Oxidation of glycerol using supported Pt, Pd and Au catalysts

Article abstract of DOI:10.1039/b212047j

The oxidation of aqueous solutions of glycerol is described and discussed for Pd, Pt and Au nanoparticles supported on graphite and activated carbon. The oxidation in a batch reactor at 60°C and 1 bar pressure using air as oxidant was initially investigated. Under these conditions, supported Pd and Pt catalysts give some selectivity to glyceric acid, but the main reaction products are considered to be non-desired C₁ by-products, e.g. CO₂, HCHO and HCOOH. In addition, under these conditions, supported Au catalysts were totally inactive. Using an autoclave with pure oxygen at 3 bar pressure gave a significant improvement in reactivity and, for Pt and Au catalysts, the formation of C₁ by-products was eliminated when NaOH was added. In particular, it was noted that, in the absence of NaOH, the Au/C catalyst was inactive. For 1 wt.% Au/graphite or activated carbon, 100% selectivity to glyceric acid at high conversion was readily achieved. The role of the base is discussed and it is proposed that the base aids the initial dehydrogenation via H-abstraction of one of the primary OH groups of glycerol and, in this way, the rate limiting step in the oxidation process is overcome.

Full text of DOI:10.1039/b212047j

View Article Online / Journal Homepage / Table of Contents for this issue
Oxidation of glycerol using supported Pt, Pd and Au catalysts
Silvio Carrettin,a Paul McMorn,a Peter Johnston,b Ken Griffin,b Christopher J. Kielyc and
Graham J. Hutchings*a
a
Department of Chemistry, Cardiff University, P.O. Box 912, Cardiff, UK CF10 3TB.
E-mail: hutch@cardiff.ac.uk
Johnson Matthey, Orchard Road, Royston, Herts, UK SG8 5HE
Department of Materials Science and Engineering, Lehigh University, 3 East Packer Avenue,
Bethlehem, PA 18015-3195, USA
b
c
Received 4th December 2002, Accepted 24th January 2003
First published as an Advance Article on the web 10th February 2003
The oxidation of aqueous solutions of glycerol is described and discussed for Pd, Pt and Au nanoparticles
supported on graphite and activated carbon. The oxidation in a batch reactor at 60 ꢀC and 1 bar pressure using
air as oxidant was initially investigated. Under these conditions, supported Pd and Pt catalysts give some
selectivity to glyceric acid, but the main reaction products are considered to be non-desired C1 by-products, e.g.
CO2 , HCHO and HCOOH. In addition, under these conditions, supported Au catalysts were totally inactive.
Using an autoclave with pure oxygen at 3 bar pressure gave a significant improvement in reactivity and, for Pt
and Au catalysts, the formation of C1 by-products was eliminated when NaOH was added. In particular, it was
noted that, in the absence of NaOH, the Au/C catalyst was inactive. For 1 wt.% Au/graphite or activated
carbon, 100% selectivity to glyceric acid at high conversion was readily achieved. The role of the base is
discussed and it is proposed that the base aids the initial dehydrogenation via H-abstraction of one of the
primary OH groups of glycerol and, in this way, the rate limiting step in the oxidation process is overcome.
Introduction
Oxidation reactions are important for the synthesis of fine che-
micals. However, at present many oxidations that are carried
out commercially involve stoichiometric oxidants, e.g. per-
manganate or chromic acid, and these entail the formation
of significant quantities of non-desired by-products. There is,
therefore, immense scope for the replacement of stoichiometric
oxidants with catalytic processes using dioxygen, and the
environmental benefits for this have been outlined by Sheldon
and Dakka.1 In recent years, this has spurred the investigation
of a range of oxidation reactions using gas or liquid phase
reagents to produce valuable intermediates or products for
the chemicals industry.2–13 In particular, the oxidation of
alcohols and polyols to chemical intermediates represents
a demanding target.12 Supported platinum and palladium nano-
particles are generally acknowledged as effective catalysts for
the oxidation of polyols, for example in carbohydrate chemis-
try for the oxidation of glucose to glucinic acid.9,11 Glycerol is
a highly functionalised molecule that is readily available from
biosustainable sources, for example it can be obtained as a by-
product of the utilisation of rape seed and sunflower crops.
This makes glucose a particularly attractive starting point for
the synthesis of intermediates, and a large number of products
can be obtained from glycerol oxidation (Scheme 1). One of
the key problems is the potential complexity of the products
that can be formed and so control of the reaction selectivity
by careful design of the catalyst is required. Glycerol oxida-
tion, in aqueous solution, has been extensively studied using
supported platinum and palladium catalysts by Gallezot and
co-workers14–16 and Kimura and co-workers.17,18 In general,
palladium catalysts were found to be more selective than
platinum and high selectivities to either dihydroxyacetone18
or glyceraldehyde14 could be obtained by controlling the
Scheme 1
DOI: 10.1039/b212047j
Phys. Chem. Chem. Phys., 2003, 5, 1329–1336
1329
This journal is # The Owner Societies 2003

View Article Online
constant by the addition of aqueous NaOH (30 wt.%) using
reaction conditions and, in particular, the pH of the solution.
However, it has been noted that formic acid, probably result-
ing from the formation of oxalic acid, was also formed. It is
worth noting that, in all these previous studies, mixtures of
most of the potential products were formed.
In recent years, there has been immense interest in the use of
gold catalysts for oxidation reactions.19 Prati and co-work-
ers20–23 have shown that supported gold nanoparticles can be
very effective catalysts for the oxidation of alcohols, including
diols. Most recently, in our earlier communication,24 we have
shown that supported gold catalysts can be effective for gly-
cerol oxidation and, in particular, glycerol can be oxidised to
glyceric acid with 100% selectivity to glyceric acid. In this
paper, we extend this earlier study and compare and contrast
non-promoted supported palladium, platinum and gold cata-
lysts for the selective oxidation of glycerol.
a Denver Instruments Titrator 290. In a typical experiment,
the catalyst (glycerol/metal mol ratio ¼ 500) was suspended
in water (10 ml) and the slurry was heated to the required tem-
perature with stirring. Glycerol (4.605 g, 99%, Lancaster; in
40 ml water) was then added to obtain a final glycerol solution
(1 mol 1ꢂ1 glycerol). After 10 min, air was passed through the
reactants at a controlled flow rate. Samples were removed
from the reaction mixture at regular intervals for analysis.
Autoclave reactor studies. These reactions were carried out
using a 50 ml Parr autoclave. The catalyst was suspended in
an aqueous solution of glycerol (0.6 mol lꢂ1, 20 ml). The auto-
clave was pressurised to the required pressure with oxygen and
heated to 60 ꢀC. The reaction mixture was stirred (1500 rpm)
for 3 h, following which the reaction mixture was analysed.
Product analysis. Analysis was carried out using HPLC with
ultraviolet and reflective index detectors. Reactant and pro-
ducts were separated on an ion exclusion column (Alltech
QA-1000) heated at 70 ꢀC. The eluent was a solution of
H2SO4 (4 ꢃ 10ꢂ4 mol lꢂ1). Reaction mixture samples (10 ml)
were diluted with a solution of an internal standard (100 ml,
0.2 mol lꢂ1 isobutanol) and 20 ml of this solution was analysed.
It is essential that a standard is added so that the carbon mass
balance can be determined. Attempts to find a suitable internal
standard that could be added prior to reaction were found to
be unsuccessful as they were readily oxidised under the reac-
tion conditions. Hence, it was found necessary to add the
standard immediately following the reaction as described.
Experimental
Catalyst preparation
5 wt.% platinum catalysts supported on activated carbon and
graphite were prepared as follows. An aqueous solution of a
platinum chloride salt (5% Pt metal by weight of support)
was added to the stirred carbon slurry and reduced with for-
maldehyde. The slurry was allowed to settle and was filtered,
washed free of chloride. The catalyst was dried at 105 ꢀC for
16 h.
5 wt.% palladium catalysts supported on activated carbon
and graphite were prepared using method Pd-A. An aqueous
solution of a palladium chloride salt (5% Pd metal by weight
of support) was added to a stirred carbon slurry and reduced
with formaldehyde. The slurry was allowed to settle, filtered
and washed free of chloride. The catalyst was dried at 105 ꢀC
for 16 h.
A sample of 5 wt.% Pd supported on activated carbon was
also prepared by the alternative method Pd-B according to
Garcia et al.14 A solution of PdCl2ꢁ2HCl (23.68% Pd, Johnson
Matthey, 21.11 g) in demineralised water (100 ml) was added
to activated carbon (99.7 g, Johnson Matthey) and the suspen-
sion was stirred for 5 h at 25 ꢀC. After cooling to 0 ꢀC, the
slurry was reduced with formaldehyde and, after stirring for
15 min, three successive additions of aqueous KOH (15 ml,
30%) were made and the slurry was stirred at 25 ꢀC for a
further 16 h. The catalyst was recovered by filtration and
washed with water until the washings were pH neutral. The
catalyst was dried at 105 ꢀC for 16 h.
1 wt.% gold catalysts supported on activated carbon and
graphite were prepared as follows. The carbon support (gra-
phite or activated carbon, Johnson Matthey, 113.2 g) was stir-
red in demineralised water (1 l) for 15 min. An aqueous
solution of HAuCl4 (41.94% Au, Johnson Matthey, 2.38 g)
in water (70 ml) was slowly added dropwise over a period of
30 min. The slurry was then refluxed for 30 min, cooled and
reduced with formaldehyde over a period of 30 min. The slurry
was then refluxed for 30 min and, following cooling, the cata-
lyst was recovered by filtration and washed with water until the
washings contained no chloride. The catalyst was dried for 16
h at 105 ꢀC. This method was also used to prepare 0.25 wt.%
Au/C and 0.5 wt.% Au/C catalyst samples using smaller
amounts of chloroauric acid.
Catalyst characterisation. Samples were structurally charac-
terised in a JEOL 2000 EX high resolution electron microscope
operating at 200 kV. The catalyst powders were made suitable
for TEM examination by grinding them in high purity ethanol
using an agate pestle and mortar. A drop of the suspension was
then deposited onto, and allowed to evaporate, on a holey
carbon grid.
Results and discussion
Oxidation of glycerol using the glass reactor at 1 bar
Gallezot and co-workers14–16 have reported that Pd/C cata-
lysts can oxidise glycerol to glyceric acid with high selectivity
at 1 bar and, hence, glycerol oxidation was investigated using
the 5 wt.% Pd/C catalysts at 60 ꢀC with air at 1 bar using the
glass slurry reactor. The results are shown in Table 1. At a low
air flow rate (1 ml minꢂ1) no reaction was observed with 5
wt.% Pd/carbon prepared by method Pd-A for the standard
reaction conditions (glycerol/Pd mol ratio ¼ 500). Increasing
the catalyst mass did, however, result in low levels of glycerol
conversion, but the mass balance to C3 and C2 products was
extremely low (ca. 7%). At higher air flow rates with the 5
wt.% Pd/carbon catalyst, the conversion of glycerol increased,
as did the mass balance. However, the carbon mass balance for
C3 and C2 products remained low (<40%). Interestingly, the
normalised C3 and C2 product selectivities, i.e. the selectivity
calculated assuming these are the only reaction products, give
very high selectivities to glyceric acid ( > 70%) and these are
very similar to the selectivities reported in the literature14–16
for non-promoted catalysts. Similar results were obtained for
the 5 wt.% Pd/graphite catalyst prepared by method Pd-A.
In view of these results, a catalyst was prepared by method
Pd-B according to the procedure of Garcia et al.14 and the
results are also shown in Table 1. Again, at low air flow rates,
no reaction is observed. Increasing the air flow rate leads to
increased glycerol conversion, and the normalised selectivity
to glyceric acid is similar to that observed for the catalyst
Catalyst testing and characterisation
Glass reactor studies. Experiments at 1 bar pressure were
carried out using two glass reactors (300 and 50 ml volume)
equipped with an overhead stirrer (ca. 500 rpm). Air was
passed into the reactor at a controlled flow rate via a calibrated
mass flow controller, through a glass frit inserted into the
stirred reactants. The pH of the reactants was maintained
1330
Phys. Chem. Chem. Phys., 2003, 5, 1329–1336

View Article Online
Table 1 Oxidation of glycerol using 5 wt.% Pd/C catalystsa
Product selectivityc (%)
Glyceric
Catalyst
mass/g
Airb /
ml minꢂ1
Glycerol
conversion (%)
Tartronic
acid
Oxalic
acid
Mass
balanced (%)
Catalyst
acid
Glyceraldehyde
Pd/carbone
Pd/carbone
Pd/carbone
Pd/carbone
Pd/carbone
Pd/carbone
Pd/carbonf
Pd/carbonf
Pd/carbonf
Pd/graphiteg
0.320
0.320
0.320
0.320
1.230
2.430
0.320
0.320
0.320
0.320
1
10
30
100
1
0
–
–
–
–
–
17.9
48.3
57.0
9.0
6.6 (85.7)
–
–
1.1 (14.3)
2.4 (6.9)
3.2 (18.6)
1.9 (27.8)
2.0 (26.7)
–
7.5
34.7
22.4
7.1
7.5
–
32.3 (93.1)
18.0 (76.3)
2.8 (38.9)
3.5 (46.6)
–
–
–
1.2 (5.1)
2.4 (33.3)
2.0 (26.7)
–
–
–
1
19.0
0
–
1
–
30
100
100
22.5
73.5
54.0h
39.0 (94.2)
62.0 (75.6)
4.0 (67)
1.3 (3.1)
2.0 (4.9)
1 (16.5)
1.1 (2.7)
16.0 (19.5)
–
–
41.4
80.0
–
–
1 (16.5)
a
b
c
Reaction conditions: 60 ꢀC, 1 bar, 21 h, pH 11, aqueous glycerol (50 ml, 1 mol lꢂ1). Air flow rate through reaction. Product selectivity on
basis of using internal standard. Figures in parenthesis are normalised for the observed C2 and C3 products. d Determined on the basis of observed
C2 and C3 products. e 5 wt.% Pd/activated carbon prepared by method Pd-A. f 5 wt.% Pd/activated carbon prepared according to the method of
Garcia et al.14 g 5 wt.% Pd/graphite prepared using same procedure as (e). Conversion after 8 h reaction.
h
prepared by method Pd-A. At the highest air flow rate inves-
tigated (100 ml), the C3 and C2 carbon mass balance was
80% and the normalised product selectivities were similar to
those reported by Garcia et al.14. However, all these studies
show that there is appreciable formation of compounds other
than the expected C3 and C2 products. This will be discussed
subsequently.
formaldehyde) or, secondly, the catalyst leaches Pd into
solution and this subsequently acts on an active homoge-
neous oxidation catalyst. Following the reactions, the solu-
tion was analysed and no Pd was found to have been
leached, hence the first explanation appears to be preferred.
Following catalyst preparation, the catalyst was reduced at
room temperature and it is possible that the catalyst reduc-
tion was not complete under such conditions. Glycerol can
act as a reducing agent and, during the initial reaction
period, the reduction process could be completed leading
to the observation of an induction period. However, it is
unlikely that glycerol is a stronger reducing agent towards
Pd2+ compared with formaldehyde, hence it is more likely
that the initial induction period noted with this preparation
is related to the removal of inhibitors.
The 5 wt.% Pd/carbon catalysts derived by method Pd-A
and Pd-B (Garcia et al. method) were compared by transmis-
sion electron microscopy and typical micrographs are pre-
sented in Fig. 2. The Pd-A derived catalyst, shown in
Fig. 2(a) and at higher resolution in Fig. 2(b), comprises a very
fine dispersion of metallic Pd particles which exhibit a distinct
bimodal size distribution. The majority of Pd particles were in
the 2–5 nm site range (mean 4 nm), whereas larger particles in
the 8–12 nm size range were also occasionally present. The
Pd-B derived catalyst (Fig. 2(c)) also exhibited a bimodal size
distribution of Pd, with most particles being in the 3–7 nm
range (mean 5 nm) along with a significant number of larger
ones in the 10–20 nm range.
There are significant differences that can be noted in the
catalytic performance of the 5 wt.% Pd/carbon catalysts pre-
pared by methods Pd-A and Pd-B. First, for the Pd-A
derived catalyst, the major products comprise glyceric acid
with glyceraldehyde and oxalic acid whereas, for the Pd-B
derived catalyst, the products comprise glyceric acid with
glyceraldehyde and tartronic acid. It is possible that tartro-
nic acid is oxidised to oxalic acid, and the difference in activity
of the two catalysts may account for this. However, more
importantly, the two catalysts show different conversion
with time-on-line profiles for glycerol oxidation (Fig. 1).
The catalyst prepared by method Pd-A exhibits the expected
dependence of conversion with reaction time. The catalyst
prepared by method Pd-B, however, shows a marked induc-
tion period, during which no glycerol conversion is observed.
This behaviour was noted for all the reaction conditions for
which this catalyst was investigated. There are two potential
reasons for this observation. First, the catalyst surface is
initially incompletely reduced or is covered by an inhibitor
or poison (e.g. chloride or high molecular weight products
originating from reactions of the initial reducing agent,
The effect of in situ reduction with hydrogen (99%, 100 ml
minꢂ1, 60 ꢀC, 30 min) of the Pd/C catalyst prepared method
Pd-A was investigated and the results are shown in Fig. 3.
For the 5 wt.% Pd/graphite catalyst, the in situ reduction step
did not affect the glycerol conversion notably (Fig. 3(a))
whereas, for the 5 wt.% Pd/carbon catalyst, the in situ reduc-
tion did increase the conversion significantly. In both cases, the
selectivity to glyceric acid was improved (Fig. 3(b)) and, during
the initial stages of the reaction, very high mass balances to C3
and C2 products are observed (Fig. 3(c)). It is considered that
the in situ hydrogen reduction treatment removes traces of
deleterious surface PdO species that form readily after catalyst
preparation.12
5 wt.% Pt/carbon was investigated for glycerol oxidation
under the same conditions and the results are given in Table
2 and Fig. 4. The Pt catalyst was found to be more active than
the Pd catalysts and significant glycerol conversion was
observed even for the lowest air flow rate. Interestingly, the
only C3 or C2 product observed is glyceric acid, together with
minute traces of glyceraldehyde, but the C3 and C2 carbon
Fig. 1
Effect of reaction time on glycerol conversion using 5 wt.%
Pd/C catalysts. Key : ˘ Pd/C prepared method Pd-A, L Pd/C
prepared using method Pd-B of Garcia et al.14 Reaction conditions :
60 ꢀC, 1 bar, pH 11.
Phys. Chem. Chem. Phys., 2003, 5, 1329–1336
1331

View Article Online
Fig. 3 Effect of in situ reduction on 5 wt.% Pd/C catalysts. Key : K
Pd/carbon, X Pd/graphite; closed symbols no in situ reduction; open
symbols with in situ reduction. Reaction conditions : 60 ꢀC, 1 bar,
pH 11. (a) Conversion, (b) glyceric acid selectivity, (c) C2 and C3
carbon mass balance.
range was 1.5–3 nm with most particles being sub-2 nm in
diameter.
The low mass balance for the formation of C2 and C3 pro-
ducts using 5 wt.% Pd/C and 5 wt.% Pt/C catalysts could
be due to three causes : (a) the polymerisation of reaction
products to polymers not observable by the hplc analytical
method, (b) the adsorption of reaction products onto the
Fig. 2 (a) Bright field image and (b) high resolution lattice image of
the 5 wt.% Pd/carbon catalyst derived from method Pd-A; (c) shows a
bright field image of the 5 wt.% Pd/carbon catalyst derived from
method Pd-B.
Table 2 Glycerol oxidation over 5 wt.% Pt/C catalysta
Product selectivityb (%)
Glycerol
conversion
(%)
Mass
balancec
(%)
Air/
ml minꢂ1
Glyceric
acid
mass balance was very low for all conditions examined
(<50%). The effect of reduction was also investigated for the
Pt/carbon catalyst (Fig. 4). During the initial stages of the
reaction, the use of in situ or ex situ reduction does not have
a significant effect, although the eventual glycerol conversion
observed with the in situ reduced catalyst was significantly
higher. Experiments with 5 wt.% Pt/graphite for the oxidation
of glycercol gave similar results.
Representative electron micrographs of the 5 wt.% Pt/car-
bon sample are shown in Fig. 5. This catalyst shows a highly
dispersed and very uniform distribution of Pt nanoparticles
over the activated carbon support. The measured particle size
Glyceraldehyde
1
10
29.6
47.0
60.0
89.0
6.7 (100)
17.4 (100)
47.5 (100)
41.0 (100)
tr
tr
0
6.7
17.4
47.5
41.0
30
100
tr
a
5 wt.% Pt/activated carbon; reaction conditions: 60 ꢀC, 21 h, 1 bar,
pH 11, 0.320 g catalyst, (50 ml, 1 mol lꢂ1). b Product selectivity deter-
mined on basis of analysis using internal standard Figures in paren-
c
thesis are normalised for the observed C3 products. Determined on
the basis of C2 and C3 products observed.
1332
Phys. Chem. Chem. Phys., 2003, 5, 1329–1336

View Article Online
Fig. 5 (a) Bright field image and (b) high resolution lattice image of
the 5 wt.% Pt/carbon catalyst.
dihydroxyacetone is the most stable. These results are in agree-
ment with a previous study of 1,3-butanediol oxidation by Lin
and Sun25 in which CO2 was observed as a reaction product,
however, this study involves the electrocatalytic oxidation in
acidic medium and, consequently, operates via a different
mechanism than that in basic medium, the present reaction
conditions. The formation of CO2 as a by-product of alcohol
oxidation using Pt and Pd catalysts has been noted by Prati
and Rossi.26 In view of this, experiments were conducted to
investigate the formation of CO2 , and it was confirmed that
CO2 constitutes a major factor of the carbon balance for these
oxidation reactions. In addition, it is considered that carbon
monoxide and formic acid may also be formed. The determina-
tion of a closed mass balance is complicated by the solubility
of CO2 and other C1 products, e.g. formaldehyde and formic
acid, in the reaction medium since sodium carbonate and other
soluble products will be formed.
Fig. 4 Glycerol oxidation over Pt/carbon catalysts showing effect of
reduction conditions. Key : L unreduced; ˘ reduced ex situ; S reduced
in situ. Reaction conditions, 60 ꢀC, 1 bar, pH 11. (a) Conversion, (b)
glyceric acid selectivity, (c) C2 and C3 carbon mass balance.
carbon supports or (c) the sequential oxidation of the C3 and
C2 oxidation products to CO2 and formic acid.
To examine the possible formation of polymers, reactions
were carried out using mixtures of dihydroxyacetone, glyceric
acid and glyceraldehyde with the 5 wt.% Pd/carbon catalyst.
The reaction mixtures were analysed for polymeric products
using hplc and no appreciable formation of oligomers and
polymers was observed.
To determine if reaction products were being adsorbed onto
the carbon support during the reaction, the catalysts were
investigated after reaction using thermal gravimetric analysis
(TGA). Only the desorption of water, the reaction solvent,
was observed and, hence, it is concluded that the adsorption
of the partial oxidation products is not occurring on the
catalyst.
A series of experiments were carried out using the glass
slurry reaction with Au/C catalysts. No conversion was
observed under the reaction conditions at the low oxygen
partial pressure. It was, therefore, decided to investigate the
oxidation of glycerol at a higher oxygen pressure using an
autoclave reactor.
It was, therefore, considered that the low mass balance was
most likely to be due to the sequential oxidation of the C3
partial oxidation products. To examine this possibility, the
stability of the reaction products was investigated under the
standard reaction conditions, and the results are shown in
Table 3. It is clear that glyceric acid and glyceraldehyde
are particularly unstable under these conditions. Hydroxy
pyruvic acid and oxalic acid are much more stable, and
Oxidation of glycerol using an autoclave reactor
The oxidation of glycerol with dioxygen was investigated using
an autoclave reactor at elevated pressure with both supported
Pt and Pd catalysts, and the results are shown in Table 4. As
expected under these conditions, the carbon and graphite sup-
ports were inactive. For the 5 wt.% Pt/carbon catalyst, in the
Phys. Chem. Chem. Phys., 2003, 5, 1329–1336
1333

View Article Online
Table 3 Product stability over 5 wt.% Pt/C catalystsa
Productsb (%)
Substrate
Reaction time/h
Conversion (%)
Hydroxy pyruvic acid
Oxalic acid
Glyceric acid
Glyceraldehyde
Dihydroxyacetonec
Glyceric acidd
Glyceraldehydee
Hydroxy pyruvic acidf
Oxalic acidg
21
4
19.4
100
100
100
90
13.7
tr
2.1
tr
6.7
–
2.8
0
21
21
7
0
0
0
–
–
91.0
–
0
0
0
0
0
a
5 wt.% Pt/activated carbon; reaction conditions: 60 ꢀC, pH 11, 1 bar, air 100 ml minꢂ1, water as solvent. b Observed products by hplc, CO2 and
HCOOH not quantified. Calcium salt, 0.05 mol lꢂ1
.
1 mol 1ꢂ1
.
0.1 mol lꢂ1
.
0.2 mol lꢂ1
.
0.5 mol lꢂ1
.
c
d
e
f
g
Table 4 Oxidation of glycerol using Pt/C catalysts in an autoclave reactora
Product selectivityb (%)
Glycerol
Mass
balancec
(%)
NaOH/glycerol Glycerol/metal conversion Glyceric
mol ratio
Oxalic
Glyceraldehyde acid
Tartronic
acid
Catalyst
mol ratio
(%)
acid
None
None
1
–
0
–
–
–
–
–
–
–
–
1
–
0
–
–
–
–
Carbon
Graphite
1
–
0
–
–
–
–
1
–
0
–
–
–
–
5 wt.% Pt/carbon
5 wt.% Pt/carbon
5 wt.% Pt/carbon
5 wt.% Pt/carbon
5 wt.% Pt/carbon
10 wt.% Pt/carbon
5 wt.% Pt/graphite
5 wt.% Pd/carbon
0
500
500
500
500
100
500
500
500
500
27.0
48.0
63.1
56.0
88.0
53.0
58.6
57.2
53.0
27.8 (32.6)
45.7 (85.1)
0 (0)
6.0 (11.2)
57.4 (17.4)
2.0 (3.7)
0 (0)
0 (0)
0 (0)
0 (0)
0 (0)
0 (0)
0 (0)
0 (0)
0 (0)
3.1 (3.2)
0 (0)
23 (29.7)
0 (0)
1.6 (1.9)
85.2
53.7
98.0
99.0
77.5
54.0
94.2
0.5
1.0
2.0
1.0
1.0
1.0
1.0
73.7 (75.2) 21.0 (21.4)
69.0 (69.7) 30.0 (30.3)
54.5 (70.3)
50.0 (96.6)
65.7 (69.7) 26.9 (28.6)
0 (0)
4.0 (7.4)
67.0 (78.3)
50.0 (77.1)
6.4 (7.5)
7.5 (11.4)
12.2 (14.2) 85.6
7.5 (11.4) 66.0
5 wt.% Pd/graphite 1.0
a
Reaction conditions: 60 ꢀC, 3 h, water (20 ml), catalyst (50 mg), pO 3 bar, stirring speed 1500 rpm. b Product selectivity determined on basis of
2
analysis using internal standard. Figures in parenthesis are normalised for the observed C2 and C3 products. c Determined on basis of observed C2
and C3 products.
absence of sodium hydroxide, the main product was oxalic
acid formed by oxidative cleavage, although an appreciable
selectivity to glyceric acid was also observed. The presence of
sodium hydroxide markedly decreased the selectivity to oxalic
acid. The principal products were now glyceric acid and glycer-
aldehyde. In particular, the C2 and C3 carbon mass balance
was now typically 98–99% indicating that, under these condi-
tions, the oxidative cleavage to C1 products (i.e. CO2 ,
H2CO, HCOOH) had essentially been eliminated. Interest-
ingly, increasing the catalyst concentration, so that the
glycerol/metal ratio was 100, instead of 500, decreased the
C2 and C3 carbon mass balance, as did increasing the metal
loading of the Pt to 10 wt.% Pt/carbon. This indicates that
the catalyst metal loading and concentration require careful
optimisation for this complex oxidation reaction to enable
the formation of non-desired C1 by-products to be avoided.
A range of alternative bases was also investigated and the
results are given in Table 5. It is apparent that the highest gly-
cerol conversion and highest selectivity to glyceric acid is
observed with NaOH. Interestingly, the use of NaNO3 also
leads to a high C2 and C3 carbon mass balance, but the selec-
tivity is almost exclusively to oxalic acid. However, with
NaNO3 , there is no significant effect on glycerol conversion
when compared with the reaction in the absence of base. This
Table 5 Oxidation of glycerol using 5 wt.% Pt/C catalysts in the presence of different basesa
Product selectivityb (%)
Base
Glycerol conversion
Glyceric acid
Glyceraldehyde
Oxalic acid
Tartronic acid
Mass balancec (%)
None
27.0
48.9
63.1
36.0
35.4
37.0
52.0
27.8 (22.6)
52.0 (74.2)
73.7 (75.4)
14.0 (14.6)
87.0 (88.8)
13.7 (75.3)
18.0 (100)
0 (0)
57.4 (67.4)
0 (0)
0 (0)
0 (0)
0 (0)
85.2
70.0
97.8
99.0
98.0
18.2
18.0
LiOH
NaOH
NaNO3
KOH
19.0 (25.8)
21.0 (21.4)
0 (0)
3.1 (3.2)
0 (0)
0 (0)
85.0 (85.4)
0 (0)
0 (0)
11.0 (11.2)
4.5 (24.7)
0 (0)
RbOH
CsOH
0 (0)
0 (0)
0 (0)
a
5 wt.% Pt/activated carbon; reaction conditions: 3 h, 60 ꢀC, pO 3 bar, water (20 ml), glycerol/metal mol ratio ¼ 500, base/glycerol mol
2
b
ratio ¼ 1./0, catalyst (50 mg), stirring speed 1500 rpm. Product selectivity determined on basis of analysis using internal standard. Figures in
c
parenthesis are normalised on the basis of observed C2 and C3 products. Determined on basis of observed C2 and C3 products.
1334
Phys. Chem. Chem. Phys., 2003, 5, 1329–1336

View Article Online
Table 6 Oxidation of glycerol using 1 wt.% Au/C catalystsa
Selectivity (%)
Glyceric
Glycerol/
metal
Glycerol/mmol pO /bar mol ratio NaOH/mmol conversion (%) acid
Glycerol
Tartronic
Catalyst
Glyceraldehyde acid
2
1 wt.% Au/activated carbon 12
3
3
6
6
3
3
3
3
6
6
6
538b
538b
538b
538b
540c
540c
214d
214d
214d
214d
214d
12
12
12
24
12
6
56
54
72
58
56
43
59
69
58
91
0
100
100
86
0
0
2
0
0
0
0
0
0
0
0
0
1 wt.% Au/graphite
1 wt.% Au/graphite
1 wt.% Au/graphite
1 wt.% Au/graphite
1 wt.% Au/graphite
1 wt.% Au/graphite
1 wt.% Au/graphite
1 wt.% Au/graphite
1 wt.% Au/graphite
1 wt.% Au/graphite
12
12
12
6
12
3
97
93
7
6
80
63
82
20
12
18
33
6
6
6
6
12
6
6
67
92
6
12
0
6
a
b
c
d
60 ꢀC, 3 h, H2O (and 20 ml), stirring speed 1500 rpm. 220 mg catalyst. 217 mg catalyst. 450 mg catalyst.
indicates that the addition of OHꢂ is important for the activa-
tion of glycerol and this will be commented upon subsequently.
The results also indicate that Na+ prevents the subsequent oxi-
dation of oxalic acid to carbon oxides. The other group 1 metal
cations are not as effective as Na+, although KOH did pro-
mote the formation of glyceric acid with very high selectivity
(87%).
and Au supported catalysts were characterised using TEM.
Representative electron micrographs of the 1 wt.% Au/
graphite and 1 wt.% Au/carbon materials are shown in
Figs. 6(a) and 6(b) respectively. In the 1 wt.% Au/graphite spe-
cimen, Au particles as small as 5 nm and as large as 50 nm in
diameter were detected. The majority, however, were about 25
nm in size and were multiply twinned in character. Decreasing
the loading to 0.5 wt.% or 0.25 wt.% did not appreciably
change the particle size distribution; the particle number den-
sity per unit area was observed to decrease proportionately
however, which may be correlated to the decrease in glycerol
Supported Pd catalysts (Table 4) also give high selectivities
to glyceric acid, but the selectivities and C2 and C3 carbon
mass balance was much lower than with the corresponding
supported Pt catalyst. Interestingly, under the same condi-
tions, the 5 wt.% Pt and Pd catalysts supported on either acti-
vated carbon or graphite give very similar glycerol conversion.
The oxidation of glycerol with dioxygen was also investi-
gated using 1 wt.% Au/C catalysts in an autoclave and the
results are given in Table 6. NaOH was added as a base since,
in the absence of NaOH, no glycerol conversion was observed.
In addition, as noted earlier, the carbon supports were also
inactive for glycerol oxidation under these conditions, even
when NaOH was present. For all the data presented in Table
6, the carbon mass balance was 100% indicating that, under
these conditions, supported Au/C catalysts are extremely
selective for this reaction and no C1 or C2 by-products were
detected. In addition, it is apparent that the selectivity to gly-
ceric acid and the glycerol conversion are very dependent upon
the glycerol/NaOH ratio. In general, with high concentrations
of NaOH, exceptionally high selectivities to glyceric acid
can be observed. However, decreasing the concentration of
glycerol, and increasing the mass of the catalyst and the con-
centration of oxygen, leads to the formation of tartronic acid
via consecutive oxidation of glyceric acid. Interestingly, this
product is stable with these catalysts. It is apparent that, with
careful control of the reaction conditions, 100% selectivity to
glyceric acid can be obtained with 1 wt.% Au/C. For compar-
ison, the supported Pd/C and Pt/C always gave other C3 and
C2 products in addition to glyceric acid and, in particular, also
gave some C1 by-products. In a final set of experiments, cata-
lysts with lower Au concentrations were investigated. For cat-
alysts containing 0.25 or 0.5 wt.% Au supported on graphite,
lower glycerol conversions were observed (18% and 26%
respectively as compared to 54% for 1 wt.% Au/graphite under
the same conditions) and lower selectivities to glyceric acid
were also observed. The previous studies for diol oxidation
by Prati and co-workers20–23 have also shown that the conver-
sion is dependent on the Au loading upon the support. This is
possibly due to a particle size effect of the Au nanoparticles on
the support. For gold as a CO oxidation catalyst, it has been
shown that the activity is highly dependent on the particle size,
and the optimum size is ca. 2–4 nm.19 In view of this, the Pd, Pt
Fig. 6 Bright field transmission electron micrographs of (a) 1 wt.%
Au/graphite and (b) 1 wt.% Au/carbon.
Phys. Chem. Chem. Phys., 2003, 5, 1329–1336
1335

View Article Online
H. E. J. Hendriks, B. F. M. Kuster and G. B. Marin, Carbohydr.
Res., 1990, 204, 121.
T. Mallat and A. Baiker, Catal. Today, 1994, 19, 247.
F. R. Venema, J. A. Peters and H. van Bekkum, J. Mol. Catal.,
1992, 77, 75.
A. Abbadi and H. van Bekkum, Appl. Catal. A, 1995, 124, 407.
A. Abbadi and H. van Bekkum, J. Mol. Catal. A, 1995, 97, 111.
P. Gallezot, S. Tretjak, Y. Christidis, G. Mattioda and A.
Schouteeten, J. Catal., 1993, 142, 729.
3
conversion and selectivity to glyceric acid. The 1 wt.% Au/car-
bon specimen (Fig. 6(b)) showed a considerably narrower par-
ticle size distribution (15–30 nm) with the 15 nm size range
being predominant. It is interesting to note that, for both types
of carbon supports, the Au particles were considerably larger
than the 2–4 nm size that is considered to be optimal for low
temperature CO oxidation, yet they were still highly active
for the glycerol to glyceric acid reaction.
4
5
6
7
8
The role of OHꢂ requires further comment. It is noted that,
for supported Pd and Pt catalysts, the addition of OHꢂ, as
NaOH, increases the conversion of glycerol. However, for
Au/C catalysts, the presence of OHꢂ is essential to observe
any glycerol oxidation. It is proposed that, in the absence of
a base, the initial dehydrogenation of glycerol, via H-abstrac-
tion, which must be the first step in the oxidation process, is
not possible for the Au/C catalyst and is slow for the Pd/C
and Pt/C catalysts. In the presence of base, the H+ is readily
abstracted from one of the primary hydroxyl groups of gly-
cerol for form glycerate, thereby overcoming the rate limiting
step of the alternative oxidation pathway. Hence, it is
proposed that the oxidation mechanism of glycerol with Au
catalysts in the presence of base proceeds via an initial
deprotonation pathway as has been previously observed.26–28
`
M. Besson, P. Gallezot, F. Lahmer, G. Fleche and P. Fuertes,
9
J. Catal., 1995, 152, 116.
10 P. Fordham, M. Besson and P. Gallezot, Appl. Catal. A, 1995,
133, L179.
11 J. C. Beziat, M. Besson and P. Gallezot, Appl. Catal. A, 1996,
135, L7.
12 M. Besson and P. Gallezot, Catal. Today, 2000, 57, 127.
13 P. C. C. Smits, B. F. M. Kuster, K. van der Wiele and H. S. van
der Baan, Appl. Catal., 1987, 33, 83.
14 R. Garcia, M. Besson and P. Gallezot, Appl. Catal. A, 1995,
127, 165.
15 P. Fordham, R. Garcia, M. Besson and P. Gallezot, Stud. Surf.
Sci. Catal., 1996, 101, 161.
16 P. Gallezot, Catal. Today, 1997, 37, 405.
17 H. Kimura, K. Tsuto, T. Wakisaka, Y. Kazumi and Y. Inaga,
Appl. Catal. A, 1993, 96, 217.
18 H. Kimura, Appl. Catal. A, 1993, 105, 143.
19 G. C. Bond and D. Thompson, Cat. Rev.-Sci. Eng., 1999, 41, 319.
20 L. Prati and M. Rossi, J. Catal., 1998, 176, 552.
21 F. Porta, L. Prati, M. Rossi, S. Colluccia and G. Martra, Catal.
Today, 2000, 61, 165.
Acknowledgements
22 C. Bianchi, F. Porta, L. Prati and M. Rossi, Top. Catal., 2000, 13,
231.
23 L. Prati, Gold Bull., 1999, 32, 96.
24 S. Carrettin, P. McMorn, P. Johnston, K. Griffin and G. J.
Hutchings, Chem. Commun., 2002, 696.
We thank Johnson Matthey for financial support and provi-
sion of a scholarship for Silvio Carrettin.
25 N. H. Lin and S. G. Sun, J. Electroanal. Chem., 1998, 448, 5.
26 L. Prati and M. Rossi, J. Catal., 1998, 176, 552.
27 K. G. Griffin, P. Johnston, S. Bennett and S. Kaliq, Abstracts 4th
World Congress Oxidation Catalysts, Berlin/Potsdam, September
2001, Dechema, Frankfurt, 2001, vol. 1, p. 423.
28 T. Mallat and A. Baiker, Catal. Today, 1994, 19, 1247.
References
1
2
R. A. Sheldon and J. Dakka, Catal. Today, 1994, 19, 215.
K. Heyns and H. Paulsen, Angew. Chem., 1957, 69, 600.
1336
Phys. Chem. Chem. Phys., 2003, 5, 1329–1336