Screening of Pd and Ni supported on sol-gel derived oxides for dichloromethane hydrodechlorination

Article abstract of DOI:10.1016/j.molcata.2004.08.007

Several alumina, titania, zirconia and silica xerogels modified with approximately 0.05 wt.% Pd or 0.05 wt.% Ni were synthesized by cogellation and conventional impregnation method using Pd and Ni acetylacetonate as metal precursors. Thermolysis of supported complexes was conducted in static air at 300 and 350°C for Pd and Ni loaded samples, respectively. The resulting catalysts were characterized by XRD, XRF, BET, TPR, and tested in the vapor phase hydrodechlorination of dichloromethane (DCM) with molecular hydrogen. Prior to reaction, catalysts were reduced at 300°C in 5% H₂/He. The synthetic mixture consisted of 1200 ppm CH₂Cl₂, 1.2% H₂ and balance He at a space velocity of 0.005 g min ml^-1. Palladium supported catalysts were more active than those containing nickel, i.e. DCM conversions at 200°C over Pd(I)-Al₂O₃ almost doubled those obtained over Ni(I)-Al₂O₃ and no intermediate reaction products were observed over Pd supported catalysts whereas CH₃Cl was formed over nickel catalysts. Comparison of the specific activities of selected catalysts expressed as TOF, indicate that the observed differences in activity are related to the nature of the support and the accessibility of the metal function to the fluid phase, even though textural properties also appear to play a role. The most active catalyst, palladium impregnated on sol-gel titania (Pd(I)-TiO₂) had an initial specific activity of 0.36 s^-1 but, it strongly deactivated after a 48 h exposure to 1200 ppm DCM or by thermolysis of the Pd precursor in static air at 600°C. Although palladium impregnated in sol-gel alumina exhibited lower activity it was much more stable than Pd(I)-TiO₂. Xerogel alumina, titania, zirconia and silica modified with Pd and Ni by impregnation and cogellation were examined for the gas phase dichloromethane hydrodechorination at 200°C. The observed differences in activity are related to the nature and texture of the support and the accessibility of the metal function to the fluid phase.

Full text of DOI:10.1016/j.molcata.2004.08.007

Journal of Molecular Catalysis A: Chemical 222 (2004) 189–198
Screening of Pd and Ni supported on sol–gel derived oxides
for dichloromethane hydrodechlorination
a
a
b
´
´
´
Beatriz Aristizabal , Carlos Andres Gonzalez , Izaskun Barrio ,
Mario Montesb, Consuelo Montes de Correaa,∗
a
Chemical Engineering Department, Universidad de Antioquia, calle 67 No. 53-108, A.A. 1226, Medell´ın, Colombia
Applied Chemistry Department, Universidad del Pa´ıs Vasco/Euskal Herriko Unibertsitatea, P◦ Manuel de Lardiza´bal,
3E-20018 San Sebastia´n, Espan˜a
b
Received 23 July 2003; received in revised form 31 July 2004; accepted 4 August 2004
Available online 12 September 2004
Abstract
Several alumina, titania, zirconia and silica xerogels modified with approximately 0.05 wt.% Pd or 0.05 wt.% Ni were synthesized by
cogellation and conventional impregnation method using Pd and Ni acetylacetonate as metal precursors. Thermolysis of supported complexes
was conducted in static air at 300 and 350 ◦C for Pd and Ni loaded samples, respectively. The resulting catalysts were characterized by
XRD, XRF, BET, TPR, and tested in the vapor phase hydrodechlorination of dichloromethane (DCM) with molecular hydrogen. Prior to
reaction, catalysts were reduced at 300 ◦C in 5% H2/He. The synthetic mixture consisted of 1200 ppm CH2Cl2, 1.2% H2 and balance He at
a space velocity of 0.005 g min ml−1. Palladium supported catalysts were more active than those containing nickel, i.e. DCM conversions
at 200 ◦C over Pd(I)–Al2O3 almost doubled those obtained over Ni(I)–Al2O3 and no intermediate reaction products were observed over Pd
supported catalysts whereas CH3Cl was formed over nickel catalysts. Comparison of the specific activities of selected catalysts expressed as
TOF, indicate that the observed differences in activity are related to the nature of the support and the accessibility of the metal function to
the fluid phase, even though textural properties also appear to play a role. The most active catalyst, palladium impregnated on sol–gel titania
(Pd(I)–TiO2) had an initial specific activity of 0.36 s−1 but, it strongly deactivated after a 48 h exposure to 1200 ppm DCM or by thermolysis
of the Pd precursor in static air at 600 ◦C. Although palladium impregnated in sol–gel alumina exhibited lower activity it was much more
stable than Pd(I)–TiO2.
© 2004 Elsevier B.V. All rights reserved.
Keywords: Dichloromethane; Sol–gel catalysts; Pd supported catalysts; Ni supported catalysts; Catalytic hydrodechlorination; Catalyst deactivation
1. Introduction
cinogenic properties [2]. Catalytic hydrodechlorination of
ClVOCs could be a safe alternative for the treatment of these
wastes since ClVOCs are transformed into harmless non-
chlorinated compounds and hydrogen chloride at relatively
mild reaction conditions [3–5]. The resulting VOCs products
can be recovered or burned, and hydrogen chloride can be
separated by alkaline washing. Furthermore, the production
of harmful products: phosgene, Cl2 or fragments of ClVOCs,
are minimized [6]. In this sense, catalytic hydrodechlorina-
tion has potential economic and environmental advantages
over thermal incineration and catalytic oxidation.
Chlorinated volatile organic compounds (ClVOCs) play
an important role in the chemical industry both as reactants
and solvents. The increasing amounts of ClVOCs released
into the environment have increased the demand for find-
ing effective methods of their disposal and destruction [1].
The emission of chlorinated compounds to the atmosphere
contributes to ozone depletion, smog formation and global
warming. Besides, ClVOCs have suspected toxicity and car-
Catalytic hydrodechlorination of chlorinated aromat-
ics and chlorinated olefins have been more studied than
cloroalkane hydrodechlorination. Hagh and Allen com-
∗
Corresponding author. Tel.: +57 42105537; fax: +57 42638282.
E-mail address: cmontes@quimbaya.udea.edu.co (C. Montes de Cor-
rea).
1381-1169/$ – see front matter © 2004 Elsevier B.V. All rights reserved.
doi:10.1016/j.molcata.2004.08.007

190
B. Aristiza´bal et al. / Journal of Molecular Catalysis A: Chemical 222 (2004) 189–198
pared Pd and hydrotreatment (Ni, Mo) catalysts for the
hydrodechlorination of chloro- and dichlorobenzene [7].
Though the activity of both catalysts was similar for
dichlorobenzene hydrodechlorination, the selectivity to
chlorobenzene was higher over Pd than over Ni, Mo. Sup-
ported noble metals especially Pd, Pt and Rh, have been found
active at lower temperatures and pressures than hydrotreat-
ment catalysts for hydrodechlorination of most aliphatic
ClVOCs [8]. Among precious metals, systems containing Pd
are the most active. Although Ni supported catalysts have
proved useful in hydrodehalogenation processes their hy-
drodechlorination ability is rather limited and they are easily
poisoned by HCl produced during reaction. Alumina sup-
ported Pd catalysts show better performance than those sup-
ported on carbon [8–10]. A disadvantage of the alumina sup-
port in the hydrodechlorination of ClVOCs is its low resis-
tance against hydrogen chloride. Surface area decrease and
increased surface acidity due to the formation of Al halides
which would intensify the trend to coke deposition has been
of major concern [11]. Other alternative supports as AlF3,
MgO, ZrO2 or modified zeolites have been shown to be un-
stable or exhibit poor performance [12,13]. Therefore, the
selection of an appropriate support is an important issue in
the development of hydrodechlorination catalysts.
Lately, the sol–gel process has been used to prepare sup-
ported metal catalysts and catalyst supports with high thermal
stability, resistance to deactivation and flexibility to control
catalyst properties such as particle size, surface area and pore
size distribution. Numerous authors have shown the ability
to disperse catalytic metals on gels whose texture is finely
controlled by the sol gel route [14]. When the metal of in-
terest is introduced into the initial solution in the form of a
salt (cogellation), it can be incorporated into the framework
or partially buried at the surface of the support. In impreg-
nated samples the metal is spread over the carrier surface and
is more easily accessible to reactant molecules [15]. In this
study we synthesized alumina, titania, zirconia and silica xe-
rogelsmodifiedwithPdorNibycogellationandconventional
impregnation method using Pd or Ni acetylacetonate as metal
precursors. The nature of supported Pd or Ni species gen-
erated after pretreatment of the acetylacetonate (acac) pre-
cursors was investigated. The resulting catalysts were tested
in the hydrodechlorination of dichloromethane (DCM). This
halomethane was chosen as a model compound since it is the
less reactive low molecular weight chlorinated hydrocarbon
[16].
col (HG; Aldrich) and the mixture stirred at 120 ◦C for 3 h
under reflux. About 0.02 g of palladium or 0.03 g of nickel
acetylacetonate (Merck, Aldrich) dissolved in acetone was
then added and stirred for 5 h. After lowering the tempera-
ture to 100 ◦C, 4.1 ml of water was added dropwise and the
system stirred during 4 h. The solvent from gels was removed
in a rotavap at 50 ◦C for 2 h. The solid materials were dried
at 100 ◦C for 1 h, and calcined at 700 ◦C for 1 h. Catalysts are
denoted as Pd(C)–Al2O3 and Ni(C)–Al2O3.
In the preparation of cogelled silica supported Pd or Ni cat-
alysts, 10 ml of tetraethyl orthosilicate (TEOS; Sigma) and
6.8 ml of ethanol were mixed with 0.04 g palladium or 0.06 g
of nickel acetylacetonate (Merck, Aldrich) dissolved in ace-
tone. Then 6.37 ml of water and NH4OH (pH = 9) was added
and the solution stirred at 45 ◦C for 24 h. Afterwards, the tem-
perature was raised to 80 ◦C and the gel maintained at this
temperature for 1 h. The solvent was removed from gels in
a rotavap at 70 ◦C for 2 h. Solids were dried at 100 ◦C for
1 h and calcined at 450 ◦C during 4 h. Catalysts are coded as
Pd(C)–SiO2, Ni(C)–SiO2.
Cogelled titania supported Pd or Ni catalysts were syn-
thesized by mixing 8 ml titanium isopropoxide (TIP; Acros)
and 19.2 ml isopropanol (IP, Merck) at 50 ◦C. Then, 0.03 g of
palladium or 0.05 g nickel acetylacetonate (Merck, Aldrich)
dissolved in acetone was added. After adjusting the pH to 9
with NH4OH (Merck), 11.5 ml deionized water was added
dropwise. The gel was stirred under reflux for 23 h at 50 ◦C.
Finally, the temperature was raised to 70 ◦C and the system
stirred at this temperature during 6 h. The solvent was re-
moved from gels in a rotavap at 70 ◦C for 2 h. Solids were
dried at 100 ◦C for 1 h and calcined at 600 ◦C during 4 h.
Catalysts are denoted as Pd(C)–TiO2 and Ni(C)–TiO2.
Zirconia supported Pd or Ni catalysts obtained by cogella-
tion were prepared as follows: 3.2 g of zirconium isopropox-
ide (ZI, Alfa Aesar) was added to a solution of 10.32 ml of dry
isopropanol (IP, Aldrich) and 0.26 ml of HNO3 (70 wt.%, J.T.
Baker). Then, 0.015 g of palladium or 0.024 g nickel acety-
lacetonate (Merck, Aldrich) dissolved in acetone was added.
After stirring for 10 min, a solution of 5.16 ml of isopropanol
(Merck) and 0.5 ml deionized water was added at once and
the mixture stirred for 2 h at room temperature. The solvent
was then removed from the gels at 80 ◦C for about 2 h in
a rotavap and the resulting solids dried at 100◦ C for 12 h
and calcined at 600 ◦C during 4 h. Catalysts are denoted as
Pd(C)–ZrO2 and Ni(C)–ZrO2.
Pdandnickelcontentsoftheabovecatalystsasdetermined
by X-ray fluorescence (XRF) were about 0.6 wt.% ( 0.1%).
Unmodified supports: alumina, silica, titania and zirconia xe-
rogels were synthesized following the above procedures but,
without the addition of the precursor salts. Pd or Ni impreg-
nated on the xerogel oxide supports were prepared by con-
ventional incipient wetness impregnation using acetone solu-
tions of palladium or nickel acetylacetonate (Aldrich) to ob-
tain 0.5 wt.% Pd or 0.5 wt.% Ni on the xerogel oxide supports
and calcining the resulting materials at 300 and 350 ◦C, re-
spectively, during 1 h. Catalysts are coded as Pd(I)-oxide and
2. Experimental
2.1. Catalyst preparation
Catalysts were prepared by modifications of previously
reported methods [17–20]. Cogelled alumina supported cat-
alysts containing Pd or Ni were prepared by adding 5 g of
aluminum isopropoxide (AIP; Aldrich) to 2.56 g hexilengli-

B. Aristiza´bal et al. / Journal of Molecular Catalysis A: Chemical 222 (2004) 189–198
191
locity was 0.005 g min ml−1. Catalysts were stabilized in the
reaction mixture at room temperature. After getting stable
DCM concentrations (1–3 h), the reaction temperature was
increased up to 200 ◦C and the reaction products monitored
byGC-ECDatthistemperature;reactortemperaturewascon-
stant to within 1 ◦C. Conversions were obtained from the
disappearance of DCM. Each catalytic run was repeated up
to three times using fresh samples from the same batch of
catalyst and results were reproducible to within 8%.
During catalytic runs, DCM concentrations varied signif-
icantly so the experimental reactor was considered to be an
Ni(I)-oxide depending on the type of support used. For com-
parison purposes 0.5 wt.% Pd was also impregnated on com-
mercial alumina (Alpha Aesar 80–120 m2/g). It was coded
Pd(I)–Al2O3 (comm.).
2.2. Catalyst characterization
X-ray diffraction (XRD) studies were carried out on a
Philips PW 1710 X-ray diffractometer with Cu K␣ radia-
˚
tion (=1.5406 A) and Ni filter. The X-ray tube was oper-
ated at 30 kV and 20 mA. Samples were scanned from 2θ
= 5◦–85◦ and the X-ray diffraction line positions were de-
termined with a step size of 0.05◦ and a counting time of
2.5 s per step. The surface area and pore volume of catalysts
were determined by N2 adsorption–desorption at −195 ◦C in
a Micromeritics ASAP 2000 equipment. Samples were pre-
viously evacuated overnight at 450 ◦C under high vacuum.
The adsorption data were treated with the full BET equation.
N2 adsorption–desorption experiments were repeated twice
and the values of BET surface areas were reproducible within
5%. The pore size distributions were calculated applying
the Barret–Joyner–Halenda (BJH) method to the desorption
branches of the isotherms [21]. The assessments of microp-
orosity were made from t-plot constructions [22].
˜
integral reactor. Recently Ordonez and coworkers [23] found
that the kinetics of single organochlorinated compounds over
a commercial Pd/Al2O3 catalyst can be fitted to a pseudo-first
order equation with a reasonable accuracy. Although reaction
rate depends on catalyst type, we assumed a first-order rate
expression as a first approximation to compare the activity of
the synthesized materials and used integral analysis to obtain
initial reaction rates [24].
3. Results and discussion
3.1. X-ray diffraction
Chlorine content of selected samples was determined by
X-ray fluorescence (XRF) in a Siemens SRS 3000 sequential
spectrophotometer with a rhodium tube as the source of ra-
diation. Palladium dispersions and temperature-programmed
reduction (TPR) experiments were carried out using a Ther-
moquest TPD/R/O 1100 instrument. To measure palladium
dispersions by chemisorption of H2 at 80 ◦C, the catalysts
were first reduced in flowing 5% H2/Ar heating from room
temperature to 300◦ C at a rate of 2 ◦C/min and then main-
tainedatthistemperaturefor2 h. Aftersubsequentevacuation
at 300 ◦C the samples were cooled down to 80 ◦C under Ar
and subjected to H2 chemisorption (5% H2/Ar) using a pulse
titration procedure. Prior to TPR experiments, the samples
were first treated at 120 ◦C in flowing Ar. Afterwards, they
were cooled down to 50 ◦C in flowing Ar. The reduction was
conducted from 50 to 700 ◦C at a linear ramp of 10 ◦C/min in
a 5% H2/Ar stream. The apparatus was equipped with TCD
andMSdetectorsandwithadriertoremovewaterfromthere-
duction process. Chemisorption and TPR experiments were
repeated up to three times and the hydrogen uptake values
were reproducible to within 5%.
Fig. 1 shows the X-ray diffraction patterns (XRD) of the
unmodified xerogel supports used in this work: Al2O3, SiO2,
TiO2 and ZrO2, respectively. Trace (a) shows the crystallo-
graphic phase characterized by peaks at 2θ = 37.44◦, 42.66◦,
67.25◦ corresponding to ␥-alumina [25]. Trace (b) shows a
broad peak at 2θ = 24◦ which is characteristic of amorphous
silica. Anatase is the only phase obtained in titania supported
catalysts (see trace (c)) exhibiting high intensity peaks at 2θ
= 25.19◦, 37.69◦ and 47.62◦ [26]. A mixture of about 63%
tetragonal and 37% monoclinic phases was observed for the
zirconia support as shown in trace (d) [27]. No additional
phases from palladium or nickel oxides were observed in the
XRD patterns of the Pd or Ni supported catalysts [28] sug-
gesting that the metal loading was quite low for the sensitivity
limits of this technique.
2.3. Catalytic tests
Catalytic experiments were performed in a fixed bed pyrex
tubular reactor at atmospheric pressure. Prior to each reac-
tion the catalyst was pretreated in flowing 5% H2 while being
heatedatarateof2 ◦C/minto300 ◦Candthenheldatthistem-
perature during 1 h. For a typical reaction, a catalyst weight
of 100 mg (particle sizes in the range 250–425 ␮m) and a to-
tal flow of 20 ml/min of a gas stream containing 1200 ppm
CH2Cl2, 1.2% H2 and He (balance) were used. The space ve-
Fig. 1. X-ray diffraction patterns of the unmodified xerogel supports: (a)
␥-alumina; (b) silica; (c) titania; (d) zirconia.

192
B. Aristiza´bal et al. / Journal of Molecular Catalysis A: Chemical 222 (2004) 189–198
Table 1
BET surface area, pore volume, and average pore diameter of cogelled and impregnated Ni and Pd xerogel oxides
BET surface area (m2/g)
Pore volume (cm3/g)
Pore diameter (A)
Micropore area (m2/g)
˚
Catalyst
Al2O3
Pd(C)–Al2O3
Ni(C)–Al2O3
Pd(I)–Al2O3
Ni(I)–Al2O3
Pd(I)–Al2O3 (commercial)
Ni(I)–Al2O3 (commercial)
320
367
369
335
213
77
1.16
0.80
1.15
0.87
0.57
0.42
0.40
150
65–140
95
100
100
450
350
40
29
40
24
8.1
7.6
8.1
82
SiO2
496
297
249
505
500
1.0
400
40
400–500
350
600–800
105
57
66
101
92
Pd(C)–SiO2
Ni(C)–SiO2
Pd(I)–SiO2
Ni(I)–SiO2
0.30
0.55
0.97
1.1
TiO2
30
9
48
33
33
0.17
0.03
0.18
0.12
0.12
150–1100
90–900
85–1100
80–1100
75–1100
–
Pd(C)–TiO2
Ni(C)–TiO2
Pd(I)–TiO2
Ni(I)–TiO2
1.39
0.27
–
–
ZrO2
33
52
44
35
23
0.06
0.23
0.17
0.075
0.06
40
1.25
7.3
0.74
0.73
0.56
Pd(C)–ZrO2
Ni(C)–ZrO2
Pd(I)–ZrO2
Ni(I)–ZrO2
30–140
35–650
40–70
50
3.2. BET surface area and pore size distribution
size distribution in the mesopore range but, microporosity as
well as mesoporosity was observed in Pd or Ni supported
on xerogel SiO2. Most xerogel silica and titania supported
samples presented H3 or H4 type hysteresis usually found
on solids consisting of aggregates or agglomerates of parti-
cles forming slit shaped pores (plates or edged particles like
cubes), with nonuniform or uniform size and/or shape [29].
Pd(C)–TiO2, Ni(I)–SiO2 and Ni(C)–SiO2 samples showed a
reduced hysteresis loop. This is the case of blind cylindrical,
cone-shaped and wedge-shaped pores [29]. Ni(C)–ZrO2 ex-
hibited isotherm type II, and bimodal pore size distribution
exhibiting pores in the meso- and macro-pore range and low
microporosity. Pd(I)–ZrO2 exhibited type IV isotherms. Co-
gelled Ni or Pd zirconia catalyst exhibited type H3 hysteresis
while impregnated ones showed H4 type hysteresis.
Table 1 compiles the results of BET surface area, pore
volume, and average pore diameter of cogelled and impreg-
nated Ni and Pd xerogel oxides. Alumina based xerogel cata-
lysts exhibited high surface areas. For comparison Pd and Ni
were also impregnated in commercial alumina samples (Al-
pha Aesar 80–120 m2/g). The surface areas of xerogel alu-
mina supported catalysts are larger than those supported on
commercial Al2O3. The latter contain larger particles and less
microporosity than xerogel alumina supported catalysts. No
significant differences were found in the texture of cogelled
and impregnated Pd or Ni alumina supported catalysts. The
nitrogen adsorption isotherms for the alumina supported ma-
terials (not shown) correspond to type IV according to the
IUPAC classification and are typical of mesoporous solids.
Desorption isotherms for the alumina supported catalysts (not
shown) exhibit hysteresis of the Type H1 which is the charac-
teristic of solids consisting of spheroidal particles of uniform
size and shape [29]. The addition of Pd or Ni did not sig-
nificantly change the superficial area or the isotherm type
and shape but, decreased the pore diameter. Microporosity
was confirmed in xerogel alumina supported catalysts as in-
ferred from Table 1 and from the so-called t-plot in which
the straight line does not pass through the origin of axes. In
all cases the c constant from the BET equation is 120–150
confirming that the BET model is in the optimal range 50 <
c < 200.
3.3. Hydrodechlorination reactions
Fig. 2(a) shows DCM conversion as a function of H2/DCM
ratio over Pd and Ni xerogel alumina supported catalysts
at 200◦ C. The conversion profile of 0.5% Pd impregnated
on commercial alumina (Pd(I)–Al2O3 (comm.)), is also
shown in Fig. 2(a). DCM conversion slightly increases with
H2/DCM ratio on Pd loaded catalysts while it is rather con-
stant over Ni supported alumina catalysts. Ni loaded cata-
lysts and Pd cogelled (Pd(C)–Al2O3) samples were much
less active than Pd impregnated samples (Pd(I)–Al2O3). At
H2/DCM ratios lower than 10 the activity of Pd(I)–Al2O3
(comm.) is higher than Pd(I)–Al2O3 but, they are close to
each other at H2/DCM = 10. Therefore, this ratio was chosen
for further experiments. The temperature profiles of DCM hy-
Silica, and titania xerogel supported catalysts exhibited
type II isotherms observed for adsorbents containing macro-
pores. Pd or Ni supported on xerogel titania showed pore

B. Aristiza´bal et al. / Journal of Molecular Catalysis A: Chemical 222 (2004) 189–198
193
Fig. 3. Initial DCM reaction rates at 200 ◦C over Pd or Ni loaded on different
xerogel supports. H2/DCM ratio of 10.
nation at 200 ◦C than Pd(I)–Al2O3 and Pd(I)–TiO2, respec-
tively. The difference in activity for cogelled xerogel alu-
mina supported catalysts, Pd(C)–Al2O3 and Ni(C)–Al2O3,
is less significant indicating that in cogelled catalysts the
metal function may be inaccessible or difficult to reach by
the reactant stream. DCM activity at 200 ◦C over Pd im-
pregnated catalysts varies in the following order: Pd(I)–TiO2
> Pd(I)–Al2O3 > Pd(I)–ZrO2 > Pd(I)–SiO2 while on co-
gelled ones, Pd(C)–ZrO2 ∼ Pd(C)–TiO2 > Pd(C)–SiO2 ∼
Pd(C)–Al2O3. These results are in agreement with those of
Fig. 2. (a) DCM conversion as a function of H2/DCM ratio over Pd and
Ni xerogel alumina supported catalysts at 200 ◦C; (b) temperature profiles
of DCM hydrodechlorination over Pd(C)–Al2O3, Ni(C)–Al2O3 and of the
homogeneous reaction at H2/DCM = 10.
´
Lopez et al. [32] who reported that in cogelled ruthenium
catalysts a large amount of metal atoms rests in the bulk of
the carrier leading to low turnover rates for several reactions.
Fig. 4 compares the effect of exposing selected catalyst
samples to the reaction mixture during 1–3 h or 48 h before
DCM hydrodechlorination. Interestingly, Fig. 4 shows that
the reaction rate at which DCM disappears over cogelled
alumina supported catalysts increased after 48 h pretreatment
withDCM. Thisobservation, combinedwithfurtherevidence
presented in the literature [33], leads us to suggest that chlo-
ride ions in the alumina support induce agglomeration of Pd
clusterswhichmaymigratetothesupportsurfaceandbecome
more accessible to reactants. As shown in Fig. 4, the activ-
ity of Pd(I)–TiO2 dramatically decreased when exposed 48 h
to 1200 ppm DCM before reaction with hydrogen. The de-
crease observed on Pd(I)–Al2O3 or Ni(I)–Al2O3 under simi-
lar conditions was much lower than on Pd(I)–TiO2. However,
durability tests in the present study indicate that DCM hy-
drodechlorination is more stable over Pd(I)–Al2O3 than over
Ni(I)–Al2O3, i.e. DCM conversion over Pd(I)–Al2O3 was
constant at about 44% at 200 ◦C during 48 h while a con-
version decrease from 40% to about 20% after 36 h was ob-
served over Ni(I)–Al2O3 under the same conditions. Table 2
shows chloride contents of selected samples as determined
by XRF. Table 2 suggests that Pd is less prone to deacti-
vation by chlorine than Ni. Choi and Lee reported that ad-
sorbed HCl modified the crystal structure of metallic Ni to
NiCl2 leading to irreversible catalyst deactivation and metal
drodechlorination at H2/DCM = 10 over Pd(C)–Al2O3 and
Ni(C)–Al2O3, as well as, that of the homogeneous reaction
are compared in Fig. 2(b). It can be observed that a cata-
lyst is required for DCM hydrodechlorination and enables
comparison of the metal character on DCM conversion. Pd
loaded catalysts appear to be more active than those contain-
ing Ni. Formation of other organochlorides was not observed
on Pd based catalysts. In contrast, selectivity towards CH3Cl
slightly increased over Ni supported xerogel alumina cata-
lysts. The low selectivity of Ni catalysts may be explained
by the high chlorine content on the Ni surface, found by
others [30]. In contrast, a PdCx solid solution has been de-
tected during CH2Cl2 hydrodechlorination over a Pd/Al2O3
catalyst [31] suggesting that the mildly carbided palladium
surface binds intermediate species less strongly, resulting in
their more facile hydrogenative desorption [31].
Initial reaction rates (200 ◦C, H2/DCM = 10) of Pd and
Ni supported on xerogel alumina, silica, titania and zir-
conia are compared in Fig. 3. Among the catalysts tested
Pd(I)–TiO2 shows the highest activity. The initial DCM re-
action rate over impregnated xerogel titania and alumina
supported catalysts appears to be a clear function of the
type of metal loaded. Ni(I)–Al2O3 and Ni(I)–TiO2 cata-
lysts exhibited much less activity for DCM hydrodechlori-

194
B. Aristiza´bal et al. / Journal of Molecular Catalysis A: Chemical 222 (2004) 189–198
Fig. 4. The effect of pretreatment (1–3 h or 48 h) on initial DCM reaction rates at 200 ◦C, H2/DCM ratio of 10 over selected catalyst samples.
sintering [34]. On the other hand, previous findings indicate
that chlorine coverage under hydrodechlorination conditions
decreases in the series Ru > Rh > Pd postulating that HCl
formed on the Pd surface during DCM hydrodechlorination
either desorbs to the gas phase or migrates to alumina [35].
Notwithstanding, the inhibition of palladium metal supported
catalyst by the HX formed in the hydrodehalogenation of or-
ganic halides is generally accepted [36]. It can also be seen
from Table 2 that spent Pd(I)–TiO2 adsorbed less chlorine
than spent Pd(I)–Al2O3 but, the former deactivated faster
when exposed 48 h to a DCM stream at room temperature.
The oxygen vacancies of titania are known to capture chlo-
ride species, so exposure to the chlorinated reactant appears
to alter the nature of DCM and/or hydrogen activation step
by altering the electronic structure of the adsorption site(s).
Pd(I)–TiO2 also deactivated after the first catalytic run. The
BET surface area of a spent Pd(I)–TiO2 sample decreased
from 33 to 14 m2/g and hydrogen chemisorption was inhib-
ited after DCM hydrodechlorination reaction, vide infra. The
latter result is in agreement with Dow and Huang who re-
ported that reduction of the titania support is difficult when
chorine atoms are incorporated into the oxygen vacancies of
titania [37].
3.4. Hydrogen chemisorption and
temperature-programmed reduction (TPR)
As a first approximation, hydrogen chemisorption up-
takes, metal dispersions, particle sizes and TOF of selected
Pd supported on Al2O3 and TiO2 is provided in Table 3.
Hydrogen chemisorption on Ni supported alumina and ti-
tania catalysts was below analytical detection limits (0.05
loops). Particle sizes of Pd loaded catalysts were estimated
from metal dispersion using ds (nm) = 112/D(%) assuming
spherical particles [38]. The assignment of metal particle
size/dispersion values based on chemisorption measurements
is limited, since it assumes an exclusive H:Pd or H:Ni stoi-
chiometry and ignores possible surface coverage effects and
possible metal site blocking by residual halogen [39]. Never-
theless, diminished hydrogen chemisorption on Pd(I)–Al2O3
may be attributed to sintering of Pd particles. The same phe-
nomenon was observed on Pd(I)–TiO2 samples used (about
16 h), on a fresh Pd(I)–TiO2 sample following pretreatment in
air at 600 ◦C, and on fresh Pd(C)–TiO2 catalysts. A strong in-
hibiting effect of hydrogen chemisorption is characteristic of
strong metal–support interactions (SMSI) [40]. The suppres-
sioninhydrogenchemisorptionhasbeenascribedtothephys-
ical blockage of metal particles by the mobile TiOx phase,
which is formed due to facile reduction of the TiO2 in the
presence of metal and hydrogen [40]. Initial TOFs increased
with increasing particle size for catalysts pretreated using
standard pretreatment (calcination in static air at 300 ◦C).
Table 2
Chloride contents of selected samples
Catalyst
Cl (wt.%)
Pd(I)–Al2O3 fresh
0.04
0.8
3.9
2.4
0.21
Pd(I)–Al2O3 after reactiona
b
Pd(I)–Al2O3
Ni(I)–Al2O3
c
Pd(I)–TiO2 after reactiond
a
Previously exposed 48 h to the reaction stream at 25 ◦C.
b
Exposed 1–3 h to the reaction stream at 25 ◦C and 48 h at 200 ◦C.
Exposed 1–3 h to the reaction stream at 25 ◦C and 36 h at 200 ◦C.
Previously exposed 1–3 h to the reaction stream at 25 ◦C.
c
d

B. Aristiza´bal et al. / Journal of Molecular Catalysis A: Chemical 222 (2004) 189–198
195
Table 3
Hydrogen chemisorption of selected Pd supported samples
Catalyst
H2 chemisorption (␮mol/gcat
)
Dispersion (%)
Particle size (nm) Ds,H2
Initial TOF (s−1
)
Pd(C)–Al2O3
Pd(I)–Al2O3
Pd(I)–TiO2
5.7
4.3
2.5
0.005
0.2
–
24.5
18.1
10.7
<0.16a
1.1
<0.8a
1.6
4.6
6.2
10.5
–
102
–
0.026
0.083
0.357
–
c
Pd(C)–TiO2
b
Pd(I)–Al2O3
b
c
Pd(I)–TiO2
d
Pd(I)–TiO2
0.4
70
0.106
a
Lower than detection limit.
After reaction.
These used samples were not further tested in DCM hydrodechlorination.
Pretreated in air under static conditions at 600 ◦C.
b
c
d
However, when Pd(I)–TiO2 was calcined in air under static
conditions at 600 ◦C particle size increased and TOF also
decreased.
In order to obtain preliminary insights into the nature of
the species involved in DCM hydrodechlorination H2-TPR
runs were performed over fresh palladium and nickel sup-
ported alumina and titania catalysts. The corresponding H2-
TPR profiles are shown in Figs. 5–8. The H2-TPR profile of
a Pd(C)–Al2O3 sample (trace (a), Fig. 5) shows a negative
peak around 90 ◦C and a broad negative feature between 400
and 700 ◦C. The hydrogen evolution at 90 ◦C corresponds
to decomposition of a palladium hydride phase formed at
lower temperatures during TPR. The formation of Pd in this
oxidized sample might be due to adsorption of CO or other
organic fragments from the Pd(acac)2 precursor onto the gen-
erated PdO. This result suggests that decomposition to PdO
proceeded via the evolution of CO, CO2 and a variety of CHO
fragments. MS analysis confirms the existence of residual or-
ganics. The MS signals up to 100 amu indicate that mass 28
evolved around 550 ◦C together with hydrogen, likely due
to decomposition of carbonaceous compounds sitting deep
inside the porous alumina matrix. Interestingly, the negative
broad feature disappeared after heating a Pd(C)–Al2O3 sam-
ple at 200 ◦C in O2 for 1 h.
Fig. 6. H2-TPR profile of Pd loaded xerogel titania catalysts: (a) TiO2; (b)
Pd(I)–TiO2; (c) Pd(C)–TiO2.
TPR of Pd(I)–Al2O3 (trace (b), Fig. 5) shows a hydro-
gen consumption peak at ca. 199 ◦C with a shoulder around
290 ◦C and a broad signal under the base line from 300 to
Fig. 7. H2-TPR profile of Ni loaded xerogel ␥-alumina catalysts: (a)
Ni(C)–Al2O3; (b) Ni(I)–Al2O3; (c) Ni(I)–Al2O3 after DCM hydrodechlo-
rination.
Fig. 5. H2-TPR profile of Pd loaded xerogel ␥-alumina catalysts: (a)
Pd(C)–Al2O3; (b) Pd(I)–Al2O3.

196
B. Aristiza´bal et al. / Journal of Molecular Catalysis A: Chemical 222 (2004) 189–198
titania support. The total hydrogen consumption corresponds
to reduction of about five times the hydrogen necessary for
reduction of loaded palladium confirming SMSI.
TPR of cogelled Pd(C)–TiO2 sample (trace (c), Fig. 6)
shows peaks at 174 and 500 ◦C. Peak at 174 ◦C is assigned to
bulk PdO reduction. The large peak at 500◦ C is attributed to
H2 consumption of the titania carrier. The reduction of titania
in Pd loaded titania samples is shifted to lower temperatures
as compared to bare titania [44] and can be attributed to hy-
drogen dissociation at the surface of the Pd particles and the
successive spillover of dissociated hydrogen atoms that can
reduce TiO2 more easily. Reduction at higher temperature
give rise to SMSI between Ti3+and Pd which could stabilize
Ti3+, i.e. titania is partly reduced and suboxide phase mi-
grates onto the metal particles. Total hydrogen consumption
corresponds to reduction of about nine times the hydrogen
necessary to reduce loaded palladium. Since chemisorption
uptake was inhibited on this sample it suggests that part of
metal surface partially covered by TiOx is blocked as has
been reported [40].
The TPR patterns of fresh and used Ni(I)–Al2O3 and
fresh Ni(C)–Al2O3 samples are shown in Fig. 7. The TPR
profile of Ni(C)–Al2O3, (trace (a)), shows a negative peak
around 480 ◦C indicative of hydrogen evolution and masses
12, 15, 18, 28, 42, 53, etc. as determined by MS. The hydro-
gen consumption peak around 680 ◦C revealed in the TPR of
Ni(C)–Al2O3 is assigned to reduction of bulk NiO and cor-
responds to reduction of approximately 31% of Ni species.
The relatively high reduction temperature compared with val-
ues obtained for unsupported NiO (382 ◦C) strongly points
to the existence of interaction between Ni species and the
alumina support. Li and Chen observed that at low content,
nickelpreferentiallyoccupiedthetetrahedralsitesofalumina,
forming a nonreducible surface spinel [45]. A negative peak
around 600 ◦C is observed in the TPR profiles of fresh and
used Ni(I)–Al2O3 samples (traces (b) and (c)). As previously
reported [46] submitting Ni(acac)2/Al2O3 catalysts to a pre-
vious heating at 250 ◦C in an oxidizing atmosphere ensured
the decomposition of the acac ligands. However MS analysis
indicates that together with hydrogen, masses 12, 14, 15, 28,
52, etc. are evolved around 600 ◦C due to decomposition of
organic residues. No peaks due to reduction of Ni species
are observed in these samples. The pseudo-peak at 680 ◦C
does not correspond to H2 consumption and no clear assign-
ment can be given by MS. Decreased reducibility of nickel
in NiO/Al2O3 catalysts has been observed and attributed to
reinforced chemical interaction with the support, changes of
the NiO crystallites size and incorporation of mobile Al3+
into NiO crystallites, and formation of non-stoichiometric or
stoichiometric nickel aluminate [46].
Fig. 8. H2-TPR profile of Ni loaded xerogel titania catalysts: (a) Ni(I)–TiO2;
(b) Ni (C)–TiO2.
500 ◦C with a minimum at 440 ◦C. The peak at 199 ◦C is
assigned to bulk PdO reduction. MS analysis shows that to-
gether with the H2 signal, a mass of 28 is also released around
290 ◦C. Again, different masses (18, 28, 43, 78, etc.) are
evolved with hydrogen around 440 ◦C due to decomposition
of organic residues. TPR experiments performed on blank
xerogel alumina did not show any reducibility of the sup-
port in the 50–700 ◦C range indicating that organics used to
synthesize the support were totally oxidized during support
calcination.
Fig. 6 shows the characteristic H2-TPR profiles of Pd sup-
ported titania samples treated in Ar at 130 ◦C. Traces (a), (b)
and (c) correspond to TiO2, Pd(I)–TiO2 and Pd(C)–TiO2, re-
spectively. Commercial titania was found to be reduced at
very high temperature (around 743 ◦C) and TPR profile is
very broad, ranging from 550 to 774 ◦C [41]. Therefore, the
peak at 650 ◦C in trace (a) is attributed to the partial reduc-
tion of the titania support. The TPR profile of Pd impregnated
catalyst, trace(b), exhibitsreductionpeaksat149, 260(shoul-
der), 360 and 550 ◦C. The feature at 149 ◦C is assigned to the
one-step reduction of bulk PdO. Hydrogen consumption at
this temperature corresponds to reduction of about 53% of
hydrogen necessary for reduction of PdO to metallic palla-
dium. MS analysis shows that together with H2 consumption
at 260◦ C masses 28, 57, etc. evolved. The latter one is most
likely due to acetone. Kohler et al. [42] proposed that metal
acetylacetonates are partly decomposed by the catalytic ac-
tion of the acidic and basic sites of the support involving the
splitting of the acetylacetonate ligands and, possibly, the fur-
ther transformation to acetone and acetic acid which strongly
adsorb on the support. Arnoldy and Moulijn [43] showed that
reduction of acetone required H2 consumption and resulted
in the production of CH4. The reduction peak at 360 ◦C is
likely due to a PdO surface phase that is interacting strongly
with the support and corresponds to reduction of 83% of
the Pd species in Pd(C)–TiO2. The small broadened peak at
550 ◦C in trace (b) is attributed to the partial reduction of the
TPR profiles of Ni(I)–TiO2 and Ni(C)–TiO2 samples are
shown in Fig. 8. TPR of Ni(I)–TiO2 sample (trace (a)) ex-
hibits reduction peaks at 220, 300, 370 ◦C with a shoulder
at 440 ◦C and a broad feature between 520 and 800 ◦C. The
reduction of nickel oxide typically occurs around 320 ◦C on
physical mixtures of NiO/TiO2. Consequently, peaks below

B. Aristiza´bal et al. / Journal of Molecular Catalysis A: Chemical 222 (2004) 189–198
197
400 ◦C (trace (a), Fig. 8) are assigned to reduction of NiO
[47]. During reduction, two competitive processes are re-
ported to occur, i.e. the reduction of NiO and interdifussion
of NiO and TiO2. At 300 ◦C the relative rate of reduction is
high and a supported catalyst results, whereas at temperatures
close to 500 ◦C considerable amounts of NiTiO3 or a solid
solution are formed [47]. Therefore, the shoulder observed
at 440 ◦C in the TPR profile of a Ni(I)–TiO2 sample might
correspond to the reduction of a nickel titanate phase. At this
temperature simultaneous CO evolution from decomposition
of organic residues was also observed. Further reduction of
NiTiO3 causes segregation of the incorporated titanium ions
and TiOx islands on the surface of the nickel crystallites are
formed [47]. The broad feature between 520 and 800 ◦C may
be assigned to titanium ions transported to the surface of the
nickel crystallites by a bulk diffusion mechanism. The total
calculated hydrogen consumption is almost that required for
reduction of Ni species. In the case of Ni(C)–TiO2 samples,
reduction peaks shifted to still higher temperatures (see trace
b, Fig. 8). Geus and coworkers [47] observed that calcina-
tions of Ni/TiO2 catalysts at 750 ◦C resulted in the formation
of NiTiO3. The migratory TiOx phase may have formed in
Ni(C)–TiO2 catalysts due to the reduction of NiTiO3 which
might have been produced during catalyst calcination at
700 ◦C. The total calculated hydrogen consumption amounts
to twice the hydrogen necessary for the reduction of incor-
porated Ni, indicating reduction of the titania support and
consequently SMSI.
Most authors agree that the thermolysis of supported com-
plexes in an oxygen-containing atmosphere gives the cor-
responding metal oxides that are then reduced to generate
dispersed metal particles [48–50]. Pd and Ni catalysts, syn-
thesized in this work by impregnation, were calcined under
static air at 300 or 350 ◦C, respectively, while cogelled cat-
alysts were calcined at 700 ◦C. TPR of Pd or Ni supported
on alumina and titania indicate that residual organics were
present in all synthesized catalysts. The source of carbon in
the Pd or Ni crystallites might mostly come from the pal-
ladium acetylacetonate precursor. Similar contamination of
the Pd surface by carbon from palladium acetylacetonate
was demonstrated by Vannice and coworkers in Pd/C and
Pd/SiO2 samples during a standard hydrogen reduction pre-
treatment at 300 ◦C [51]. TPR profiles in the present study
indicate that the decomposition temperature of residual car-
bonaceous compounds under reducing conditions was higher
for cogelled than for impregnated catalysts. Therefore, in co-
gelled samples residual organics appear to be located in the
bulk, while they are more superficial in the impregnated cat-
alysts, as expected from the preparation method used. Addi-
tionally, Pd(I)–Al2O3 appears to contain much more carbon
residues than Pd(I)–TiO2. Notwithstanding, low deactivation
was observed in the former samples suggesting that carbon
deposition is not a major factor in catalyst deactivation.
Interestingly, the size of metal particles in Pd(C)–Al2O3
is the smallest one, even though a high temperature (700 ◦C)
was used to remove the organic ligands. This result suggests
that Pd crystallites in Pd(C)–Al2O3 are less prone to sinter-
ing. As demonstrated by the textural analysis, this sample
exhibits a monodisperse pore size distribution in the meso-
porous domain, so it is reasonable to assume that palladium
particles with a diameter of about 4.1 nm would be located
inside porous alumina particles whose diameter is between
6.5 and 14 nm.
The role of the support in the hydrodehalogenation of or-
ganic halides has been studied by several authors [52–55].
For example, the effect of the support porous structure on
the catalytic activity for hydrodechlorination of CFC-12 over
Pd/Al2O3 was insignificant. However, it was found to influ-
ence the selectivity [52]. The structure sensitivity of the hy-
drodehalogenation reaction over Pd catalysts is still unclear.
Although some authors have reported that the reaction is not
structure sensitive [55], others have found that an increase in
dispersion leads to a decrease in hydrodechlorination activ-
ity [9]. Our results indicate that the initial catalytic activity of
catalysts pretreated in static air at 300 ◦C is inversely related
to metal dispersion (%D). The initial observed TOF trend
´
is: Pd(I)–TiO2 > Pd(I)–Al2O3 > Pd(C)–Al2O3.Aramendıa et
al. [54], used catalysts having similar metal dispersions and
showed that the acid–base properties of the support influ-
ence the specific activity of palladium supported catalysts in
the liquid-phase hydrodechlorination of chlorobenzene. Ba-
sic supports such as ZrO2 neutralized the HCl formed in the
reaction. Our results do not show a clear trend of support acid-
ity for DCM hydrodechlorination considering that the rela-
tive acidity of traditional supports decrease in the sequence:
Al2O3 > TiO2 > ZrO2 > SiO2. The accessibility of the metal
function to the fluid phase appears to be a more important
parameter, i.e. Pd impregnated over titania and alumina was
much more active than cogelled Pd on the same supports.
Furthermore, textural properties appear to be very important
for DCM hydrodechlorination. For example, Ni–ZrO2 which
exhibits textural properties rather different than the other cat-
alysts tested in this work is the less active catalyst for DCM
hydrodechlorination.
4. Conclusions
This work allows us to draw several interesting conclu-
sions as regards to the influence of the metal function, method
of preparation and xerogel support on DCM hydrodechlori-
nation:
• Overall, Pd loaded catalysts are more active and selec-
tive than those containing nickel and impregnated cata-
lysts show superior activity than cogelled ones. However,
cogelled alumina based catalysts appear to be more re-
sistant to sintering and their activity increased after being
exposed to the reaction mixture during 48 h.
• Pd(I)–TiO2 exhibited the highest activity at 200 ◦C
but, it was easily deactivated by the reaction mixture.
Pd(I)–Al2O3 showed good activity and resistance to de-

198
B. Aristiza´bal et al. / Journal of Molecular Catalysis A: Chemical 222 (2004) 189–198
[22] B.C. Lippens, J.H. DeBoer, J. Catal. 4 (1965) 319O.
activation up to about 48 h on stream, whereas Pd(I)–ZrO2
and Pd(I)–SiO2 showed moderate activity.
[23] E. Lo´pez, S. Ordo´n˜ez, H. Sastre, F.V. D´ıez., J. Haz. Mater. B 97
(2003) 281.
[24] O. Levenspiel, Ingenier´ıa de las Reacciones Qu´ımicas, Limusa Wiley,
Me´xico, DF, 2004, p. 411.
[25] L.V. Azaroff, Elements of X-ray Crystallography, McGraw Hill, New
York, 1968.
[26] M. Schneider, M. Wildberger, M. Maciejewski, D.G. Duff, T. Malla´t,
A. Baiker, J. Catal. 148 (1994) 625.
• Strongmetal–supportinteractions(SMSI)weremainlyob-
served in titania supported Pd and Ni catalysts. Under re-
ducing conditions titania is partly reduced and part of metal
surface may partially be covered by TiOx. Also, the SMSI
effect is observed in Ni loaded alumina catalysts, particu-
larly those prepared by cogellation.
[27] L. Castro, P. Reyes, C. Montes de Correa, J. Sol. Gel. Sci. Technol.
25 (2002) 159.
[28] B.H. Aristiza´bal, Master Thesis, Chemical Sciences, UdeA, 2003, p.
141.
[29] G. Leofanti, M. Padovan, G. Tozzola, B. Venturelli, Catal. Today 41
(1998) 207.
[30] S. Ordo´n˜ez, F.V. D´ıez, H. Sastre, Communication in First European
Congress in Chemical Engineering, Florence, 1997.
[31] A. Malinowski, D. Lomot, Z. Karpinski, Appl. Catal. B 9 (1998)
79.
[32] T. Lo´pez, M. Mora´n, J. Navarrete, L. Herrera, R. Go´mez, J. Non-
Cryst. Solids 147 (1992) 773.
Acknowledgements
Financial support from Colciencias and Universidad de
Antioquia through the project 1115-05-10129 is gratefully
acknowledged. B.A. is grateful to red Vc from Cyted program
and UdeA for supporting her stay in the group of professor
Mario Montes.
[33] Z.C. Zhang, B.C. Beard, Appl. Catal. A 188 (1999) 229.
[34] Y.H. Choi, W.Y. Lee, Catal. Lett. 67 (2000) 155.
[35] S. Ordon˜ez, F.V. D´ıez, H. Sastre, Appl. Catal. B 25 (2000) 49.
[36] F.J. Urbano, M.J. Marinas, J. Mol. Catal. A 173 (2001) 329.
[37] W. Dow, T. Huang, J. Catal. 160 (1996) 229.
[38] C. Belver, M.J. Lo´pez-Mun˜oz, J.M. Coronado, J. Soria, Appl. Catal.
B 46 (1996) 497.
References
[1] E.J. Shin, M.A. Keane, Appl. Catal. B 18 (1998) 241.
[2] USEPA, Needs for Eleven TRI Organic Chemical Groups, Pollution
Prevention Research, Washington, DC, 1991.
[3] B. Coq, G. Ferrat, F. Figueras, J. Catal. 101 (1986) 434.
[4] D.I. Kim, D.T. Allen, Ind. Eng. Chem. Res. 36 (1997) 3019.
[5] L. Prati, M. Rossi, Appl. Catal. B 23 (1999) 135.
[6] M. Martino, R. Rosal, H. Sastre, F.V. D´ıez, Appl. Catal. B 20 (1999)
301.
[39] K.V. Murthy, P.M. Patterson, G. Jacobs, B.H. Davis, M.A. Keane, J.
Catal. 223 (2004) 74.
[40] Y. Li, Y. Fan, H. Yang, B. Xu, L. Feng, M. Yang, Y. Chen, Chem.
Phys. Lett. 372 (2003) 160.
[41] S.K. Maity, M.S. Rana, S.K. Bej, J. Ancheyta-Jua´rez, G. Mu-
[7] B.F. Hagh, D.T. Allen, Chem. Eng. Sci. 45 (1990) 2695.
[8] S. Ordon˜ez, F.V. D´ıez, H. Sastre, Appl. Catal. B 31 (2001) 113.
[9] W. Juszczyk, A. Malinowski, Z. Karpinski, Appl. Catal. A 166
(1998) 311.
[10] S. Ordon˜ez, F.V. D´ıez, H. Sastre, Ind. Eng. Chem. Res. 41 (2002)
505.
[11] B. Coq, J.M. Cognion, F. Figueras, D. Trournigant, J. Catal. 141
(1993) 21.
[12] D.J. Moon, M.J. Chung, K.Y. Park, S.I. Hong, Appl. Catal. A 168
(1998) 159.
[13] S.Y. Kim, H.C. Choi, O.B. Yanga, K.H. Lee, J. Chem. Soc. Chem.
Commun. 21 (1995) 2169.
[14] B. Heinrichs, F. Noville, J-P. Pirard, J. Catal. 170 (1997) 366.
[15] I.H. Cho, S.B. Park, S.J. Cho, R. Ryoo, J. Catal. 173 (1998) 295.
[16] G. Sinquin, C. Petit, S. Libs, J.P. Hindermann, A. Kinnemann, Appl.
Catal. B 27 (2000) 105.
[17] F. Mizukauni, US Patent 5,780,102 (1998).
[18] G. Pecchi, M. Morales, P. Reyes, React. Kinet. Catal. Lett. 61 (1997)
237.
rali Dhar, T.S.R. Prasada Rao, Appl. Catal.
215.
A 205 (2001)
[42] S. Kohler, M. Reiche, C. Frobel, M. Baerns, Studies in Surface
Science and Catalysis, vol. 91, Elsevier, Amsterdam, 1995, p. 1009.
[43] P. Arnoldy, J.A. Moulijn, J. Catal. 93 (1985) 38.
[44] W. Shen, M. Okumura, Y. Matsumura, M. Haruta, Appl. Catal. A
213 (2001) 225.
[45] C.H. Li, Y.W. Chen, Thermochim. Acta 256 (1995) 457.
[46] R. Molina, G. Poncelet, J. Catal. 173 (1998) 257.
[47] P.K. de Bokx, R.L.C. Bonne, J.W. Geus, Appl. Catal. 30 (1987)
33.
[48] J.C. Kenvin, M.G. White, M.B. Mitchell, Langmuir 7 (1991) 1198.
[49] M.G. White, Catal. Today 18 (1993) 73.
[50] R. Sekine, M. Kawai, K. Asakura, T. Hikita, M. Kundo, Surf. Sci.
278 (1992) 175.
[51] N. Krishnankutty, J. Li, M.A. Vannice, Appl. Catal. A 173 (1998)
137.
[52] W. Juszczyk, A. Malinowski, M. Bonarowska, Z. Karpinski, Polish
J. Chem. 71 (1997) 1314.
[19] L.A. Contreras, J.L.R. Cerda, Actas do XVII Simposio Iberoameri-
cano de Cata´lise, Porto, 2000, p. 205.
[20] A.F. Bedilo, K.J. Klabunde, J. Catal. 176 (1998) 448.
[21] E.P. Barret, L.G. Joiyner, P. Halenda, J. Am. Chem. Soc. 73 (1951)
373.
[53] F.H. Ribeiro, C.A. Gerken, G.A. Somorjai, C.S. Kellner, G.W. Coul-
ston, L.E. Manzer, L. Abrahams, J. Catal. 176 (1998) 352.
[54] M.A. Aramend´ıa, V. Bora´u, I.M. Garc´ıa, C. Jime´nez, F. Lafont, A.
Marinas, J.M. Marinas, F.J. Urbano, J. Catal. 187 (1999) 392.
[55] J.A. McCaulley, J. Phys. Chem. 97 (1993) 10372.