Synthesis of very highly dispersed platinum catalysts supported on carbon xerogels by the strong electrostatic adsorption method

Article abstract of DOI:10.1016/j.jcat.2008.10.014

Highly dispersed Pt/carbon xerogel catalysts are obtained by applying the "strong electrostatic adsorption" (SEA) of hexachloroplatinic acid to carbon xerogels (PZC = 9.4) and platinum tetraammine chloride to oxidized carbon xerogels (PZC = 2.4). After th

Full text of DOI:10.1016/j.jcat.2008.10.014

Journal of Catalysis 261 (2009) 23–33
Contents lists available at ScienceDirect
Journal of Catalysis
www.elsevier.com/locate/jcat
Synthesis of very highly dispersed platinum catalysts supported on carbon
xerogels by the strong electrostatic adsorption method
Stéphanie Lambert a,∗, Nathalie Job a, Lawrence D’Souza b, Manuel Fernando Ribeiro Pereira c, René Pirard a,
Benoît Heinrichs a, José Luis Figueiredo c, Jean-Paul Pirard a, John R. Regalbuto b
a
Laboratoire de Génie Chimique, B6a, Université de Liège, B-4000 Liège, Belgium
Chemical Engineering Department, University of Illinois at Chicago, 810 S. Clinton Street, Chicago, IL 60607, USA
Laboratório de Catálise e Materiais (LCM), Laboratório Associado LSRE/LCM, Departamento de Engenharia Química, Faculdade de Engenharia, Universidade do Porto,
b
c
Rua Dr. Roberto Frias, 4200-465 Porto, Portugal
a r t i c l e i n f o
a b s t r a c t
Article history:
Highly dispersed Pt/carbon xerogel catalysts are obtained by applying the “strong electrostatic adsorption”
(SEA) of hexachloroplatinic acid to carbon xerogels (PZC = 9.4) and platinum tetraammine chloride to
oxidized carbon xerogels (PZC = 2.4). After the reduction step, all these Pt/carbon xerogel catalysts
display a very high level of metal dispersion: very small platinum particles (1.1–1.3 nm) are observed
by TEM. Pt particle sizes obtained by CO chemisorption are in good agreement with TEM micrographs,
which shows that the metal is accessible to reactants. These Pt/carbon xerogel catalysts are very active
for the hydrogenation of benzene into cyclohexane.
Received 11 September 2008
Revised 24 October 2008
Accepted 27 October 2008
Available online 22 November 2008
Keywords:
Strong electrostatic adsorption
Platinum catalysts
© 2008 Elsevier Inc. All rights reserved.
Carbon xerogel
Aqueous sol–gel process
Benzene hydrogenation
1. Introduction
tential losses. In the latter field, nanostructured carbons constitute
an interesting alternative to carbon blacks.
Recently, carbon aerogels and xerogels were used to replace
For several years, the synthesis of Pt/C fuel cell electrocatalysts
has been in a great expansion. The catalytic layer of proton ex-
change membrane fuel cells (PEMFCs) is indeed a key element to
the cell performance; in particular, the carbon pore texture as well
as the platinum dispersion and crystalline structure are known
to strongly influence the electrode performance [1]. Most of the
air/H2 PEMFC electrodes use platinum catalysts supported on car-
bon blacks, which are composed of carbon particles (10–30 μm)
assembled together as aggregates. The packing of these aggregates,
and therefore the pore structure of the catalytic layer, depends on
the carbon black nature and on the electrode processing. Typically,
at the air-fed cathode, where oxygen, proton and water transports
are involved, high potential losses due to diffusion limitations off-
set the cell performances; this problem is compensated by increas-
ing the metal loading, to the detriment of the cost of such devices.
As detailed by Gasteiger et al. [2], the necessary reduction of plat-
inum loading must be led in two ways: (i) improvement of the
catalyst activity and efficiency, either by optimizing the metal dis-
persion or by using appropriate alloys and (ii) improvement of the
electrode structure in order to decrease the diffusion-induced po-
classical carbon supports in catalysis [3–5] and electrocatalysis [6–
9] applications. These materials are obtained either by supercritical
drying or evaporative drying of organic gels, followed by pyrol-
ysis [10,11], and their pore texture is fully controllable within a
wide range via the synthesis variables of the pristine gel. With re-
gard to active carbons or carbon blacks, the advantages of carbon
aerogels and xerogels are their high purity and their pore texture
flexibility. For example, carbon xerogels were recently used as a Pd
and Pd–Ag catalyst supports in order to eliminate diffusional lim-
itations during a gas phase reaction [12]. In previous studies [7,8],
PEMFCs cathodes were prepared from Pt catalyst supported on car-
bon aerogels and xerogels. While carbon blacks are constituted of
aggregates connected through van der Waals bonds, carbon aerogel
and xerogel powders display monolithic structures at the microme-
ter scale; as a consequence, the pore texture of a carbon aerogel or
xerogel micromonolith remains identical in the catalytic layer of a
membrane–electrode assembly [7,8]. These previous works showed
that an adequate choice of the carbon pore texture could lead to a
significant decrease of the diffusion-induced potential losses. How-
ever, the highly loaded (35 wt%) Pt catalysts used in these studies,
and prepared by direct deposition of large metal amounts on the
support, presented low metal dispersions: up to 60% of the Pt
atoms constituted large particles (10–50 nm), which decreased the
Corresponding author. Fax: +32 4366 3545.
*
E-mail address: stephanie.lambert@ulg.ac.be (S. Lambert).
0021-9517/$ – see front matter © 2008 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2008.10.014

24
S. Lambert et al. / Journal of Catalysis 261 (2009) 23–33
Pt utilization ratio. In addition, the synthesis method was rather
long and sophisticated. In order to increase the efficiency of car-
bon aerogel and xerogel supported Pt catalysts, it is thus necessary
to develop a simple procedure to obtain Pt catalysts with relatively
high metal loading and optimal dispersion.
cursor solution. The resorcinol/formaldehyde molar ratio, R/F , was
fixed at 0.5 and the dilution ratio, D, i.e. the solvent/(resorcinol
+ formaldehyde) molar ratio, was chosen to be equal to 5.7. 99 g
of resorcinol (Vel, 99%) were first mixed with 188 ml of deion-
ized water in 500 ml vessels under stirring. After dissolution, the
pH value was increased close to the chosen starting value with
concentrated sodium hydroxide solutions (5 and 2 N). 135 ml
of formaldehyde solution (37 wt% in water, stabilized with 10–
15 wt% methanol) were added to the mixture, and the pH was
then adjusted exactly to the value chosen by the addition of di-
luted sodium hydroxide solution (0.5 N). In order to cover a wide
range of pore textures, the starting pH of the four solutions was
adjusted to 5.50, 5.75, 6.00, and 6.25 respectively. The vessels were
then sealed and heated up to 343 K for 72 h for gelling and aging.
The gels obtained were then dried under vacuum according to the
following procedure: the flasks were opened and put into a drying
oven at 333 K, then the pressure was progressively decreased down
to the minimum value (1200 Pa). This step was performed over
48 h. The samples were then kept at 423 K for 72 additional h.
After drying, pyrolysis was performed at 1073 K under flowing ni-
Much progress in the preparation of heterogeneous catalyst has
been made through the postulate of Brunelle that the adsorp-
tion of noble metal complexes onto common oxide supports was
essentially coulombic in nature [13]. The hydroxyl (–OH) groups
that populate oxide surfaces become protonated and so positively
charged below a characteristic pH value, while the same hydroxyl
groups become deprotonated and negatively charged above this
characteristic pH value. This pH, at which the surface is neutral, is
termed the point of zero charge (PZC). Brunelle explained that ox-
ides placed in solutions at pH values below their PZC would adsorb
anions such as hexachloroplatinate [PtCl6]2−; at pH values above
their PZC, the same support would adsorb cations such as platinum
tetraammine [Pt(NH3)4]2+. A few years ago, Regalbuto et al. stud-
ied this adsorption process, beginning with the chloroplatinic acid
(H2PtCl6 or CPA)/alumina system [14–17]. In this system, anionic
chloride (PtCl26−) and oxy-chloride (PtCl5(OH)2−, PtCl4(OH)(H2O)
)
−
trogen and following the same temperature program, as described
−1
complexes adsorb over a positively charged alumina surface in
the low pH range. This method, called the “strong electrostatic
adsorption” (SEA) method, has also extended to the platinum am-
monium chloride (Pt(NH3)4Cl2 or PTA)/silica system at high pH [18,
in previous studies [4,25,26]: (i) ramp at 1.7 K min
to 423 K and
−1
hold for 15 min; (ii) ramp at 5 K min
to 673 K and hold for 60
to 1073 K and hold for 120 min; and
−1
min; (iii) ramp at 5 K min
(iv) cool slowly down to room temperature.
19], where cationic complexes (Pt(NH3)2+) adsorb over negatively
4
Oxidized carbon xerogels were also prepared by treating the
above-obtained carbon materials with nitric acid. 20 g of each car-
bon xerogel were oxidized in an HNO3 aqueous solution (5 N) for
48 h at room temperature. Afterwards, oxidized carbon xerogels
were washed with deionized water until the pH of the wash-
ing solutions reached 4.5. These four samples were dried under
vacuum (1200 Pa) at 298 K for 24 h, and then under helium
(0.04 mmol s−1) at 473 K for 1 h to release a great number of
micropores. According to Figueiredo et al. [21], no surface oxygen
groups are removed at 473 K.
Finally, all unoxidized and oxidized carbon xerogels were
crushed and sieved between 100 and 500 μm before their use
for the synthesis of Pt/carbon xerogel catalysts by the SEA method.
The four carbon supports are labeled as follows: the letter X
(for “xerogel”) is followed by the initial pH of the precursor so-
lution multiplied by 100. For example, X-550 is a carbon xerogel
issued from the drying and pyrolysis of a gel prepared at an initial
pH of 5.50. In the case of the corresponding oxidized carbon sup-
ports, the abbreviation “Ox” follows the names of the initial carbon
xerogels (e.g. X-550-Ox).
charged surfaces. The latter two works especially demonstrate the
practical consequence that, when strongly adsorbed at the optimal
pH, the monolayer of adsorbed coordination complexes retains its
high dispersion through the catalyst pretreatment process such as
drying and reduction steps. Finally, the SEA method was applied
to the activated carbon surfaces because controlled oxidation of a
carbon surface at mild or rigorous conditions leads to a lesser or
greater amount of oxygen functional groups on the surface [20–
23], irreversibly altering the PZC and influencing the adsorption
of Pt complexes in a way that is systematic and controllable. In
a previous work on activated carbons [24], the highest-PZC ac-
tive carbons adsorb the largest amount of anions [PtCl6]2− and the
lowest amount of cations [Pt(NH3)4]2+ while the lowest-PZC ac-
tive carbons adsorb the lowest amount of anions and the largest
amount of cations. Pt uptakes reach a maximum with respect to
pH for a given metal precursor and active carbon PZC.
The aim of the present work was to apply the same method
to carbon xerogels in order to establish whether or not highly dis-
persed Pt catalysts can be prepared by optimizing the adsorption
pH with regard to both the carbon surface chemistry and the cho-
sen precursor. In this ambit, the PZC of the carbon xerogels, before
and after oxidation treatment, was first determined. Then, the in-
fluence of PZC alteration on the control of Pt uptake, using both
anionic and cationic Pt complexes, was studied. Finally, the impact
of the impregnation method on the metal dispersion was checked.
The final goal was to develop a simple and effective way to syn-
thesize highly loaded and highly dispersed Pt/carbon materials that
could be further used as electrocatalysts in PEM fuel cells.
2.1.2. Determination of the point of zero charge (PZC) of carbon xerogels
As explained in the Section 1, when using the SEA method, it
is very important to know the PZC of the support, that is the pH
value at which the electrical charge density on the support surface
is zero. Here, the PZCs of unoxidized and oxidized carbon xerogels
were determined by the method of Park and Regalbuto (equilib-
rium pH at high loading, EpHL) [27]: the porous solid is soaked
in water solutions of various starting pHs and after stabilization,
the pH is measured again. The PZC of the solid corresponds to a
plateau in a plot of the final pH vs the initial pH. A critical operat-
ing variable in a such system is the total material surface area in
solution, i.e. the surface loading, SL (units: m2 l−1). When compar-
ing samples with different surface areas, the mass of the sample
can be adjusted so as to achieve the same SL in each case. In the
2. Experimental and methods
2.1. Synthesis
2.1.1. Carbon xerogels synthesis
Four carbon xerogels with various pore textures were syn-
thesized by the evaporative drying and pyrolysis of resorcinol–
formaldehyde gels following a method described in a previous
study [25]. Basically, the gel was obtained by polycondensation
of resorcinol with formaldehyde in water, with NaOH as basifi-
cation agent, whose role is to settle the starting pH of the pre-
−1
present study, the SL was chosen to be equal to 104 m2 l
for all
the carbon xerogels, in line with previous studies [27,28]. There-
fore, before measuring the equilibrium pH, the mass of the eight
−1
carbon xerogels was adjusted so that SL = 104 m2 l
in 25 ml of
water.

S. Lambert et al. / Journal of Catalysis 261 (2009) 23–33
25
Fig. 1. PZC determination of (a) unoxidized carbon xerogels, and of (b) oxidized
carbon xerogels.
Using a spear-tip semi-solid electrode (Electrode Accumet), the
equilibrium pH of the eight carbon xerogels was measured over a
wide range of initial pH values (from pH 1 to 13) adjusted using
HCl or NaOH solutions (Fig. 1).
2.1.3. Determination of the optimal adsorption pH
In order to determine the optimal pH condition leading to
maximum metal loading, adsorption experiments were conducted
under various pH conditions, according to the PZC value of the
supports. Chloroplatinic acid (CPA, H2PtCl6) (Aldrich, 99.9%) and
tetraammineplatinum chloride (PTA, Pt(NH3)4Cl2) (Aldrich, 99.9%)
were used for adsorption on unoxidized and oxidized supports re-
Fig. 2. (a) Final pH versus initial pH curves and CPA uptake curves of unoxidized
carbon xerogels and (b) Final pH versus initial pH curves and PTA uptake curves of
oxidized carbon xerogels.
−1
−3
spectively. Aqueous solutions of CPA and PTA (5.1 × 10
mol l
)
2.1.4. Determination of the concentration of the minimal precursor
Before synthesizing large amounts of Pt/carbon xerogel cata-
lysts, bearing in mind that platinum salts are very expensive, it is
interesting to determine the lowest initial CPA or PTA concentra-
tion that will allow the highest level of adsorbed Pt salt on carbon
xerogels to be reached after 1 h. These CPA and PTA concentrations
were determined experimentally: (i) for CPA, 0.0378 g of sample
X-625 was added to 25 ml of four CPA aqueous solutions with var-
ious concentrations varying from 1.3 to 10.2 mmol l−1, and the pH
was adjusted to between 2.5 and 3.0 (initial optimal pH for the
adsorption of CPA onto unoxidized carbon xerogels—Fig. 2a). After
1 h on the rotary shaker, the pH of the four slurries was deter-
mined again: its value was the same in all the slurries, that is
2.4–2.5 (Fig. 2a); (ii) in the case of PTA, 0.0382 g of sample X-625-
Ox was added to 25 ml of five PTA aqueous solutions with various
concentrations varying from 1.3 to 11.8 mmol l−1; then the pH was
adjusted at 11.7 (initial optimal pH for the adsorption of PTA onto
oxidized carbon xerogels—Fig. 2b). Again, the pH of the slurries
was measured after 1 h on the rotary shaker, and was found to be
equal to 11.1–11.2 in each case (Fig. 2b). Finally, the amount of ad-
sorbed Pt was determined from the difference in Pt concentration
were prepared and poured into several 50-ml flasks. The pH values
of all these solutions were adjusted using HCl or NaOH between
pH 1 and 13 (Fig. 2).
The adsorption experiments were carried out in an excess of
liquid to enable the measurement of the solution’s pH and to pre-
vent large shifts due to the “solid buffering” effect [17,24,28]. Thus
before the adsorption experiments, the mass of the eight carbon
−1
xerogels was adjusted so that SL = 103 m2 l
in 25 ml of aque-
ous solution of CPA or PTA.
Following this, contacted slurries were placed on
a
rotary
shaker for 1 h, after which the final pHs of these slurries were
measured again (Fig. 2). Furthermore, 3–4 ml of the contacted slur-
ries was withdrawn and filtered. The remaining concentration of
Pt in the solution was determined by inductively coupled plasma
(ICP) with a Perkin–Elmer Optima 2000 ICP instrument. Previous
studies with activated carbons [24,28] have shown that 1 h is more
than sufficient to achieve adsorption equilibrium over such sur-
faces. Platinum uptakes from pH 1.5 to 13 were determined from
the difference in Pt concentration between the precontacted and
postcontacted solutions (Fig. 2).

26
S. Lambert et al. / Journal of Catalysis 261 (2009) 23–33
rogel, 2 g of Pt/carbon xerogel catalysts were synthesized with PTA
−1
(initial concentration = 6.0 mmol l−1) at SL = 103 m2 l
and at
the initial pH of 11.7. These Pt/carbon xerogel catalysts are labeled
as SEA-550-Ox, SEA-575-Ox, SEA-600-Ox and SEA-625-Ox in the
sections below. Again, platinum loadings were determined from
the difference between Pt concentrations in the precontacted and
postcontacted solutions (Table 1). Afterwards, all Pt/carbon xerogel
catalysts were dried in an oven at 393 K for 12 h. For the reduction
step, Pt/carbon xerogel catalysts were heated up to 473 K under
flowing H2 (0.04 mmol s−1) and were maintained at this tempera-
ture for 1 h (using the same flow).
For comparison purposes, eight Pt/carbon xerogel catalysts were
prepared by dry impregnation of unoxidized and oxidized carbon
xerogels with solutions of CPA (for unoxidized carbon xerogels)
and PTA (for oxidized carbon xerogels). The amount of liquid was
chosen to be equal to the pore volume of the supports, V v (Ta-
ble 2). The concentrations of the CPA and PTA solutions were
thus adjusted so as to obtain the same platinum loading as that
achieved in the eight Pt/carbon xerogel catalysts synthesized by
the SEA method (Table 1). These eight Pt catalysts were labeled
in the same way as the eight Pt/carbon xerogel catalysts synthe-
sized by the SEA method, except that the abbreviation “SEA” was
replaced by the abbreviation “DI.”
2.2. Characterization
Textural properties of all the samples were obtained by nitrogen
adsorption–desorption isotherms and mercury porosimetry mea-
surements. Nitrogen adsorption–desorption isotherms were mea-
sured at 77 K with a Fisons Sorptomatic 1990 after outgassing at
Fig. 3. Determination of the optimal initial Pt concentration for (a) sample X-625
(CPA) and (b) sample X-625-Ox (PTA).
−3
10
Pa for 24 h at ambient temperature. These isotherms pro-
vided the BET specific surface area, SBET, the specific mesopore
surface area determined by the Broekhoff–de Boer method, Smeso
,
between the precontacted and postcontacted solutions, measured
by ICP (Fig. 3).
the micropore volume calculated by the Dubinin–Radushkevich
equation, VDUB, the cumulative volume of pores of width be-
tween 2 and 7.5 nm determined by the Broekhoff–de Boer the-
ory, Vcum<7.5 nm, and the pore volume calculated from the ad-
sorbed volume at saturation, V p [29]. Mercury porosimetry mea-
surements were performed with a Pascal porosimeter thermoelec-
tron and provided the pore volume corresponding to pores of a
width >7.5 nm, VHg. In the case of macroporous samples (pore
size >50 nm), the pore volume calculated from the adsorbed vol-
2.1.5. Synthesis of Pt/carbon xerogel catalysts
From the unoxidized carbon xerogels, 2 g of Pt/carbon xero-
gel catalysts were synthesized with CPA (initial concentration =
−1
4.1 mmol l−1) at SL = 103 m2 l
and at the initial pH of between
2.5 and 3.0. The corresponding Pt catalysts are labeled as SEA-550,
SEA-575, SEA-600 and SEA-625. In the case of oxidized carbon xe-
Table 1
Platinum loading, particle size and dispersion.
TEM
CO chemisorption
Sample
%PtSEA (wt%) 0.5
%PtICP (wt%) 0.1
−1
dTEM (nm)
σTEM (nm)
nCO (mmol gPt
)
0.05
DPt (%)
5
dchem (nm) 0.1
SEA-550
SEA-575
SEA-600
SEA-625
8.0
8.0
10.0
7.5
1.1
1.2
1.2
1.1
0.1
0.1
0.1
0.1
2.75
–
2.55
2.70
85
–
80
85
1.3
–
1.4
1.3
a
a
a
a
–
–
a
9.0
9.2
SEA-550-Ox
SEA-575-Ox
SEA-600-Ox
SEA-625-Ox
1.0
2.0
4.0
5.0
1.2
1.2
1.2
1.3
1.3
0.2
0.1
0.2
0.1
2.70
–
2.45
2.40
85
–
75
75
1.3
–
1.5
1.5
a
a
a
a
–
–
a
4.0
a
a
a
DI-550
DI-575
DI-600
DI-625
8.0
8.0
9.0
8.0
8.0
1.1
1.3
1.1
1.2
0.1
0.1
0.1
0.1
–
–
–
a
a
a
a
–
–
–
–
–
a
a
a
a
–
–
–
7.3
2.65
85
1.3
a
a
a
DI-550-Ox
DI-575-Ox
DI-600-Ox
DI-625-Ox
1.0
2.0
4.0
5.0
0.7
4 to 25
4 to 25
4 to 25
4 to 25
–
–
–
–
–
–
–
a
a
a
a
–
–
–
–
–
a
a
a
a
–
–
–
4.1
0.05
<5
>20
%PtSEA: platinum loading determined as the difference in Pt concentrations in the precontacted and postcontacted solutions; %PtICP: platinum loading determined by ICP-AES
on the final Pt/C catalysts after drying and reduction; dTEM: mean diameter of platinum particles measured by TEM; σTEM: standard deviation associated with dTEM;nCO
:
amount of chemisorbed CO per Pt unit mass; DPt: platinum dispersion measured by chemisorption; dchem: mean platinum diameter derived from dispersion D.
a
Not measured.

S. Lambert et al. / Journal of Catalysis 261 (2009) 23–33
27
Table 2
Textural properties of carbon xerogels.
SBET
(m2/g)
5
Smeso
(m2/g)
5
dp,max
(mn)
1
V p
VDUB
Vcum<7.5
(cm3/g)
0.01
VHg
V v
Sample
nm
(cm3/g)
0.1
(cm3/g)
0.01
(cm3/g)
0.1
(cm3/g)
0.1
X-550
X-575
X-600
X-625
635
660
660
660
110
150
245
300
90
68
32
18
0.8
0.9
1.0
1.1
0.25
0.25
0.24
0.25
0.04
0.05
0.09
0.14
2.8
2.5
1.9
1.3
3.1
2.8
2.2
1.7
X-550-Ox
X-575-Ox
X-600-Ox
X-625-Ox
625
640
640
655
90
145
240
285
93
72
35
20
0.9
1.0
1.0
1.0
0.20
0.21
0.20
0.20
0.04
0.04
0.09
0.13
2.7
2.5
1.7
1.4
2.9
2.7
2.0
1.7
SBET: specific surface area obtained by BET method; Smeso: specific mesopore surface area obtained by the Broekhoff–de Boer method; dp,max: the pore diameter limit under
which smaller pores represent 95% of the total pore volume; V p : total pore volume calculated from the adsorbed volume at saturation; VDUB: micropore volume calculated
by the Dubinin–Radushkevich equation; Vcum<7.5 nm: cumulative volume of pores of diameter between 2 and 7.5 nm determined by Broekhoff–de Boer theory; VHg: specific
pore volume measured by mercury porosimetry; V v : pore volume obtained by addition of VDUB, Vcum<7.5
and VHg.
nm
ume at saturation is usually very imprecise or incomplete; the
total void volume, V v , was calculated from the combination of ni-
trogen adsorption–desorption and mercury porosimetry measure-
ments [30]:
2 h outgassing at 303 K, a second isotherm, which includes ph-
ysisorption only, was measured. Both isotherms were determined
−8
−1
kPa. The difference be-
in the pressure range of 10
tween the first and second isotherms provided the CO chemisorp-
tion isotherm.
to 2 × 10
V v = VDUB + Vcum<7.5 nm + VHg
.
(1)
2.3. Catalytic experiments: Benzene hydrogenation
In the case of the micro/mesoporous samples, the maximum
pore diameter, dp,max, i.e. the pore diameter limit under which
smaller pores represent 95% of the total pore volume, was deduced
from pore size distribution provided by the Broekhoff–de Boer
method, assuming an open cylinder geometry [29]. In the case of
macroporous samples, the pore size distribution must be deduced
from mercury porosimetry data. Because macroporous samples un-
dergo intrusion only during an increase in mercury pressure, as
already observed in other studies [4,25,26], the pore size distribu-
tion was calculated here following Washburn’s theory [31].
The actual platinum content of the catalysts after adsorp-
tion, drying and reduction was measured by inductively coupled
plasma–atomic emission spectrometry (ICP-AES), using an Iris ad-
vantage Thermo Jarrel Ash device. For analysis, the Pt solution was
prepared as follows: first, 0.1 g of catalyst were digested by 20 ml
of H2SO4 (95 wt%) and 10 ml of HNO3 (70 wt%), and the resulting
solution was then heated at 573 K until clear (after about 16 h).
After complete dissolution of the carbon support and evaporation,
3 ml of HNO3 (70 wt%) and 9 ml of HCl (38 wt%) were added. The
obtained solution was transferred into a 100-ml calibrated flask,
which was finally filled with deionized water.
The metal particles were examined by transmission electron
microscopy (TEM) and CO chemisorption. TEM measurements were
performed on the reduced catalyst samples. A few mg of cat-
alyst sample were dispersed in isopropanol and were sonicated
for 10 min. A drop of the suspension was then placed onto
a carbon-coated copper grid purchased from SPI supplies, USA.
Z-contrast Imaging was carried out by using an electron micro-
scope (JEM-2010F FasTEMm FEI) manufactured by JEOL, USA op-
erated at 200 kV and with an extracting voltage of 4500 V. For
statistical purposes, around 400 Pt particles were considered for
the particle size distribution and the average particle size.
Samples SEA-550, SEA-625, SEA-550-Ox and SEA-625-Ox were
tested for the hydrogenation of benzene at atmospheric pressure.
The catalysts were first crushed and sieved again between 100
and 200 μm. Due to the fact that benzene conversion is very high
on highly loaded Pt/carbon xerogel catalysts, the amount of cata-
lyst used inside the reactor was too low to form a complete cat-
alytic bed. To solve this problem, Pt/carbon xerogel catalysts were
diluted with their original support to prepare 1 wt% Pt/carbon
xerogel catalysts. Afterwards, 0.05 g of non-reduced diluted cat-
alyst was loaded into the reactor, an autoclave Engineers’ BTRS.
Before the reaction, the catalyst was reduced in situ under H2
(0.04 mmol s−1) at 473 K for 1 h.
The C6H6 flow rate, controlled by a set of saturators, was
0.002 mmol s−1. The H2/C6H6 molar ratio was fixed at 11. All gas
flows were adjusted by Bronkhorst mass flow controllers. The re-
actor was enclosed in a convection oven with the temperature
controlled and programmed to within 0.1 K using a thermocouple
located inside the reactor envelope and a PID controller. The tem-
perature of the catalyst was also measured by a second thermocou-
ple located just above the sample because benzene hydrogenation
is highly exothermic [33]. The feed and the product streams were
analyzed on-line by a gas chromatograph equipped with a flame
ionization detector.
Each catalytic test started at 393 K. This temperature level was
maintained until a quasi-stationary state was observed (i.e., after
about 2 h). Then if the benzene conversion was <50%, the tem-
perature was increased from 393 to 423 K. By contrast, if the
conversion was >50%, the temperature was decreased from 393 K
to 353 K. The reactor was kept for 30 min at each temperature
step; in each case, a quasi-stationary state was observed.
The activity of the catalysts was calculated as the amount of
benzene (mmol) converted into cyclohexane per gram of Pt and
per second. Because this reaction is structure-insensitive on Pt-
based catalysts supported on activated carbon [34–36], the activity
of the catalyst is directly linked to the number of accessible metal
atoms. Consequently, catalytic activity is an indirect measurement
of metal dispersion, defined as the ratio between the number of
surface Pt atoms and the total number of Pt atoms in the cata-
lyst. Finally the activation energy of the reaction catalyzed by Pt
was determined from measurements performed at various temper-
atures.
A
volumetric static method was used for CO chemisorp-
tion [32]. Measurements were obtained on a Fisons Sorptomatic
1990 equipped with a turbomolecular vacuum pump that allows
−3
a high vacuum of 10
Pa. The non-reduced catalyst was first re-
duced in situ under hydrogen flow (0.04 mmol s−1); the samples
−1
were heated from 298 to 473 K at a rate of 5 K min
and were
kept at 473 K for 1 h. Afterwards, the samples were outgassed
under vacuum at 463 K for 1 h. A double adsorption method
was used: (i) a first adsorption isotherm, which includes both ph-
ysisorption and chemisorption, was determined; (ii) then after a

28
S. Lambert et al. / Journal of Catalysis 261 (2009) 23–33
−1
after Pt adsorption. Moreover, for samples
3. Results
to 525 and 0.18 cm3 g
X-600 and SEA-600, X-625 and SEA-625, X-600-Ox and SEA-600-
3.1. PZC of carbon xerogels and determination of Pt uptakes
Ox, X-625-Ox and SEA-625-Ox, the mesopore surface area, Smeso
,
−1
decreases too: from 300 to 250 m2 g
for samples X-625 and
for samples X-600-Ox and
−1
The PZCs of unoxidized and oxidized carbon xerogels are shown
in Figs. 1a and 1b respectively. These curves clearly show a plateau
for both types of support: the PZC value equals 9.4 in the case of
unoxidized carbon xerogels, and is equal to 2.4 when the supports
are oxidized.
Fig. 2 shows the Pt uptake curves: the Pt surface density, de-
fined as the amount of Pt per m2 of the support, is plotted against
the final pH value of the slurries after 1 h on the rotary shaker.
In Fig. 2a, a sharp maximum is observed in the narrow final pH
range of 2.4–2.5 (initial pH range between 2.5 and 3.3 for CPA
solutions). At the adsorption maximum of CPA, i.e. in the ’strong
adsorption’ condition, the uptake is quite similar for all the unoxi-
dized samples, that is about 0.91 μmol m−2, which corresponds to
a Pt loading, defined as the percentage of the Pt mass per the to-
tal mass of the catalyst, of about 11.6 wt%. In Fig. 2b, the uptake
curves display a maximum for each oxidized xerogel in the pH
SEA-625, and from 240 to 195 m2 g
SEA-600-Ox. By contrast, Smeso does not change much in the case
−1
for sam-
of supports with large pore sizes: for instance, Smeso = 110 m2 g
−1
for samples X-550 and SEA-550 and Smeso = 150 m2 g
ples X-550 and SEA-575-Ox before and after Pt adsorption. These
findings indicate that platinum deposition affects micropores and
small mesopores only, as already observed in [37] in the case of
Pt/carbon xerogel catalysts prepared by impregnation.
3.3. Platinum loading in Pt/carbon xerogel catalysts
Platinum loadings were determined by ICP-AES on the final
Pt/carbon xerogel catalysts after drying and reduction (%PtICP) in
order to check whether the platinum loadings calculated from the
difference between Pt concentrations in the precontacted and post-
contacted solutions (%PtSEA) were correct (Table 1). These results
show that Pt loadings determined by both methods are quite sim-
ilar.
range of 11.1–11.2. (initial pH 11.7 for PTA solutions). Depending on
−2
the support, the uptake varies from 0.60 μmol m
(X-625-Ox) to
−2
0.38 μmol m
(X-550-Ox), that is 7.7 to 4.6 wt%.
The Pt loadings of samples SEA-550, SEA-575, SEA-600 and
SEA-625 obtained from the solution concentration differences,
%PtSEA, range from 8 to 10 wt%, which shows that the amount of
adsorbed CPA is similar whatever the support, although the in-
crease of mesopore surface seems to slightly favor adsorption. This
result is in good agreement with Fig. 2a, in which at the adsorp-
tion maximum of CPA (final pH range of 2.4–2.5), the Pt uptake
is quite similar for all unoxidized samples: 0.91 μmol m−2, which
corresponds to a Pt loading of about 11.6 wt%. For samples issued
from oxidized supports, SEA-550-Ox, SEA-575-Ox, SEA-600-Ox and
SEA-625-Ox, %PtSEA ranges from 1.0 to 5.0 wt% (Table 1); this is
in good agreement with Fig. 2b in which the amount of adsorbed
PTA increases from sample X-550-Ox to sample X-625-Ox. Never-
theless, the amounts of adsorbed PTA on little amounts of oxidized
carbon xerogels (about 0.038 g of oxidized carbon xerogels were
used to establish Fig. 2b) are proportionally higher (Pt loadings
range from 4.6 wt% for X-550-Ox to 7.7 wt% for X-625-Ox) than
the amounts of Pt salts adsorbed on greater amounts of oxidized
carbon xerogels (2 g of each oxidized carbon xerogel were used
to synthesize samples SEA-550-Ox, SEA-575-Ox, SEA-600-Ox and
SEA-625-Ox with Pt loadings from 1.0 to 5.0 wt% Table 1), assum-
ing diffusional limitations.
The lowest initial CPA or PTA concentration that will allow the
highest level of adsorbed Pt salt to be reached on carbon xero-
gels was determined from the adsorption curves obtained at the
optimal pH for various Pt precursor concentrations. These curves
are presented in Fig. 3a for CPA adsorbed on sample X-625, and in
Fig. 3b for PTA adsorbed on sample X-625-Ox. The Pt uptake shows
−1
that increasing the Pt concentration beyond 4.1 and 6.0 mmol l
in the case of CPA or PTA precursors, respectively, does not lead to
the adsorption of larger amounts of Pt complex. As a consequence,
for the synthesis of Pt/carbon xerogel catalysts, the initial Pt con-
centration of CPA and PTA solutions was chosen to be equal to 4.1
−1
and 6.0 mmol l
respectively.
3.2. Textural properties of carbon xerogels and Pt/carbon xerogel
catalysts
The textural properties of unoxidized and oxidized carbon xero-
gels are shown in Table 2. In agreement with previous studies [25,
26], the texture of unoxidized carbon xerogels greatly varies with
the initial pH of the resorcinol–formaldehyde solution: when the
synthesis initial pH increases from 5.50 to 6.25, the total pore vol-
ume, V v , decreases from 3.1 to 1.7 cm3 g−1, and the mesopore
surface area, Smeso, increases from 110 to 300 m2 g−1. The increase
Concerning the increase of Pt loadings from sample X-550-Ox
to sample X-625-Ox (Fig. 2b), assuming that the number and the
nature of surface oxygen groups are similar for all oxidized carbon
xerogels, since their total material surface in solution was the same
(SL = 103 m2 l−1) and since their PZC is identical, the only differ-
ence between the four samples is the surface distribution between
micro-, meso-, and macropores. Indeed, Smeso increases from 90
of Smeso is due to the decrease of the meso/macropore size, dp,max
,
from 90 to 18 nm. Similar trends are observed for the oxidized
series: when the initial pH is increased from 5.50 to 6.25, V v de-
creases from 2.9 to 1.7 cm3 g−1, while Smeso increases from 90 to
285 m2 g−1. The increase of Smeso is also due to the decrease of the
meso/macropore size, dp,max, from 93 to 20 nm. So the oxidation
step of carbon xerogels does not change the size of the mesopores.
Indeed, as the standard deviation associated with dp,max is about
1 nm, a variation from 18 to 20 nm is not significant.
−1
to 285 m2 g
when the synthesis pH of the pristine gel increases
from 5.50 to 6.25, due to the decrease of the pore size (Table 2).
Fig. 4 shows that, in the case of oxidized supports, the amount of
adsorbed PTA, and thus the Pt weight loading increases linearly
After platinum adsorption by the SEA method, drying and re-
duction steps, the textural characteristics of all Pt/carbon xerogel
catalysts vary with regards to those of the original supports. In-
deed, the specific surface area, SBET, and the micropore volume,
VDUB, decrease significantly compared to the SBET and VDUB of cor-
responding carbon xerogels used as supports for Pt/carbon xerogel
catalysts: for instance, for samples X-550 and SEA-550, SBET and
with Smeso.
3.4. Platinum dispersion in Pt/carbon xerogel catalysts
Table 1 shows the platinum particle sizes determined by TEM
and CO chemisorption measurements.
TEM micrographs show that, in all Pt/carbon xerogel catalysts
synthesized by the SEA method, the size of the platinum particles
is quite homogeneous, and is about 1.2 nm (Figs. 5a and 5b). In or-
der to compare the Pt dispersion obtained using the SEA method
on unoxidized and oxidized carbon xerogels with that resulting
−1
VDUB decrease respectively from 635 and 0.25 cm3 g
before Pt
after Pt adsorption. Again for
samples X-575-Ox and SEA-575-Ox, SBET and VDUB decrease re-
−1
adsorption to 500 and 0.22 cm3 g
−1
−1
before Pt adsorption
spectively from 640 m2 g
and 0.21 cm3 g

S. Lambert et al. / Journal of Catalysis 261 (2009) 23–33
29
from the classical dry impregnation (DI) method on the same sup-
ports, micrographs of catalysts synthesized by the DI method are
also shown in Figs. 5c and 5d. Concerning Pt/carbon xerogel cata-
lysts synthesized with CPA using either the SEA method (Fig. 5a) or
the DI method (Fig. 5c), all the samples are very highly dispersed
(Table 2). By contrast, Pt/carbon xerogel catalysts synthesized with
PTA using the SEA method are very highly dispersed (Fig. 5b),
while the same samples synthesized with PTA using the DI method
present very large platinum particles from 5 to 25 nm (Fig. 5d).
From CO chemisorption measurements on Pt/carbon xero-
gel catalysts synthesized with the SEA method, the amount of
chemisorbed CO per Pt unit mass, nCO
, ranges from 2.40 to
−1
2.75 mmol gPt (Table 1). From the adsorbed amount of CO, one
can calculate the platinum dispersion, DPt, i.e. the ratio between
the number of Pt atoms located at the surface of the metal parti-
cles and the total number of Pt atoms in the catalyst, and the mean
diameter of Pt crystallites, dchem, as explained in [37]. Following
Fig. 4. Platinum weight loading as a function of the total mesoporous surface area,
Smeso, of oxidized carbon xerogels for samples SEA-550-Ox (1), SEA-575-Ox (E),
SEA-600-Ox (P) and SEA-625-Ox (!). Adsorption of PTA (6.0 mmol l−1) with an
initial pH = 11.7.
Fig. 5. STEM micrographs of samples: (a) SEA-600, (b) SEA-625-Ox, (c) DI-600, and (d) DI-625-Ox.

30
S. Lambert et al. / Journal of Catalysis 261 (2009) 23–33
Table 3
Catalytic test results.
353, 363, 383 and 393 K for samples SEA-550-Ox and SEA-625-Ox.
Ea was found to range from 38 to 55 kJ mol−1, in agreement with
−1
−1
−1
values reported in the literature (30–70 kJ mol
[37,38]).
rben (mmol gPt
s
)
0.05
Sample
353 K
363 K
383 K
393 K
403 K
413 K
423 K
4. Discussion
a
a
a
SEA-550
SEA-625
–
–
–
1.15
1.05
1.60
1.50
2.05
1.90
2.85
2.50
a
a
a
–
–
–
The PZC of carbons can be changed irreversibly by oxidizing the
surface. In addition, in previous studies [22,24], it has been demon-
strated that the Pt anion and cation uptakes can be modified on
this basis. In the present study, the main difference between unox-
idized and oxidized carbon xerogels is the number of acidic groups
on the surface. Concerning unoxidized carbon xerogels, their PZC
equals 9.4 (Fig. 1a), which indicates that these carbon surfaces ini-
tially contain a very low number of acidic groups. Indeed, the PZC
of a carbon seems to be determined mainly by the concentration
of the surface acid sites: the higher the number of acidic groups
a
a
a
SEA-550-Ox
SEA-625-Ox
1.00
0.60
1.65
0.90
2.25
2.10
2.75
2.30
–
–
–
a
a
a
–
–
–
rben: apparent reaction rate for benzene hydrogenation per Pt mass unit.
a
Not measured.
previous studies conducted on highly dispersed Pt catalysts [37,
38], the chemisorption mean stoichiometry XPt–CO, that is, the
mean number of Pt atoms on which one CO molecule is adsorbed,
was chosen equal to 1.61. Values of DPt and dchem are given in Ta-
ble 1. For all samples, one has to note the agreement between TEM
and chemisorption, assuming that all the Pt atoms located at the
surface of the metal particles are accessible to CO.
on a carbon surface, the lower its PZC [22]. When carbon xerogels
−1
are oxidized with 5 mol l
nitric acid for 48 h at room tempera-
ture, their PZC decreases to 2.4 (Fig. 1b), as a consequence of the
increase of the number of their acidic surface sites, as explained
extensively in previous studies [22,40].
3.5. Catalytic activity of Pt/carbon xerogel catalysts
In contrast to silica and alumina [24], carbon xerogels of vary-
ing PZC differ markedly in their capacity to adsorb Pt anion and
cation complexes (Fig. 2). In Fig. 2a, the uptake of CPA adsorp-
tion on unoxidized carbon xerogels (PZC = 9.4) versus final pH
shows a sharp maximum in the narrow pH range of 2.4–2.5.
At the maximum adsorption of CPA, i.e. the ‘strong adsorption’
condition, the uptake is quite similar for all unoxidized samples,
about 0.91 μmol m−2, which corresponds to a Pt loading of about
11.6 wt%. It is for this reason that, in Table 1, the platinum load-
ings, %PtSEA, are very similar for samples SEA-550, SEA-575, SEA-
600 and SEA-625, with Pt loadings from 8.0 to 10.0 wt%. In con-
trast, in Fig. 2b, the uptake curve displays a maximum for each
oxidized xerogel in the pH range of 11.1–11.2. Depending on the
Catalytic test results are presented in Table 3. Benzene hydro-
genation on Pt/carbon xerogel catalysts resulted in very high con-
version ratios, even with low amounts of diluted catalyst (0.05 g
with Pt loading = 1 wt%). The apparent reaction rate, rben, was cal-
culated at each temperature for each catalyst, taking into account
an integral isothermal reactor working in a stationary state and
a first-order reaction, rben = k[C6H6] [39]. Indeed, for these cat-
alytic tests, the concentration of hydrogen is always much greater
than the concentration of benzene ([H2]/[C6H6] = 11). Thus, be-
cause the hydrogen partial pressure is essentially constant during
the reaction, the rate observed depends only on the concentration
of benzene:
Fben
support, the maximum uptake varies from 0.38 (X-550-Ox) to
rben = −
ln(1 − f ),
(2)
−2
W
0.60 μmol m
(X-625-Ox). Fig. 4 shows that, in the case of ox-
idized supports and at an initial pH 11.7 (to obtain a final pH
11.1–11.2 after mixing and stirring for 1 h; Fig. 2b), the amount
of adsorbed PTA, and thus the Pt weight loading increases linearly
where Fben is the benzene flowrate at the inlet of the reactor
(mmol s−1), W is the Pt mass inside the reactor (gPt), and
f is
the conversion of benzene into cyclohexane.
−1
The reaction rate observed in the present study is an apparent
reaction rate: it equals the intrinsic reaction rate when no diffu-
sional limitation occurs and when the reactor is differential only.
Following the method used in previous studies [12,37], the Weisz
modulus, which compares the reaction rate to the diffusion rate,
was calculated to determine the importance of diffusional limita-
tions. In all cases, the Weisz modulus was ꢀ1, indicating that the
Pt/carbon xerogel catalysts were operating in a chemical regime;
the effect of the diffusion of reactants on the observed kinetics
was negligible.
with Smeso, this latter parameter increasing from 90 m2 g
for
for sample X-625-Ox (Table 2).
−1
sample X-550-Ox to 285 m2 g
These results can be explained by the fact that it is not possible for
the large Pt ammine complexes, which are believed to retain two
hydration sheaths and the size of which is estimated equal to 1.5
nm [28,41], to penetrate inside the micropores of carbon xerogels,
the size of which has been calculated to be equal to 0.8 nm [25].
Indeed, in previous studies concerning silica, alumina and active
carbons [16,24], it has been explained that the maximum Pt up-
take of both CPA and PTA complexes appears to be dictated by
a steric maximum, which can be calculated as a closely packed
monolayer of complexes that retain either one hydration sheath in
the case of the anionic chloride complexes (their size has been es-
timated to be equal to 0.6 nm [16,28]), or two sheaths in the case
of the cationic ammine complexes (their size has been estimated
to be equal to 1.5 nm [28,41]). Thus for microporous supports, it is
very difficult to adsorb cationic ammine complexes because their
steric volume is too high compared to the size of micropores in the
support. In this case, the ammine complex adsorbs at the surface
of the mesopores and, as a consequence, the Pt uptake increases
linearly with the mesopore surface (Fig. 4). Therefore the present
study shows again how carbon xerogels can be engineered to over-
come the steric hindrance limitation with PTA.
−1
−1
for sam-
At 393 K, rben is equal to 1.15 and 1.05 mmol gPt
s
ples SEA-550 and SEA-625 respectively, and rben is equal to 2.75
−1
−1
for samples SEA-550-Ox and SEA-625-Ox
and 2.30 mmol gPt
s
respectively. The values of samples SEA-550 and SEA-625 are very
similar to those presented in a previous study [37]. Nevertheless,
Pt/carbon xerogel catalysts synthesized from oxidized carbon xero-
gels present higher values for rben
.
Considering an integral isotherm reactor working in a station-
ary state and a first-order reaction, the relationship between the
conversion, f , and the apparent activation energy of the reaction,
Ea, can be written as in [39]:
ꢀ ꢀ
ꢁꢁ
ln ln 1/(1 − f ) = ln C − (Ea/RT ),
(3)
where C is a constant, R is the perfect gas constant, and T is the
temperature. The apparent activation energy for benzene hydro-
genation on Pt was calculated from experimental data obtained at
393, 403, 413 and 423 K for samples SEA-550 and SEA-625, and at
The trends in anionic and cationic Pt adsorption on unoxidized
and oxidized carbon xerogels seen in the present study are con-
sistent with the adsorption of CPA or PTA on carbon blacks and
activated carbons [20,22,23,42]. Indeed, surface acidic groups (and

S. Lambert et al. / Journal of Catalysis 261 (2009) 23–33
31
correspondingly low PZCs) have been shown to favor the anchoring
of [Pt(NH3)4]2− on carbon blacks, whereas less severe or no oxi-
dation (and correspondingly high PZCs) leads to a greater uptake
of anionic Pt complexes onto various forms of carbons. Therefore
the results presented in the present study indicate that: (i) the Pt
complex (anionic or cationic) must be chosen according to the PZC
of the given carbon xerogel support; (ii) an optimal pH, at which
the metal complex–support interactions are the strongest and the
Pt uptake is the highest, exists for each carbon xerogel/Pt complex
pair, and that the optimal pH depends on the carbon xerogel/Pt
complex pair selection (Fig. 2).
As explained in previous studies dealing with the preparation
of Pt/active charcoal catalysts by impregnation, metal dispersion
essentially depends on two parameters: the pore texture of the
support and its surface composition [43–45]. On the one hand,
the pore texture mostly impacts the contact between the carbon
surface and the solution containing the precursor. In this respect,
carbon xerogels seem more favorable to homogeneous carbon–
solution contact due to the broad size distribution of meso- and
Fig. 6. Relationship at 393 K between the apparent reaction rate, rben, expressed by
unit mass of Pt, and the platinum dispersion, DPt for unoxidized carbon xerogels
(F), for oxidized carbon xerogels (2) and for Pt/carbon xerogel catalysts synthe-
sized by the Wet Impregnation method (P) in a previous study [37].
groups at the surface of the support [22] coupled with an inade-
quate pH of PTA aqueous solutions during the impregnation step
(pH 8.7, i.e. low electrostatic interactions between PTA and the
oxidized carbon support cationic Pt complexes and the carbon sup-
port, Fig. 2b) provide so many reasons for obtaining of a very low
Pt dispersion.
A very important concern regarding Pt/carbon xerogel catalysts
is the accessibility of the active centers. Because platinum is lo-
cated inside the carbon support, there is a risk that it may not
be accessible. The values of V v in Table 2 show that drying un-
der vacuum allows large pore volumes to be retained, but they do
−1
macropores (the total pore volume is >1.7 cm3 g
in Table 2)
and their pore texture homogeneity [25,26]. The pore texture of
active charcoals is usually much less homogeneous than that of
carbon xerogels and the volume of the large pores (meso- and
macropores) is low: as a consequence, the carbon inner surface,
mostly related to the microporosity, is not easily reached by the
solvent, and thus by the metal precursor. On the other hand, the
surface composition determines not only the interactions between
the support and the precursor of the species to be deposited,
but also the interactions between the support and the solvent.
For instance, due to the high pyrolysis temperature, active char-
coals are usually hydrophobic [43], and further oxidation is often
necessary to increase their surface acidic group content and the
water–carbon affinity. Note that, in the case of activated charcoals,
surface acidic groups seem to have a negative effect on platinum
dispersion after reduction [22,43]: this is attributed to a delocal-
ization of the aromatic cycles π-electrons of the carbon, which are
responsible for the anchoring of the reduced metal particles, by
the oxygen of the acidic surface groups. By contrast, carbon xero-
gels are very hydrophilic after pyrolysis, despite their low acidic
group content [25]. Indeed, a carbon xerogel prepared by pyrolysis
at 1073 K absorbs an amount of water corresponding to 95% of its
pore volume within a few minutes, whatever the carbon texture.
As a consequence, carbon xerogels should then lead more easily to
a good metal dispersion than active charcoals.
In the present study, Pt/carbon xerogel catalysts synthesized
with the use of dilute CPA aqueous solutions by both the SEA
(samples SEA-550, SEA-575, SEA-600, SEA-625) and the DI (sam-
ples DI-550, DI-575, DI-600, DI-625) methods present very high Pt
dispersion as a consequence of very small Pt particles (1.1–1.3 nm).
These two methods give very similar results because the pH used
in the SEA method for the strong electrostatic adsorption of CPA
(pH 2.5) is very close to the pH value of CPA solutions (pH 1.9)
after the dissolution of H2PtCl6 salt in water, used for the DI im-
pregnation.
not prove the accessibility of platinum. Nevertheless, it can be ob-
∼
served in Table 1 that dTEM
d
for all the samples: the mean
=
chem
Pt particle size obtained by CO chemisorption measurements is in
each case consistent with TEM measurements. The convergence of
TEM and chemisorption measurements tends to indicate that the
Pt particles located inside the carbon support are accessible.
Benzene hydrogenation was carried out with samples SEA-550,
SEA-625, SEA-550-Ox and SEA-625-Ox to check: (i) whether the
very highly dispersed Pt particles located inside the carbon support
are accessible for large molecules such as C6H6; (ii) whether a rela-
tionship exists between the high Pt dispersion of Pt/carbon xerogel
catalysts and the benzene activity. In previous studies [34–36],
it has been observed that benzene hydrogenation is a structure-
insensitive reaction on Pt-based catalysts supported on activated
carbons. In other words, the reaction rate is directly proportional
to the accessible Pt surface (i.e., to the metal dispersion). In Fig. 6,
the catalytic activity at 393 K, expressed in mmol of benzene
transformed into cyclohexane per unit mass of Pt and per sec-
ond, rben, is presented as a function of the Pt dispersion, DPt
for the four samples SEA-550, SEA-625, SEA-550-Ox and SEA-625-
Ox and for three Pt/carbon xerogel catalysts synthesized by the
Wet Impregnation method (aqueous solution in excess) in a previ-
ous study [37]. In Fig. 6, it can be observed for samples SEA-550,
SEA-625 and the three Pt/unoxidized carbon xerogels synthesized
in [37] that rben is directly proportional to the Pt dispersion, DPt
,
By contrast, Pt catalysts supported on oxidized carbon xerogels
present much higher dispersion values when the SEA method is
used instead of the classical Dry Impregnation technique. Indeed,
while the Pt particle size ranges from 4 to 25 nm for samples DI-
550-Ox, DI-575-Ox, DI-600-Ox and DI-625-Ox (Table 1), dSTEM does
not exceed 1.2 nm in the case of samples SEA-550-Ox, SEA-575-
Ox, SEA-600-Ox and SEA-625-Ox. In this particular case, the SEA
method provides the best solution for obtaining very highly dis-
persed platinum catalysts from an oxidized carbon support, i.e. the
use of a cationic Pt complex and the pH optimization during the
impregnation step. Indeed, in samples DI-550-Ox, DI-575-Ox, DI-
600-Ox and DI-625-Ox, the presence of a large number of acidic
which is consistent with the structure-insensitive nature of the
benzene hydrogenation on Pt.
However, samples SEA-550-Ox and SEA-625-Ox, synthesized
from oxidized carbon xerogels, are much more active than all the
Pt/unoxidized carbon xerogel catalysts. According to previous re-
sults for benzene hydrogenation over Pt supported on active car-
bons [35], better values of rben (mmol g−Pt1 s−1) are obtained when
active carbons are oxidized by nitric acid solutions. In the above-
cited study [35], this result is explained by: (i) a better dispersion
obtained on the oxidized supports in the considered synthesis con-
ditions; (ii) the significant change in the surface area and pore
size distribution after oxidation of the carbon supports. In the

32
S. Lambert et al. / Journal of Catalysis 261 (2009) 23–33
present study, both items are inapplicable: (i) all Pt/C xerogel cat-
alysts synthesized from the SEA method are very highly dispersed
(dTEM = 1.1–1.3 nm in Table 1); (ii) the textural properties of both
the unoxidized and oxidized carbon xerogels are very similar (Ta-
ble 2).
5. Conclusions
Highly dispersed Pt/C catalysts were obtained by applying the
“strong electrostatic adsorption” method to carbon xerogels. Four
supports with various maximal pore sizes (18, 32, 68 and 90 nm)
were used and their PZC was determined as being equal to 9.4.
These carbon xerogels were also oxidized in nitric acid in order
to alter irreversibly their surface and to increase the number of
surface acidic groups, with the consequence of a strong decrease
of their PZC to 2.4. Unoxidized carbon xerogels placed in solu-
tions at pH values below their PZC become protonated, positively
charged and then adsorb platinum anions such as hexachloroplati-
nate [PtCl6]2−, while oxidized carbon xerogels placed in solutions
at pH values above their PZC become deprotonated, negatively
charged and adsorb platinum cations such as platinum tetraam-
mine [Pt(NH3)4]2+. Platinum uptake curves reach a maximum with
respect to the pH for each platinum precursor (final pH 2.4–2.5 for
CPA on unoxidized carbons and final pH 11.0–11.2 for PTA on oxi-
dized supports). At these maxima, Pt/C catalysts are obtained with
platinum loadings of about 8 to 10 wt% for CPA and about 1 to
5 wt% for PTA. Indeed, the present study shows that, in the case
of oxidized supports, the amount of adsorbed PTA, and thus the Pt
weight loading increases linearly with Smeso. These results can be
explained by the fact that it is not possible for the large Pt ammine
complexes, which are believed to retain two hydration sheaths, to
penetrate inside the micropores of carbon xerogels. After the re-
duction step at 473 K for 1 h, all these Pt/carbon xerogel catalysts
preserve a high dispersion and very small platinum particles (1.1–
1.3 nm) are observed by TEM. These Pt particles are accessible and
the Pt/carbon xerogel catalysts are very active for the hydrogena-
tion of benzene into cyclohexane.
Assuming that the benzene hydrogenation reaction remains
structure-insensitive, the fact that the data obtained with SEA-550-
Ox and SEA-625-Ox do not fall within the same linear relationship
as the other samples (Fig. 6) can be explained in two general
ways: (i) a first explanation may be offered in terms of electronic
effects. It is well known that the presence of surface functional
groups can affect the activity and selectivity of the active phase
via metal–support interactions, as previously reported for Pt/C cat-
alysts in selective hydrogenations [46–48]. A remarkable increase
in the activity for methanol electrooxidation was also observed
when Pt–Ru catalysts were supported on oxidized carbon xero-
gel [6]. However, the precise nature of this effect has not yet been
clearly elucidated. The nature of the metal–support interaction in-
volves a change in the electronic structure of Pt particles, altering
the catalytic properties of the Pt surface atoms. For instance, it has
been found that the turnover frequency (TOF) of tetralin hydro-
genation on zeolite supported Pt catalysts strongly depends on the
composition of the support [49]. In the present case, it may be en-
visaged that the presence of oxygen groups in the vicinity of small
Pt crystallites will generate a polarization within the Pt particle,
which will lead to a change in the electron affinity of the Pt 5d
valence orbitals. This may favor the adsorption of the planar ben-
zene ring. Furthermore, the competition of the carbon support for
benzene adsorption will decrease in the presence of oxygen sur-
face groups [50]. Both effects will enhance the catalytic activity
of Pt; (ii) a second explanation may be offered by the fact that
the metal dispersion of the catalyst calculated from CO chemisorp-
tion is not correct. The calculation of DPt could be distorted by
a partial poisoning of the Pt particle surface. Indeed, if the re-
duction treatment is not sufficient, chlorine atoms coming from
the Pt complex could still be present in the catalyst, either on
the support or on the Pt particles: chlorine would then block ad-
sorption sites for benzene. Due to the fact that six Cl atoms are
contained in one CPA molecule while two Cl atoms are present
in one PTA molecule, it seems possible that the chlorine content
of Pt catalysts supported on unoxidized carbon xerogels will be
higher if the reduction treatment does not eliminate all the chlo-
rine coming from the Pt complex. One can argue that, since the
CO chemisorption and TEM results are in good agreement, the
values of the dispersion are correct. Indeed in this work, it was
assumed [37,38] that the chemisorption mean stoichiometry, that
is, the mean number of Pt atoms on which one CO molecule is
adsorbed, XPt–CO, equals to 1.61: lower values of XPt–CO would
lead to lower dispersions, i.e. to lower accessible Pt surfaces and
higher ‘equivalent diameter’ (dCO) values, which would be in agree-
ment with Cl poisoning. Another possibility is the removal of Cl
by CO during chemisorption: assuming that XPt–CO = 1.61 is the
right stoichiometry, chemisorption and TEM results would then
match, but the true metal dispersion, i.e. the fraction of accessible
Pt atoms in the catalyst, would be overestimated by chemisorption.
Indeed, during the hydrogenation reaction, Cl poisoning would still
block some Pt atoms, decreasing the activity of the catalyst. This
would explain why Pt catalysts prepared with CPA (unoxidized
supports) would be less active than those obtained with PTA (ox-
idized supports). In fact, a study is in progress to explain the
difference in activity for the hydrogenation of benzene between
unoxidized and oxidized carbon xerogels, focusing on Cl poison-
ing.
These results obtained by applying the SEA method to prepare
very highly dispersed Pt/carbon xerogel catalysts are very encour-
aging, demonstrating the usefulness of this method for heteroge-
neous catalysis. Research perspectives are numerous. In particular,
a study concerning successive “strong electrostatic adsorptions” to
achieve Pt loading up to 20 wt%, while keeping the Pt particle size
below 3 nm is in progress. Finally, these Pt/carbon xerogel cata-
lysts are now being tested as PEM fuel cell electrocatalysts for the
oxygen reduction reaction.
Acknowledgments
The authors thank the Fonds de Bay, the Fonds de Recherche
Fondamentale Collective, the Ministère de la Région Wallonne Di-
rection Générale des Technologies, de la Recherche et de l’Energie
and the Interuniversity Attraction Pole (IAP-P6/17) for their finan-
cial support. S.L. and N.J. are grateful to the F.R.S.-F.N.R.S. for their
postdoctoral research positions. S.L. also acknowledges the Com-
mission for Educational Exchange between The United States of
America, Belgium and Luxembourg for its financial help (Fulbright
grant) in facilitating her postdoctoral secondment to the Chemi-
cal Engineering Department at the University of Illinois at Chicago.
We also acknowledge the assistance of Vera Santos in the catalytic
tests undertaken in the present study.
References
[1] H.A. Gasteiger, W. Gu, R. Makharia, M.F. Mathias, B. Sompalli, in: W. Vielstich,
A. Lamm, H.A. Gasteiger (Eds.), Handbook of Fuel Cells—Fundamentals, Tech-
nology and Applications, vol. 3, Wiley, Chichester, 2003, pp. 593–610.
[2] H.A. Gasteiger, S.S. Kocha, B. Sompalli, F.T. Wagner, Appl. Catal. B 56 (2005) 9.
[3] C. Moreno-Castilla, F.J. Maldonado-Hódar, Carbon 43 (2005) 455.
[4] N. Job, B. Heinrichs, F. Ferauche, F. Noville, J. Marien, J.-P. Pirard, Catal. To-
day 102–103 (2005) 234.
[5] B. Heinrichs, S. Lambert, N. Job, J.-P. Pirard, in: J.R. Regalbuto (Ed.), Catalyst
Preparation: Science and Engineering, CRC Press/Taylor & Francis Group, Boca
Raton, 2007, p. 163.

S. Lambert et al. / Journal of Catalysis 261 (2009) 23–33
33
[6] J.L. Figueiredo, M.F.R. Pereira, P. Serp, P. Kalck, P.V. Samant, J.B. Fernandes, Car-
bon 44 (2006) 2516.
[7] J. Marie, S. Berthon-Fabry, M. Chatenet, E. Chainet, R. Pirard, N. Cornet, J. Appl.
Electrochem. 37 (2007) 147.
[8] N. Job, J. Marie, S. Lambert, S. Berthon-Fabry, P. Achard, Energy Convers. Man-
age. 49 (2008) 2461.
[9] H. Du, B. Li, F. Kang, R. Fu, Y. Zeng, Carbon 45 (2007) 429.
[10] R.W. Pekala, J. Mater. Sci. 24 (1989) 3221.
[31] E.W. Washburn, Proc. Natl. Acad. Sci. USA 7 (1921) 115.
[32] G. Bergeret, P. Gallezot, in: G. Ertl, H. Knözinger, J. Weitkamp (Eds.), Handbook
of Heterogeneous Catalysis, Wiley–VCH, Weinheim, 1997, p. 439.
[33] R.J. Farrauto, C.H. Bartholomew, Fundamentals of Industrial Catalytic Processes,
Blackie Academic & Professionals/Chapman & Hall, London, 1997, p. 436.
[34] A.E. Aksoylu, M. Madalena, A. Freitas, J.L. Figueiredo, Appl. Catal. A 192 (2000)
29.
[35] A.E. Aksoylu, M. Madalena, A. Freitas, M.F.R. Pereira, J.L. Figueiredo, Carbon 39
(2001) 175.
[36] A.E. Aksoylu, J.L. Faria, M.F.R. Pereira, J.L. Figueiredo, P. Serp, J.-C. Hierso, R.
Feurer, Y. Kihn, P. Kalck, Appl. Catal. A 243 (2003) 357.
[11] S.A. Al-Muhtaseb, J.A. Ritter, Adv. Mater. 15 (2003) 101.
[12] N. Job, B. Heinrichs, S. Lambert, J.-F. Colomer, B. Vertruyen, J. Marien, J.-P. Pi-
rard, AIChE J. 52 (2006) 2663.
[13] J.P. Brunelle, Pure Appl. Chem. 50 (1978) 1211.
[14] J.R. Regalbuto, A. Navada, S. Shadid, M.L. Bricker, Q. Chen, J. Catal. 184 (1999)
335.
[37] N. Job, M.F. Ribeiro Pereira, S. Lambert, A. Cabiac, G. Delahay, J.-F. Colomer,
J. Marien, J.L. Figueiredo, J.-P. Pirard, J. Catal. 240 (2006) 160.
[38] S.D. Lin, M.A. Vannice, J. Catal. 143 (1993) 539.
[15] W.A. Spieker, J.R. Regalbuto, Stud. Surf. Sci. Catal. 130 (2000) 203.
[16] W.A. Spieker, J.R. Regalbuto, Chem. Eng. Sci. 56 (2001) 3491.
[17] X. Hao, J.R. Regalbuto, J. Colloid Interface Sci. 267 (2003) 259.
[18] M. Schreier, J.R. Regalbuto, J. Catal. 225 (2004) 190.
[19] J.T. Miller, M. Schreier, A.J. Kropf, J.R. Regalbuto, J. Catal. 225 (2004) 203.
[39] C.N. Satterfield, Mass Transfer in Heterogeneous Catalysis, MIT Press, Cam-
bridge, 1970.
[40] P. Vinke, M. van der Eijk, M. Verbree, A.F. Voskamp, H. van Bekkum, Carbon 32
(1994) 675.
[41] N. Santhanam, T.A. Conforti, W. Spieker, J.R. Regalbuto, Catal. Today 21 (1994)
141.
[42] G.C. Torres, E.L. Jablonski, G.T. Baronetti, A.A. Castro, S.R. de Miguel, O.A. Scelza,
M.D. Blanco, M.A. Pena Jiménez, J.L.G. Fierro, Appl. Catal. A 161 (1997) 213.
[43] F. Rodríguez-Reinoso, Carbon 36 (1998) 159.
[44] M. Che, O. Clause, C. Marcilly, in: G. Ertl, H. Knözinger, J. Weitkamp (Eds.),
Handbook of Heterogeneous Catalysis, Wiley–VCH, Weinheim, 1997, p. 191.
[45] R. Schlögl, in: G. Ertl, H. Knözinger, J. Weitkamp (Eds.), Handbook of Heteroge-
neous Catalysis, Wiley–VCH, Weinheim, 1997, p. 138.
[20] A. Sepúlveda-Escribano, F. Coloma, F. Rodríguez-Renoso, Appl. Catal.
(1998) 247.
A 173
[21] J.L. Figueiredo, M.F. Pereira, M.M.A. Freitas, J.J.M. Orfao, Carbon 37 (1999) 1379.
[22] M.A. Fraga, E. Jordao, M.J. Mendes, M.M.A. Freitas, J.L. Faria, J.L. Figueiredo,
J. Catal. 209 (2002) 355.
[23] H.E. van Dam, H. van Bekkum, J. Catal. 131 (1991) 335.
[24] X. Hao, L. Quach, J. Korah, W.A. Spieker, J.R. Regalbuto, J. Mol. Catal. A 219
(2004) 97.
[25] N. Job, R. Pirard, J. Marien, J.-P. Pirard, Carbon 42 (2004) 619.
[26] N. Job, A. Théry, R. Pirard, J. Marien, L. Kocon, J.-N. Rouzaud, F. Béguin, J.-P.
Pirard, Carbon 43 (2005) 2481.
[27] J. Park, J.R. Regalbuto, J. Colloid Interface Sci. 175 (1995) 239.
[28] J.R. Regalbuto, in: J.R. Regalbuto (Ed.), Catalyst Preparation: Science and Engi-
neering, CRC Press/Taylor & Francis Group, Boca Raton, 2007, p. 297.
[29] A.J. Lecloux, in: J.R. Anderson, M. Boudart (Eds.), Catalysis: Science and Tech-
nology, vol. 2, Springer-Verlag, Berlin, 1981, p. 171.
[46] F. Coloma, A. Sepúlveda-Escribano, F. Rodríguez-Reinoso, Appl. Catal.
(1997) 165.
[47] M.L. Toebes, F.F. Prinsloo, J.H. Bitter, A.J. van Pillen, K.P. de Jong, J. Catal. 214
(2003) 78.
[48] M.L. Toebes, Y. Zhang, J. Hájek, T.A. Nijhuis, J.H. Bitter, A. Jos van Pillen, D.Y.
Murzin, D.C. Koningsberger, K.P. de Jong, J. Catal. 226 (2004) 215.
[49] D.E. Ramaker, J. de Graaf, J.A.R. van Veen, D.C. Koningsberger, J. Catal. 203
(2001) 7.
A 150
[30] C. Alié, S. Lambert, B. Heinrichs, J.-P. Pirard, J. Sol–Gel Sci. Technnol. 26 (2003)
827.
[50] M. Domingo-Garcia, F.J. Lopez-Garzon, M. Perez-Mendoza, J. Colloid Interface
Sci. 222 (2000) 233.