle
size. Broadly speaking, there is an inverse relationship for
furfuryl alcohol and tetrahydrofurfuryl alcohol selectivity
with particle size compared to the A series; as the particle
size increases furfuryl alcohol selectivity decreases and tet-
rahydrofurfuryl alcohol selectivity increases. It is apparent
that there is a collaborative effect between the solvent sys-
tem and temperature of colloidal preparation, which di-
rects the catalytic performance. It is our assertion that the
ACS Paragon Plus Environment

ACS Catalysis
Page 6 of 10
competing influences of sol immobilization - rate of reduc- further particle growth when performed at the higher tem-
1
2
3
4
5
6
7
8
tion, PVA solubility, PVA/solvent/colloid interaction – are
responsible for the trends observed. The difference in se-
lectivity can be ascribed to the different binding modes of
furfural on the Pd surfaces. The CO chemisorption studies
for the A and B series show a clear difference in linear CO:
bridged CO ratio, with the B series having a far greater pro-
portion of linear sites. Where CO binds linearly to Pd edge
and corner sites, previous studies have
perature. However, for the Pd catalysts prepared in
H2O:EtOH mixture, the temperature of reaction has a
larger influence for PdB1, in which the average particle size
increases from 1.4 to 2.0 nm and 2.7 nm when the reaction
was performed at 25 and 50°C, respectively and these re-
sults are in agreement with the observed initial decrease in
catalytic activity. The average Pd particle diameter is still
very small and STEM/HAADF was used to identify the
presence of Pd clusters in the PdA1 and PdB1 catalysts after
performing the reaction at 25°C (figure S8). It is clear that
very small Pd clusters are still preserved during catalysis, a
combined result of the mild reaction conditions and pro-
tective stabilization role of PVA.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 5. Catalytic activity for Recycling studies of cata-
lysts PdA1 and PdB1.
indicated that these sites tend to bind furfural in a perpen-
dicular orientation, rather than flat across a surface, in
agreement with the recent work of Medlin et al.24 In these
instances we observe improved selectivity to furfuryl alco-
hol. Moreover, the adsorption sites that correspond to
bridge-bonded or three-fold CO adsorption are able to
bind furfural through both the aldehyde functionality and
the furan ring and promote the complete reduction to tet-
rahydrofurfuryl alcohol. Catalysts used for sustainable
technologies need to be robust and durable, so recycling
studies were performed. The recycling tests were per-
formed on A1 and B1 catalysts at different temperatures
(Tables S4 and S5, and figure 5), by reusing the same cata-
lyst without any pretreatment. These data evidenced a sig-
nificant deactivation of both catalysts when the reaction
was performed at 25°C, whereas at 50°C the deactivation
was less pronounced. Assessing the reaction profile (Fig-
ures S5 and S6) as a function of time, evidences that
stronger deactivation phenomena occur for all the cata-
lysts when the reaction was performed at 25°C. To explain
the origin of the deactivation the used catalysts were also
characterized.
Figure 6. Normalized XANES spectra of PdA1 and PdB1 af-
ter catalysis and a Pd0 reference.
The used PdA1 and PdB1 catalysts tested at 25°C (after 5 re-
action cycles) were assessed using XAFS spectroscopy, with
the XANES data shown in figure 6. The initial observation
from the XANES is the presence of Pd2+ and Pd0. The height
of the main edge of the fresh and used catalysts are greater
than that of Pd foil. This difference in height can be at-
tributed to the presence of oxidized Pd in the Pd/TiO2 cat-
alysts. Elsewhere, the maximum of the second XANES peak
is seen to move to lower energy for the used catalysts, com-
pared to the Pd foil. The shift in the position of the second
maximum is observed for both hydride and carbide forms
of Pd, however, the broadening of the first peak is only ob-
served for carbidic Pd.43-45 In this instance changes in the
first peak maximum are complicated by the separate con-
tribution of oxidized forms of Pd. Further evidence of car-
bide formation was sought by assessing the EXAFS (figures
7 and S9 and table S6), which is able to probe changes in
Pd-Pd spacing.
Used Catalyst Characterization. TEM was used to calcu-
late the average Pd particle size after the recycling experi-
ments were performed at both 25 and 50°C. The particle
size distribution (figure S7) indicate that for the samples
prepared in water, there is a small increase in average size
(2.5 to 2.7 nm for PdA1) when performed at 25°C, with no
Analysis of the EXAFS data confirms an increase of Pd-Pd
spacing after reaction from ~2.74 Å to 2.77 Å. An increase
in Pd-Pd spacing is encountered for the formation of Pd
carbide and hydride,39, 46 however, hydride is known to
ACS Paragon Plus Environment

Page 7 of 10
ACS Catalysis
readily desorb from the Pd lattice under normal atmos- DFT calculations. To assess the effect of carbidization on
1
2
3
4
5
6
7
8
pheric conditions. Considering this, it is apparent that this
expansion could not be caused by the formation of hydride.
The expansion of the Pd lattice from 2.74 Å to 2.79 Å has
been previously reported for bulk (PdCx, where x = 0.13) Pd
carbide formation, suggesting that only partial carbidiza-
tion has occurred.46 We propose that the deactivation en-
countered at lower reaction temperature is a result of the
transformation to Pd carbide, expansion of the Pd-Pd dis-
tance, and the subsequent effect on hydrogen dissociation
and the η2 adsorption of C=O and C=C functionalities. In-
deed, a greater degree of deactivation is observed for cata-
lysts A1Pd (464 → 394 mol (mol Pd)-1 h-1) compared to B1Pd
(491 → 450 mol (mol Pd)-1 h-1), which correlates with the
extent of carbidization observed. Although we experience
deactivation upon carbidization, other catalytic processes
experience different effects. Previous work investigating
the effect of carbide on the catalytic activity of acetylene
hydrogenation, concluded that both an increased, or de-
creased activity can be observed, depending on the type of
carbidized Pd site (flat = increase, step = decrease).47-48
the adsorption of furfural, DFT calculations were per-
formed. We first considered two different configurations of
furfural, parallel and perpendicular, on a pristine Pd(111)
surface, (figure 8 (a-b)). In the case of the perpendicular
configuration, we considered the stability of the system on
top of a Pd atom (Pdtop), on the bridge site between two Pd
atoms (Pdbridge) and on the three-fold hollow site (Pdhollow).
The calculated adsorption energies for the parallel config-
uration (Pdparallel) is -2.160 eV, which is slightly more nega-
tive than the earlier reported value of -1.830 eV, which can
be accounted for by the different versions of DFT-D ap-
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
proaches used.21 The adsorption energies for Pdtop, Pdbridge
,
and Pdhollow configurations are -0.632 eV, -0.940 eV and -
2.185 eV, respectively, with relaxed structures of Pdtop and
Pdbridge retaining a perpendicular orientation. However, it
is interesting to note that for the Pdhollow site the furfural
molecule attained a parallel configuration to the surface af-
ter relaxation (Pdhollow-parallel), and is comparable in energy
to the Pdparallel configuration. The adsorption energies ob-
tained confirm that the parallel configuration of furfural
binding is more stable than the perpendicular orientation.
Subsequently, we investigated the increase in Pd–Pd bond
distances, as a result of carbidization, in bulk Pd (by
0.095Å) and sub-surface of Pd(111) (by 0.057Å) (Figures S10
and S11), which were consistent with experimentally deter-
mined values of partially carbidized Pd nanoparticles.43-44
We found that the increase in Pd-Pd distance correlated
with an increased adsorption energy for furfural. The cal-
culated adsorption energy for this system is -1.613 eV, con-
firming that adsorption of furfural is more favourable on a
pristine Pd(111), compared to a PdC(111). Furthermore, on
comparing the geometry of furfural on pristine and carbide
systems we find that in both cases parallel adsorption is
preferred, although there are some minor differences in
the planarity of the furfural adsorption (Table S7).
Figure 7. FT χ2 weighted EXAFS data of fresh (black line)
and used (red line) PdA1.
Figure 8 (a) Parallel, (b) Perpendicular configurations of furfural on Pd(111) surface, (c) apDOS for pristine Pd(111) and (d)
apDOS for furfural on PdC(111) surface. Black and blue lines represent the signatures for Pd d orbitals close to C and O
atoms respectively. Red, aqua and green lines represent the p-orbitals due to C, O and interstitial C atoms respectively.
ACS Paragon Plus Environment

ACS Catalysis
Page 8 of 10
To assess this change in stability we analyzed the atom pro- Pd, reduces the binding energy of furfural, leading to a low-
1
2
3
4
5
6
7
8
jected partial density of states (apPDOS) of the C and O
atoms of furfural molecule and Pd atoms of the Pd(111) and
PdC(111) surfaces closest to each other. For convenience,
the Fermi energy (EF) is shifted to 0, which is represented
by a dotted line in Figure 8 (c-d). From the analysis of ap-
PDOS we see that around -7.0 eV to -5.5 eV (Figure 8 c-d)
there is a strong interaction between the p-orbitals, of C
and O atoms of furfural, with the d-orbital signatures of Pd
atoms on Pd(111) surface. However, on incorporating an in-
terstitial C-atom, another sharp signature due to C p or-
bital appears around this region, which has a higher con-
tribution than the p-orbital signatures due to C and O at-
oms of the furfural molecule. The interaction of the C p-
signatures, due to interstitial C-atoms, interact more
strongly with the nearby Pd atoms and result in a weaken-
ing of the interaction of furfural C and O atoms with the
Pd(111) surface. As a consequence, adsorption of furfural is
more favoured on pristine Pd(111), as compared to the
PdC(111).
ering of catalytic activity. To our knowledge this is the first
report of this proposed deactivation pathway for the hy-
drogenation α, β-unsaturated aldehydes.
AUTHOR INFORMATION
Corresponding Author
* Email: alberto.villa@unimi.it
* Email: p.p.wells@soton.ac.uk
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Author Contributions
S.M.R. synthesized catalysts, performed catalyst characteri-
zation (TEM, IR and XAFS) and had major contributions to
the writing of the manuscript. M.P. helped to synthesize the
catalysts and performed TEM. A.V. and C.E.C.T. conducted
all of the liquid phase hydrogenation experiments. N.J. and
R.E.P. conducted HAADF STEM. A.C., C.R.A.C., and A.T per-
formed all of the computational studies. N.D and C.R.A.C.
were involved in project discussions and supervised the pro-
ject. P.P.W and A.V. had major contributions to the writing
of the manuscript, were involved in all of the project discus-
sions and supervised the project. All authors discussed the
results and commented on the manuscript
Conclusions.
We demonstrate the tandem optimization of the colloidal
preparation of Pd nanoparticles, through adapting a simple
and systematic combination of choice of solvent and tem-
perature of reduction. These parameters allow us to tune
both the size domains of Pd nanoparticles and available re-
action sites, which directs the performance towards the hy-
drogenation of furfural. Reduced Pd NP size was achieved
by lowering the temperature of colloidal reduction; for the
catalyst prepared at 1°C in water the average particle size
was found to be 2.5 nm, however, we were also able to de-
tect clusters of Pd <20 atoms. For the B series prepared at
-30°C, an average particle size of 1.4 nm was produced,
which is smaller than has been previously reported for Pd
NPs prepared through a PVA/NaBH4 colloidal route. Site
selective catalysis was enabled by adjusting the solvent of
colloid preparation; the Pd/TiO2 catalysts prepared using a
water-ethanol solvent were found to have a larger propor-
tion of available corner and edge sites, which were able to
direct the selectivity of the furfural hydrogenation prod-
ucts. We propose that the greater number of available cor-
ner and edge sites adsorb furfural perpendicular to the na-
noparticle surface, resulting in greater selectivity to form
furfuryl alcohol over the complete reduction product, tet-
rahydrofurfuryl alcohol. We also find that the selectivity
profile does not correspond solely to particle size; for the B
series furfuryl alcohol selectivity decreases with increasing
particle size, whereas for the A series the inverse relation-
ship is found. The available sites are therefore not just a
result of particle size, but also a consequence of the inter-
action of the protecting agent with the NPs prepared. This
influence was also manifested in recycling studies, where
those catalysts prepared in water-ethanol, were more ro-
bust than those prepared solely in water. By assessing the
EXAFS of the used catalysts we were able to identify a Pd-
Pd lattice expansion. This is attributed to the formation of
Pd carbide, which we propose is responsible for the deac-
tivation observed. This assertion is supported by computa-
tional modelling studies, which show the carbidization of
Funding Sources
The UK Catalysis Hub Consortium is funded by EPSRC (port-
folio grants EP I019693, EP/K014706/1, EP/K014668/1,
EP/K014854/1 and EP/K014714/1). Via our membership of the
UKs HEC Materials Chemistry Consortium, which is funded
by EPSRC(EP/L000202), this work used the ARCHER UK Na-
tional Supercomputing Service.
ASSOCIATED CONTENT
Supporting Information.
TEM HAADF images with supporting Pd particle size distri-
bution histograms, EXAFS fitting parameters, CO-
chemisorption IR spectra, micro-wave plasma-atomic emis-
sion spectroscopy, computational studies and additional cata-
lytic data for furfural hydrogenation. This material is available
free of charge via the Internet at http://pubs.acs.org.
ACKNOWLEDGMENT
The authors wish to acknowledge the Diamond Light Source
for provision of beamtime (SP10306). Diamond Light Source
are also thanked for the outstanding support of the staff on
B18: Prof Andrew Dent, Dr Giannantonio Cibin, Dr Diego Gi-
anolio, Dr Stephen Parry and Mr Phil Robbins. The RCaH are
acknowledged for use of facilities and support of their staff.
Kristina Penman is thanked for performing the MP-AES meas-
urements. UK Catalysis Hub is kindly thanked for resources
and support provided via our membership of the UK Catalysis
Hub Consortium.
REFERENCES
1. Hu, X.; Westerhof, R. J. M.; Dong, D.; Wu, L.; Li, C. Z. ACS Sus-
tain. Chem. Eng., 2014, 2, 2562-2575.
2. Lange, J. P.; Van der Heide, E.; Van Buijtenen, J.; Price, R.
Chemsuschem, 2012, 5, 150-166.
3. Huber, G. W.; Iborra, S.; Corma, A. Chem. Rev., 2006, 106, 4044-
4098.
ACS Paragon Plus Environment

Page 9 of 10
ACS Catalysis
4. Li, G.; Li, N.; Li, S.; Wang, A.; Cong, Y.; Wang, X.; Zhang, T.
Chem. Comm., 2013, 49, 5727-5729.
Prati, L.; Dimitratos, N.; Villa, A.; Wells, P. P. ACS Catal., 2015, 5,
4377-4384.
1
5. Corma, A.; Iborra, S.; Velty, A. Chem. Rev., 2007, 107, 2411-2502.
6. Yan, K.; Wu, S.; Lafleur, T.; Jarvis, C. Renew. Sust. Energ. Rev.,
2014, 38, 663-676.
7. Nakagawa, Y.; Nakazawa, H.; Watanabe, H.; Tomishige, K.
ChemCatChem, 2012, 4, 1791-1797.
8. Panagiotopoulou, P.; Vlachos, D. G. Appl. Catal. A, 2014, 480,
17-24.
9. Nakagawa, Y.; Takada, K.; Tamura, M.; Tomishige, K. ACS
Catal., 2014, 4, 2718-2726.
10. Biradar, N. S.; Hengne, A. M.; Birajdar, S. N.; Niphadkar, P. S.;
Joshi, P. N.; Rode, C. V. ACS Sustain. Chem. Eng., 2014, 2, 272-281.
11. Rao, R.; Dandekar, A.; Baker, R. T. K.; Vannice, M. A. J. Catal.,
1997, 171, 406-419.
12. Nagaraja, B. M.; Padmasri, A. H.; Raju, B. D.; Rao, K. S. R. J.
Mol. Catal. A: Chem., 2007, 265, 90-97.
27. Wang, Z. W.; Toikkanen, O.; Yin, F.; Li, Z. Y.; Quinn, B. M.;
Palmer, R. E. J. Am. Chem. Soc., 2010, 132, 2854-+
28. Jian, N.; Palmer, R. E. J. Phys. Chem. C, 2015, 119, 11114-11119.
29. Ravel, B.; Newville, M. J. Synch. Rad, 2005, 12, 537-541.
30. Newville, M. J. Synch. Rad., 2001, 8, 322-324.
31. Grimme, S. J. Comput. Chem., 2006, 27, 1787-1799.
32. Blöchl, P. E. Phys. Rev. B, 1994, 50, 17953-17979.
33. Kresse, G.; Hafner, J. Phys. Rev. B, 1994, 49, 14251-14269.
34. Kresse, G.; Hafner, J. Phys. Rev. B, 1993, 47, 558-561.
35. Kresse, G.; Furthmüller, J. Phys. Rev. B, 1996, 54, 11169-11186.
36. Perdew, J. P.; Burke, K.; Ernzerhof, M. Phys. Rev. Lett., 1996,
77, 3865-3868.
37. Singh-Miller, N. E.; Marzari, N. Phys. Rev. B, 2009, 80, 9.
38. Su, R.; Tiruvalam, R.; He, Q.; Dimitratos, N.; Kesavan, L.; Ham-
mond, C.; Lopez-Sanchez, J. A.; Bechstein, R.; Kiely, C. J.; Hutch-
ings, G. J.; Besenbacher, F. ACS Nano, 2012, 6, 6284-6292.
39. Wells, P. P.; Crabb, E. M.; King, C. R.; Wiltshire, R.; Billsbor-
row, B.; Thompsett, D.; Russell, A. E. Phys. Chem. Chem. Phys.,
2009, 11, 5773-5781.
40. Lear, T.; Marshall, R.; Lopez-Sanchez, J. A.; Jackson, S. D.;
Klapotke, T. M.; Baumer, M.; Rupprechter, G.; Freund, H. J.; Len-
non, D. J. Chem. Phys., 2005, 123, 13.
41. Hadjiivanov, K. I.; Vayssilov, G. N. Adv. Catal., 2002, 47, 307-
511.
42. Zhu, H. Q.; Qin, Z. F.; Shan, W. J.; Shen, W. J.; Wang, J. G. J.
Catal., 2004, 225, 267-277.
43. Tew, M. W.; Nachtegaal, M.; Janousch, M.; Huthwelker, T.;
Van Bokhoven, J. A. Phys. Chem. Chem. Phys., 2012, 14, 5761-5768.
44. McCaulley, J. A. J. Phys. Chem., 1993, 97, 10372-10379.
45. Bugaev, A. L.; Guda, A. A.; Lazzarini, A.; Lomachenko, K. A.;
Groppo, E.; Pellegrini, R.; Piovano, A.; Emerich, H.; Soldatov, A.
V.; Bugaev, L. A.; Dmitriev, V. P.; Van Bokhoven, J. A; Lamberti,
C. Catal. Today. 2017, 283, 119-126.
46. McCaulley, J. A. Phys. Rev. B, 1993, 47, 4873-4879.
47. Yang, B.; Burch, R.; Hardacre, C.; Headdock, G.; Hu, P. J. Catal.
2013, 305, 264-276.
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
13. Rao, R. S.; Baker, R. T. K.; Vannice, M. A. Catal. Lett., 1999, 60,
51-57.
14. Sitthisa, S.; Resasco, D. E. Catal. Lett., 2011, 141, 784-791.
15. Sitthisa, S.; Sooknoi, T.; Ma, Y.; Balbuena, P. B.; Resasco, D. E.
J. Catal., 2011, 277, 1-13.
16. Sitthisa, S.; An, W.; Resasco, D. E. J. Catal., 2011, 284, 90-101.
17. Aldosari, O. F.; Iqbal, S.; Miedziak, P. J.; Brett, J. L.; Jones, D.
R.; Liu, X.; Edwards, J. K.; Morgan, D. J.; Knight, D. K.; Hutchings,
G. J. Catal. Sci. Technol., 2016, 6, 234-242.
18. Iqbal, S.; Liu, X.; Aldosari, O. F.; Miedziak, P. J.; Edwards, J. K.;
Brett, J. L.; Akram, A.; King, G. M.; Davies, T. E.; Morgan, D. J.;
Knight, D. K.; Hutchings, G, J. Catal. Sci. Technol., 2014, 4, 2280-
2286.
19. King, G. M.; Iqbal, S.; Miedziak, P. J.; Brett, J. L.; Kondrat, S. A.;
Yeo, B. R.; Liu, X.; Edwards, J. K.; Morgan, D. J.; Knight, D. K.;
Hutchings, G. J. ChemCatChem, 2015, 7, 2122-2129.
20. Shekhar, R.; Barteau, M. A.; Plank, R. V.; Vohs, J. M. J. Phys.
Chem. B, 1997, 101, 7939-7951.
21. Vorotnikov, V.; Mpourmpakis, G.; Vlachos, D. G. ACS Catal.,
2012, 2, 2496-2504.
22. Bhogeswararao, S.; Srinivas, D. J. Catal., 2015, 327, 65-77.
23. Schoenbaum, C. A.; Schwartz, D. K.; Medlin, J. W. Acc. Chem.
Res., 2014, 47, 1438-1445.
24. Pang, S. H.; Schoenbaum, C. A.; Schwartz, D. K.; Medlin, J. W.
Nat. Commun., 2013, 4, 2448, 1-6.
25. Pang, S. H.; Schoenbaum, C. A.; Schwartz, D. K.; Medlin, J. W.
ACS Catal., 2014, 4, 3123-3131.
26. Rogers, S. M.; Catlow, C. R. A.; Chan-Thaw, C. E.; Gianolio, D.;
48. Yang, B.; Burch, R.; Hardacre, C.; Hu, P.; Hughes, P. Surf. Sci.
2016, 646, 45-49.
Gibson, E. K.; Gould, A. L.; Jian, N.; Logsdail, A. J.; Palmer, R. E.;
ACS Paragon Plus Environment

ACS Catalysis
Page 10 of 10
TOC graphic:
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
10
ACS Paragon Plus Environment