Applied Catalysis A: General 499 (2015) 89100

Contents lists available at ScienceDirect

Applied Catalysis A: General

journal homepage: www.elsevier.com/locate/apcata

Catalytic behaviour of four different supported noble metals in the

crude glycerol oxidation
a,b,

Elzbieta
Skrzynska
, Soraya Zaid b , Jean-Sbastien Girardon b , Mickal Capron b ,
b,c,
Franck Dumeignil
a

Faculty of Chemical Engineering and Technology, Cracow University of Technology, Ul. Warszawska 24, 31-155 Cracow, Poland
CNRS UMR8181, Unit de Catalyse et Chimie du Solide, UCCS, Universit de Lille 1 Sciences et Technologies, F-59655 Villeneuve dAscq, France
c
Institut Universitaire de France, Maison des Universits, 103, Boulevard Saint-Michel, 75005 Paris, France
b

a r t i c l e

i n f o

Article history:
Received 25 January 2015
Received in revised form 30 March 2015
Accepted 7 April 2015
Available online 15 April 2015
Keywords:
Crude glycerol oxidation
Supported noble metal catalysts
Gold
Platinum
Palladium
Silver

a b s t r a c t
The activity of four different noble metals (Ag, Au, Pd and Pt) in the liquid phase oxidation of pure
glycerol was confronted with the results obtained with a crude glycerol fraction, received from a largescale biodiesel production plant. The catalysts were characterized by numerous techniques, giving insight
into actual metal loading (elemental analysis by ICP and XRF), surface morphology (nitrogen absorption
methodsBET and porosity), chemical state of both the support and the metal (XRD and XPS), and, nally,
the metal particle size distribution (TEM microscopy). A good dispersion of totally reduced noble metal
particles of a nanometric size (an average metal diameters were equal 3.5 nm, 4.2 nm, 4.7 nm and 21.2 nm
for respectively Pd, Pt, Au and Ag) was accompanied with a comparable values of total metal loadings
on the alumina support (from 0.95 and 0.96 wt.% for Pt and Pd, up to 0.98 and 1.13 wt.% for Au and
Ag supported catalysts, respectively). In terms of initial reaction rate, the most active sample was the
Au/Al2 O3 catalyst, both using pure (12976 mol h1 molAu 1 ) or crude glycerol (1230 mol h1 molAu 1 ).
However, comparison of the selectivities and conversions after 12 h shows that the most robust and
resistant catalyst toward the impurities present in crude glycerol is Pd/Al2 O3 , with a loss of conversion
less than 50% (in respect to analogous reaction using pure glycerol) and almost unchanged high selectivity
to glyceric acid (close to 8090%). Ag/Al2 O3 also showed a relatively high resistance to impurities in terms
of glycerol conversion, but with a drastic modication of its selectivity. The activity of the two other
catalysts was dramatically affected with a conversion divided by ca. 4 and even 10 for the Pt and the
Au catalysts, respectively, when using crude glycerol instead of pure glycerol. Finally, the effect of each
main impurity (MONG-NM, i.e., matter organic non-glycerol and non-methanol; ash; methanol; sulphur
compounds) was independently studied. In any case, the sulphur compounds and MONG-NM were the
impurities the most detrimental for the performances of catalysts. Thus, they should be removed in
priority from crude glycerol fractions before reaction, while ashes and methanol should not be considered
as completely undesirable.
2015 Elsevier B.V. All rights reserved.

1. Introduction
Nowadays, glycerol is considered as an important bio-derived
platform molecule and a feedstock for bioreneries [15]. Its

Corresponding author at: Faculty of Chemical Engineering and Technology, Cracow University of Technology, Ul. Warszawska 24, 31-155 Cracow, Poland.
Corresponding author at: CNRS UMR8181, Unit de Catalyse et Chimie du Solide,
UCCS, Universit de Lille 1 Sciences et Technologies, F-59655 Villeneuve dAscq,
France.

E-mail addresses: eskrzynska@pk.edu.pl (E. Skrzynska),

franck.dumeignil@univ-lille1.fr (F. Dumeignil).
http://dx.doi.org/10.1016/j.apcata.2015.04.008
0926-860X/ 2015 Elsevier B.V. All rights reserved.

upgrading by chemical or biochemical transformations has thus

attracted much attention, and especially in the last decade [119].
Even if the use of glycerol as a raw material for producing a
large variety of chemicals is well documented in the literature
[1,4,6], only a very few applications have reached industrialization mainly due to the cost of puried glycerol. To reach viability,
many processes would need to be adapted to the use of cheap
dirty waste by-produced with biodiesel (the so-called glycerine
or crude glycerol), which currently is the main source of glycerol
on the market [16,20]. As purication of glycerine is a relatively
expensive multistep process [1,4,6,20], the possibility of direct
transformation of crude glycerol could be very attractive from an

90

E. Skrzy
nska et al. / Applied Catalysis A: General 499 (2015) 89100

industrial point of view. Indeed, in the liquid phase glycerol oxidation reaction, in addition to the obvious direct economic benet
brought when using a cheaper raw material, some other indirect
economic as well as environmental advantages can be listed: (i)
the unreacted base used in the conventional process of biodiesel
production (i.e., basic transesterication) could be directly used in
a downstream oxidation process, which is often also conducted
under basic conditions; (ii) the global amount of waste mineral salts
(from base neutralization) should thus be reduced; (iii) the number
of purication steps would also be reduced with only one purication downstream for recovering the nal product. Unfortunately,
as the most of the scientic papers dealing with catalytic conversion of glycerol to value-added products are based on the use of a
puried raw material [1,1019]. Hence, the effect of the different
impurities present in crude glycerol streams is still not well known.
The studies where the crude glycerol fractions were used as
a raw material are rare [21], especially in the eld of glycerol
oxidation in the liquid phase. Recently, we compared different
grades of glycerine and investigated in details the effect of each
major impurity on the performances of platinum supported on alumina in both basic and neutral reaction media [20]. The organic
matter non-glycerol and non-methanol (called MONG-NM), comprising mainly of various fatty acid derivatives, was identied as
the most problematic contaminant of crude glycerol fraction, showing the highest detrimental effect on the glycerol conversion. The
same conclusion was reported later by Chan-Thaw et al. [22], who
worked on oxidation of raw glycerol (from edible rapeseed vegetable oil transesterication) using supported Au-Pd nanoparticles.
The highest initial activity at 50 C using non-treated raw glycerol (1672 mol/mol h) was achieved over a 1% Au6 Pd4 /AC catalyst,
which also showed the highest activity in a process using commercial pure glycerol (3205 mol/mol h). Gil et al. [23] attempted the
oxidation of partially puried glycerol fractions (minimum purity
not less than 95.5%) at 60 C under 5 bar of O2 over carbonaceous
materials-supported gold catalysts. They reported almost 50% of
glycerol conversion after 10 h of test with 95.5% glycerol. The performances were much better when using almost pure, neutralized
97.1% glycerol with 60% of conversion after 4 h. Essentially the same
conversion was observed using a high-purity 99.5% commercial
anhydrous product, but with a higher selectivity to glyceric acid
(65% vs. 45% when using the neutralized fraction). Sullivan and
Burnham [24] used model titanium-supported gold catalysts. They
proved that such catalytic system is very sensitive to impurities
present in a crude (68.5%) glycerol fraction. The authors obtained
less than 20% conversion after 24 h at 60 C under air ow at atmospheric pressure; a remarkably enhanced production of formic
acid from crude glycerol fraction in comparison to analogous test
with pure glycerol and puried fractions (after partial removal of
potassium, phosphorous and fatty acid derivatives) was observed
without any tentative explanation [24]. Finally, Kondamudi et al.
[25] studied photooxidation over a titanium di-silicide catalyst
(TiSi2 ) at 65 C under atmospheric pressure, where almost 64% conversion with 100% selectivity to glyceric acid were achieved after
6 h of reaction, using crude glycerol of an unspecied composition.
In the present paper, we evaluated the impact of glycerol
purity on the performances of four different noble metals-based
catalysts, namely: Ag/Al2 O3 , Au/Al2 O3 , Pd/Al2 O3 and Pt/Al2 O3 .
Three of them, i.e., gold, palladium and platinum supported
catalysts, are commercially available and well known for their
high activity in the partial oxidation of pure glycerol in the liquid
phase. In numerous studies, various research groups investigated
the effect of such parameters as: the reaction temperature, the
oxygen pressure, the glycerol concentration, the presence of base,
the catalyst synthesis method, etc. [1,5,10,1220]. Nevertheless,
the effect of glycerol purity was investigated only over gold and
platinum catalysts [20,22,24]. As each of the aforementioned

research groups used different glycerol fractions, further using different reaction conditions, it seems quite difcult to draw reliable
comparative conclusions on the role of each impurity on catalytic
performances modulation over each type of catalyst. From the
best of our knowledge, there are no articles concerning oxidation
of crude glycerol in the liquid phase using other monometallic
catalytic systems, especially those based on palladium and silver.
Herein, the abovementioned catalysts were tested under identical reaction conditions (0.3 M glycerol concentration in the reaction
mixture, 60 C, 5 bars of oxygen, NaOH/glycerol molar ratio equal
4 and glycerol/catalyst weigh ratio of 11), using both commercial
anhydrous glycerol and a crude glycerol fraction received from
a biodiesel plant. Then, in order to decouple the effect of each
impurity and to avoid misinterpretation due to possible interactions and synergism, each of identied and quantied impurity was
separately added to pure glycerol solution and the results were
compared in terms of glycerol conversion and selectivity to main
products. We believe that the outcomes of this study will give
elements to decide which type of catalyst should be used for the
oxidation of crude glycerol fractions.
2. Experimental
2.1. Materials
Anhydrous glycerol 99% from Sigma-Aldrich and crude glycerol
from Orlen Poudnie S.A. were used for the catalytic tests. The main
composition of the crude glycerol fraction is given elsewhere [20]
and consists of 47.4 wt.% of glycerol, 29.1 wt.% of methanol (HPLC),
8.6 wt.% of water (Karl Fischer titration), 1.3 wt.% of ash (gravimetric
method) and 13.6 wt.% of matter organic, non-glycerol and nonmethanol (MONG-NM, calculated according to IUPAC guidelines
[26]). The sulphur concentration in the crude fraction (0.1 wt.%) was
further determined by portable XOS Sinide OTG analyzer based on
Monochromatic Wavelength Dispersive X-Ray Fluorescence (MWD
XRF).
Sodium sulfate (99.0%, ACS reagent from Sigma-Aldrich),
methanol (99.9% HPLC grade from Aldrich) and thioglycolic acid
(98.0%, pure from Fluka) were used as received without any further purication. The MONG-NM used for supplementary tests was
obtained by physical separation from the crude glycerol fraction
(hydrophobic top layer of the fraction).
Silver nitrate (99.0%, ACS reagent), methanol (99.9% CH3 OH),
formaldehyde (37 wt.% in H2 O) and sodium hydroxide (purum)
were all purchased from Sigma Aldrich, and alumina oxide powder
(activated basic Al2 O3 , Merck) was used for preparation of silversupported catalyst.
Commercial 1 wt.% Au/Al2 O3 catalyst (AUROliteTM from Sterm
Chemicals) and 1 wt.% Pt/Al2 O3 from Sigma-Aldrich were grounded
and sieved to obtain fraction 50125 m, while 1 wt.% Pd/Al2 O3
(powder from Sigma-Aldrich) was used as received. No additional
pretreatment procedure was applied to these catalysts before
testing.
2.2. Catalyst preparation method
The silver supported alumina oxide catalyst was prepared by
chemical reduction in the liquid phase. The alumina support powder (fraction 50125 m) was suspended in a methanol solution
of silver nitrate and, after having adjusted the pH of the reaction
mixture to 8 with sodium hydroxide, a 2 M aqueous solution of
formaldehyde was used as a reducing agent. The suspension was
mixed and heated under the reux for 90 min. Then, the solid was
separated by ltration, washed with distilled water and dried at
110 C for 24 h prior testing.

E. Skrzy
nska et al. / Applied Catalysis A: General 499 (2015) 89100

2.3. Catalysts characterization

3. Results and discussion

The specic surface area (BET method) and pore size distribution
(BJH method) were measured by the nitrogen adsorption technique
using a TriStar II 3020 apparatus from Micromeritics.
XRD analysis was performed at ambient temperature on a D8
Advanced apparatus from Bruker AXS instrument. The samples
were scanned at a rate of 0.02 per step over the 5 2 90 range
(scan time = 3 s step1 ). The diffractograms were indexed using the
JCPDS database.
The oxidation state of the metals on the surface of fresh
and used catalysts was determined by XPS analysis (VG ESCALab
220XL from Thermo Fisher Scientic) using a monochromatized
aluminium source (AlK = 1486.6 eV). The high-resolution spectra
were recorded with a 40 eV pass energy, and the value of the C1s
core level (284.6 eV) was used for calibration of the energy scale.
Curve tting was performed using the CasaXPS software taking into
account a Shirley-type background subtraction, symmetric peak
shape of silver, gold and palladium, and an asymmetric peak shape
of platinum LA(1.2,85,70) [27].
For transmission electron microscopy analysis two apparatus
were used. The rst one, a TEM FEI Tecnai G2-20 twin electronic
microscope, working with an accelerating potential of 200 kV,
enabled observation of samples with a very high resolution (25 nm
scale), while the second one, a Philips CM 30 Transmission electron microscope, working with an accelerating potential of 300 kV
and equipped with an energy dispersive spectrometer (EDS) for the
local chemical analysis (spatial resolution 105 mm3 ), enabled also
operating in a transmission scan mode (STEM mode), which is useful for mapping of the catalyst surface composition. The samples for
TEM were prepared from a diluted suspension of catalyst in ethanol.
A drop of suspension was placed on a Lacey carbon-coated grid and
allowed to dry in air. The particle size distribution was calculated
by counting over 400 particles over multiple areas using the Visilog
6.5 software.
The metal loading in fresh platinum, palladium and gold catalysts was determined by inductively coupled plasma-atomic
emission spectroscopy (ICP-AES, Vista Pro Varian). The samples
were prepared by Hot Block/HCl/HNO3 digestion [28]. For the
silver-supported catalyst, due to low sensitivity and accuracy to
Ag+ by ICP, X uorescence spectroscopy was applied using a S2
Ranger Bruker spectrometer equipped with Pd X-ray tube.

3.1. Catalysts characterization

2.4. Catalytic performances evaluation

A typical experiment of the liquid phase glycerol oxidation
was carried out in a 300 cm3 semi-batch stainless steel reactor
equipped with a gas-induced turbine, 4 bafes, a thermocouple,
and a thermo-regulated oxygen supply system. In each experiment,
200 cm3 of a pure or crude glycerol solution in water were heated
to 60 C, and the reaction was started when the calculated amount
of NaOH with selected catalyst (0.5 g) were ushed into the reactor and pressurized with oxygen (5 bar). In all the experiments,
the initial concentration of glycerol was 0.3 M, and the amount of
base was adjusted to give a NaOH/glycerol molar ratio of 4. For
the supplementary studies, well-controlled quantities of selected
and desired amount of additive (Na2 SO4 , methanol, MONG-NM,
thioglycolic acid) were introduced in the reactor together with
the reagents. The products were periodically sampled (1 cm3 samples directly acidied to quench the reaction) and analyzed with
an Agilent 1200 HPLC equipped with a Rezex ROA-Organic Acid
H+ column (300 7.8 mm) and a reective index detector (RID). A
solution of H2 SO4 (0.0025 M) in deionized water (0.5 cm3 min1 )
was used as an eluent. The identication and quantication of the
obtained products was done by comparison with the corresponding
calibration curves.

91

First, the metal loading in the fresh catalysts was determined.

Results presented in Table 1 show that the real metal loading in all
the tested catalysts was close to the 1 wt.%, for the Au, Pd and Pt
supported materials (metal loading guaranteed by producer), and
also for the Ag/Al2 O3 material prepared by the chemical reduction
method.
Due to the relatively low metal loadings used, the surface properties of each catalyst, i.e., the specic surface area and the pore
size distribution (Table 1, Fig. 1), undoubtedly reected the character of the corresponding alumina support. While information on
the aluminas used for the preparation of commercial catalysts is
unavailable, this can be veried for the lab-prepared silver supported catalyst (Table 1, Figs. 1 and 2).
Combining the results of nitrogen adsorptiondesorption analysis with XRD analysis (Fig. 2), we can see that the Au/Al2 O3
and Pt/Al2 O3 catalysts were prepared using gamma alumina
supports with respectively high (>200 m2 g1 ) and moderate
(<100 m2 g1 ) specic surface areas. Both solids showed similar average pore diameter, close to 8 nm, with a quite uniform
and sharp pore size distribution (Fig. 1c and e). The shape of
the nitrogen adsorption hysteresis loop suggested the presence
of non-interconnected inkbottle pores [29,30] for the Au and the
Pt samples, slightly distorted for the latter. For the Ag/Al2 O3 and
Pd/Al2 O3 catalysts, in spite of similar specic surface areas (about
150 m2 g1 ), quite different pore structures were observed. The
nitrogen adsorptiondesorption prole of the palladium-based catalyst (Fig. 1d) shows a hysteresis loop reecting both H1 and H3
types, with a broad distribution of the pore diameters (from 3 to
80 nm). This suggests the presence of aggregates of plate-like particles giving rise to slit-shaped pores as well as to non-interconnected
inkbottle pores with random size distributions of cavities and necks
[29,30]. The Pd/Al2 O3 sample was also the only material where Al2 O3 , accompanied with -alumina and Boehmite (hydrated form
of -alumina), was detected by XRD (Fig. 2). Finally, the Ag/Al2 O3
material, which is based on a partially hydrated gamma alumina
(Boehmite and Gibbsite phases detected by XRD for both the catalyst and the original support used for preparation), exhibited a
narrow distribution of pore diameters (centred on 4.39 nm for the
silver catalyst and on 4.34 nm for the original alumina support),
with a distorted hysteresis loop suggesting possible interconnections of the cavities.
The relatively low metal loading and the high specic surface
areas of the catalysts are elements that can explain why metallic phases could not be clearly detected by XRD (Fig. 2). Metallic
phases are also hardly detected when they are present as small

Table 1
Metal (Me) loading estimated by elemental analysis (EA) and results of nitrogen
adsorption analysis, i.e., specic surface area (SSABET ), BJH desorption average pore
diameter (dpores ) and BJH desorption cumulative pore volume (Vpores ).
Catalyst

Al2 O3
Ag/Al2 O3
Au/Al2 O3
Pd/Al2 O3
Pt/Al2 O3

Real loadinga
(Me wt.%)a

(mmolMe g1 )

1.13
0.98
0.96
0.95

0.10
0.05
0.09
0.05

SSABET
(m2 g1 )

dpores (nm)

Vpores
(cm3 g1 )

196
158
229
143
88

4.34
4.39
8.23
10.09
7.67

0.235
0.217
0.701
0.826
0.257

a
For the Au, Pd and Pt supported commercial catalysts, the elemental composition
was determined by ICP-AES, while for the lab-made Ag/Al2 O3 sample, XRF was used.
b
Support used for preparing the lab-made silver supported catalyst.

92

E. Skrzy
nska et al. / Applied Catalysis A: General 499 (2015) 89100

Fig. 1. Nitrogen adsorption/desorption proles (insets) and pore diameter distribution calculated by the BJH method of fresh alumina-supported catalysts: (a) Ag/Al2 O3 and
(b) corresponding bare alumina support; (c) Au/Al2 O3 ; (d) Pd/Al2 O3 ; (e) Pt/Al2 O3 .

metallic particles very well dispersed on the support surface. The

TEM images (Fig. 3) are on line with this assumption.
The commercial catalysts exhibited a relatively narrow particle
size distribution, and almost 30% of all the particles had a diameter
of 3 nm. For the Pt and Pd catalysts, the biggest measured particle
had a diameter of 16 nm and 12 nm, respectively, while for the Au
material, single particles with a diameter of 1826 nm were also
found. In the case of the lab-made Ag catalyst, the particle size distribution was less uniform, but 30% of all the particles were below
10 nm and almost 40% had a diameter between 16 and 26 nm. Thus,

the main particle diameter calculated as an arithmetic average of

all the particles, increased in the order:
Pd/Al2 O3 (3.5 1.5 nm) < Pt/Al2 O3 (4.2 2.5 nm) < Au/Al2 O3
(4.7 2.6 nm) < Ag/Al2 O3 (21.2 18.6 nm).
Finally, XPS analysis was performed in order to determine the
oxidation state of the supported particles (Fig. 4). For all the samples, a totally reduced state of the metal particles was conrmed.
The corresponding photoelectron peak binding energy (BE) values
were similar to those reported in the literature for metallic species,
i.e., 368.2 eV for Ag 3d5/2 , 83.4 eV for Au 4f7/2 , 335.8 eV for Pd 3d5/2

E. Skrzy
nska et al. / Applied Catalysis A: General 499 (2015) 89100

93

and 71.2 eV for Pt 4f7/2 [27,3033]. In each case, the value of the C
1s core level (284.6 eV) was used for calibration of the energy scale
and the binding energy of Al 2p level was accordingly essentially
the same, equal 74.6 0.2 eV.
The XPS analysis repeated for the used catalysts after oxidation
tests with pure and crude glycerol conrmed a good stability of the
oxidation state of the metallic particles under operating conditions.
Irrespective of the type of glycerol used in the oxidation tests, the BE
of each metallic core level remained almost unchanged with very
small variations at max of +0.1 eV or 0.15 eV (Table 2). Moreover,
the ratio between the metal and alumina measured before and after
oxidation tests remained practically the same for each type of the
catalyst.
3.2. Catalytic testsoxidation of pure and crude glycerol

Fig. 2. X-ray diffractograms of the fresh catalysts with the position of the main
diffraction lines expected for the metallic particles: Ag (JCPDS card 01-089-3722),
Au (JCPDS card 00-004-0784), Pd (JCPDS card 01-087-0637) and Pt (JCPDS card
00-001-1194). Symbols correspond to: () gamma-Al2 O3 (JCPDS 00-010-045); ()
theta-Al2 O3 (JCPDS 01-086-1410); (#) Boehmite, gamma-Al2 O3 H2 O (JCPDS 00-0011283); and (*) Gibbsite, Al(OH)3 (JCPDS 00-033-0018).

The performance of the catalysts was assessed in the liquid

phase glycerol oxidation reaction with pure and crude streams.
Each of the catalysts exhibited different behaviours (different activities and selectivities), but a clear relation between the purity of the
reactant feed and the reaction rate was observed (Table 3).
For the oxidation of pure glycerol, under our conditions the
most active catalyst was the Au catalyst, which enabled obtaining full conversion after only 30 min of reaction (Fig. 5). The main
products observed at the beginning of the reaction were: glyceric acid (60% selectivity), glycolic acid (about 20% selectivity)
and formic acid (1012% selectivity). The selectivity to these two
latter compounds remained constant during the progress of the
experiment, while the glyceric acid selectivity decreased due to
further oxidation to tartronic acid. That stays in agreement with

Fig. 3. TEM images and histograms of the metallic particle size distributions observed over the catalysts: (a) Ag/Al2 O3 ; (b) Au/Al2 O3 ; (c) Pd/Al2 O3 ; and (d) for Pt/Al2 O3 .

94

E. Skrzy
nska et al. / Applied Catalysis A: General 499 (2015) 89100

Fig. 4. XPS spectra of the Ag 3d, Au 4f, Pd 3d and Pt 4f levels with corresponding BE values observed for fresh catalysts: (a) Ag/Al2 O3 ; (b) Au/Al2 O3 ; (c) Pd/Al2 O3 ; and (d)
Pt/Al2 O3 .
Table 2
Comparison of the BEs estimated for each metallic core levels (Ag 3d, Au 4f, Pd 3d and Pt 4f) and selected relations between surface atomic concentrations from XPS analysis
performed for fresh and spent catalysts.
Relative atomic concentration

Catalyst*

Corresponding supported
metal core level BE (eV)

metal (100)/Al

Na/Al

C/Al

Ag/Al2 O3 fresh
Ag/Al2 O3 after test with pure glycerol
Ag/Al2 O3 after test with raw glycerol

368.2
368.1
368.2

1.5
1.5
1.6

0
<0.01
0

0.34
0.26
0.30

Au/Al2 O3 fresh
Au/Al2 O3 after test with pure glycerol
Au/Al2 O3 after test with raw glycerol

83.4
83.3
83.4

0.1
0.1
0.1

0
0.02
<0.01

0.25
0.32
0.70

Pd/Al2 O3 fresh
Pd/Al2 O3 after test with pure glycerol
Pd/Al2 O3 after test with raw glycerol

335.8
335.7
335.6

0.6
0.5
0.5

0
<0.01
0.01

0.21
0.13
0.22

Pt/Al2 O3 fresh
Pt/Al2 O3 after test with pure glycerol
Pt/Al2 O3 after test with raw glycerol

71.2
71.1
71.1

0.3
0.3
0.3

0
<0.01
0

0.18
0.15
0.30

* Catalysts recovered after processes described in Fig. 5, i.e., with pure or crude glycerol solution (0.3 M), at 60 C under 5 bar of O2 , with NaOH/glycerol = 4 (molar ratio). After
each process, the catalysts were separated from the reaction mixture, washed with water (200 cm3 ) and dried at 110 C before analysis.

the generally accepted reaction pathway for glycerol oxidation in

the liquid phase [34]. Platinum- and palladium-based catalyst were
less active but more selective to glyceric acid, which can be easily seen from comparison of the selectivities to main products at
30% of glycerol isoconversion, i.e., 74% and 85.8% selectivity to glyceric acid using respectively Pt and Pd catalyst, vs. 60.4% selectivity
over the Au supported material (Table 3). Only the Ag catalyst

showed a different pattern concerning the main products distribution, as the highest selectivity was recorded for glycolic acid,
about 45% after 30 min of reaction, while the selectivity to glyceric acid, which was initially of 34% decreased after a few minutes
of reaction to 2326% and then remained constant until the end
of the experiment. In the same experiment repeated under analogous reaction conditions but for 5 h, 43% of glycerol conversion was

E. Skrzy
nska et al. / Applied Catalysis A: General 499 (2015) 89100

95

Table 3
Comparison of the initial reaction rates and the selectivities to main products at 10% and 30% isoconversion for the oxidation of pure and crude glycerol over Ag, Au, Pd and
Pt supported catalysts. The reaction conditions are the same as those in Fig. 5.
Type of catalyst

Type of glycerol

Initial reaction
rate*
(mol h1 molMe 1 )

Selectivity to main products (%) at:

10% of glycerol isoconversion

30% of glycerol isoconversion

GA

GLYCA

FA

GA

GLYCA

FA

TARTA

OXALA

Ag/Al2 O3

Pure
Crude

866
384

27.8
49.5

35.9
27.0

35.3
17.6

27.2
n.a.

44.8
n.a.

28.0
n.a.

0
n.a.

0
n.a.

Au/Al2 O3

Pure
Crude

12976
1230

59.8
43.4

19.7
13.2

10.2
39.6

60.4
n.a.

20.7
n.a.

12.5
n.a.

0.9
n.a.

0.2
n.a.

Pd/Al2 O3

Pure
Crude

1091
226

91.7
81.2

2.8
4.9

0
2.8

85.8
n.a.

2.6
n.a.

1.0
n.a.

5.7
n.a.

0.3
n.a.

Pt/Al2 O3

Pure
Crude

2898
828

75.6
60.5

10.7
3.3

11.0
36.1

74.0
n.a.

9.9
n.a.

8.1
n.a.

5.1
n.a.

2.6
n.a.

* For the calculation of the specic initial reaction rate, we used the metal loadings estimated by elemental analysis; n.a.not analyzed/not applicable as the calculation
of selectivities at 30% glycerol isoconversion would require too extensive extrapolation of the experimental results; Abbreviations: GAglyceric acid, GLYCAglycolic acid,
FAformic acid, TARTAtartronic acid, OXALAoxalic acid.

achieved with similar selectivities to glycolic acid (42%) and glyceric

acid (23%).
Such a big difference between the activities of the catalysts can
be rst explained in the light of the metallic particles distribution
on the catalysts surface (Table 1). That would follow the concept
of the structure-sensitive behaviour introduced for this reaction
by Demirel et al. [35], and further conrmed by the works of Prati
et al. [36] and Carretin et al. [37], and also further discussed by Dimitratos et al. [16]. It is generally accepted that small particles with
narrow size distribution usually means better metal dispersion on
the catalysts surface and higher activity. Indeed, as it can be seen
from Figs. 3 and 5, silver supported catalyst with the highest average diameter and widest particle size distribution (21.2 18.6 nm)
was the least active in the pure glycerol oxidation. It also promoted an oxidative cleavage of glycerol molecule leading directly
to glycolic and formic acids (respectively 35.9% and 35.3% selectivity at 10% of glycerol conversionTable 3), while the other metals
favoured a non-destructive oxidation of hydroxyl group to tartronic
acid via glyceric acid. Further transformation of primary products
should proceed by deep oxidation reactions, leading to mesoxalic
acid, oxalic acid, formic acid and nally to carbon dioxide in the
form of carbonate under basic reaction conditions. Possible reaction
pathways were discussed in details in a previous paper [34].
Although a complex and thoroughgoing studies are required to
explain the nature of silver active sites in the glycerol oxidation
process at liquid phase, very high selectivity to glycolic acid seems
to be characteristic for this metal. Analogous behaviour was previously reported by Ketchie et al. [38], who tested gold and silver
powders in the glycerol oxidation reaction. They concluded that not
only particle size distribution plays a signicant role in the glycerol
oxidation process and other parameters must also be considered.
That also stay in agreement with our experiences with the blank
tests and quasi-homogenous glycerol oxidation over gold nanoparticles [34], as well as on the results of catalysts screening from
samples prepared by different methods, where the mean diameter of the particles (platinum, gold and silver) varied between 2 nm
and 260 nm. As it can be seen from Fig. 5, the purity of glycerol is
another parameter which greatly affects the process.
Application of the crude glycerol fraction as a raw material for
the oxidation process evidently blocked the catalysts activity. In any
case, the decrease in the initial reaction rate was signicant, from a
factor 2 in the case of Ag/Al2 O3 to a 10 fold decrease over Au/Al2 O3
(Table 3). The highest resistance exhibited the catalysts with both
the lowest and the highest average diameter of the particles (3.5 nm
for Pd and 21.2 nm for Ag, respectively). Except for the palladium

catalyst, which enabled obtaining 24% of conversion after 2 h at

60 C, the conversions were much below 20%. The selectivities to
the main products were then compared at 10% of glycerol isoconversion (Table 3). Note that while the conversion of pure glycerol
over the gold catalyst was 27% after 3 min of reaction, comparison
with data at 10% conversion in the case of using crude glycerol was
reliable due to the relatively stable selectivities to main products
(i.e., glyceric, glycolic and formic acids) observed during the rst
20 min of reaction.
A decrease in selectivity to glyceric acid in favour of formic
acid was observed over the Pt and Au catalysts when shifting to
the crude glycerol feed. As methanol was present in the crude
glycerol fraction (29.1 wt.%), its conversion to formic acid should
be expected. However, the contribution of this side reaction was
subtracted when calculating glycerol conversion and selectivities
(method explained in details elsewhere [20]). Thus, overproduction of formic acid could only be explained by a change in glycerol
reactivity over the solids placed under crude glycerol conditions.
The dominating reaction pathways (i.e., selective oxidation of
hydroxyl groups leading successively to glyceric and tartronic
acids, and oxidative cleavage directly to glycolic and formic acids)
can compete with destructive, deep oxidation of glycerol molecule
to C1 derivatives. Possibility of both reaction pathways was veried
previously by kinetic calculations and proposed for both the noncatalytic oxidation at basic medium and the quasi-homogenous
oxidation of pure glycerol using unsupported gold nanoparticles
[34]. What is meaningful is that the production of formic acid is
only marginally increased from 0% to 2.8%, over the Pd catalyst
(Table 3), with a similar trend for the selectivity to glycolic acid
(small increase from 2.8% to 4.9%). In the case of the Ag/Al2 O3
catalyst, a completely opposite situation was observed, as the
selectivities to C1 and C2 derivatives decreased in favour of that of
glyceric acid. Thus, it seems reasonable that the impurities present
in the crude glycerol fraction blocked the sites responsible for
oxidative cleavage of glycerol, keeping active the sites responsible
for the non-destructive oxidation of hydroxyl groups. Thus, in
spite of poisoning, due to limited concentration of impurities in
the crude glycerol fraction, an increase of the overall number of
available metal particles (i.e. increase of Ag active sites responsible
for unique high selectivity to glycolic acid) should improve the
contribution of oxidative cleavage reactions in the crude glycerol
oxidation process and make the decrease of glycolic acid selectivity
less important. Additional experiments using the silver catalyst
with a substantially larger loading (2.3 wt.% Ag/Al2 O3 ) and a particle size distribution comparable to that of the 1.13 wt.% Ag/Al2 O3

96

E. Skrzy
nska et al. / Applied Catalysis A: General 499 (2015) 89100

Fig. 5. Glycerol conversion () and selectivity to main products: glyceric acid (), glycolic acid (), oxalic acid (), tartronic acid (), formic acid () in the oxidation of pure
glycerol (ad) and crude glycerol (eh). Catalyst: Ag/Al2 O3 (a and f); Au/Al2 O3 (b and f); Pd/Al2 O3 (c and g); Pt/Al2 O3 (d and h). Reaction conditions: 200 cm3 of pure or crude
glycerol solution with nominal 0.3 M concentration of glycerol; NaOH/glycerol molar ratio of 4; 0.5 g of catalyst; 60 C; 5 bars of O2 ; and a stirring speed of 1500 rpm.

E. Skrzy
nska et al. / Applied Catalysis A: General 499 (2015) 89100

Fig. 6. Glycerol conversion () and selectivity to main products: glyceric acid (),
glycolic acid () and formic acid () in the oxidation of pure glycerol (a) and crude
glycerol (b) over 2.3 wt.% Ag/Al2 O3 . Reaction conditions the same as at Fig. 5.

sample (mean particle diameters of 20.9 16.2 nm and

21.2 18.6 nm, respectively) proved that a higher concentration of active sites on the catalyst surface enables maintaining
relatively high selectivity to glycolic acid (4047%) during the
whole experiment with crude glycerol fraction (Fig. 6).
While the conversion of the crude fraction was again lower in
comparison to tests with pure glycerol (initial reaction rate equal,
respectively to 305 and 492 mol h1 molAg 1 ), a signicant outcome is that at comparable 15% isoconversion the selectivity to
formic acid decreased from 36% to 10% in favour of glyceric acid
(growth from 14% to 35% using pure and crude glycerol, respectively), and essentially the same selectivity to glycolic acid was
maintained, i.e., 41% for oxidation of pure glycerol and 42% while
using the crude fraction. Thus, the ratio between the impurities
and the silver active sites present on the catalyst surface has to
be controlled and kept at adequate low level, and/or higher metal
loading has to be applied, in order to maintain high conversion
rates and achieve protable selectivity to more valuable C2 and C3
derivatives.
3.3. Catalytic tests with addition of single impurities
The reasonable questions which arose from tests with crude
glycerol fraction were: (i) what is the impact of separate impurities?; (ii) Which of them should be limited to the minimum?; and
(iii) is there any key enabling proper choice of the catalyst for crude
glycerol oxidation? In our previous work, we identied the answer
to the rst question applied for the platinum supported catalyst

97

[20]. Namely, MONG-NM, which in majority consists of various

fatty acid derivatives (i.e., fatty acid methyl esters left after incomplete separation of biodiesel, as well as mono-, di- and triglycerides
coming from unconverted fatty raw material used for biodiesel
productionmore detailed discussion on this topic can be found
elsewhere [20]), was found to be the most problematic impurity,
as its deposition on the catalyst surface physically blocked the
access of glycerol to the active sites. While the performances of the
catalysts exhibiting a high activity might be restored by a simple
washing procedure with an organic solvent, the presence of small
quantities of MONG-NM in the reaction mixture strongly hindered
the glycerol conversion. The mineral salts had only a small detrimental effect on the process, while a low concentration of methanol
promoted the oxidation by improved oxygen solubility in the reaction mixture. Only the effect of the organic sulphur derivatives was
not assessed. The presence of such components was recently evidenced in the crude glycerol fraction, using sulphur sensitive GC
chromatography. Their content was estimated by MWD XRF measurements. The origin of such impurities in crude glycerol fraction
is not obvious. They can be formed during biodiesel production process, i.e., during neutralization of the basic catalyst conventionally
used for transesterication, as well as during acidic esterication
of low quality vegetable oils and other free fatty acids rich-raw
materials. However, they can also be present in the initial fatty acid
raw material, especially when non-conventional raw materials are
used, like: discards from vegetable oils industry, waste fats from
food processing plants, gastronomy or leather manufacture, etc.
Also, a non-rened oil obtained from crops demanding high levels
of sulphur fertilization (including rapeseed oil) can be responsible
for the presence of organic sulphur derivatives in crude glycerol
fractions [3941]. Our experiences with Au/Al2 O3 suggested that
strong poisoning of the active sites by organic sulphur derivatives
might be expected also in the processes with other noble metals
supported catalysts [42]. Thus, a series of catalytic experiments
with pure glycerol and small amounts of single additives, representing the impurities present in crude glycerol fractions, including
organic sulphur derivatives, were carried out. On the basis of our
previous experiments with crude glycerol using supported platinum [20] and also gold catalysts [42], we decided to add into the
mixture comparable amount of methanol, higher concentration of
sodium sulphate (due to its tinny effect on the conversion), and as
small as possible amount of MONG-NM and thioglycolic acid additives, as they were expected to be the most powerful and the most
problematic at the same time (both components should be avoided
in the mixtures injected on the HPLC column, and also possible
strong poisoning of both the catalyst and the walls of reactor might
appear). Composition of each prepared and tested reaction mixture
can be found in Table 4.
Clear similarities between the behaviour of the catalysts can be
observed (Fig. 7). The presence of methanol or sodium sulphate
affected the glycerol conversion in the lowest extent. A small promotional effect of CH3 OH, previously attributed to an increase in
the oxygen solubility in the reaction mixture [20], was evidenced
only when platinum and palladium catalysts were used, although
their selectivity was affected in different ways (Table 5).
The palladium catalyst practically did not exhibit any change
in performances, while the destructive oxidation pathway was
enhanced over the supported platinum catalyst. At 10% isoconversion, the selectivity to glyceric acid dropped from 75.6% to 48% in
favour of formic acidfrom 11% to almost 40%. For the other two catalysts, a small detrimental effect of methanol was observed, which
indicated that a purely physical phenomenon of change in oxygen
dissolution in the reaction composition is not the most important factor affecting process with methanol addition over these
two catalysts. Indeed, the results of pure glycerol oxidation under
oxygen pressure varying from 3 to 10 bars [34] showed that the

98

E. Skrzy
nska et al. / Applied Catalysis A: General 499 (2015) 89100

Table 4
Composition of each reaction mixture used to assess the effect of impurities in crude glycerol fractions. Solutions of nominal 0.3 M glycerol concentration were prepared
from crude glycerol (11.656 g of crude fraction in 200 cm3 ) or pure glycerol (5.53 g of pure, anhydrous glycerol in 200 cm3 ) with single, selected additive.
Type of impurity identied in crude
glycerol/additive used

Methanol
Ash/Na2 SO4
MONG-NMb
OSDc /thioglycolic acid
a
b
c

Tests with crude glycerola

concentration of impurities (wt.%)

Tests with pure glycerol

Amount of additive
used

Concentration in the
reaction mixture (wt.%)

4.28 cm3
2.432 g
0.271 g
1 l

1.7
1.2
0.1
7 ppm

1.7
0.1
0.8
66 ppm

No additives, impurities originated from crude fraction.

MONG-NMmatter organic non-glycerol and non-methanol [20].
OSDorganic sulphur derivatives.

Fig. 7. Changes in conversion during the oxidation glycerol in the presence of various additives: 2.432 g of sodium sulphate (); 4.28 ml of methanol (); 0.271 g of MONGNM (); 1 l of thioglycolic acid (). Conversion of pure glycerol (, full line) and crude glycerol (, dotted line) without additives are given for comparison. Catalysts:
Ag/Al2 O3 (a); Au/Al2 O3 (b); Pd/Al2 O3 (c); Pt/Al2 O3 (d). The reaction conditions are the same as those in Fig. 5.

performance of the gold nanoparticles practically did not change,

which was also the case for the process in the absence of catalyst. Thus, the effect of a small increase in the oxygen dissolution
caused by the presence of small amounts of methanol, i.e., physical aspect, might be less important than that of the chemical one,
especially if we remark that selectivities of both catalysts changed
in comparison to the process performed without methanol addition (Tables 4 and 5). For the gold catalyst, cleavage to formic was
much more important (selectivity to formic acid shifted from 10.2%
to 31.5%), while for the silver catalyst, the high selectivity to glycolic acid dropped from 35.9% to 14.5% in favour of non-destructive
oxidation to glyceric acid (of which the selectivity increased from
27.8% to 46.5%).

Similarly, the effect caused by addition of sodium sulphate was

more important over silver and gold catalysts than over platinum
and palladium catalysts. In the case of the two latter ones, the
selectivity remained almost unchanged. In contrast, over the gold
material, again a small increase in selectivity to formic acid was
observed (10.227.7%). Over the silver catalyst, a drastic increase in
glyceric acid selectivity was evidenced (from 27.8% to 46.8%), while
the selectivity to glycolic acid remained substantially unchanged
(36.3% vs. 35.9%), and nally, the presence of sodium sulphate eradicated the formic acid formation route (decrease in selectivity from
35.3% to 0%). The difference in the initial reaction rate observed
upon addition of sodium sulphate was the less pronounced compared to that observed using other additives. Moreover, the amount

E. Skrzy
nska et al. / Applied Catalysis A: General 499 (2015) 89100

99

Table 5
Comparison of the initial reaction rates over Ag, Au, Pd and Pt catalysts and selectivities to main products at 10% isoconversion in the presence of various additives: 2.432 g
Na2 SO4 , 4.28 ml CH3 OH, 0.271 g of MONG-NM and 1 l of thioglycolic acid. Reaction conditions are the same as those in Fig. 7.
Type of catalyst

Ag/Al2 O3

Au/Al2 O3

Pd/Al2 O3

Pt/Al2 O3

Type of additive

Na2 SO4
CH3 OH
MONG-NM
HSCH2 COOH
Na2 SO4
CH3 OH
MONG-NM
HSCH2 COOH
Na2 SO4
CH3 OH
MONG-NM
HSCH2 COOH
Na2 SO4
CH3 OH
MONG-NM
HSCH2 COOH

Initial reaction rate

(mol h1 molMe 1 )

673
550
385
234
9551
8176
1230
492
1149
1389
718
722
1449
2957
468
483

Selectivity to main products (%) at 10% of

glycerol isoconversiona
Glyceric acid

Glycolic acid

Formic acid

46.8
46.5
44.3
59.5a
58.3
57.1
65.7
60.5a
84.1
81.9
76.9
87.7
75.3
48.0
75.0
72.2a

36.3
14.6
25.0
7.1a
12.7
9.1
6.8
8.7a
3.2
1.8
2.3
1.3
7.4
10.1
4.3
2.9a

0
21.2
16.4
0a
27.7
31.5
13.3
4.1a
0.2
1.2
0
0
14.5
39.5
16.1
0a

a
For the tests with thioglycolic acid addition using silver, gold and platinum, the catalysts selectivities presented in this table are given for the highest observed values of
glycerol conversion, namely 4.1% for Ag/Al2 O3 , 6.8% for Au/Al2 O3 and 7.3% for Pt/Al2 O3 .

of Na2 SO4 used in the additional tests was 10-fold larger than the
amount of ash actually present in the crude glycerol streams used
for the tests. Hence, it is obvious that this component was not
responsible for any important decrease in the glycerol conversion,
especially when the process was carried out over the palladiumbased catalyst.
The third type of tested impurity was a hydrophobic, dark
brown, complex mixture of fatty acids derivatives, present in
crude glycerol fractions after biodiesel separation process [20]. We
previously showed that this mixture, recognized as matter organic
non-glycerol and non-methanol (MONG-NM), blocks the access
to active sites of Pt/Al2 O3 catalysts by reversible adsorption on
the catalyst surface. Regeneration of such deactivated catalysts is
however possible, but it requires extensive washing with a solvent
capable of dissolving the adsorbed hydrophobic fatty acid derivatives [20]. Concerning silver and palladium catalysts, the glycerol
conversion observed when adding MONG-NM to pure glycerol
was very similar to that observed when directly using the crude
glycerol fraction (Fig. 7). Interestingly, the same catalysts were also
the least sensible ones to organic sulphur poisoning among the
tested samples. Indeed, the ratio between the glycerol conversion
in the presence of thioglycolic acid to the glycerol conversion with
pure glycerol after 2 h was 0.35 for Pd/Al2 O3 , 0.15 for Ag/Al2 O3 , and
only 0.10 and 0.07 for Pt/Al2 O3 and Au/Al2 O3 , respectively. Exactly
the same sequence can be found when using the initial reaction
rates ratios as a criterion (calculated from results in Tables 3 and 5):

0.66Pd/Al2 O3 > 0.27Ag/Al2 O3 > 0.17Pt/Al2 O3 > 0.03Au/Al2 O3

As almost 66 ppm of sulphur derivatives were detected in
the reaction mixture prepared from the crude glycerol fraction
(Table 2), it seems reasonable to conclude that it was the main
reason why the results of crude glycerol oxidation over gold and
platinum catalysts were worse from that arising from test with
single MONG-NM addition (Fig. 7). All these observations clearly
indicate that MONG-NM and organic sulphur derivatives have the
most detrimental effect on the glycerol conversion in the oxidation process. Although removal of hydrophobic MONG-NM might
be realized by simple two-phase separation (less expensive gravitational or induced by centrifugation, tangential ow separation,

etc.), the removal (or deactivation) of organic sulphur derivatives

would require more sophisticated and expensive methods. Thus,
an improvement of the catalysts resistance to sulphur should be
considered.
Decoupling of the effects of the impurities allowed us to deduce
some specic trends concerning selectivity modications. However, these conclusions cannot be straightforwardly transposed to
real streams, as the complexity of crude glycerol composition is
such that some synergistic or antagonistic effects can be obtained
in the simultaneous presence of various impurities families. Moreover, the differences in the catalyst morphology can also play a
role in the response of the catalytic system to separate impurities,
even if the present data do not allow us to conclude on that specic
aspect. However, under the conditions used in the present paper,
the Pd-based catalyst suffered the least of selectivity modication.
Gold and platinum catalysts should not be used in the oxidation
of crude glycerol in the presence of large amounts of methanol, as
undesirable conversion of glycerol to formic acid was increased in
such a case. Analogous effect was also reported by Sullivan [24] for
crude glycerol oxidation over Au/TiO2 catalyst. On the other hand,
the presence of organic sulphur derivatives efciently blocked this
undesirable reaction, and selectivity to formic acid was then negligible (or very low) for all the tested catalysts. This additive had
also the most marked inuence in the change of selectivity over
the Ag catalyst from glycolic to glyceric acid, i.e., at 4.1% conversion
these selectivities were 7.1% and 59.5%, respectively. For comparison, during the test without any additives the selectivity to glycolic
acid at initial stage of the process (the lowest analyzed conversion
of 6.3%) was 26.9%, while selectivity to glyceric acid was only 33.7%,
and it further stabilized at the level of 2527% (Fig. 5a). Unique
high selectivity to glycolic acid was maintained only in the presence of mineral salt and MONG-NM (respectively 36.3% and 25.0%
at 10% isoconversion). Interestingly, the general performances of
the Ag catalyst were almost identical when using pure or MONGNM-added glycerol.
4. Conclusions
The results presented in this paper highlight the impact of
impurities present in crude glycerol fractions in the reaction of
liquid phase glycerol partial oxidation over four different types of

100

E. Skrzy
nska et al. / Applied Catalysis A: General 499 (2015) 89100

monometallic noble catalysts supported on alumina. The complexity of crude glycerol composition and its variety depending on its
origin (biodiesel production conditions from which it is issued),
as well as possible synergetic/antagonistic effects between the
impurities, makes it difcult to strictly draw any denitive conclusion. Nevertheless, based on the present results, some rationales
in the choice of proper catalysts for crude glycerol oxidation can
be given. Hence, monometallic gold catalysts, known from their
excellent activity in pure glycerol oxidation, should be avoided
when using unpuried waste fractions from biodiesel production.
This catalyst was the most sensitive to all the tested impurities.
Among the tested monometallic catalysts, palladium seems to be
the most resistant to both methanol, mineral salts, MONG-NM
and even to organic sulphur derivatives, which easily poison gold
and platinum catalysts. Pt/Al2 O3 can be used for oxidation of fractions rich in ash (mineral salts) and methanol, while an increase
in formic acid production might be expected in the presence
of this latter. Finally, the silver catalyst, with a high selectivity to glycolic acid (close to 50% at higher reaction rates) can
be used for processes with elevated MONG-NM concentrations.
Other impurities can drastically change the selectivity (glyceric
acid then becomes the main product), which can be an interesting and non-conventional method of controlling the products
distribution.
The results presented in this paper clearly show that the issue
of crude glycerol impurities is very complex and that the effect
of each impurity on the performance of various noble metal catalysts can be substantially different in terms of both the activity and
the selectivity. This opens interesting perspectives in the further
studies on bimetallic catalytic systems. One should also consider
catalytic systems with enhanced sulphur resistance, as this impurity together with MONG-NM of which a sticky amalgam can be
physically separated before reaction had the most detrimental
effect on the overall glycerol conversion performances.
Acknowledgments
This work was performed, in partnership with the SAS PIVERT,
within the frame of the French Institute for the Energy Transition (Institut pour la Transition Energtique (ITE) P.I.V.E.R.T.
(www.institut-pivert.com) selected as an Investment for the Future
(Investissements dAvenir). This work was supported, as part of
the Investments for the Future, by the French Government under
the reference ANR-001.
The authors would like to acknowledge Orlen Poudnie S.A. for
the crude glycerol samples kindly provided for testing.
Maxence Vandewalle (Universit de Lille 1 Sciences et Technologies) is acknowledged for XRF analysis.
The Fonds Europen de Dveloppement Rgional (FEDER),
CNRS, Rgion Nord-Pas-de-Calais and Ministre de lEducation
Nationale de lEnseignement Suprieur et de la Recherche are
acknowledged for fundings of X-ray diffractometers and XPS/LEIS/
ToF-SIMS spectrometers within the Ple Rgional dAnalyses de
Surface.
The authors would like to thank the European Union through the
Maopolska Regional Operational Programme (MROP) 2007-2013,
for funding of portable XOS Sinide OTG analyser.

References

[1] B. Katryniok, H. Kimura, E. Skrzynska,

J-S. Girardon, P. Fongarland, M. Capron,
R. Ducoulombier, N. Mimura, S. Paul, F. Dumeignil, Green Chem. 13 (2011)
19601979.
[2] B. Wang, Y. Shen, J. Sun, F. Xu, R. Sun, RSC Adv. 4 (2014) 1891718923.
[3] R. Ciriminna, C.D. Pina, M. Rossi, M. Pagliaro, Eur. J. Lipid Sci. Technol. 116 (2014)
14321439.
[4] M. Pagliaro, Glycerol The Platform Biochemical of the Chemical Industry, Simplicissimus Book Farm, 2013, ISBN: 9788863699647.
[5] Y. Zheng, X. Chen, Y. Shen, Chem. Rev. 108 (2008) 52535277.
[6] M. Pagliaro, M. Rossi, The Future of Glycerol, 2nd ed., RSC Green Chemistry Book
Series, Cambridge, 2010.
[7] A. Villa, S. Campisi, K.M.H. Mohammed, N. Dimitratos, F. Vindigni, M. Manzoli, W. Jones, M. Bowker, G.J. Hutchings, L. Prati, Catal. Sci. Technol. 5 (2015)
11261132.
[8] A. Corma, S. Iborra, A. Velty, Chem. Rev. 107 (2007) 24112502.
[9] S. Carrettin, P. McMorn, P. Johnston, K. Ch. Grifn, J. Kiely, G.J. Hutchings, Phys.
Chem. Chem. Phys. 5 (2003) 13291336.
[10] Ch-H. Zhou, J.N. Beltramini, Y.-X. Fan, G.Q. Lu, Chem. Soc. Rev. 37 (2008)
527549.
[11] M.O. Guerrero-Perez, J.M. Rosas, J. Bedia, J. Rodriguez-Mirasol, T. Cordero, Rec.
Pat. Chem. Eng. 2 (2009) 1121.
[12] B.N. Zope, R.J. Davis, Green Chem. 12 (2013) 34843491.
[13] S.E. Davis, M.S. Ide, R.J. Davis, Green Chem. 15 (2013) 1745.
[14] Y. Zheng, X. Chen, Y. Shen, Chem. Rev. 108 (2008) 52535277.
[15] A. Behr, J. Eilting, K. Irawadi, J. Leschinki, F. Lindner, Green Chem. 10 (2008)
1330.
[16] N. Dimitratos, A. Villa, C.L. Bianchi, L. Prati, M. Makkee, Appl. Catal. A 311 (2006)
185192.
[17] G.J. Hutchings, Catal. Today 100 (2005) 5561.
[18] N. Mimura, N. Hiyoshi, T. Fujitani, F. Dumeignil, RSC Adv. 4 (63) (2014)
3341633423.
[19] N. Mimura, N. Hiyoshi, M. Date, T. Fujitani, F. Dumeignil, Catal. Lett. 144 (2014)
21672175.

[20] E. Skrzynska,
A. Wondoowska-Grabowska, M. Capron, F. Dumeignil, Appl.
Catal. A 482 (2014) 245257.
[21] S. Bagheri, N.M. Julkapli, W.A. Yehye, Renew. Sustain. Energy Rev. 41 (2015)
113127.
[22] C.E. Chan-Thaw, S. Campisi, D. Wang, L. Prati, A. Villa, Catalysts 5 (2015)
131144.
[23] S. Gil, M. Marchena, C.M. Fernndez, L. Snchez-Silva, A. Romero, J.L. Valverde,
Appl. Catal. A 450 (2013) 189203.
[24] J.A. Sullivan, S. Burnham, Catal. Commun. 56 (2014) 7275.
[25] N. Kondamudi, M. Misra, S. Banerjee, S. Mohapatra, S. Mohapatra, Appl. Catal.
B 126 (2012) 180185.
[26] A. Hautfenne, Pure Appl. Chem. 52 (1980) 19391954.
[27] X-ray Photoelectron Spectroscopy (XPS) Reference Pages, Elements: Platinum,
Silver, Palladium, Gold, Home page: http://www.xpstting.com (access 15th
Jan. 2015).
[28] NIOSH Manual of Analytical Methods (NMAM), 4th ed., Elements by ICP (Hot
Block/HCl/HNO3 Ashing): Method 7303, Issue 1, dated 15 March 2003.
[29] A. Grosman, C. Ortega, Langmuir 24 (2008) 39773986.
[30] IUPAC, Pure Appl. Chem. 57 (1985) 603619.
[31] C.D. Wagner, W.M. Riggs, L.E. Davis, J.M. Moulder, G.E. Mulienberg (Eds.), Handbook of X-Ray Photoelectron Spectroscopy, Perkin-Elmer Corp, USA, 1979.
[32] A.M. Ferraria, A.P. Carapeto, A.M. Botelho do Rego, Vacuum 86 (2012)
19881991.
[33] M.P. Casaletto, A. Longo, A. Martorana, A. Prestianni, A.M. Venezia, Surf. Interface Anal. 38 (2006) 215218.

[34] E. Skrzynska,
J. Ftouni, J.-S. Girardon, M. Capron, L. Jalowiecki-Duhamel, J.-F.
Paul, F. Dumeignil, ChemSusChem 10 (20012) 20652078.
[35] S. Demirel-Gulen, M. Lucas, P. Claus, Catal. Today 102103 (2005) 166172.
[36] F. Porta, L. Prati, J. Catal. 224 (2004) 397403.
[37] S. Carretin, P. McMorn, P. Johnston, K. Grifn, C. Kiely, A. Attard, G.J. Hutchings,
Top. Catal. 27 (2004) 131136.
[38] W.C. Ketchie, Y.-L. Fang, M.S. Wong, M. Muryama, R.J. Davis, J. Catal. 250 (2007)
94101.
[39] B.B. He, J.H. Van Gerpen, J.C. Thompson, Appl. Eng. Agric. 25 (2009) 223226.

[40] D. Ostrowska, S. Pietkiewicz, M. Ciesinski,

K. Kucinska,
D. Gozdowski, World J.
Agric. Sci. 4 (2008) 133136.
[41] G. Ahmad, A. Jan, M. Arif, M.T. Jan, R.A. Khattak, J. Zhejiang Univ. Sci. B 8 (2007)
731737.

[42] E. Skrzynska,
J. Ftouni, A.-S. Mamede, A. Addad, M. Trentesaux, J.-S. Girardon,
M. Capron, F. Dumeignil, J. Mol. Catal. A 382 (2014) 7178.