Ruthenium Nanoparticles Supported on Carbon: An Active Catalyst for the Hydrogenation of Lactic Acid to 1,2-Propanediol

Article abstract of DOI:10.1021/acscatal.5b00625

The hydrogenation of lactic acid to form 1,2-propanediol has been investigated using Ru nanoparticles supported on carbon as a catalyst. Two series of catalysts which were prepared by wet impregnation and sol-immobilization were investigated. Their activity was contrasted with that of a standard commercial Ru/C catalyst (all catalysts comprise 5 wt % Ru). The catalyst prepared using sol-immobilization was found to be more active than the wet impregnation materials. In addition, the catalyst made by sol-immobilization was initially more active than the standard commercial catalyst. However, when reacted for an extended time or with successive reuse cycles, the sol-immobilized catalyst became less active, whereas the standard commercial catalyst became steadily more active. Furthermore, both catalysts exhibited an induction period during the first 1000 s of reaction. Detailed scanning transmission electron microscopy, X-ray photoelectron spectroscopy and X-ray absorption fine structure analysis data, when correlated with the catalytic performance results, showed that the high activity can be ascribed to highly dispersed Ru nanoparticles. Although the sol-immobilization method achieved these optimal discrete Ru nanoparticles immediately, as can be expected from this preparation methodology, the materials were unstable upon reuse. In addition, surface lactide species were detected on these particles using X-ray photoelectron spectroscopy, which could contribute to their deactivation. The commercial Ru/C catalysts, on the other hand, required treatment under reaction conditions to change from raft-like morphologies to the desired small nanoparticle morphology, during which time the catalytic performance progressively improved.

Full text of DOI:10.1021/acscatal.5b00625

Page 1 of 34
ACS Catalysis
1
2
3
4
Ruthenium nanoparticles supported on carbon ꢀ an active catalyst for the hydrogenation of
lactic acid to 1,2ꢀpropanediol
5
6
7
8
Sarwat Iqbal,a Simon A. Kondrat,a Daniel R. Jones,a Daniël Schoenmakers,a Jennifer K. Edwards, a
Li Lu,b Benjamin R. Yeo, a Peter P. Wells,c Emma K. Gibson,c David J. Morgan,a Christopher J.
Kielyb and Graham J. Hutchingsa,*
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
*hutch@cardiff.ac.uk
a Cardiff Catalysis Institute, School of Chemistry, Main Building, Cardiff University, Park Place,
Cardiff, CF10 3AT, UK
b
Department of Materials Science and Engineering, Lehigh University, 5 East Packer Avenue,
Bethlehem, PA 18015ꢀ3195, USA
c
The UK Catalysis Hub, Research Complex at Harwell, Harwell, Oxon, OX11 0FA, UK and,
Kathleen Lonsdale Building, Department of Chemistry, University College London, Gordon Street,
London, WC1H 0AJ, UK
1
ACS Paragon Plus Environment

ACS Catalysis
Page 2 of 34
1
2
Abstract
3
4
5
6
7
8
9
The hydrogenation of lactic acid to form 1,2ꢀpropanediol has been investigated using Ru
nanoparticles supported on carbon as a catalyst. Two series of catalysts were investigated which
were prepared by wet impregnation and solꢀimmobilisation. Their activity was contrasted with that
of a standard commercial Ru/C catalyst (all catalysts comprise 5 wt% Ru). The catalyst prepared
using solꢀimmobilisation was found to be more active than the wet impregnation materials. In
addition, the catalyst made by solꢀimmobilisation was initially more active than the standard
commercial catalyst. However, when reacted for an extended time or with successive reꢀuse cycles
the solꢀimmobilised catalyst became less active, whereas the standard commercial catalyst became
steadily more active. Furthermore, both catalysts exhibited an induction period during the first
1000s of reaction. Detailed scanning transmission electron microscopy, Xꢀray photoelectron
spectroscopy and XAFS data, when correlated with the catalytic performance results, showed that
the high activity can be ascribed to highly dispersed Ru nanoparticles. While the solꢀimmobilisation
method achieved these optimal discrete Ru nanoparticles immediately, as can be expected from this
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
preparation methodology, the materials were unstable upon reꢀuse. In addition, urface lactide
S
species were detected on these particles using Xꢀray photoelectron spectroscopy which could
contribute to their deactivation. The commercial Ru/C catalysts on the other hand, required
treatment under reaction conditions to change from raftꢀlike morphologies to the desired small
nanoparticle morphology, during which time the catalytic performance progressively improved.
Keywords: Catalytic hydrogenation, lactic acid, Ru/C, catalyst reuse, propane diol, propylene
glycol
2
ACS Paragon Plus Environment

Page 3 of 34
ACS Catalysis
1
2
Introduction
3
4
5
6
7
8
9
1,2ꢀpropanediol (PDO) is an attractive commodity chemical, frequently used as a deꢀicing fluid and
antiꢀfreeze, as well as in the production of cosmetics and pharmaceuticals.1 The production of PDO
currently requires the selective oxidation of propene and involves environmentally unfriendly
processes e.g. hydroperoxidation chemistry or the chlorohydrin process which utilises hypochlorous
acid. PDO can also, in principle, be generated from lactic acid or lactate esters, providing an
alternative and much greener production pathway than the conventional commercial synthesis.
Fermentation of glucose can produce large quantities of lactic acid and its cost is expected to
decrease as advances are made in fermentation and separation techniques. Hence, lactic acid can be
considered to be a platform material that is derived from biomass which can form the starting point
for the manufacture of a series of derivatives. We consider that 1,2ꢀpropanediol is a viable target
molecule and if a suitable catalyst can be identified, an economically viable green route for the
production of 1,2ꢀpropanediol from biomass would become available.
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
The catalytic hydrogenation of lactic acid to PDO requires the formation of an alcohol by the
hydrogenation of a carboxyl group, without removal of the αꢀhydroxyl group. The hydrogenolysis
of the αꢀhydroxyl group would result in the formation of propionic acid.2 Previous studies have
indicated that relatively harsh reaction conditions are required for the hydrogenation of lactic acid
and its esters, due to the intrinsically low reactivity of carboxylic acid groups with hydrogen.3ꢀ6
Thermodynamic considerations also suggest that the formation of propionic acid is favoured
compared to 1,2ꢀpropanediol.2 Cortright et al.7 have experimentally shown that the lower
temperature, and higher hydrogen pressure can favour the formation of PDO from lactic acid. The
first catalytic hydrogenation of lactic acid was reported by Broadbent et al.8 over an unsupported Re
black catalyst at 150°C, and 27 MPa pressure of H2 in which they achieved an 80% yield of PDO.
Copper/chromium oxide and Raney nickel catalysts have also been reported to give yields of PDO
using ethyl lactate as a substrate. The yield of PDO was as high as 80% at 250°C and 20ꢀ30 MPa of
H2.9 Several recent studies have reported Ruꢀbased catalysts as being highly active for lactic acid
hydrogenation.10ꢀ16 For instance, Zhang et al.10 found that hydrogenation of lactic acid to PDO is
possible over a Ruꢀcontaining catalyst at 150°C and 14.5 MPa of H2. In subsequent studies, the
vapour phase hydrogenation of lactic acid was performed over a Cu/SiO2 catalyst which led to high
conversion and good selectivity for PDO even at atmospheric pressure.7 In addition, ꢀ
Cu/SiO2,7Pd/SiO2, Ni/SiO2 and Fe/SiO216 catalysts have all been shown to be efficient catalysts for
this reaction, but only under relatively harsh reaction conditions. Lei et al.17 described a novel
method of lactic acid hydrogenation utilising syngas over a Au/ZrO2 catalyst. Overall however the
supported Ruꢀbased catalysts have generated increasing attention in the literature for this reaction
3
ACS Paragon Plus Environment

ACS Catalysis
Page 4 of 34
1
2
because of their superior intrinsic activities.2 Ru supported on activated carbon, Al2O3 and TiO2
3
4
10,18
have all been shown to be more active than Raney nickel, copper chromite and Ni/Al2O3
A biꢀ
5
6
7
8
functional Ru/TiO2 catalyst has been proposed for this reaction at 150°C, where the small Ru
nanoparticles serve to dissociate molecular hydrogen and the TiO2 support activates the carbonyl
group.18 There are some reports also available describing the effects of promoters such as Sn, B, Co,
Fe and Zn on Ruꢀbased catalysts for the hydrogenation of lactic acid.16,17
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
We consider the efficient synthesis of 1,2ꢀpropanediol from lactic acid is an area of considerable
interest in modern day green chemistry. To some extent, selective synthesis has already been
achieved with several catalysts including Ruꢀbased catalysts. The RuꢀC materials are the most
widely studied catalysts for this reaction, but most of the literature is focussed on using
commercially available RuꢀC materials for either (i) characterizing the reaction parameters, or (ii)
for comparing different promoters. In the current study, we report that various different types of
carbon can be used for supporting the Ru nanoparticles and that they show different activities under
moderate reaction conditions. We also compare the activity of a solꢀimmobilized Ru/C catalyst with
a commercial RuꢀC catalyst and reconcile variations in their catalytic performance to nanostructural
differences deduced using a complementary selection of advanced materials characterization
techniques.
Experimental Methods
Materials
A number of activated carbon supports (Darco G60, Cabot Vulcan XC72R, Norit GCN3070, and
Norit ROX 0.8) were obtained for use in this study. Graphene oxide was purchased from Sigmaꢀ
Aldrich. Solutions of RuNO(NO3)3 (Sigma Aldrich, 1.5 wt. % Ru), and RuCl3 (Sigma Aldrich, 45ꢀ
55 wt % Ru) were used as ruthenium precursors. Lactic acid (Sigma Aldrich, 98 %) was used asꢀ
received. Possible liquid products (1,2ꢀpropanediol, propionic acid, methanol, ethanol, 1ꢀpropanol
and 2ꢀpropanol) were purchased from Sigma Aldrich and possible gas products (methane, ethane,
and propane) were purchased from BOC and these were used as standards for calibration and
determination of analytical purity. A commercial 5 wt% Ru/C (Sigma Aldrich) catalyst was used as
a baseline to compare the activity of the synthesised catalysts.
4
ACS Paragon Plus Environment

Page 5 of 34
ACS Catalysis
1
2
Catalyst preparation
3
4
5
6
7
8
9
Catalysts were prepared by two methods, namely wet impregnation (WI) and sol immobilization
(SI). For WI catalysts, a solution of RuNO(NO3)3 (5 wt% with respect to activated carbon) was
added to the support in order to obtain a paste. The catalyst was then dried (110°C, 16 h) and
subsequently heat treated in N2 (400°C, 3h, 20°C/min ramp rate). SI catalysts were prepared starting
from a solution of PVA (10 mg) and RuCl3 (0.04g) in water (800 ml). A freshly prepared solution of
NaBH4 (RuCl3/NaBH4 = 1/3.3 mole %) was then added to generate the sol. After 30 min, the
activated carbon (1.98g) was added and the solution was acidified to pH 2 with sulphuric acid. The
catalyst was then filtered and dried (110°C, 16 h).
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
Catalyst characterization
Xꢀray powder diffraction (XRPD) was performed using a PANalytical X'Pert Pro diffractometer
fitted with an X'Celerator detector and a Cu Kα Xꢀray source operated at 40 kV and 40 mA.
Temperature programmed reduction (TPR) analysis was performed with a Quantachrome ChemBet
system equipped with a cold trap using 10% H2/Ar (5 ml/min), with a ramp rate of 20°C/min.
Samples (0.5g) were preꢀtreated at 100oC (ramp rateꢀ20oC/min) under He for 1h prior to analysis.
Samples for examination by scanning transmission electron microscopy (STEM) were prepared by
dispersing the dry catalyst powder onto a holey carbon film supported by a 300 mesh copper TEM
grid. Samples were then subjected to bright field (BF) and high angle annular dark field (HAADF)
imaging using an aberration corrected JEOL ARM 200CF microscope operating at 200kV.
Microwave plasma atomic emission spectroscopy (MPꢀAES) was performed using an Agilent 4100
MPꢀAES. The Ru content was analysed with two emission lines with 349.894 nm and 371.803 nm.
The samples were introduced to the nitrogen plasma using a single pass spray chamber at a pressure
of 120 kPa without air injection. The instrument was calibrated with 1, 2, 4, 7, 10 ppm standards in
10% HCl along with a 5% lactic acid blank. The samples were tested in triplicate and the average
result was used. A lactic acid rinse solution (5%) was introduced between samples for 15 s to
ensure there was no sample contamination.
Samples were characterized by Xꢀray photoelectron spectroscopy (XPS) using a Kratos Axis Ultraꢀ
DLD photoelectron spectrometer, utilising monochromatic Al Kα radiation. High resolution and
survey scans were performed at pass energies of 40 and 160 eV respectively. Spectra were
calibrated to the C (1s) signal at 284.5 eV.
5
ACS Paragon Plus Environment

ACS Catalysis
Page 6 of 34
1
2
3
4
5
6
7
Surface area and pore size analysis were determined by N2 adsorption at 77 K using a
Quantachrome Autosorb AS1 system. Samples were degassed for a minimum of 24 h at 120°C
before the analysis. Surface areas were evaluated using the BET method and the pore size
distribution was determined by applying the nonꢀlocal density functional theory (NLDFT) method. .
8
9
Ru K edge XAFS studies were carried out on the B18 beamline at the Diamond Light Source,
Didcot, UK. Measurements were performed using a QEXAFS setꢀup with a fastꢀscanning Si (311)
double crystal monochromator. The time resolution of the spectra reported herein was 5
min/spectrum (kmax = 14, step size 0.5 eV). On average three scans were acquired to improve the
signalꢀtoꢀnoise level of the data for transmission measurements. All solid reference samples were
diluted with cellulose and pressed into pellets to optimise the effective edgeꢀstep of the XAFS data
and measured in transmission mode using ion chamber detectors. All XAFS spectra were acquired
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
concurrently with the appropriate foil placed between It and
Iref. XAFS data processing was
performed using IFEFFIT19 with the Horae package20
(i.e. Athena and Artemis methods). The
amplitude reduction factor, ꢀꢁꢂ, was derived from EXAFS data analysis of a known reference
compound, namely Ru metal foil, which was determined to be 0.85 and used as a fixed input
parameter.
Lactic acid hydrogenation
The hydrogenation of lactic acid was carried out using a stainless steel stirred autoclave (50 ml, Parr
Instruments, Model 5500HP) equipped with a Teflon liner using catalyst (0.025 g), lactic acid (0.50
g) and deionised water (9.50 g). The sealed autoclave was purged three times with N2, and three
times with H2 before being pressurised to 35 bar with H2. The autoclave was heated to 120oC and
stirred at 2000rpm. When the reaction was completed, the mixture was cooled, filtered and the
products were analysed using an HPLC column (Agilent Technologies 1260 Infinity) equipped with
a Varian MetaCarb 67H capillary column (0.65 x 30 cm) and a refractive index detector. Gas
analysis was carried out using gas chromatography (Varian 450ꢀGC) equipped with flame
ionisation and thermal conductivity detectors, a methaniser and a CPꢀSiL5CB column (50m,
0.33mm) using He as a carrier gas.
All the experiments were performed at least three times and the catalytic results were compared
using the standard deviation formula. These results are provided in terms of error bars.
Results and discussion
6
ACS Paragon Plus Environment

Page 7 of 34
ACS Catalysis
1
2
Lactic acid hydrogenation
3
4
5
6
7
8
9
Lactic acid (LA) hydrogenation was investigated over a series of 5wt% Ru/C catalysts prepared
using wet impregnation with a range of different carbon supports. The activity was compared with
that of a commercial 5wt% Ru/C catalyst which was used as a standard for this study. The results
are shown in Table 1 and the only product observed for all catalysts was PDO at the reaction
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
o
temperature we studied of 120 C. This is an interesting observation and contrasts with previous
observations for this reaction.7, 21 In other studies a number of byꢀproducts have been observed such
as propanol, and propionic acid with Cu/SiO2 catalysts,7 and lower hydrocarbons with Ru/C catalyst
under harsher reaction conditions (e.g., high pressure).21 However, using our Ru catalysts we have
not observed any product other than propylene glycol. This we consider is due to the different
preparation method, carbon types and reaction conditions we have utilised.
Table 1. LA conversion and selectivity to PDO for 5wt% Ru catalysts supported on carbon
prepared by the wet impregnation method.
Catalyst (Ru/C)
LA conversion (%) PDO selectivity (%)
Commercial
ROX0.8
GCN3070
G60
Graphene oxide
XC72R
19
13
10
9
2
15
100
100
100
100
100
100
Reaction conditions: 5 wt% LA (0.5g) in water (9.5g), 120oC, 35 bar H2 , 2.5 h.
All the catalysts prepared by wet impregnation were active for the lactic acid hydrogenation
reaction, but all were considerably less active than the commercial standard catalyst. Graphene
oxide was found to be the least suitable support of all those tested as it only exhibited a conversion
of 2%. However, the catalyst prepared using XC72R carbon as the support gave the best results and
so this was selected for use as a support to investigate an alternative method of metal deposition. A
5wt% Ru on XC72R carbon catalyst was prepared using solꢀimmobilisation (designated Ru/C
XC72RꢀSI). When investigated for lactic acid hydrogenation, this solꢀimmobilised material was
more active than the commercial 5wt% Ru/C standard catalyst. It was also more active than the
Ru/C XC72R prepared by wet impregnation method, showing that the type of carbon as well as Ru
deposition method have a significant effect on catalytic activity. Using reaction conditions that were
7
ACS Paragon Plus Environment

ACS Catalysis
Page 8 of 34
1
2
3
4
5
6
the same as those used for the wet impregnation catalysts, the Ru/C XC72RꢀSI catalyst gave 23%
conversion compared with 19% for the commercial catalyst (Table 1) after 2.5 h of reaction. Again
only PDO was observed as a product.
7
8
9
We decided to investigate the Ru/C XC72RꢀSI and commercial catalysts in more detail. Figure 1
shows the effect of the reaction time on the LA conversion. PDO was the only product for the
reactions performed over longer reaction times. No by products were detected both in liquid and gas
phases. It is apparent that there are three features characteristic of the reaction profile. First there is
an induction period for both catalysts; second the catalyst prepared by solꢀimmobilisation is initially
more active than the commercial Ru catalyst and subsequently the solꢀimmobilised catalyst
becomes deactivated, but the commercial Ru catalyst retains activity and after 5 h of reaction
demonstrates higher activity.
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 1. Effect of the reaction time on lactic acid conversion. Key ● 5 wt% Ru/XC72RꢀSI; ▲ 5
O
wt% Ru/C commercial catalyst. Reaction conditions: 120 C, 35 bar H2, 5%LA/H2O, catalyst
(0.025 g).
The induction period of ca. 1000 s is observed for both catalysts and Figure 2 shows this more
clearly. There are several reasons why an induction period may be observed. First it is possible that
Ru is leached from the catalyst and homogeneous catalysis then ensues. Reaction mixtures were
analysed for Ru by MPꢀAES and no leaching was detected. Hence any changes in the reaction rate
cannot be attributed to loss of Ru from the catalyst or due to a homogeneous reaction. Second the
8
ACS Paragon Plus Environment

Page 9 of 34
ACS Catalysis
1
2
3
4
5
6
7
8
effect could be due to the initial reduction of the catalyst. To investigate this we preꢀreduced both
catalysts (5wt% Ru/CXC72ꢀSI and 5 wt% Ru/C commercial) in H2 ex-situ. However, these
catalysts were less active than the nonꢀreduced catalysts (after 2.5 h: 23% LA conversion for the
nonꢀreduced versus 18% LA conversion for the preꢀreduced 5wt% Ru/CXC72ꢀSI catalyst and 19%
LA conversion for the nonꢀreduced versus 13% LA conversion for the preꢀreduced 5 wt% Ru/C
commercial). The induction period may therefore involve the creation of the active surface which
involves the presence of LA and H2.
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
Following the induction period and between 1000 – 9000 s the reaction is first order in LA with the
measured first order rate constants being 4.5 x 10ꢀ4 sꢀ1 (for 5wt% Ru/CXC72ꢀSI) and 3.7 x 10ꢀ4 sꢀ1
(for the commercial 5wt% Ru/C catalyst). The turnꢀover frequencies (TOF) were calculated and the
data provided in literature10,22 was used to calculate the TOF for a comparison purpose. For the
5wt% Ru/CXC72ꢀSI the TOF is 4116 hꢀ1 (productivity = 1.84 mol LA gcatalystꢀ1 hꢀ1), compared with
2312 hꢀ1 (productivity = 1.1 mol LA gcatalystꢀ1 hꢀ1) for the commercial catalyst. These values for our
Ru catalysts are far higher than that reported by Miller et al10; 405 hꢀ1 (productivity = 0.20 mol LA
gcatalystꢀ1 hꢀ1) and Jang et al22 ; 1435 hꢀ1 (productivity = 0.71 mol LA gcatalystꢀ1 hꢀ1) who also used a
5%Ru/C catalyst together with more severe reaction conditions (150°C and 50 bar H2).
Figure 2. Effect of reaction time during the initial reaction period. Key: ▲ commercial 5wt% Ru/C;
● 5wt% Ru/CXC72ꢀSI. Reaction conditions: 120 °C, 35 bar H2, 5%LA/H2O, catalyst (0.025 g).
9
ACS Paragon Plus Environment

ACS Catalysis
Page 10 of 34
1
2
3
4
5
6
7
8
An important feature of any catalyst performance is that it should be reꢀusable. In view of this we
have conducted a series of reꢀusability experiments using the same catalyst mass in each
experiment. The catalysts were used for 2.5 h then recovered washed with water, dried and then reꢀ
used. The results for the commercial 5wt% Ru/C catalyst are shown in Figure 3 for four successive
reꢀuse cycles.
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
35
30
25
20
15
10
5
0
1
2
3
4
Cycles
Figure 3. Effect of number reꢀuse cycles on the LA conversion for the commercial 5wt% Ru/C
catalyst. Reaction conditions: LA (0.5 g) in water (0.95 g), 120oC, 35 bar H2, 2.5 h reaction time
catalyst (0.025 g).
10
ACS Paragon Plus Environment

Page 11 of 34
ACS Catalysis
1
2
3
4
5
6
7
8
9
35
30
25
20
15
10
5
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
0
1
2
3
4
Cycles
Figure 4. Effect of number of reꢀuse cycles on the LA conversion for the 5 wt% Ru/C XC72RꢀSI
catalyst. Reaction conditions: LA (0.5 g) in water (0.95 g), 120oC, 35 bar H2, 2.5 h reaction time,
catalyst (0.025 g).
An increase in conversion occurs with successive catalyst reꢀuse cycles. PDO was the only product
detected in all these experiments. We then conducted reꢀuse experiments for the 5 wt% Ru/C
XC72RꢀSI material which showed a distinctly different reꢀuse profile with a steady decrease in
activity being observed on each subsequent reꢀuse, again with PDO being the only product detected
(Figure 4). These trends are in direct agreement with the reaction profile noted in Figure 1 since the
four successive reꢀuses simulate the reaction over the first 10 h reaction period. For the commercial
catalyst during this reaction time the activity increases whereas the solꢀimmobilised catalyst initially
has a higher activity that then declines. Further, we studied the time on line effect on both catalysts
after their first use and the data are presented in Figure 5. Interestingly, no change in induction
period was observed for 1000 s with both of the used catalysts. A comparison of the time online
data of the used catalysts with the fresh catalysts showed that there is an increase in catalytic
activity of the used commercial 5wt% Ru/C catalyst after about 3600 s and the activity of the Ru/C
XC72RꢀSI catalyst remained similar to the fresh catalyst. This trend is further observed in the
successive reuse cycles (Figures 3 and 4) where the commercial 5wt% Ru/C catalyst showed an
improved activity and a decline in activity of the Ru/C XC72RꢀSI catalyst was observed. To
understand these effects observed in the catalytic hydrogenation reaction we have characterised
both catalysts in a considerable detail.
11
ACS Paragon Plus Environment

ACS Catalysis
Page 12 of 34
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 5. Effect of the reaction time during the initial reaction period. Key: ▲ commercial 5wt%
Ru/C fresh; ● 5wt% Ru/CXC72ꢀSI fresh; Δ commercial 5wt% Ru/C after one use; ○ 5wt%
°
Ru/CXC72ꢀSI after one use. Reaction conditions: 120 C, 35 bar H2, 5%LA/H2O, catalyst (0.025
g).
Catalyst characterisation
The BET surface areas of the bare carbon supports and the Ru/C samples were determined along
with the pore size distributions of the two most active catalysts. The data (see supporting data,
Tables S1, S2, S3 & Figures S1, S2, and S3) show that the addition of Ru did not affect the overall
surface area for any particular carbon support. Furthermore although there are significant variations
in surface areas and pore size distributions between the different carbons tested, there is no
correlation with the observed catalyst activity. XRD patterns for the Ruꢀcatalysts supported on
various carbons did not show any reflections that could be assigned to Ru (see supporting data,
Figure S4), suggesting that the supported particles are smaller than 3ꢀ4nm in size. However, the Ru
surface area as determined by CO chemisorption for commercial 5wt% Ru/C catalyst and the 5wt%
Ru/CXC72ꢀSI sample were found to be similar (3 m2gꢀ1).
The 5 wt% Ru catalysts on different carbon supports were analysed using TPR (see supplemental
data, Figures S6 and S7). The commercial catalyst, and Ru supported on GCN3070 show two
distinct reduction peaks, one below 100°C and the other above 200°C. The catalysts prepared by
12
ACS Paragon Plus Environment

Page 13 of 34
ACS Catalysis
1
2
3
4
5
6
7
8
wet impregnation and solꢀimmobilization methods using XC72 carbon show two reduction peaks
below 100 °C, whereas the catalyst prepared by solꢀimmobilization method has an additional broad
peak at 140 °C. The first reduction peaks at a lower temperature can be assigned to the reduction of
Ru3+to Ru2+ while the other peak at higher temperature could be due to second step reduction of
Ru2+ to Ru1+ however, there is also the possibility of reduction of Ru2+ to the metallic state.23, 24 Ru
supported on G60 and ROX 0.8 show only one reduction peak just below 100°C, but these catalysts
are less active. In comparison, the graphene oxide supported Ru catalyst exhibits a higher reduction
temperature (~150°C). None of the catalysts were reduced prior to our catalytic testing experiments,
hence we expect that all the catalysts, apart from the one supported on graphene oxide, will be
reduced during the initial phase of the reaction which takes place at 120oC under 35 bar of H2. This
may account for the abnormally low catalytic activity exhibited by the graphene oxideꢀsupported
Ru catalyst where the lactic acid hydrogenation temperature is always kept below the 150oC value
needed for Ru reduction on the GO surface.
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
We decided to examine the structure of the most active solꢀimmobilised catalyst (Ru/C XC72RꢀSI)
and the standard commercial catalyst using Xꢀray Absorption Fine Structure (XAFS). This is a
powerful technique for determining the local structure for materials such as nanoparticles that do
not display long range order, with the near edge structure (XANES) also providing information on
the oxidation state. Figure 6 shows the XANES data for the commercial 5wt% Ru/C and 5wt%
Ru/C XC72RꢀSI catalysts, alongside reference spectra taken from RuO2.xH2O, anhydrous RuO2 and
Ru metal foil. The Ru/C catalyst samples have an edge position of 22130 eV as do both the standard
ruthenium oxide samples, which is higher in energy than the edge position of the Ru foil at 22117
eV, as would be expected for a shift of valence state from Ru0 to Ru4+. Other groups have reported
the same edge position (i.e. 22130eV) for anhydrous RuO2 and RuO2.xH2O, even though some Ru3+
centres are known to be present in the hydrated sample.25 As such we can also not rule out the
presence of a small Ru3+ component in addition to Ru4+ in our Ru/C catalyst materials. The double
maxima feature observed beyond the absorption edge is associated with anhydrous forms of RuO2,
with completely hydrated forms having only one maximum.26 The XANES spectra of RuO2.xH2O,
the 5 wt% Ru/C commercial catalyst and 5 wt% Ru/C XC72RꢀSI catalyst all show varying degrees
of hydration, with the 5 wt% Ru/C commercial catalyst showing the most pronounced double
maxima feature, implying that it therefore has the greatest anhydrous character.
The Fourier Transform of the χ2 weighted extended Xꢀray absorption fine structure (EXAFS)
spectra of anhydrous RuO2, RuO2.xH2O, the 5wt% Ru/C commercial catalyst and the 5wt% Ru/C
XC72RꢀSI catalysts are shown in Figure 7. The simulated EXAFS fits are also shown superimposed
for the two Ru/C catalyst samples. By studying the EXAFS data it is possible to differentiate
13
ACS Paragon Plus Environment

ACS Catalysis
Page 14 of 34
1
2
3
4
5
6
7
8
between hydrated forms of RuO2. When RuO2 is hydrated its native rutile structure is disturbed,
which disrupts the RuꢀRu scattering paths in the EXAFS Fourier transform. This phenomenon has
been attributed to chains of RuO6 octahedra, where the three dimensional network of these chains
does not extend out as far as for anhydrous RuO2.25 This effect can be seen by the lack of features
after 3Å for the RuO2.xH2O sample in the Fourier transform data shown in Figure 7. The EXAFS
fitting parameters (Table 2) for the two Ru/C catalyst samples show that the standard commercial
material is less hydrated than the solꢀimmobilised Ru/CꢀXC72RꢀSI material. This is evidenced by
the presence of an additional RuꢀRu coordination shell at 3.56 Å for the former catalyst, with both
samples exhibiting an initial RuꢀRu distance at 3.11 Å, Both the 3.11 Å and 3.56 Å distances are
characteristic of plane spacings present in the rutile structure expected of anhydrous RuO2. It can
therefore be deduced that the chains of RuO6 octahedra have a more extended threeꢀdimensional
structure in the commercial Ru/C sample, although not quite as extended as that for purely
anhydrous RuO2. We therefore consider that the differences in the initial activities we observe with
these two Ru/C catalysts might be associated with their different initial levels of hydration.
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
14
ACS Paragon Plus Environment

Page 15 of 34
ACS Catalysis
1
2
3
4
5
6
7
8
1.4
1.2
1.0
0.8
0.6
0.4
0.2
0.0
Ru foil
RuO .xH O
2
2
9
RuO
2
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
5 wt% Ru/C commercial
5 wt% Ru/C (XC72R)
22080
22100
22120
22140
22160
22180
Energy / eV
Figure 6. Normalised XANES spectra for (i) the standard commercial 5wt% Ru/C catalyst, (ii) the
most active 5wt% Ru/C XC72RꢀSI catalyst, (iii) RuO2.xH2O, (iv) anhydrous RuO2 and (v) metallic
Ru foil. The inset figure shows the expanded XANES region for these materials.
15
ACS Paragon Plus Environment

ACS Catalysis
Page 16 of 34
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 7. The Fourier Transform of the χ2 weighted EXAFS spectra of (i) anhydrous RuO2, (ii)
RuO2.xH2O, (iii) the 5wt% Ru/C commercial catalyst and (iv) 5wt% Ru/C XC72RꢀSI catalyst.
Simulated EXAFS fits are also shown as red and purple dashed lines for the two Ru/C catalyst
samples.
16
ACS Paragon Plus Environment

Page 17 of 34
ACS Catalysis
1
2
Table 2. EXAFS fitting parameters for the standard commercial 5wt% Ru/C catalyst and the 5 wt%
3
4
Ru/XC72R_SI catalyst shown in Figure 6.
5
6
7
8
Sample
Abs Sc
Ru ꢀ O
Ru ꢀ O
N
R
/ Å
2σ2 / Å2
0.008 (1)
0.012 (5)
E
f / eV
Rfactor
5 wt%
1.8 (4)
3.7(4)
1.85 (2)
2.02 (1)
ꢀ5 (1)
0.02
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
Ru/XC72RꢀSI
catalyst
Ru ꢀRu
2.0 (6)
3.12 (1)
0.009(2)
5wt% Ru/C
commercial
catalyst
Ru – O
Ru – O
1.9 (4)
4.0 (5)
1.87 (2)
2.00 (1)
0.003 (2)
0.003 (1)
ꢀ4 (2)
0.04
Ru – Ru
Ru ꢀ Ru
2.2 (6)
2.0 (6)
3.11 (2)
3.56 (2)
0.008 (2)
0.006 (1)
2
Fitting parameters: S0 = 0.85 as deduced from a Ru foil standard; Fit range 2.5<k<14, 1<R<4;
number of independent points = 22.
As the two catalysts showed intriguingly different reaction profiles either in continuous use (Figures
1 and 2) or on successive reꢀuse cycles (Figures 3 and 4), we subjected them to detailed
characterisation. No discernible change in surface area from BET analysis was observed, indicating
that no significant coking that blocked access to the Ru particles was occurring. Also MPꢀAES of
the reaction effluents revealed no Ru leaching. STEM analysis was performed for the fresh and used
batches of both catalysts (commercial Ru/C and Ru/CXC72RꢀSI). Figures 8(a) and (b) show a
complementary pair of BFꢀ and HAADFꢀSTEM images of the unused commercial Ru/C catalyst
material. It is clear that there are discrete ~2nm particles present as well as elongated lines, about
1.0ꢀ1.5 nm in width and up to 10nm in length, of Ruꢀcontaining material. These two morphologies
are more clearly visible at a higher magnification in Figures 8(c) and (d) respectively. The discrete
particles are seen to be semiꢀcrystalline, whereas the raftꢀlike structures are fairly disordered in
nature and may be indicative of partially converted Ru precursor material or hydrated RuOx. After
the first use of the catalyst, see Figure 9(a), the ~2nm particles remain, but the flat raft linear
17
ACS Paragon Plus Environment

ACS Catalysis
Page 18 of 34
1
2
3
4
5
6
7
8
structures appear to have broken up into a population of ~1nm particles. The mean size of this new
overall particle distribution is 1.4nm as illustrated in Figure 9(b). Higher magnification micrographs
of these particles (Figures 9(c) and (d)) show them to be poorly crystalline in character, suggesting
they are RuOx. After four usage cycles, the particle structure in this commercial catalyst shows
further modification as illustrated in Figure 10, with the mean particle size increasing to 2.2nm and
a considerably broader size distribution being evident (Figure 10(b)). High magnification HAADFꢀ
STEM imaging also indicated (Figure 10(d)) that these particles are now more crystalline in nature,
exhibiting interplanar spacings and angles that are more consistent with metallic Ru.
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
The Ru/C XC72RꢀSI catalyst series prepared by solꢀimmobilization display some marked
differences to the commercial Ru/C materials when examined by STEM. As shown in Figs. 11 (a),
(c) and (d) the unused Ru/C XC72RꢀSI catalyst material shows only a homogeneous dispersion of
crystalline Ruꢀcontaining nanoparticles from the outset. These have a mean size of 1.25nm and a
rather narrow size distribution (Figure 11(b)). After just one usage cycle there is some evidence of
limited particle sintering as the mean size increases slightly to ~1.5nm (see Figures 12(a) and (b)).
In addition, many of the larger particles seem to be crystalline Ru, whereas the smaller ones appear
to be more disordered and RuOxꢀlike in character (see Figures 12(c) and (d)). After the Ru/C
XC72R catalyst had been used three times (Figures 13(a) and (b)), the gradual increase of mean
particle size due to sintering was seen to continue, reaching a value of ~1.8nm and the crystallinity
of the particles was largely retained (Figures 13(c) and (d)).
18
ACS Paragon Plus Environment

Page 19 of 34
ACS Catalysis
1
2
3
4
a
b
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
c
d
Figure 8. Low magnification (a) BFꢀSTEM and (b) HAADFꢀSTEM images showing the typical
particle size and shape distribution in the unused commercial Ru/C catalyst: (c) and (d) are higher
magnification HAADF images showing greater structural detail in individual particles.
19
ACS Paragon Plus Environment

ACS Catalysis
Page 20 of 34
1
2
3
4
a
a
b
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
d
c
Figure 9. (a) Low magnification HAADFꢀSTEM image and (b) corresponding particle size
distribution from the commercial Ru/C catalyst used once: (c) and (d) are higher magnification
HAADF images showing greater structural detail in individual particles.
20
ACS Paragon Plus Environment

Page 21 of 34
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
c
d
Figure 10 (a) Low magnification HAADFꢀSTEM image and (b) corresponding particle size
distribution from the commercial Ru/C catalyst used four times: (c) and (d) are higher magnification
BF and HAADF images respectively showing greater structural detail in individual particles.
21
ACS Paragon Plus Environment

ACS Catalysis
Page 22 of 34
1
2
3
4
a
b
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
c
d
Figure 11. (a) Low magnification HAADFꢀSTEM image and (b) corresponding particle size
distribution from the unused sol immobilized Ru/C XC72R catalyst: (c) and (d) are higher
magnification HAADF images showing greater structural detail in individual particles.
22
ACS Paragon Plus Environment

Page 23 of 34
ACS Catalysis
1
2
3
4
a
b
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
c
d
Figure 12. (a) Low magnification HAADFꢀSTEM image and (b) corresponding particle size
distribution from the solꢀimmobilized Ru/C XC72RꢀSI catalyst used once: (c) and (d) are higher
magnification HAADF images showing greater structural detail in individual particles.
23
ACS Paragon Plus Environment

ACS Catalysis
Page 24 of 34
1
2
3
4
a
c
b
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
c
d
Figure 13. (a) Low magnification HAADFꢀSTEM image and (b) corresponding particle size
distribution from the solꢀimmobilized Ru/C XC72RꢀSI catalyst used three times: (c) and (d) are
higher magnification HAADF images showing greater structural detail in individual particles.
24
ACS Paragon Plus Environment

Page 25 of 34
ACS Catalysis
1
2
3
4
5
6
7
8
9
XPS analysis of the Ru/C catalyst systems is inherently difficult due to the overlap of the C(1s) and
Ru(3d) coreꢀlevels and the asymmetric nature of the Ru core level line shape; however, chemical
state information may still be obtained from observing a combination of both the Ru(3d) and the
lesser studied Ru(3p) coreꢀlevels. Figures 14 and 15 show the Ru(3p) and C(1s)/Ru(3d) coreꢀlevel
spectra respectively, for both the commercial and 5% Ru/C catalysts prepared by solꢀ
immobilisation using the XC72R carbon as support. For the used catalysts there is a clear shift in
the binding energies for both Ru(3p) and Ru(3d) towards lower energy. For the both commercial
and solꢀimmobilised catalysts, the Ru(3p) energies (463.0 and 463.5 eV respectively) are consistent
with RuO2, and the 0.5 eV difference between them we attribute largely to the different levels of
hydration and consistent with analysis of bulk hydrated and anhydrous ruthenium oxides;27,28 this
interpretation is supported by the XAFS data presented earlier.
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
Reduction of the supported Ru species with successive reꢀuse cycles is evidenced by the downward
shift in binding energy for both the 3p and 3d levels by ~1 eV which is in excellent agreement with
STEM results that show transformation to a more crystalline state. Interestingly for the commercial
Ru/C sample, this shift in binding energy is only pronounced after the 4th successive use, whereas it
is apparent after a single use for the solꢀimmobilized catalyst. Although such downward shifts in
energy can in principle be caused by an increase in mean particle size, the average diameter changes
revealed by STEM are too small for this magnitude of change. The final Ru(3p) energy of 462.2 eV
is consistent with the formation of metallic ruthenium or RuOx/Ru and further supported by the
Ru(3d) value of 280.3 eV29 which is again in agreement with the STEM results.
Although the ruthenium states undergo a similar reduction, there is a marked contrast in the C(1s)
and O(1s) signals between the commercial and solꢀimmobilised prepared Ru/C samples. For the 5
wt% Ru/C XC72RꢀSI samples, concomitant with the reduction of the Ru species on each successive
reꢀuse cycle, is the development of prominent new C(1s) (286.7 and 289.2 eV) and O(1s) (532.5
and 533.6 eV) features (see Figures 15(b) and 16(b)). These values are consistent with the
development of ꢀCꢀO containing species, such as ether or hydroxyl groups30 and –C=O containing
species, such as ketone, carboxylic, anhydride and lactone groups.30 Such binding energies would
be consistent with retention of lactic acid or the formation of a poly/cyclicꢀlactide at the surface.31
These lactide species could inhibit reactivity by covering Ru active sites and their absence in the
commercial catalyst could explain why this catalyst does not deactivate. However, the increase in
activity, with use, of the commercial catalyst cannot be explained by this phenomenon. In addition,
the reason that these species are not observed on the commercial catalyst must be considered. A
25
ACS Paragon Plus Environment

ACS Catalysis
Page 26 of 34
1
2
3
4
5
6
7
8
potential explanation is that the surface species observed in the solꢀimmobilised catalyst is
associated with a reaction with the PVA stabilising ligand of the 5 wt% Ru/C XC72RꢀSI. It is
known that some of the PVA will be removed from the surface of solꢀimmobilised catalysts on
treatment with water at elevated temperatures.32 However, only partial removal of the PVA is
achieved, but this is sufficient to permit access of reactants to the catalyst surface. It is possible that
the residual surface adsorbed PVA can react with LA to form a surface layer leading to
deactivation. This effect would be absent in the commercial catalyst. In addition the initial partial
removal of the PVA by dissolution in the water solvent used in LA hydrogenation would explain
the origin of the induction period observed with the solꢀimmobilised catalyst (both in fresh and the
used catalyst) since this would be expected to occur during an exposure of the catalyst to the water.
This data also supports our observations with respect to the time online activity data for fresh and
used catalysts presented in Figure 5.
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
A further O(1s) species is noted at ca. 530 eV, whose abundance decreases with increasing catalyst
use; typically peaks around this energy are attributed to metal oxides, with the decreasing intensity
supporting the reduction of ruthenium oxide to metallic Ru. Although this signal remains for both
commercial and solꢀimmobilised samples (Figure 16(a)ꢀ(b)), this would be in agreement with
RuOx/Ru as the samples are airꢀhandled prior to insertion in the spectrometer. The signal at 535.5
eV, which comprises approximately 5ꢀ8% of the total oxygen signal can be attributed to
chemisorbed water.33 In comparison, the commercial 5wt% Ru/C samples exhibit a similar oxygen
profile, but the total concentration of oxygen and carbon functionalities is greatly reduced, as
evidenced by the absence of a prominent signal ca. 289 eV in the C(1s)/Ru(3d) spectra (Figure
15(a)).
The preꢀreduced commercial 5wt% Ru/C catalyst (reduced in 5% H2/Ar mixture, at 300 °C, 3 h)
was also subjected to XPS analysis (before and after use). From the O(1s) and Ru(3d) spectra
reduction has a significant influence on the catalyst surface, with a marked change in the
concentration and the type of oxygen species present (supporting information Figure S8).
Moreover, reduction causes the Ru(3d5/2) signal to shift downward by 0.6 eV to 280.3 eV, with an
equal magnitude shift for the Ru(3p) region. The final binding energy is identical to that observed
for the successively used catalysts, however the activity is reduced and we largely attribute this to
the dehydration of RuOx, which is well known to reduce the surface area of the oxide and hence
have an adverse influence on the activity34,35 indeed other authors have noted a complete reduction
of RuO2 to Ru at temperatures of 477 oC 36 above the reduction temperature employed herein. This
dehydration is supported by the decrease in the width of the 3p peaks, indicating the loss of a higher
26
ACS Paragon Plus Environment

Page 27 of 34
ACS Catalysis
1
2
binding energy species and subtraction of scaled spectra confirms the presence of a peak at ca.464
3
4
eV which we attribute to Ru(OH)3 (Supporting information Figure S9).
5
6
7
8
9
Taken together, all the characterisation data on the series of sequentially reꢀused catalysts can be
used to provide underlying reasons for the observed differing trends in catalytic behaviour; namely
the observation of an induction period (with the fresh and used catalysts) and that the standard
commercial sample increases in activity whereas the catalyst prepared by solꢀimmobilisation
becomes steadily less active when compared at longer reaction times. In the case of the standard
commercial catalyst the Ru is steadily being reduced as the reaction progresses, with the particle
morphology changing from raftꢀlike structures to discrete particles and becomes progressively more
active. For this material the induction period is most likely to be due to the initial reduction coupled
with morphology changes. For the solꢀimmobilised catalysts the induction period is considered to
be related to partial removal of the PVA stabilising ligands coupled with initial reduction.
Following the induction period with fresh catalysts, the solꢀimmobilised catalyst is the more active
catalyst as it is in a more reduced state from the outset by virtue of the mode of catalyst preparation.
Slight sintering is observed, but the particle size remains below that of the commercial catalyst.
Upon use, the XPS shows that lactide species are present on the surface and this could explain the
observed partial deactivation as the reaction progresses.
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
27
ACS Paragon Plus Environment

ACS Catalysis
Page 28 of 34
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 14. Ru(3p) coreꢀlevel spectra for (a) the commercial 5wt% Ru/C sample, (b) sol
immobilised 5wt% Ru/C XC72RꢀSI sample after successive usage cycles.
28
ACS Paragon Plus Environment

Page 29 of 34
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 15. C(1s)/Ru(3d) coreꢀlevel spectra for (a) the commercial Ru/C sample, and (b) the sol
immobilised 5wt% Ru/C XC72RꢀSI sample. Insets show the Ru(3d5/2) fitting; only the Ru fitting is
showed for clarity.
29
ACS Paragon Plus Environment

ACS Catalysis
Page 30 of 34
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 16. O(1s) coreꢀlevel spectra for (a) the commercial 5wt% Ru/C sample, and (b) the sol
immobilised 5wt% Ru/C XC72RꢀSI sample.
30
ACS Paragon Plus Environment