Glycerol oxidation in liquid phase: Highly stable Pt catalysts supported on ion exchange resins

Article abstract of DOI:10.1016/j.apcata.2015.04.026

The glycerol oxidation reaction in liquid phase was studied using platinum supported on ion exchange resin as catalysts, with the objective of obtaining glyceric acid. In particular, the effect of the competitor anion, which is the ion used during the support pre-treatment, was addressed. It was found that the competitor ion affects the radial distribution of the active metal in the resin sphere. As the adsorption equilibrium constant of the competitor ion increases, the penetration of platinum in the resin particle also increases. Nevertheless, this difference did not affect the catalytic performance. However, the presence of iodide as competitor ion affected the platinum chemical environment, thus modifying the platinum active sites in this catalyst. Such effect was not observed when chloride or citrate ions were used for the resin pretreatment. Therefore, the competitor ion effect on the platinum catalysts performance is related with the ion nature more than with its affinity for the resin exchange sites. Iodide is adsorbed by platinum, and consequently, it affects its electronic state improving the selectivity. These catalysts showed very good activity, and what is also very important, very good stability, being possible to maintain the conversion and selectivity for many reaction cycles.

Full text of DOI:10.1016/j.apcata.2015.04.026

G Model
APCATA-15363; No. of Pages9
ARTICLE IN PRESS
Applied Catalysis A: General xxx (2015) xxx–xxx
Contents lists available at ScienceDirect
Applied Catalysis A: General
journal homepage: www.elsevier.com/locate/apcata
Glycerol oxidation in liquid phase: Highly stable Pt catalysts
supported on ion exchange resins
∗
Martín S. Gross, Bárbara S. Sánchez, Carlos A. Querini
Instituto de Investigaciones en Catalisis y Petroquímica (INCAPE) – FIQ – UNL – CONICET, Santiago del Estero 2654, Santa Fe, S3000AOJ, Argentina
a r t i c l e i n f o
a b s t r a c t
Article history:
The glycerol oxidation reaction in liquid phase was studied using platinum supported on ion exchange
resin as catalysts, with the objective of obtaining glyceric acid. In particular, the effect of the competitor
anion, which is the ion used during the support pre-treatment, was addressed. It was found that the
competitor ion affects the radial distribution of the active metal in the resin sphere. As the adsorption
equilibrium constant of the competitor ion increases, the penetration of platinum in the resin particle also
increases. Nevertheless, this difference did not affect the catalytic performance. However, the presence of
iodide as competitor ion affected the platinum chemical environment, thus modifying the platinum active
sites in this catalyst. Such effect was not observed when chloride or citrate ions were used for the resin
pretreatment. Therefore, the competitor ion effect on the platinum catalysts performance is related with
the ion nature more than with its affinity for the resin exchange sites. Iodide is adsorbed by platinum, and
consequently, it affects its electronic state improving the selectivity. These catalysts showed very good
activity, and what is also very important, very good stability, being possible to maintain the conversion
and selectivity for many reaction cycles.
Received 27 December 2014
Received in revised form 14 April 2015
Accepted 18 April 2015
Available online xxx
Keywords:
Glycerol
Glyceric acid
Platinum
Ion exchange resin
©
2015 Elsevier B.V. All rights reserved.
1
. Introduction
inorganic compounds (such as potassium permanganate, chromic
acid or nitric acid) is a known process but it has negative effects
on the environment due to the formation of undesired waste [6].
Therefore, heterogeneous catalytic oxidation of glycerol is a high
potential and environmentally friendly alternative for producing
these compounds of high economic value. Among them, the glyc-
eric acid (2,3-dihydroxypropanoic acid) is widely used in medicine
since it acts as glucose metabolite. It is also used as precursor for
aminoacids synthesis, such as serine.
The growing production of biodiesel by triglycerides transesteri-
fication has generated an excess of glycerol in the world market,
since approximately 100 kg of glycerol is produced for each ton
of biodiesel. This led to a decrease in the price of glycerol, and in
some cases it has been disposed by incineration [1–3]. Therefore,
it is important to develop new applications for glycerol in order
to obtain chemicals with added value, what would significantly
improve the profitability of the biodiesel production process.
Glycerol is a highly functionalized molecule which could be used
to obtain many useful compounds through oxidation, hydrogenol-
ysis, dehydration, esterification, and/or polymerization [4,5]. One
of the routes that have been most intensively studied over the past
few years is the oxidation. Selective oxidation of glycerol leads to
many valuable oxygenated compounds such as glyceric, tartronic,
glycolic and hydroxypyruvic acids and dihydroxyacetone.
The catalytic oxidation of glycerol in liquid phase has been
extensively studied in the last years, using both mono and bimetal-
lic catalysts on different supports. The metals used in these studies
included Pt [6,9–11], Au [6,12–17], Pd [6,18–21], Pt-Bi [22–25], Au-
Pd [20,21,26–30] and Au-Pt [12,31] supported on carbonaceous
materials (carbon black, graphite, activated carbon, carbon nano-
tubes) and oxides (TiO , CeO , Al O ). It has been demonstrated
2
2
2
3
that the activity and selectivity of the reaction strongly depend on
the reaction conditions (temperature, pH, metal/glycerol ratio).
The pH of the reaction media plays a key role during glycerol
oxidation. It is generally accepted that the presence of a base is
beneficial, because higher reaction rates are obtained in this con-
dition [11]. It has been reasonably established that in basic media,
a proton is removed from the glycerol primary hydroxyl groups, as
the first step of the reaction mechanism [6,9]. Nevertheless, other
The oxidation products of glycerol are actually synthesized from
chemical processes [6] or through low-productivity fermentation
processes [7,8]. In the first method, oxidation of glycerol with
∗ Corresponding author. Tel.: +54 342 4533858; fax: +54 342 4531068.
E-mail address: querini@fiq.unl.edu.ar (C.A. Querini).
http://dx.doi.org/10.1016/j.apcata.2015.04.026
0
926-860X/© 2015 Elsevier B.V. All rights reserved.
Please cite this article in press as: M.S. Gross, et al., Appl. Catal. A: Gen. (2015), http://dx.doi.org/10.1016/j.apcata.2015.04.026

G Model
APCATA-15363; No. of Pages9
ARTICLE IN PRESS
M.S. Gross et al. / Applied Catalysis A: General xxx (2015) xxx–xxx
2
anionexch + (PtCl6)sol2− ↔ 2anionsol− + (PtCl6)exch2−
−
(2)
authors have succeeded in performing the glycerol oxidation in a
non-basic media. Au-Pt nanoparticles supported on H-mordenite
2
During the exchange step, that takes 30 min, it is important to
maintain the system agitated to assure homogeneity in the catalyst
particles. After this step, the metal was reduced with hydrazine
N H ) in the presence of NaOH at a pH of 14. The reactions that
take place during this step and the standard potentials of each of
them are the following:
[
32], MgO [33] or carbonaceous materials [34] were found to be
active for glycerol oxidation in the absence of a base. More recently,
Tongsakul et al. [35] also reported that Pt and Pt-Au nanoparticles
supported on hydrotalcites were able to oxidize glycerol to glyceric
acid without the use of a basic medium.
(
2 4
As regards the support, most authors agree that carbon-
supported catalysts are more active than most oxide-supported
catalysts [36,37]. Ion exchange resins offer extensive possibili-
ties for the deposition of metals in a controlled manner both for
monometallic and bimetallic catalysts. Moreover, this support has
the advantage of allowing the use of fixed bed reactor configuration,
which is of particular interest for the possible industrial application
of the process. In addition, the ion exchange resin is highly stable in
a wide range of pH. However, there is only one article in the litera-
ture in which this support was employed for the catalytic oxidation
of glycerol [12].
−
+ (PtCl6)exch2− ↔ 2Clsol− + (PtCl4)exch2−
◦
E
2
e
= 0.755 V
(3)
−
2−
↔ 4Clsol− + Pt0
◦
2e + PtCl
4)exch
(
E
= 0.68 V
= 1.16 V
(4)
(5)
−
−
◦
E
N H + 4HO ↔ 4H O + N + 4e
2
4
2
2
According to the potential values in Eqs. (3)–(5), it can be
inferred that the reduction of platinum with hydrazine is spon-
taneous.
Finally, after the reduction step, the catalyst was rinsed with
distilled water until neutral pH of the eluted solution.
In this work, we studied platinum catalysts supported on weak
anion-exchange resin (Mitsubishi, WA-30) for the liquid phase oxi-
dation of glycerol to obtain glyceric acid. In a previous work [38]
it was reported that the concentration of the ion used during the
support treatment, previous to the exchange step, affects the metal
dispersion and its radial distribution along the resin particle. Such
ion is known as the competitor ion, since it competes with the anion
containing the active metal for the ion-exchange sites. Besides the
concentration, the type of competitor anion modifies the metal
distribution on the resin sphere, due to a different adsorption con-
stant of each anion. In this work, the effect of the competitor anion
as well as the reaction conditions in the catalytic performance of
Pt(1%)/WA30 catalysts for glycerol oxidation reaction, was studied.
2
.2. Support and catalysts characterization
Total exchange capacity. It is the number of exchange sites
present in the resin. The determination was performed using two
different methods: (a) following the methodology described in
ASTM D-2187 standard. This test method consists of converting the
sample to the chloride form, elution of chloride with sodium nitrate,
followed by determination of chloride ion in the eluted solution; (b)
−
by titration of the resins in the HO form with hydrochloric acid.
Water retention capacity. It corresponds to the amount of water
retained in the interstices of the polymer network. The determina-
tion was carried out following the methodology described in ASTM
D-2187. It consists in the determination of the mass lost during
drying at 377 ± 2 K for 18 ± 2 h.
2
. Experimental
Metal loading. The metal content in the catalyst was determined
by inductively coupled plasma atomic emission spectroscopy
2.1. Catalyst preparation
(ICP-OES) after digesting the samples with a mixture of nitric
and perchloric acids. Measurements were performed with a
PerkinElmer Optima 2100 DV equipment. In addition, the metal
content of the aqueous solution at the end of the exchange step
was also determined using the ICP technique, in order to verify the
effectiveness of the ion exchange step.
Microscopic analysis. The catalysts were analyzed using optical
microscopy, scanning electron microscopy (SEM), and transmis-
sion electron microscopy (TEM). The optical microscope used was
a Leica, model DM2500M. The electron microscope was a JEOL JSM
Platinum supported on ion exchange resin, prepared using hex-
achloroplatinic acid (H PtCl ), was used as catalyst for glycerol
2
6
oxidation. The resin used in this work was Mitsubishi WA-30, which
is a macroporous weak anion exchange resin, with tertiary amine
functional groups. The nominal platinum loading was 1 wt%, based
on the mass of wet resin. The resin is provided in a basic form,
−
i.e. with HO ions in the exchange sites. Before adding the metal,
the resin was pretreated with different competitor anions in order
to replace the hydroxyl anion. This step is represented by Eq. (1),
which shows the equilibrium between the anions present in the
3
5C model. The observation was made in the mode of secondary
−
−
electron images using an accelerating voltage of 20 kV. Elemental
chemical analyses were also carried out by X-ray, using an electron
probe micro-analysis (EPMA) technique with energy-dispersive
system (EDAX), attached to the scanning electron microscope. With
this technique it was possible to determine the radial profile of the
metal concentration in the catalyst, taking a measurement every
resin (HOexch ) and those present in the liquid media (anionsol ),
which are the competitor anion. It has already been reported that
the type and concentration of the anion used in the pretreat-
ment influence the radial distribution of the metal in the catalyst
[
38,39], this being the main reason for this step. The anions used as
competitor in this study were chloride (hydrochloric acid), iodide
potassium iodide) and citrate (citric acid). The concentrations of
1
0 ␮m starting from the edge of the particle. The TEM analyses
(
were carried out with a JEOL microscope model 100CXII operated
at 100 kV. The Pt dispersion was calculated applying the equation
all these solutions were 1.4 M, using one volume of solution per
volume of treated resin. This concentration was chosen according
to the results previously obtained when the effect of the competi-
tor ion concentration was studied [38]. These anions incorporated
into the resin, were then exchanged in a second step by the anion
[
40]:
0
Dva
.821
DPt =
(6)
containing the noble metal (PtCl6)2 . Before adding the solution
containing the metal precursor, the resin was drained and two vol-
umes of deionized water per volume of resin were added. Eq. (2)
describes the equilibrium condition between the competitor anion
in the resin and the precursor anion present in the liquid phase.
−
where DPt is the dispersion of platinum particles and Dva is a
volume–area average size defined as follows:
ꢀ
3
ni · d
i
i
i
Dva = ꢀ
(7)
2
i
ni · d
−
−
−
−
HOexch + anionsol ↔ HOsol + anionexch
(1)
being n the number of particles with diameter d .
i
i
Please cite this article in press as: M.S. Gross, et al., Appl. Catal. A: Gen. (2015), http://dx.doi.org/10.1016/j.apcata.2015.04.026

G Model
APCATA-15363; No. of Pages9
ARTICLE IN PRESS
M.S. Gross et al. / Applied Catalysis A: General xxx (2015) xxx–xxx
Table 1
3
XPS: The catalysts were analyzed by X-ray Photoelectron Spec-
troscopy (XPS) using a multitechnique system (SPECS), equipped
with a dual Mg/Al X-ray source and a hemispherical PHOIBOS 150
analyzer operating in the fixed analyzer transmission (FAT) mode.
The spectra were obtained with pass energy of 30 eV and an Al
anode operated at 100 W. The pressure was maintained below
Selectivity coefficients of anions compared to the hydroxyl ion on functionalized
styrene–divinylbenzene anion exchange resins (Eq. (1)).
Competitor ion
K
Chloride
Iodide
Citrate
22
175
220
−
8
2
× 10 mbar during the analysis. The samples were supported on
Cu double sided tape and evacuated in the prechamber for 12 h
before analysis. The spectra were processed using the software Casa
XPS (Casa Software Ltd., UK). The binding energy was calibrated
with respect to the C1s peak (284.6 eV) of contaminated carbon.
measuring the amount of Pt in the liquid phase by ICP at differ-
ent times, in order to quantify the metal losses by leaching in the
reaction medium. This was corroborated by measuring the metal
loading in the catalyst, determined by ICP, at the start and at the
end of the reaction cycles.
2.3. Activity tests
Glycerol oxidation reactions were carried out in liquid phase,
3
. Results and discussion
using a 150 mL capacity stainless steel reactor. This reactor is dis-
continuous in the liquid phase and continuous in the gas phase.
Pure oxygen was used as oxidant, which was fed from the reac-
tor bottom with a pressure of 150 kPa. The glycerol/metal molar
3.1. Total exchange capacity and water retention capacity.
The total exchange capacity of the resin was 2.54 meq g −1. This
◦
ratio (GOH/M ) was varied between 700 and 1400, and the sodium
result is comparable with that reported in a previous work, in which
the thermal stability of the ion exchange sites was analyzed [38].
The water retention capacity of the resin was also determined.
This parameter is related to the crosslinking degree of the poly-
mer chains; the higher the water retention capacity, the higher
the crosslinking degree. The value found was 51.7%. Monitoring
these parameters as the catalyst is being used in consecutive reac-
tion cycles is useful to monitor the stability of the support. A
decrease in the water retention capacity can be expected if the
macropores become clogged by the impurities present in the reac-
tion media, and/or the polymer degrades loosing the crosslinking
degree. Therefore, these two parameters were also determined in
the catalysts used in this study. The results were similar to those
found for the fresh resin, indicating that the reduced metal neither
clogs the exchange sites, nor it affects the macropores volume, even
after 17 cycles of reaction, as it will be shown in Section 3.4.
hydroxide/glycerol molar ratio (NaOH/GOH) was varied between 2
and 3. Reaction temperature was varied between 303 K and 323 K.
The concentration of the reactants and products were followed by
high pressure liquid chromatography (HPLC), using a BioRad HPX-
8
7H column with UV and refractive index (RI) detector. The mobile
−
1
phase used was 5 mM H SO4 with a flow rate of 0.6 ml min . The
2
products were identified and quantified by comparison with pure
standards, after obtaining the calibration curves.
The conversion was determined by monitoring the glycerol con-
centration with time:
0
t
M
− M
GOH
GOH
xGOH
=
(8)
0
M
GOH
where xGOH is the glycerol conversion; M0
is the glycerol initial
GOH
molar concentration; Mt
is the glycerol molar concentration at t
GOH
time.
The selectivity to compound i was determined as indicated in
Eq. (9):
3.2. Catalysts preparation and characterization
Mit
Different competitor ions have been used during the catalysts
ni
S = ꢁ
ꢂ ·
(9)
preparation. In all cases, the platinum content measured by ICP,
was very close to the nominal value. Moreover, the metal con-
tent of the aqueous phase obtained at the end of the exchange
step (determined by ICP) was undetectable, indicating that all the
platinum initially present in the liquid solution had been effec-
tively exchanged and incorporated to the resin exchange sites. The
equilibrium constants values of the exchange reaction (Eq. (1)) for
different competitor ions are shown in Table 1. These values were
provided by the manufacturer of the commercial resin (WA30-
Mitsubishi). A larger value of the equilibrium constant indicates
a major preference of the resin for this ion. Therefore, in the next
step of the preparation procedure, in which the anion of the metal
precursor is added to the system, it will find more resistance to
replace the anions previously exchanged. As a consequence, it is
expected to penetrate deeper into the resin particle. Fig. 1 shows
the platinum concentration profiles for the catalysts prepared with
different competitor anions, obtained using an energy-dispersive
X-ray analysis (EDX) system attached to the SEM instrument. Sam-
ples have been conditioned before the analysis, by rinsing with a
concentrated solution of sodium chloride, to elute the various anion
exchanged after the reduction step. This operation was performed
to unify the ion against which the metal was quantified, because
this technique only detects elements with molecular weight higher
than 23 (sodium). Assuming that all of the exchange sites were
exchanged with chloride, and knowing the density of exchange
sites, the metal concentration could be referred to the resin
i
0
t
3
M
− M
GOH
GOH
t
where M is the molar concentration of compound i at time t, and
i
ni is the carbon atoms number of the corresponding compound.
The turn over frequency (TOF) was calculated by dividing the
number of glycerol molecules consumed per unit time and per atom
of platinum exposed:
0
M
· (x
− xGOH1)
2
GOH2
GOH
TOF =
·
1
(10)
CPt · DPt · mcat · (t − t )
where CPt is the Pt concentration in the catalyst, mcat is the catalyst
mass charged in the reactor and t is the time. This TOF is an average
value between the reaction times t1 and t2.
In order to evaluate the reproducibility of the activity tests, a set
of five identical experiments was performed. The standard devia-
tion of the conversion and selectivity to glyceric acid obtained was
0
.014 and 1.214, respectively.
2
.4. Catalyst stability
The catalyst stability was determined by performing 17 consec-
utive reaction cycles, and comparing the activity and selectivity to
glyceric acid obtained in each test. Between cycles, the reaction
system was rinsed with deionized water. Additionally, the chem-
ical stability of the metal function in the catalyst was studied by
Please cite this article in press as: M.S. Gross, et al., Appl. Catal. A: Gen. (2015), http://dx.doi.org/10.1016/j.apcata.2015.04.026

G Model
APCATA-15363; No. of Pages9
ARTICLE IN PRESS
4
M.S. Gross et al. / Applied Catalysis A: General xxx (2015) xxx–xxx
1
1
1
4
2
0
1
citrate
iodide
chloride
0
.5
8
6
4
2
0
0
.5
0.4
0.3
0.2
0.1
1.0
0.8
0.6
0.4
0.2
0
r/Rº
Fig. 1. Metal radial distribution in the ion exchange resin particles.
Fig. 3. (A) SEM micrograph of the catalyst border showing the characteristic resin
surface; (B) TEM micrograph of the Pt particles in the resin supported catalyst.
Table 2
Average metal particle sizes, TEM particle size (Dva) and calculated dispersion for
catalysts prepared with different competitor ions.
Competitor ion
Arithmetic mean (nm)
Dva (nm)
DPt (%)
Chloride
Iodide
Citrate
1.93
1.68
1.94
2.08
1.73
2.02
40
48
41
Fig. 2. Optical micrograph of a hemisphere of catalyst, prepared with iodide as
competitor ion.
volume, i.e., in wt/vol% (g of Pt/100 mL of resin) through a series
of calculations based on the experimental data.
Fig. 1 clearly shows that the higher the equilibrium constant, the
greater the metal penetration into the resin. In the case in which
chloride was used as competitor ion, the metal penetrated only
platinum in the catalyst did not alter the morphology of the sup-
port. TEM micrographs showed a very good dispersion of metallic
nanoparticles in the resin, as it can be observed in Fig. 3B. Table 2
shows that similar average particle sizes were obtained when the
competitor ion was chloride or citrate, but when the catalyst was
prepared using iodide as competitor ion the average particle size
was smaller, although the difference was only 15%. According to
Fig. 4, the particle size distribution, calculated from TEM results,
was narrow, with very small platinum particles of around 2 nm.
Because the samples contain platinum particles of a very narrow
distribution it is possible to apply equation 6 with very little error.
Dispersion values of platinum particles are shown in Table 2. The
three catalysts have similar values, being slightly higher for the
catalyst prepared using iodide as competitor ion.
Catalyst surfaces were analyzed by XPS. Fig. 5 shows the XPS
spectra obtained for catalysts pretreated with chloride and iodide
in the Pt 4f region. The spectra corresponding to the catalyst pre-
treated with citrate were very similar to that obtained for the
catalyst pretreated with chloride and hence, they are not shown
in this figure. Table 3 presents the B.E. values for the Pt 4f peaks.
According to these data, the Pt 4f signals of the catalysts pretreated
with citrate and chloride are slightly higher than the expected
for metallic Pt (71.1 eV). However, according to Ryabenkova et al.
[41] the presence of small nanoparticles causes an apparent higher
5
0% of the particle, while when iodide or citrate was used, the metal
reached the center of the particle. The metal concentration near the
particle center was slightly higher for the case in which citrate was
used. For the three catalysts, a metal-rich area was observed, corre-
◦
sponding to a dimensionless radius (r/R ) range of approximately
1
–0.9. At this distance, the metal concentration was lower than
◦
half the value of its concentration at the particle surface (r/R = 1).
In this portion of the catalyst particle, the metal concentration was
significantly higher than the nominal load, i.e., an egg- shell distri-
bution was obtained. The metal-rich area is clearly visible, even to
the naked eye. The micrograph shown in Fig. 2 was obtained using
an optical microscope (Leyca DM2500M) to observe the internal
surface of a catalyst particle hemisphere. The micrograph shown
here corresponds to the catalyst prepared with iodide as competitor
ion, but similar micrographs were obtained for the other catalysts,
prepared with different competitor ions.
SEM micrographs showed that the catalysts retained the resin
morphology, as shown in Fig. 3A. This observation is in agree-
ment with the results previously discussed, i.e. the presence of 1%
Please cite this article in press as: M.S. Gross, et al., Appl. Catal. A: Gen. (2015), http://dx.doi.org/10.1016/j.apcata.2015.04.026

G Model
APCATA-15363; No. of Pages9
ARTICLE IN PRESS
M.S. Gross et al. / Applied Catalysis A: General xxx (2015) xxx–xxx
5
anions on the surface of Pt (1 1 1) [43] suggested that while chloride
adsorption is too small, iodide forms two commensurate adlayer
structures. According to Lucas et al. [43], chloride only interacts
with Pt when a positive potential is applied to this metal, while
iodide presents a stronger interaction, and the adsorption occurs
even in the absence of an applied electrical potential. The adsorp-
2
tion of iodide may occur after the reduction of the (PtCl ) −, which
6
takes place in the presence of the competitor ion that is released to
the solution during the ion-exchange, as indicated in Eq. (2).
3.3. Catalytic activity tests
All the reactions were performed at a relatively low oxygen pres-
sure (150 kPa). When higher pressures were used, it was found that
glycerol over-oxidation occurred, giving products such as acetic
acid, and decreasing the selectivity to glyceric acid. The oxygen
pressure used for this study allows obtaining high glyceric acid
yields while minimizing the formation of over-oxidized products.
Fig. 6 shows the glycerol conversion and average TOF obtained
with different catalysts. Results shown in Fig. 6A correspond to the
Fig. 4. Pt particle size distribution for catalysts prepared with different competitor
ions.
◦
following reaction conditions: GOH/M = 700, NaOH/GOH = 3, and
T = 303 K; and in Fig. 6B the conditions were similar except that the
temperature was increased to 323 K. In both figures it is observed
that the catalysts prepared with chloride and citrate as competitor
ions had similar activity levels. These catalysts have the same metal
particle size and similar dispersion values, as was shown in Table 2,
being the only difference the metal radial distribution in the poly-
mer matrix. This indicates that the metal distribution within the
catalyst particle has no significant influence on the activity thereof,
and therefore it can be inferred that there are no intraparticle mass
transfer limitations. The catalyst prepared with iodide anion as
competitor had reduced activity in spite of having a metal distribu-
tion similar to that of the catalyst prepared with citrate. Moreover,
the iodide catalyst is the one with the highest metal dispersion:
48% vs. 40–41% for the other catalysts (citrate, chloride). As above
mentioned, iodide can be adsorbed on Pt surface, thus reducing the
activity for the oxidation reaction. During the reaction in an alkaline
media, additional iodide ions can be liberated from the exchange
sites, by the hydroxide ions fed to the system in order to adjust
the pH. Thus, these iodide ions can be further adsorbed on the Pt
surface, reaction that is kinetically controlled [43], leading to faster
deactivation and lower activity. This phenomena does not occur
with the other competitor ions (chloride and citrate) since these
ions do not adsorb on Pt, as iodide does.
Comparing Fig. 6A and B it can be seen that for all catalysts,
a temperature increase led to a better conversion, although the
change in activity at the end of the reaction was only 6% approx-
imately for the catalysts prepared using chloride and citrate as
competitor ions, while it was 27% in the case of using iodide as
competitor ion. Also, during the 8 h of reaction, the differences in
conversion and TOF between the two temperatures were always
lower than expected considering only a kinetic effect according
to the Arrhenius Law. It has to be taken into account that in this
reaction system, the reaction in liquid phase depends both on the
concentration of glycerol and oxygen. As the temperature increases,
the oxygen solubility decreases, and this effect partially compen-
sates the increase in the kinetic constant.
iodide
chloride
8
0
78
76
74
72
70
68
Binding energy (eV)
Fig. 5. XPS spectra of the iodide and chloride catalysts. Pt 4f region.
binding energy when compared with the bulk material. There-
fore, the values measured for our catalysts are consistent with Pt0
nanoparticles. On the other hand, for the catalyst prepared with
the iodide ion as competitor, the Pt 4f signal can be deconvoluted
into two pairs of doublets, as shown in Fig. 5. The most intense
doublet, with binding energies of 71.0 eV (Pt 4f7/2) and 74.4 eV (Pt
f5/2), is assigned to metallic Pt. The other doublet, with peaks at
2.9 eV (Pt 4f7/2) and 76.2 eV (Pt 4f5/2) could be attributed to Ptı+
42]. According to the XPS results, 69.8% of the platinum exists as
Pt , whereas 30.2% is in the form of Pt . These are important results,
and indicate that the competitor ion not only may affect the Pt dis-
tribution inside the resin particle but also its chemical state, and
therefore, it can modify the catalytic performance. The reason of
this difference between the catalysts prepared using iodide and
chloride (or citrate) ions as competitor is not evident. However,
according to studies carried out regarding the adsorption of halides
4
7
[
0
ı+
Fig. 7 shows the glycerol conversion and TOF curves versus time
for the experiments carried out at 303 K and 323 K (Fig. 7A and
B respectively); with molar ratios GOH/M of 700, and NaOH/GOH
Table 3
Binding energies (B.E.) of the Pt 4f peaks for catalysts prepared with different com-
petitor ions.
◦
ratio of 2. It can be seen that an increase of the reaction temperature
has no significant influence on the conversion of glycerol, as also
observed in the results obtained using a NaOH/GOH ratio of 3. It
can also be seen that despite the conversions at 1 h reaction time
are similar for the three catalysts, the TOF obtained for the catalyst
pretreated with iodide is lower, since its dispersion was slightly
Competitor ion
B.E. (eV)
Pt 4f 5/2
Pt 4f 7/2
Chloride
Citrate
Iodide
74.92
74.61
74.4–76.2
71.58
71.30
71.0–72.9
Please cite this article in press as: M.S. Gross, et al., Appl. Catal. A: Gen. (2015), http://dx.doi.org/10.1016/j.apcata.2015.04.026

G Model
APCATA-15363; No. of Pages9
ARTICLE IN PRESS
6
M.S. Gross et al. / Applied Catalysis A: General xxx (2015) xxx–xxx
1
.0
.8
.6
citrate
chloride
iodide
citrate
chloride
iodide
6
4
2
0
0
0
.4
0
.2
0
A
0
0
1
2
3
4
5
6
7
8
0
1
2
3
4
5
6
7
8
t (h)
t (h)
1
.0
citrate
chloride
iodide
citrate
chloride
iodide
6
4
2
0
0
.8
.6
0
0.4
0.2
B
0
0
1
2
3
4
5
6
7
8
0
1
2
3
4
5
6
7
t (h)
8
t (h)
◦
Fig. 6. Conversion curves vs time for reactions carried out at GOH/M 700, NaOH/GOH 3, with different catalyst; (A) 303 K; (B) 323 K.
higher, as shown in Table 2. In addition, it is evident that the TOF
corresponding to this catalyst considerably decreases within the
first three hours of reaction. This suggests that a deactivation is
taking place and it is related to the presence of iodide in the system,
as above discussed.
Comparing the curves of Fig. 7A with Fig. 6A it can be inferred
that decreasing the NaOH/GOH ratio has no significant influence on
the activity of the catalysts prepared with chloride or citrate, since
the glycerol conversion and TOF values are identical in both figures.
In the case of the catalyst prepared with iodide ion as competitor,
a decrease of the NaOH/GOH ratio led to a decrease in the glyc-
erol conversion when the reaction was carried out at 323 K. Garcia
et al. [9] suggested that at a basic pH, an increment of the hydroxyls
concentration (higher NaOH/GOH ratio) favors the deprotonation
of the hydroxyl groups of glycerol (first step in the reaction mech-
anism) or the acid desorption from the metal sites (final step in the
reaction mechanism), leading to an increment in the reaction rate.
The initial value of TOF for the catalyst prepared with iodide was
1.0
0.8
0.6
citrate
chloride
iodide
citrate
6
chloride
iodide
4
0.4
2
0
0
.2
0
A
0
1
2
3
4
5
6
7
t (h)
8
0
1
2
3
4
5
6
7
8
t (h)
1
0
0
.0
.8
.6
citrate
citrate
chloride
iodide
6
4
2
0
chloride
iodide
0
.4
0.2
B
0
0
1
2
3
4
5
6
7
8
0
1
2
3
4
5
6
7
t (h)
8
t (h)
◦
Fig. 7. Conversion curves vs time for reactions carried out at GOH/M 700, NaOH/GOH 2, with different catalyst; (A) 303 K; (B) 323 K.
Please cite this article in press as: M.S. Gross, et al., Appl. Catal. A: Gen. (2015), http://dx.doi.org/10.1016/j.apcata.2015.04.026

G Model
APCATA-15363; No. of Pages9
ARTICLE IN PRESS
M.S. Gross et al. / Applied Catalysis A: General xxx (2015) xxx–xxx
7
1.0
Glycerol
0
0
0
.4
.3
.2
Glyceric A.
0
0
0
0
.8
.6
.4
.2
Formic A.
Glycolic A.
0.1
0
Tartronic A.
Lactic A.
Oxalic A.
0
0
1
2
3
4
5
6
7
8
t (h)
Fig. 9. Reactant liquid phase composition vs time for Pt 1% (chloride) catalyst at
◦
3
03 K, NaOH/GOH ratio of 3 and GOH/M ratio of 700; triangles: final composition
Fig. 8. Selectivity evolution with time. Reaction conditions: 323 K, GOH/NaOH 3,
GOH/M 700. Catalyst: Pt(1%) (chloride).
◦
◦
obtained for the test done at the same conditions but GOH/M ratio of 1400.
higher for the case where the lower ratio NaOH/GOH was used, but
with the progress of the reaction these values became identical.
This result can also be explained considering the effect of iodine
on the platinum activity. Under oxidizing conditions, the following
reaction may occur:
60%, at 323 K, NaOH/GOH molar ratio of 3, and not dependent on
the GOH/M ratio in the range used in this work. It can also be
◦
observed that the selectivity to compounds of 2 or 1 carbon atoms
was favored with a NaOH/GOH molar ratio of 2 and low tem-
perature (303 K). Similar results were obtained with the catalyst
prepared with citrate as competitor ion: the best selectivity to glyc-
eric acid was obtained at 323 K, NaOH/GOH ratio of 3; and the major
selectivity to 1C and 2C compounds was obtained at the lowest
temperature and NaOH/GOH ratio (303 K and 2 respectively). These
results suggest that at lower temperature, the residence time of the
intermediates compounds on the surface is longer and the coverage
is higher, and therefore, the consecutive reactions occurs reaching
the final step of C–C bond breaking.
−
−
K = 1 × 10−6
O + 4I + 2H O ↔ 2I + 4HO
(11)
2
2
2
Iodine can also adsorb on the Pt surface, thus leading to a faster
deactivation. This reaction explains the effect of the NaOH/GOH
ratio. As the NaOH concentration decreases, the reaction shifts
towards the products, thus increasing the concentration of iodine
and consequently, the catalyst deactivation. Note that the value
of the TOF (Figs. 6 and 7) as reaction time progresses decreases,
indicating that the iodide-exchanged catalyst notably deactivates
compared to the other catalysts.
The catalyst prepared with iodide ion as competitor showed
selectivities to glyceric acid significantly higher than the corre-
sponding to the other catalysts, which is related to the differences
between the Pt active sites discussed in the XPS characterization
section, and the possible iodide adsorption. The best selectivity val-
ues to glyceric acid were obtained at temperatures and NaOH/GOH
ratios different from the other catalysts, but with the lowest
In order to assess whether any of the reaction products is
adsorbed in the ion-exchange resin, thus masking the activity
results, ion exchange experiments were performed using aliquots
of the liquid obtained at the end of the reaction and fresh resin.
It was concluded that none of the compounds were exchanged on
the exchange resin sites, since the chromatographic analysis of the
liquid sample prior to contact and after 24 h contact with the resin,
yielded identical concentration for all of the reaction products.
The selectivity values for different compounds remained con-
stant during the reaction for the catalysts tested at different
reaction conditions, as shown in Fig. 8 for the catalyst pre-
pared using chloride in the pretreatment. This tendency was also
observed with the other catalysts, and at different reaction condi-
tions, i.e., in all the experiments performed the selectivity remained
constant with reaction time, and thus, with the conversion. This
indicates that during the reaction there were not changes on the
catalyst surface, or if there were modifications such as iodine/iodide
adsorption on the Pt, it only led to a change in activity. Table 4
shows the selectivity values to the different substances at the end
of the reaction (8 h). It is emphasized that the selectivity remained
constant throughout the reaction, i.e., it does not depends on the
conversion or time, as shown in Fig. 8. For clarity reasons, the com-
pounds are grouped according to the number of carbon atoms of
the molecule. Because there is particular interest in the produc-
tion of glyceric acid, substances with 3 carbon atoms are detailed,
showing the selectivities to tartronic, glyceric and lactic acids. The
◦
GOH/M ratio. The higher selectivity to 1C and 2C substances was
observed in the experiments performed at lower temperatures and
lower NaOH/GOH ratio, which is a similar result to that obtained
with chloride. It was possible to obtain selectivity to glyceric acid
close to 80%.
Fig. 9 shows the evolution of compositions in the reacting phase
with time corresponding to an experiment performed with the cat-
alyst prepared with chloride, at 303 K, NaOH/GOH ratio of 3 and
◦
GOH/M ratio of 700. It is also shown in this figure the composition
obtained at the end of the reaction (8 h) in the same experimen-
◦
◦
tal conditions, except for the GOH/M ratio (triangles, GOH/M
of 1400). Triangles were placed at 5 h reaction time, in order to
◦
show that the reaction mechanism is not affected by the GOH/M
ratio, since almost the same product distribution (selectivity) was
obtained in the two experiments at the same conversion level.
◦
Therefore, the GOH/M ratio only influences the reaction rate, and
does not affect the selectivity, being therefore possible to increase
the amount of catalyst, to speed up the process, maintaining the
selectivity.
3.4. Catalyst stability.
1
C column corresponds to formic acid, which was the only 1C com-
pound found in our experiments, while the compounds in column
Fig. 10 shows the glyceric acid yield obtained in 17 consecutive
reactions performed with the citrate treated catalyst. The results
obtained reflect the remarkable stability of the platinum/resin cat-
alysts prepared in this work, contrary to what was found in other
2
C include oxalic and glycolic acids.
Using the catalyst prepared with chloride ion as competitor, it
was possible to achieve selectivity to glyceric acid of approximately
Please cite this article in press as: M.S. Gross, et al., Appl. Catal. A: Gen. (2015), http://dx.doi.org/10.1016/j.apcata.2015.04.026

G Model
APCATA-15363; No. of Pages9
ARTICLE IN PRESS
M.S. Gross et al. / Applied Catalysis A: General xxx (2015) xxx–xxx
8
Table 4
Conversion (x) and selectivity of Pt/resin catalysts at different reaction conditions.
Catalyst
Reaction conditionsa
Selectivity (%)
x
1
Cb
2Cc
3C
Tartronic
Glyceric
Lactic
303-3-700
303-3-1400
303-2-700
303-2-1400
323-3-700
323-3-1400
323-2-700
323-2-1400
8.7
9.9
10.3
11.4
7.4
7.5
10.0
9.5
15.6
14.9
17.5
17.5
12.4
11.3
16.2
14.4
11.5
6.6
8.7
5.8
9.6
8.3
10.2
7.3
56.9
62.3
58.3
58.9
60.8
61.8
52.4
58.5
7.3
6.4
5.3
6.3
9.8
11.1
11.2
10.3
0.84
0.69
0.88
0.62
0.89
0.62
0.90
0.72
Chloride
303-3-700
303-3-1400
303-2-700
303-2-1400
323-3-700
323-3-1400
323-2-700
323-2-1400
7.4
9.6
9.9
10.5
7.8
6.6
12.8
15.5
15.5
17.5
15.0
10.3
15.7
13.3
12.6
7.6
8.6
6.1
8.2
5.9
10.6
6.4
61.3
60.4
61.4
60.8
63.0
62.1
55.8
58.6
5.8
7.1
4.6
5.2
6.1
15.1
9.4
13.6
0.85
0.61
0.83
0.58
0.91
0.44
0.91
0.68
Citrate
8.6
8.3
303-3-700
303-3-1400
303-2-700
303-2-1400
323-3-700
323-3-1400
323-2-700
323-2-1400
6.0
6.4
7.8
8.1
5.4
6.5
5.7
7.9
6.4
6.7
8.1
9.3
6.4
7.2
5.6
7.8
5.0
5.6
5.6
4.5
5.4
6.8
3.9
5.3
78.7
75.3
72.3
73.0
77.5
68.5
79.4
68.6
3.9
5.9
6.2
5.2
5.3
11.1
5.4
10.4
0.48
0.35
0.43
0.30
0.59
0.58
0.47
0.54
Iodide
a
◦
Temperature (K) – NaOH/GOH molar ratio – GOH/M .
Formic acid.
Oxalic acid and glycolic acid.
b
c
competitor ion used previous to the metal exchange affects the
platinum penetration into the resin particle: the higher the equilib-
rium adsorption constant of the competitor anion, the greater the
platinum penetration.
Three different competitor ions were used in this study: chlo-
ride, citrate and iodide. The catalysts prepared with chloride and
citrate as competitor ions showed similar activity for the glycerol
oxidation reaction in spite of their different metal distributions. On
the other hand, the catalyst prepared using iodide anion as com-
petitor, which has a similar metal distribution to that of the citrate
one, showed lower activity. It is concluded that the catalytic activ-
ity is not significantly affected by the platinum distribution on the
resin particle.
Fig. 10. Glyceric acid selectivity (dashed bars) and glycerol conversion (solid bars)
The reason for the different performance of the catalyst pre-
pared by pretreating with iodide ions was attributed to the different
chemical state of platinum in this catalyst, as shown by XPS. More-
over, not only the activity was affected but also the selectivity,
since this catalyst showed selectivities to glyceric acid significantly
higher than the corresponding to the other catalysts. This behavior
is due to the iodide adsorption on Pt that modifies the electron den-
sity. According to this investigation, the nature of the competitor
ion can modify the platinum active sites, thus influencing the activ-
ity and selectivity of the resin supported catalyst. It is an important
finding, that the Pt modification by iodide–iodine adsorption may
substantially improve the selectivity in the glycerol oxidation to
glyceric acid.
obtained in 17 consecutive reactions. Reaction conditions: Pt 1% (citrate) catalyst,
◦
3
23 K, NaOH/GOH 3, GOH/M 700.
studies, in which it was claimed that platinum based catalysts
are prone to suffer deactivation due to oxygen poisoning [44–46].
Moreover, the liquid phase was analyzed by ICP after reaction (in
selected experiments), in order to quantify the possible metal losses
by leaching in the reaction medium. In all cases, the platinum con-
tent was undetectable. The metal loading of the fresh catalyst,
before the first reaction cycle and on the catalyst used in 17 cycles
was identical and close to nominal value. Thus, it is concluded that
on the prepared catalysts, the platinum function neither deacti-
vates by oxygen poisoning, nor by leaching on the reaction medium.
This is a very important result, since the catalyst stability is a key
parameter in the process, being possible to reuse the catalyst in
many reaction cycles.
Acknowledgements
The authors wish to acknowledge the financial support received
from CONICET (PIP 2010-093), UNL (PACT 69) and ANPCyT (PICT
4
. Conclusions
2
010-1526). In addition, we thank ANPCyT for the purchase of
Platinum catalysts supported on ion-exchange resins were used
for the selective oxidation of glycerol. It was found that the
multi-technical analysis instrument SPECS (PME8-2003)
Please cite this article in press as: M.S. Gross, et al., Appl. Catal. A: Gen. (2015), http://dx.doi.org/10.1016/j.apcata.2015.04.026

G Model
APCATA-15363; No. of Pages9
ARTICLE IN PRESS
M.S. Gross et al. / Applied Catalysis A: General xxx (2015) xxx–xxx
9
References
[24] N. Wörz, A. Brandner, P. Claus, J. Phys. Chem. C 114 (2010) 1164–1172.
[
[
[
[
[
[
25] W. Hu, B. Lowry, A. Varma, Appl. Catal. B 106 (2011) 123–132.
26] A. Villa, C. Campione, L. Prati, Catal. Lett. 115 (2007) 133–136.
27] L. Prati, A. Villa, C. Campione, P. Spontoni, Top. Catal. 44 (2007) 319–324.
28] W.C. Ketchie, M. Murayama, R.J. Davis, J. Catal. 250 (2007) 264–273.
29] A. Villa, D. Wang, G.M. Veith, L. Pratti, J. Catal. 292 (2012) 73–80.
[
[
[
[
[
[
1] X.L. Chen, Y.G. Zheng, Y.C. Shen, Chem. Rev. 107 (2007) 1777–1830.
2] J.J. Bozell, Science 329 (2010) 522–523.
3] C.H. Zhou, J.N. Beltramini, Y.X. Fan, G.Q. Lu, Chem. Soc. Rev. 37 (2008) 527–549.
4] F. Jérôme, Y. Pouilloux, J. Barrau, ChemSusChem 1 (2008) 586–615.
5] A. Corma, S. Iborra, A. Velty, Chem. Rev. 107 (2007) 2411–2502.
6] S. Carrettin, P. McMorn, P. Johnston, K. Griffin, C.J. Kiely, G.J. Hutchings, Phys.
Chem. Chem. Phys. 5 (2003) 1329–1336.
30] A.A. Rodriguez, C.T. Williams, J.R. Monnier, Appl. Catal.
61–168.
A 475 (2014)
1
[
[
[
31] S. Demirel, K. Lehnert, M. Lucas, P. Claus, Appl. Catal. B 70 (2007) 637–643.
32] A. Villa, G.M. Veith, L. Prati, Angew. Chem. Int. Ed. 49 (2010) 4499–4502.
33] G.L. Brett, Q. He, C. Hammond, P.J. Miedziak, N. Dimitratos, M. Sankar, A.A.
Herzing, M. Conte, J.A. Lopez-Sanchez, C.J. Kiely, D.W. Knight, S.H. Taylor, G.J.
Hutchings, Angew. Chem. Int. Ed. 50 (2011) 10136.
[
[
7] S. Varga, R. Bauer, N. Katsikis, D. Hekmat, Chem. Eng. Technol. 76 (2004) 1306.
8] H. Habe, T. Fukuoka, D. Kitamoto, K. Sakaki, Appl. Microbiol. Biotechnol. 84
(
2009) 445–452.
[9] R. Garcia, M. Besson, P. Gallezot, Appl. Catal. A 127 (1995) 165–176.
[
[
34] M. Zhang, D. Liang, R. Nie, X. Lu, P. Chen, Z. Hou, Chin. J. Catal. 33 (2012)
[
10] D. Liang, J. Gao, J. Wang, P. Chen, Z. Hou, X. Zheng, Catal. Commun. 10 (2009)
586–1590.
11] D. Liang, J. Gao, H. Sun, P. Chen, Z. Hou, X. Zheng, Appl. Catal. B 106 (2011)
23–432.
1340–1346.
1
35] D. Tongsakul, S. Nishimura, C. Thammacharoen, S. Ekgasit, K. Ebitani, Ind. Eng.
Chem. Res. 51 (2012) 16182–16187;
[
4
D. Tongsakul, S. Nishimura, K. Ebitani, ACS Catal. 3 (2013) 2199–2207.
36] L. Prati, M. Rossi, J. Catal. 176 (1998) 552.
37] S. Demirel, P. Kern, M. Lucas, P. Claus, Catal. Today 122 (2007) 292–300.
38] M.S. Gross, M.L. Pisarello, K.A. Pierpauli, C.A. Querini, Ind. Eng. Chem. Res. 49
[
[
12] A. Villa, C.E. Chan-Thaw, L. Pratti, Appl. Catal. B 96 (2010) 541–547.
13] P. Suramanee, S. Poompradub, R. Rojanathanes, P. Thamyongkit, Catal. Lett. 141
[
[
[
(
2011) 1677–1684.
14] S. Gil, M. Marchena, L. Sánchez-Silva, A. Romero, P. Sánchez, Chem. Eng. J. 178
2011) 423–435.
15] E.G. Rodrigues, M.F.R. Pereira, X. Chen, J.J. Delgado, J.J.M. Órfão, J. Catal. 281
2011) 119–127.
[
[
(
2010) 81–88.
(
[
39] S. Oeckl, F. Martinola, P. Thomas, U.S. Patent 4,853,135 (1989).
[
40] M.S. Zanuttini, C.D. Lago, C.A. Querini, M.A. Peralta, Catal. Today 213 (2013)
(
9–17.
[
[
16] E.G. Rodrigues, M.F.R. Pereira, J.J.M. Órfão, Appl. Catal. B 115–116 (2012) 1–6.
17] S. Gil, M. Marchena, C.M. Fernández, L. Sánchez-Silva, A. Romero, J.L. Valverde,
Appl. Catal. A 450 (2013) 189–203.
[
[
41] Y. Ryabenkova, Q. He, P.J. Miedziaka, N.F. Dummer, S.H. Taylor, A.F. Carleya, D.J.
Morgana, N. Dimitratos, D.J. Willocka, D. Bethell, D.W. Knight, D. Chadwickc,
C.J. Kiely, G.J. Hutchings, Catal. Today 203 (2013) 139–145.
[18] N. Dimitratos, F. Porta, L. Prati, Appl. Catal. A 291 (2005) 210–214.
[19] D. Wang, A. Villa, F. Porta, L. Prati, D. Su, J. Phys. Chem. C 112 (2008) 8617–8622.
[20] D. Wang, A. Villa, F. Porta, D. Su, L. Prati, Chem. Commun. (2006) 1956–1958.
[21] N. Dimitratos, J.A. Lopez-Sanchez, D. Lennon, F. Porta, L. Prati, A. Villa, Catal.
Lett. 108 (2006) 147–153.
42] R. Nie, D. Liang, L. Shen, J. Gao, P. Chen, Z. Hou, Appl. Catal. B 127 (2012)
212–220.
[
[
43] C.A. Lucas, N.M. Markovic-acute, P.N. Ross, Phys. Rev. B 55 (1997) 7964.
44] L. Prati, A. Villa, C.E. Chan-Taw, R. Arrigo, D. Wang, D.S. Su, Faraday Discuss. 152
(
2011) 353–365.
[
[
22] H. Kimura, K. Tsuto, T. Wakisaka, Y. Kazumi, Y. Inaya, Appl. Catal. A 96 (1993)
[
[
45] W.C. Ketchie, M. Murayama, R.J. Davis, Top. Catal. 44 (2007) 307–317.
46] C.L. Bianchi, P. Canton, N. Dimitratos, F. Porta, L. Prati, Catal. Today 102–103
2
17–228.
23] A. Brandner, K. Lehnert, A. Bienholz, M. Lucas, P. Claus, Top. Catal. 52 (2009)
78–287.
(
2005) 203–212.
2
Please cite this article in press as: M.S. Gross, et al., Appl. Catal. A: Gen. (2015), http://dx.doi.org/10.1016/j.apcata.2015.04.026