Influence of chlorine on the catalytic properties of supported rhodium, iridium and platinum in ring opening of naphthenes

Article abstract of DOI:10.1016/j.apcata.2013.05.009

Pt, Ir and Rh were deposited on SiO₂ or Al₂O ₃ using chlorinated precursors and various amounts of HCl in the impregnation medium. The Br?nsted and Lewis acidities increased with the chlorine content of the alumina supported catalysts. The silica-supported catalysts only presented Lewis acid sites. The catalysts were evaluated in methylcyclopentane (MCP) and methylcyclohexane (MCH) ring-opening (RO) under pressure (2.85 and 3.95 MPa, respectively), from 200 to 425 C. For MCP conversion, the acidity of the alumina support had no sensitive effect on the activity and selectivity to RO products, and few effects on the distribution of RO products. No isomerization or hydrocracking products were observed, confirming that these reactions occurred mainly on the metal function, which was not modified by the presence of chlorine. The nature of the support, SiO ₂ or Al₂O₃, had a strong effect on both the activity (1.9 against 0.5 mol h^-1 g^-1_metal for Ir/Al₂O₃ and Ir/SiO₂, respectively at 225 C) and selectivity to RO products (99.6% against 97.5% for Ir/Al₂O ₃ and Ir/SiO₂, respectively, at 80% of MCP conversion) for Ir catalysts only. Interestingly, the Rh/SiO₂ exhibited a high selectivity for converting MCP to RO products, similar to Ir/Al ₂O₃, i.e. 99.6% at 80% of conversion. Depending on the metal and the supports, three types of behavior were observed for MCH ring-opening: (i) a direct ring-opening on the metal function whatever the support for Ir, (ii) a first step of isomerization, and then a need of a sufficiently acidic support, for Pt and (iii) an intermediate behavior for Rh, which was able to either directly convert MCH in absence of acidic support or favor a bifunctional mechanism on chlorinated alumina.

Full text of DOI:10.1016/j.apcata.2013.05.009

Applied Catalysis A: General 462–463 (2013) 207–219
Contents lists available at SciVerse ScienceDirect
Applied Catalysis A: General
journal homepage: www.elsevier.com/locate/apcata
Influence of chlorine on the catalytic properties of supported
rhodium, iridium and platinum in ring opening of naphthenes
P. Samoilaa,b, F. Epron , P. Marécot , C. Especel
a,∗
a
a
a
Institut de Chimie des Milieux et Matériaux de Poitiers, Université de Poitiers, UMR7285 CNRS, B27, 4 rue Michel Brunet, 86022 Poitiers Cedex, France
Petru Poni, Institute of Macromolecular Chemistry, 41A Grigore Ghica Voda Alley, 700487 Iasi, Romania
b
a r t i c l e i n f o
a b s t r a c t
Article history:
Pt, Ir and Rh were deposited on SiO2 or Al O using chlorinated precursors and various amounts of
2
3
Received 29 January 2013
Received in revised form 6 May 2013
Accepted 7 May 2013
HCl in the impregnation medium. The Brønsted and Lewis acidities increased with the chlorine content
of the alumina supported catalysts. The silica-supported catalysts only presented Lewis acid sites. The
catalysts were evaluated in methylcyclopentane (MCP) and methylcyclohexane (MCH) ring-opening (RO)
Available online xxx
◦
under pressure (2.85 and 3.95 MPa, respectively), from 200 to 425 C. For MCP conversion, the acidity
of the alumina support had no sensitive effect on the activity and selectivity to RO products, and few
effects on the distribution of RO products. No isomerization or hydrocracking products were observed,
confirming that these reactions occurred mainly on the metal function, which was not modified by the
presence of chlorine. The nature of the support, SiO2 or Al2O3, had a strong effect on both the activity (1.9
Keywords:
Ring-opening
Ir
Pt
−1
−1 ◦
metal for Ir/Al2O3 and Ir/SiO2, respectively at 225 C) and selectivity to RO products
Rh
against 0.5 mol h
g
Methylcyclopentane
Methylcyclohexane
Chlorine
(99.6% against 97.5% for Ir/Al2O3 and Ir/SiO2, respectively, at 80% of MCP conversion) for Ir catalysts
only. Interestingly, the Rh/SiO2 exhibited a high selectivity for converting MCP to RO products, similar
to Ir/Al2O3, i.e. 99.6% at 80% of conversion. Depending on the metal and the supports, three types of
behavior were observed for MCH ring-opening: (i) a direct ring-opening on the metal function whatever
the support for Ir, (ii) a first step of isomerization, and then a need of a sufficiently acidic support, for Pt
and (iii) an intermediate behavior for Rh, which was able to either directly convert MCH in absence of
acidic support or favor a bifunctional mechanism on chlorinated alumina.
©
2013 Elsevier B.V. All rights reserved.
1
. Introduction
considered as two model molecules for selective ring-opening
19–21].
On the one hand, MCP ring-opening can occur on metallic,
[
Selective ring-opening (SRO) of naphthenic molecules to
alkanes is a potential solution for significantly improving cetane
number of light-cycle oils or the octane number of gasoline
after hydrogenation of aromatics [1–6]. This reaction consists
in the selective hydrogenolysis of one endocyclic C C bond. It
is accompanied by a number of reactions such as isomeriza-
tion, deep hydrogenolysis and cracking leading to a decrease
in the number of carbon atoms in the molecules, dehydrogena-
tion and aromatization. This reaction is generally carried out on
noble metal catalysts supported on Al O , SiO , TiO , zeolites
acidic or bifunctional catalysts [1,3,22–29]. On acidic alumina [24],
ring-opening reaction proceeds at high temperature (>350 C) by
◦
a direct opening of MCP ring via carbonium ion intermediate.
On metal-supported catalysts, MCP ring-opening may occur on
pure metallic sites or at the metal-support interface [30,31]. It
was shown that, on sulphur modified platinum-based mono and
bimetallic catalysts supported on alumina [32], MCP conversion
yields 2-methylpentane (2MP), 3-methylpentane (3MP) on metal
sites while cyclohexane (CH), benzene (Bz) and n-hexane (nH) are
formed on both the metal and the acidic sites. Galperin et al. [33]
prepared Pt catalysts with the same metal active sites on sup-
ports with various acid-basic properties. All the catalysts on acidic
supports favored the formation of cyclohexane and aromatics. Con-
sequently, they concluded that an acidic support is detrimental to
high selectivities in ring-opening products.
2
3
2
2
[
4,5,7–15]. Among the noble metals active for this reaction, namely
Pt, Pd, Ru, Rh and Ir, Ir monometallic catalysts have shown the
best catalytic performances in terms of activity and selectivity
to ring-opening products [2,4,16–18]. Recently, Pt Rh bimetallic
catalysts have demonstrated performances similar to Ir for methy-
cyclopentane (MCP) and methylcyclohexane (MCH) ring-opening,
On the other hand, McVicker et al. [4] demonstrated that,
whatever the metal, Pt, Ni, Ru, Ir, supported on silica or alumina,
MCH ring-opening is more difficult than MCP ring-opening,
due to the lower ring strain in six-membered compared to
∗ Corresponding author. Tel.: +33 549 454 832.
E-mail address: florence.epron@univ-poitiers.fr (F. Epron).
0
926-860X/$ – see front matter © 2013 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.apcata.2013.05.009

2
08
P. Samoila et al. / Applied Catalysis A: General 462–463 (2013) 207–219
five-membered-ring naphthenes. Only Ir is able to directly open
MCH, whereas on the other catalysts, a ring-contraction step
using an acidic catalyst is necessary for obtaining afterwards a
reasonable yield in ring-opening products. Thus, McVicker et al.
clearly established that the acidity is important for selective ring-
opening of six-membered-ring naphtenes. For this reason, except
for Ir-based catalyst, the metal catalysts are frequently supported
on acidic supports, such as zeolites, for MCH ring-opening [5,12].
In this context, the objective of the present work was to study the
effect of the acidity of a ␥-alumina support by varying the chlorine
content and to evaluate the catalytic properties of Rh, Ir and Pt
catalysts on these supports in MCP and MCH conversion, in order
to find a moderate acidity compatible with high selectivity to RO
products whatever the initial reactant. Silica supported catalysts
were also evaluated for comparison.
determination of the total amount of hydrogen adsorbed on the cat-
alyst Hc . In order to remove the hydrogen reversibly adsorbed, the
1
catalyst was degassed under pure argon at RT for 30 min and then
a second series of hydrogen pulses was introduced onto the cat-
alyst. The hydrogen adsorbed during this second step is Hc . The
2
amount of chemisorbed hydrogen Hc is given by Hc = Hc − Hc .
1
2
The results are expressed in the form of H/M values, related to the
amount of hydrogen atoms chemisorbed per atom of metal. It is
generally admitted that for Pt and Rh, there is one hydrogen atom
chemisorbed per surface metal atom, and this was checked by TEM
measurement in Ref. [20]. Thus, for the Pt and Rh catalysts, the H/M
value is similar to the dispersion value. For Ir, the stoichiometry may
change with the dispersion of the metal [14], and an increase in the
H/M value is just an indication of an increase in the dispersion.
2
.2.2. Characterization of the acid sites by FTIR of adsorbed
pyridine
The Fourier Transformed Infra-Red (FTIR) spectroscopy was
2
. Experimental
2
.1. Catalysts preparation
used to characterize the acid sites of the support by adsorption
of a basic probe molecule, i.e. pyridine. 20–25 mg of catalyst were
pressed into thin wafers with diameter of 16 mm and surface of
A ␥-alumina (Axens) and silica (Degussa) with a specific surface
2
−1
2
−1
2
◦
area (BET method) of 215 m g and 175 m g , and a total pore
2 cm . After a pretreatment under ultra-high vacuum at 450 C for
12 h, the sample was cooled to room temperature and then pyri-
3
−1
3
−1
volume of 0.51 cm g and 0.47 cm g , respectively, were used as
supports. They were crushed and sieved in order to retain particles
with sizes between 0.25 and 0.40 mm, and then calcined in flowing
2
dine vapors were injected in the cell under a pressure of 2 × 10 Pa
−
4
for 10 min. After evacuation under vacuum (6 × 10 Pa) for 1 h at
◦
◦
air at 450 C for 4 h before impregnation of the metal precursor.
150 C, spectra were recorded using a Nicolet Nexus spectrome-
The catalysts were prepared by impregnation of the support
with solutions of H PtCl , H IrCl or RhCl . The volume of the
solutions was adjusted in order to obtain a final metal content of
ter in order to determine the total number of acidic sites. In order
to estimate the strength of these acidic sites, spectra were also
recorded after evacuation at 250 and 350 C for 1 h. Band intensi-
2
6
2
6
3
◦
0
.6 wt.% on the support.
For the alumina support, three kinds of preparation conditions
ties were corrected from slight differences in sample weight and
band areas were calculated by fitting the spectral profile with
a Gaussian–Lorentzian function using IR OMNIC software. The
concentration of Lewis acid sites was calculated from the inte-
were used in order to vary the final chlorine content in the catalyst:
−
1
(1) The solution of the metal precursor salt was directly put in con-
tact with the support. In this case, the final chlorine content
depends only on the chloride content of the precursor salt.
(2) Hydrochloric acid was added to the precursor salt solution to
adjust the chlorine content to 1 wt.% in the final catalyst. Then,
the support was impregnated with this precursor salt solution.
(3) The support was put in contact with hydrochloric acid, with
grated area of the band at 1455 cm using the value of the molar
−
1
extinction coefficient of this band (1.28 cm ␮mol ) determined by
Guisnet et al. [34], in the same conditions and in the same experi-
mental device as those used in the present study. The total amount
of Lewis acid sites was calculated from the spectrum obtained at
◦
150 C. The distribution of acid strength was evaluated by differ-
ence between the amount adsorbed at a given temperature and
the amount not desorbed at higher temperature.
−
1
a concentration of 0.2 mol L in order to obtain a final pH of
. Then, the solution of the precursor salt was added. In this
catalyst series, the theoretical chlorine content is of 2.5 wt.%.
1
2.2.3. Characterization of the Brønsted acid sites by isomerization
of 3,3-dimethyl-1-butene
◦
The mixture was stirred on a sand bath at 60 C up to a com-
The skeletal isomerization of 3,3-dimethyl-1-butene (33DMB1)
was used as model reaction to characterize the Brønsted acid sites.
Kemball et al. [35] demonstrated that the Lewis centers of alumina
are not involved in this reaction and the 33DMB1 isomerization
is likely to occur through a pure protonic mechanism on Brønsted
acidic sites [36–38].
◦
plete evaporation of the solvent and then dried overnight at 120 C.
Finally, the catalysts were calcined in flowing air at 300 C (Rh and
Ir supported catalysts) or 450 C (Pt catalysts) for 4 h and reduced
in flowing hydrogen at 500 C for 4 h.
The silica-supported catalysts were prepared according to the
procedure (1).
◦
◦
◦
◦
Catalyst samples (0.1 g) were calcined at 450 C in flowing air
−
1
(
30 mL min ), in order to deactivate the metallic phase, and then
2
.2. Characterization
purged under nitrogen. The catalytic reactions were carried out
◦
at 250 C at atmospheric pressure with a reactant flow rate of
−
1
The metal and chlorine contents were determined by elemental
15.2 mmol h and a reactant partial pressure of 20 kPa. Analysis
of the reaction products was performed by gas chromatography
with a flame ionization detector (AlphaMos PR2100) on a Rtx-1
analysis involving ICP-OES technique (inductively coupled plasma
optical emission spectrometry, Perkin).
(
Restek) column (105 m × 0.53 mm × 3.00 ␮m). The detected prod-
2.2.1. Hydrogen chemisorption
ucts were 2,3-dimethyl-1-butene and 2,3-dimethyl-2-butene. The
33DMB1 disappearance rates were determined from the 33DMB1
conversion and normalized by the BET surface area.
The H2 chemisorption measurement was carried out in a pulse
◦
chromatographic system. After reduction under hydrogen at 500 C
and outgassing for 2 h in an argon stream at the same temperature,
5
00 mg of catalyst were cooled down to room temperature. Then,
2.3. Catalytic tests
pulses of pure hydrogen were introduced onto the catalyst up to
saturation. The amount of hydrogen in the effluent was continu-
ously analyzed by a thermal conductivity detector. This allowed the
The cyclohexane dehydrogenation, which occurs only on the
metal function, was performed in order to study the effect of the

P. Samoila et al. / Applied Catalysis A: General 462–463 (2013) 207–219
209
n-C6
(
a)
2
MP
+
H
2
3
MP
MCP
n-C
7
(
b)
3
EP
2
,3DMP
a
c
a
c
b
a
a
b
a
ECP
1
,2DMCP
c
a
c
b
b
b
c
b
a a
a
c
c
c
b
b
3
MH
3MH
c
MCH
b
b
a
a
a
1
,3DMCP
1
,1DMCP
c
a
b
b
b
b
b
b
c
a
a
c
c
2
,2DMP
2
MH
2,4DMP
3
,3DMP
Scheme 1. Products of (a) MCP and (b) MCH ring opening.
presence of various chlorine contents on the metallic activity for
this reaction. The test was carried out in a continuous flow reactor
at 270 C under atmospheric pressure on 10 mg of catalyst. The par-
chromatograph using a FID and equipped with a CP-Sil 5 (Varian,
50 m length, 0.53 mm i.d., 1 ␮m film thickness) and Pona (SGE, 50 m
length, 0.15 mm i.d., 0.5 ␮m film thickness) capillary column for
MCP and MCH experiments, respectively. Conversion, defined as
the percentage of MCP or MCH converted, was determined as a
function of temperature.
◦
tial pressures were 93 kPa and 7 kPa for hydrogen and cyclohexane,
−
1
−1
and the corresponding flow rates 6 L h and 2 mL h , respectively.
Injection of cyclohexane was made using a calibrated motor-driven
syringe upstream of the reactor.
The absence of intra and extragranular diffusion limitations
in these experimental conditions was checked by varying inde-
pendently the particle size (in 0.1–0.6 mm range), the amount of
catalyst (from 125 mg to 500 mg and from 500 mg to 2 g for MCP and
MCH experiments, respectively), and the reactant flow rate (from
Analysis of the reaction products was performed by gas chro-
matography with a flame ionization detector (Varian 3400X) on
a HP-PLOT Al O “KCl” column. The only detected product was
2
3
benzene. The initial activity of the catalysts was determined after
−
1
1
0 min of reaction.
0.0325 to 0.1336 mL min ).
The catalytic performances of the catalysts were evaluated in
methylcyclopentane (MCP) and methylcyclohexane (MCH) ring-
opening under pressure (Scheme 1). Reaction conditions are
summarized in Table 1. The catalysts were evaluated in MCP or MCH
Table 1
Experimental conditions of MCP and MCH ring-opening.
Reaction conditions
MCP
MCH
◦
conversion from 200 to 425 C. The reaction was started at a given
Mass of catalyst (mg)
Mass of dilutant (␣-Al2O3, mg)
Pressure (MPa)
WHSV (h
H2/MCP or H2/MCH (mol/mol)
255
245
2.85
12
7.5
1500
0
temperature chosen as a function of the catalyst activity, in order to
obtain a conversion lower than 10% and then the temperature was
3.95
2.1
8
◦
a
−1
)
increased in steps of 25 C up to a conversion of 80–90%. The dura-
tion of each step was 2 h and four measurements were performed
at each temperature. Effluent products were analyzed by an on-line
a
Weight hourly space velocity.

2
10
P. Samoila et al. / Applied Catalysis A: General 462–463 (2013) 207–219
For MCP conversion, the only reaction products were the ring-
opening products (2-methylpentane (2-MP), 3-methylpentane
3-MP), n-hexane (n-C6)), C1 to C5 products, resulting from
deep hydrogenolysis, and 2,3-dimethylbutane. The yield in 2,3-
dimethylbutane was negligible. The selectivity of the catalysts for
ring-opening (RO), i.e. the hydrogenolysis of only one endocyclic
C–C bond, may be easily estimated from the yield in C1–C5 products
resulting from deep hydrogenolysis.
For MCH conversion the products obtained can be classified
into four categories: (i) products resulting from deep hydrogenol-
ysis or hydrocracking (C1–C6 alkanes); (ii) isomerization products,
i.e. cyclic products with 7 carbon atoms, corresponding to ethyl-
cyclopentane (ECP) and three dimethylcyclopentanes (1,1DMCP,
The amount of chemisorbed hydrogen per metal atom (H/M
ratio), characterizing the metal accessibility, depends on the metal,
the support and the amount of added HCl. However, for a given
metal deposited on alumina, the change in the H/M ratio as a func-
tion of the chlorine content is of 10% at maximum. The catalysts
supported on alumina show higher metal accessibility than those
supported on silica. This difference may be explained by the iso-
electric point of the supports, i.e. the pH at which the oxide surface
is neutral [39], and by the presence of chlorine. The alumina sup-
port exhibits an isoelectric point at pH around 7 and its surface is
positively charged at pH lower than 4, by adsorption of H+ from
acidic solution. This means that, in the preparation conditions, a
good interaction with the chlorinated precursor salt, in the anionic
form, is possible, thus favoring a good dispersion of the metals. This
is not possible with the silica support, which is neutral at pH lower
than 7 [39]. The values of the isoelectric points of both supports, alu-
mina and silica, may also explain the lower chlorine content on the
silica support compared to alumina, for a same amount of chloride
introduced in solution. As far as the effect of chlorine on the disper-
sion is concerned, it has been demonstrated that the impregnation
(
1
,2DMCP and 1,3DMCP); (iii) ring-opening products, i.e. C7 alkanes
resulting from the opening of MCH or isomers, constituted of
n-heptane (n-C7), two methylhexanes (2MH and 3 MH), four
dimethylpentanes (2,2DMP, 2,3DMP, 2,4DMP and 3,3DMP) and
ethylpentane (EP); (iv) dehydrogenation product, i.e. toluene.
As already mentioned in our previous works [20,21], it was
checked on some representative catalysts that the deactivation
occurring during reaction is negligible under our experimental con-
ditions.
of H PtCl6 followed by a calcination leads to the formation of oxy-
2
chloride species, PtOxCly, on alumina, thus favoring the dispersion
of platinum particles [40,41], whereas PtClx are produced on silica
favoring the sintering of platinum [42]. However, it appears that,
on alumina, for a given metal prepared from chlorinated precursor,
the H/M value decreases as the amount of HCl added in the pre-
cursor solution, and consequently the final chlorine content of the
catalyst, increases. This result may be explained by a strong com-
petitive chloride adsorption, during the impregnation step, which
decreases the adsorption of the anionic metal complex [43], which
is then deposited during the evaporation of the solvent by sim-
ple impregnation. Finally, the highest H/M value is obtained with
iridium.
3
. Results and discussion
3.1. Characterization of the catalysts
The percentages of chlorine and the H/M ratios are reported
in Table 2 for the Ir, Pt and Rh catalysts supported onto sil-
ica or alumina, with various chlorine contents introduced in the
impregnation solution by the precursor salt and/or by addition of
hydrochloric acid. The results show that for the same amount of
chlorine introduced during the preparation of the catalysts, the
alumina support retains a larger amount of chlorine than silica,
the chlorine content on silica-supported catalysts being around
In the following, the catalyst will be named 0.6M(xCl)/S, where
M is the metal (Rh, Pt or Ir), x the chlorine content (wt.%) and S the
support (Al O3 or SiO ).
0.1 wt.%, whatever the metal deposited. On alumina, the amount
2
2
The surface acidity of the catalysts was determined by IR spec-
troscopy of adsorbed pyridine, which is one of the most largely used
basic probe molecules for surface acidity characterization [44,45].
Due to the nitrogen electron lone pair, pyridine should interact with
acidic centers in a specific way to form (i) coordinated species on
of chlorine retained on the catalyst logically increases with the
amount of chloride introduced in the preparation medium. When
chlorine is only introduced with the precursor salt (preparation
(
1), described in the experimental part), the chlorine content
varies from 0.22 wt.% for Ir/Al O to 0.46 wt.% for Pt/Al O . The
2
3
2
3
Lewis acid sites (PyL) and (ii) the pyridinium ion on protonic sites
highest chlorine contents are obtained when the precursor salt
is introduced in a solution at pH = 1, adjusted by addition of
hydrochloric acid (preparation (3)). In this case, whatever the cat-
alyst, the chlorine content is higher than 1 wt.%.
+
(
PyH ). Fig. 1 shows the spectra obtained after thermodesorption
◦
of pyridine at 150, 250 and 350 C on the 0.6Pt(1.36Cl)/Al O . On
the spectrum obtained at 150 C, two bands are found in the ␯8a
mode region, one at 1617 cm , assigned to pyridine coordinated to
2
3
◦
−
1
Table 2
Chlorine content (introduced during the preparation procedure and actually
deposited on the catalyst), H/M ratio determined by H2 chemisorption of Ir, Pt and
Rh supported on silica or alumina.
Catalyst
Cl introduced
wt.%)
Cl in the final
catalyst (wt.%)
H/M
(
0
0
0
0
.6Ir/Al2O3a
0.66
1.00
2.50
0.66
0.22
0.68
1.17
0.10
0.95
0.87
0.85
0.64
b
.6Ir/Al2O3
.6Ir/Al2O3
c
.6Ir/SiO2
0
0
0
0
.6Pt/Al2O3a
0.65
1.00
2.50
0.65
0.46
0.62
1.36
0.13
0.85
0.80
0.75
0.56
b
.6Pt/Al2O3
.6Pt/Al2O3
c
.6Pt/SiO2
0
0
0
0
.6Rh/Al2O3a
0.62
1.00
2.50
0.62
0.34
0.59
1.02
0.10
0.60
0.57
0.56
0.33
b
.6Rh/Al2O3
.6Rh/Al2O3
c
.6Rh/SiO2
a
b
c
Cl exclusively introduced via the precursor salt.
1
2
.00 wt.%. Cl initially introduced (via the precursor salt + HCl solution).
.50 wt.%. Cl initially introduced (via the precursor salt + HCl solution).
Fig. 1. FTIR spectra of adsorbed pyridine on the 0.6Pt(1.36Cl)/Al O catalyst after
thermodesorption at 150, 250 and 350 C.
2
3
◦

P. Samoila et al. / Applied Catalysis A: General 462–463 (2013) 207–219
211
Table 4
1455
1451
Activity and turnover frequency (TOF) of Ir, Pt and Rh catalysts supported on alumina
for cyclohexane dehydrogenation for various chloride contents (T = 270 C, Patm).
◦
1
624
−1 −1
TOFa (s−1
)
Catalyst
Activity (mol h
g
metal)
1617
0.02
0.6Ir(0.22Cl)/Al O
8.47
7.36
7.64
0.48
0.45
0.48
2
3
0
0
.6Ir(0.68Cl)/Al2O3
.6Ir(1.17Cl)/Al2O3
1494
0.6Pt(0.46Cl)/Al2O3
0.6Pt(0.62Cl)/Al2O3
0.6Pt(1.36Cl)/Al2O3
19.03
17.08
16.40
1.21
1.16
1.18
1576
0
2 3
.6Pt(1.36Cl)/Al O
0
2 3
.6Pt(0.46Cl)/ Al O
0
0
0
.6Rh(0.34Cl)/Al2O3
.6Rh(0.59Cl)/Al2O3
.6Rh(1.02Cl)/Al2O3
9.17
7.08
7.36
0.44
0.36
0.38
0
.6Pt(0.13Cl)/SiO
650
2
1
1600
1550
1500
1450
a
TOF determined from H/M ratio.
-
1
Wavenumber (cm )
results in an increase in the acidity of the Brønsted sites [48], even
if the chlorination is performed by simple impregnation of HCl in
water [49]. For this reason, the production of aromatic compounds
is performed in the refining and petrochemistry industry via the
dehydrocyclization of n-paraffins through a bifunctional mecha-
nism on catalysts made up of a platinum-based metal function
well dispersed on a ␥-alumina for which the acidity is promoted
by addition of around 1 wt% of chlorine [50]. As no Brønsted
acid sites were identified by FTIR of pyridine, a test reaction,
namely the isomerization of 3,3-dimethyl-1-butene (33DMB1),
which only occurs through a pure protonic mechanism on Brønsted
acidic sites [35], was used to characterize the Brønsted acidity of
◦
Fig. 2. FTIR spectra of adsorbed pyridine after desorption at 150
.6Pt(1.36Cl)/Al2O3, 0.6Pt(0.46Cl)/Al2O3 and 0.6Pt(0.13Cl)/SiO2 catalysts.
C on
0
3+
both tetrahedral and octahedral Al (medium Lewis acid site (LAS))
−1
and another one at 1624 cm corresponding to pyridine coordi-
nated to tetrahedral Al3+ (strong LAS) [45]. The band at 1576 cm−1
is ascribed to pyridine adsorbed on weak LAS. The ␯19b vibration
band, located at 1455 cm , corresponds to pyridine adsorbed on
strong LAS. This band shows a shoulder at 1451 cm , disappear-
ing after desorption at 250 C, which can be ascribed to weak Lewis
acid sites. The absence of a band at about 1540 cm indicates that
Brønsted acid sites are not strong enough to form pyridinium ions.
In order to estimate the effect of the chlorine content on the
Lewis acidity of the platinum catalysts, the FTIR of adsorbed pyri-
dine was performed on two Pt/Al O catalysts, with the highest
and lowest chlorine content, as well as on Pt/SiO . The comparison
−1
−
1
◦
−
1
the 0.6Pt(1.36Cl)/Al O , 0.6Pt(0.46Cl)/Al O and 0.6Pt(0.13Cl)/SiO
2
3
2
3
2
catalysts. Results in terms of activity are reported in Table 3.
The most chlorinated sample, 0.6Pt(1.36Cl)/Al O , exhibits a very
2
3
high activity in 3,3DMB1 isomerization, much higher than the one
obtained with 0.6Pt(0.46Cl)/Al O , whereas on the silica support,
2
3
2
3
2
there is no activity in isomerization, associated with the absence of
Brønsted acid sites.
of the acidity of the 0.6Pt(1.36Cl)/Al O , 0.6Pt(0.46Cl)/Al O and
2
3
2
3
0
.6Pt(0.13Cl)/SiO2 catalysts is shown in Fig. 2. This figure reveals
the presence of peaks attributable to Lewis acid sites only, with
no peak at 1540 cm corresponding to Brønsted acid sites. On the
alumina support the total Lewis acidity, determined after thermod-
esorption at 150 C, decreases with the decrease in chlorine content,
and that the catalyst supported on silica exhibits a very low Lewis
acidity. The results of the amount and strength of the LAS in Table 3
confirm this trend. It appears that, on alumina, the amount on weak
LAS is similar whatever the chlorine content, whereas the highest
chlorine content leads to an increase in the number of medium
and strong LAS, the effect on strong LAS being more significant.
This increase in the number of strong LAS can be explained by the
substitution of hydroxyls on alumina by chlorine species increas-
ing the strength of adjacent LAS [46,47]. As far as the Rh/Al O
In order to evaluate the metal function, catalysts were tested
in a structure insensitive reaction [51–55], the dehydrogenation of
−1
◦
cyclohexane at 270 C and at atmospheric pressure. The values of
◦
the initial activity, and turnover frequency (TOF), calculated from
the H/M values, are given in Table 4. The calculations were made
considering a stoichiometry one hydrogen atom per surface metal
atom. As for Ir, the stoichiometry of adsorbed hydrogen may be
higher than one, and may change with the particle size [16] the
TOF values may be considered as a qualitative way to compare the
activity results. It can be seen that, for a given metal, the TOF is
similar whatever the chlorine content of the catalyst, which shows
that the presence of chlorine has no sensitive effect on the metal
activity. Moreover, for iridium, the same value of TOF obtained for
the various catalysts proves that the stoichiometry of hydrogen
chemisorption is similar whatever the catalyst.
2
3
catalysts are concerned (results not shown), the amount of LAS is
similar whatever the chlorine content, in the same range as that
measured for the Pt/Al O3 catalyst with the lowest chlorine con-
2
tent, indicating that the chlorine species are either badly dispersed
on the support or more associated with the metal. The absence of
peak at 1540 cm indicates that, if these Brønsted acid sites are
present, their strength is not sufficient to protonate pyridine. It has
been demonstrated in the literature that chlorination of the surface
3.2. Influence of the chlorine content on MCP ring-opening
−1
Pt, Rh and Ir catalysts supported on alumina with the lowest and
highest chlorine contents, i.e. prepared from a chlorinated precur-
sor in the absence or in the presence of hydrochloric acid at pH = 1,
Table 3
◦
Acidity of 0.6Pt/Al2O3 and 0.6Pt/SiO2 catalysts with various coverages in chlorine atoms evaluated by isomerization of 33DMB1 at 250 C (Brønsted acid sites) and by
thermodesorption of pyridine followed by FTIR (Lewis acid sites).
Coverage in Cl per nm2
Activity in 33DMB1 isomerization
Lewis acid sites (␮mol g
−1
)
Catalyst
−
1
−2
m )
(
␮mol h
Total
Weak
Medium
Strong
0.6Pt(1.36Cl)/Al2O3
0.6Pt(0.46Cl)/Al2O3
0.6Pt(0.13Cl)/SiO2
1.070
0.360
0.123
310
73
0
342
283
21
155
148
14
79
56
7
108
79
0

2
12
P. Samoila et al. / Applied Catalysis A: General 462–463 (2013) 207–219
a) 100
on Pt(Cl)/Al2O3 and Ir(Cl)/Al2O3, respectively, also at atmospheric
pressure. In the present study, performed at high pressure, only
C1–C5 and RO products are observed, even on the most chlori-
(
8
6
4
2
0
0
0
0
nated catalyst. With this 0.6Pt(1.36Cl)/Al O3 catalyst a selectivity
2
to C1–C5 of 1% and consequently a selectivity to RO products of 99%
◦
are obtained at 400 C, and 50% of MCP conversion. The high pres-
sure, which is not favorable for dehydrogenation reactions, may
account for the results obtained. The yield in benzene, calculated
at the thermodynamic equilibrium for H /MCP = 7.5, is zero at low
2
◦
temperature and increases from a temperature of 295 C under a
pressure of 0.1 MPa and from 480 C under a pressure of 2.85 MPa,
◦
i.e. in the present experimental conditions. At the two pressures,
the calculated yield in cyclohexane is negligible in the studied
range of temperature. It should be also mentioned that for plat-
0
◦
inum, which is active from 325 C whatever the chlorine content,
2
00
225
250
275
300
325
350
375
400
425
1
.36 or 0.46 wt.%, around 75% of the C1–C5 products correspond to
Temperature (°C)
methane. This result confirms that the reaction occurs mainly on
the metal instead of the support, even if it is known that the pres-
ence of chlorine, creating acidic sites, enhances the hydrocracking
activity, especially at high conversion and temperature [56].
The distribution of the various RO products, namely 2-
methylpentane (2MP), 3-methylpentane (3MP) and n-hexane
(
b) 14
1
2
0
8
6
4
2
0
1
(
n-C6), is reported in Table 5 for the Pt/Al O catalysts with 0.46 and
2 3
1
.36% of Cl. One can see that the product distribution is not affected
by the chlorine content, and is statistical, as already described in
Ref. [19] for the most chlorinated catalysts, in accordance with
what is commonly observed for well dispersed platinum particles
on alumina support [4,32,57–59]. Some authors reported that the
acidity of the support influences the distribution of RO products.
More precisely, the formation of n-C6 may increase as the support
acidity and the metal particle size increase [31]. However, it was
observed on Pt/Al O3 that for small particle sizes, around 1.5 nm,
2
0
20
40
60
80
100
well dispersed on the support, the acidity has no effect [31,60]. In
the present study, the high dispersion of platinum can explain why
the distribution of RO products is not strongly dependent on the
chlorine content.
Conversion (%)
Fig. 3. (a) Conversion of MCP vs. temperature and (b) yield in C1–C5 products vs. MCP
conversion over 0.6Ir(1.17Cl)/Al2O3 (ꢀ), 0.6Ir(0.22Cl)/Al2O3 (♦), 0.6Rh(1.02Cl)/Al2O3
(
(
ꢁ), 0.6Rh(0.34Cl)/Al2O3 (ꢀ), 0.6Pt(1.36Cl)/Al2O3 (ꢂ) and 0.6Pt(0.46Cl)/Al2O3 (ꢃ)
The product distributions obtained on the various catalysts in
the same range of conversion are compared in Fig. 4. As for Pt/Al O ,
−1
I = 2.85 MPa, WHSV = 12 h , H2/MCP = 7.5 mol/mol).
2
3
the chlorine content has no significant effect on the product distri-
bution obtained on Ir/Al O3 and Rh/Al O , which corresponds to a
selective mechanism, with few n-C6 produced. However, it should
2
2
3
were evaluated in MCP ring-opening. The evolutions of the conver-
sion vs. temperature and of the yield in C1–C5 products vs. the MCP
conversion are reported in Fig. 3a and b, respectively.
It can be seen that, independent of the catalyst, the temperature
and the conversion, the chlorine content has no effect on the activ-
ity and on the yield in C1–C5 products, and consequently on the
selectivity in RO products. It should be noticed that neither cyclo-
hexane nor aromatic compound is produced whatever the catalyst
and the temperature.
be mentioned that among the Ir/Al O3 catalyst, the one with 1.17%
2
of chlorine (Ir1 in Fig. 4) produces 5% of n-C6, whereas the other
one with 0.22% of chlorine (Ir2 in Fig. 4) leads to 2% of n-C6 only.
3.3. Influence of the support on MCP ring-opening
Fig. 5a and b shows the evolutions of the conversion vs. tem-
perature and of the yield in C1–C5 products vs. MCP conversion,
respectively, obtained with the Pt, Rh and Ir catalysts, supported on
alumina by impregnation at pH = 1 and on silica. Fig. 5a shows that,
at a given temperature, the silica support leads to lower conversions
Galperin et al. [33] obtained 18.8% of selectivity in aromatics
and 2.1% in cyclohexane at 18.8% of conversion, at atmospheric
◦
pressure and 350 C over a 0.375Pt(1.0%Cl)/␥-Al O . At roughly the
2
3
same temperature, Dees et al. [32] obtained 7% and 15% of benzene
Table 5
Distribution in n-C6, 2MP, 3MP, and n-C6/3MP and 2MP/3MP ratios on 0.6%Pt(xCl)/Al2O3 during MCP ring-opening (P = 2.85 MPa, WHSV = 12 h−1, H2/MCP = 7.5 mol/mol).
◦
Catalyst
T ( C)
Conversion (%)
2MP (%)
3MP (%)
n-C6 (%)
n-C6/3MP
2MP/3MP
3
3
350
00
25
1
1.5
6
21.5
56
15
38
45
46
45
9
20
22
23
22
76
42
33
31
33
8.4
2.1
1.5
1.3
1.5
1.6
1.9
2.0
2.0
2.0
Pt(0.46Cl)/Al2O3
3
4
75
00
3
3
350
00
25
1
2.5
9
22
50
15
37
42
45
45
9
19
22
23
23
76
44
36
32
32
8.8
2.3
1.7
1.4
1.4
1.7
1.9
1.9
2.0
2.0
Pt(1.36Cl)/Al2O3
3
4
75
00

P. Samoila et al. / Applied Catalysis A: General 462–463 (2013) 207–219
213
Table 6
Activities and turnover frequencies (TOF) of Ir, Pt and Rh catalysts supported on alumina or silica for MCP conversion (P = 2.85 MPa, WHSV = 12 h−1, H2/MCP = 7.5 mol/mol).
◦
−1 −1
TOFa (s−1
)
Catalyst
T ( C)
Conversion (%)
Activity (mol h
g
metal)
0
0
.6Ir(1.17Cl)/Al2O3
.6Ir(0.10Cl)/SiO2
225
225
8.2
2.2
1.9
0.5
0.12
0.03
0
0
.6Pt(1.36Cl)/Al2O3
.6Pt(0.13Cl)/SiO2
350
350
9.0
4.6
2.1
1.1
0.15
0.11
0
0
.6Rh(1.02Cl)/Al2O3
.6Rh(0.10Cl)/SiO2
225
225
11.3
9.0
2.6
2.1
0.13
0.18
a
TOF determined from H/M ratio.
compared to alumina, for Pt and Ir, whereas the contrary is observed
for Rh, except at low temperature (225 C). As the accessibility of
ones obtained with the Rh/Al O3 catalyst (Fig. 5b). The influence
2
◦
of the particle size and the support on the catalytic properties
of rhodium for methylcyclopentane hydrogenolysis was already
studied in the literature [61,64,65]. Differences in performances
observed between alumina and silica-supported Rh catalysts were
generally attributed to various particle structures according to the
nature of the support, alumina being able to stabilize the cuboocta-
hedral morphology, whereas on silica, due to weaker metal-support
interactions, particles possess icosahedral structure. Consequently,
the support plays a significant role in the abundance of various
Rh crystal planes, with an impact on the catalytic properties of Rh
the metals varies strongly with the support, results, expressed in
TOF, are summarized in Table 6.
For Rh and Pt catalysts, the TOF can be considered as similar
whatever the support. It has been related in the literature [61]
that the particle size has a strong effect on the catalytic proper-
ties of rhodium for hydrogenolysis of methylcyclopentane for a
given support. The authors found that, for Rh/SiO , the highest
2
TOF are obtained at high dispersions, whereas, for Rh/Al O , the
2
3
contrary is observed, the highest TOF being measured at low disper-
sion (H/M < 0.40). Results shown in Fig. 5a are in accordance with
these results, the dispersions of Rh/SiO (H/M = 0.33) and Rh/Al O
particles. Ir/Al O3 and Ir/SiO2 present also different behaviors in
2
term of (i) yield in C1–C5 products (Fig. 5b), which is slightly higher
at high conversions for Ir/SiO2 than for Ir/Al O , and (ii) distribu-
2
2
3
(
H/M = 0.56) being the ones leading to the smallest TOF according
2
3
to Ref. [61], assuming a stoichiometry H/Rh of 1. For iridium, strong
differences in the MCP conversion are observed as a function of the
tion of RO products (Fig. 4), 5.5% of n-hexane being obtained on
support, the TOF measured on Ir/SiO representing 25% of the value
2
determined on Ir/Al O , whereas the difference in the H/M value is
(a) 100
2
3
not so significant (0.64 for Ir/SiO2 compared to 0.85 for Ir/Al O ).
2
3
The same trend was also observed for MCP ring-opening in a batch
reactor [62], and it was suggested that the acidity of the support
participate to MCP conversion. More clearly, Upare et al. [63], con-
sidered that the RO mechanism of MCP depends strongly on the
acidity of the Ir catalysts supported on various zeolites, the highest
conversion being obtained with the most acidic support.
8
6
0
0
As far as the yield in C1–C5 products is concerned, it is sim-
40
20
0
ilar, lower than 1%, for Pt supported on Al O3 or SiO2 (Fig. 5b)
2
in the range of conversion studied, which is of 50% at maximum.
One can mention that, in the case of rhodium, the selectivity to
ring-opening products depends on the nature of the support, since
the Rh/SiO2 catalyst exhibits C1–C5 yields much lower than the
2
00
225
250
275
300
325
350
375
400
425
Temperature (°C)
14
(
b)
12
10
8
6
4
2
0
0
20
40
60
80
100
Conversion (%)
Fig. 4. Distribution of RO products for
a given conversion C of MCP
for 0.6Ir(1.17Cl)/Al2O3 (Ir1), 0.6Ir(0.22Cl)/Al2O3 (Ir2), 0.6Ir(0.10Cl)/SiO2 (Ir3),
Fig. 5. (a) Conversion of MCP vs. temperature and (b) yield in C1–C5 products vs. MCP
conversion over 0.6Ir(1.17Cl)/Al2O3 (ꢀ), 0.6Ir(0.10Cl)/SiO2 (–), 0.6Rh(1.02Cl)/Al2O3
0
0
.6Rh(1.02Cl)/Al2O3 (Rh1), 0.6Rh(0.34Cl)/Al2O3 (Rh2), 0.6Rh(0.10Cl)/SiO2 (Rh3),
.6Pt(1.36Cl)/Al2O3 (Pt1), 0.6Pt(0.46Cl)/Al2O3 (Pt2) and 0.6Pt(0.13Cl)/SiO2 (Pt3)
(ꢁ), 0.6Rh(0.10Cl)/SiO2 (×), 0.6Pt(1.36Cl)/Al2O3 (ꢂ) and 0.6Pt(0.13Cl)/SiO2
(
)
−1
−1
(P = 2.85 MPa, WHSV = 12 h , H2/MCP = 7.5 mol/mol).
(P = 2.85 MPa, WHSV = 12 h , H2/MCP = 7.5 mol/mol).

2
14
P. Samoila et al. / Applied Catalysis A: General 462–463 (2013) 207–219
100
Ir/Al O3 compared to 0.5% on Ir/SiO . Do et al. [16] also observed
2
2
a strong support effect on the mechanism of ring-opening reac-
tions under high hydrogen pressure using two probe molecules, 1,2
and 1,3-dimethylcyclohexane. They proposed that the dicarbene
mechanism, corresponding to a perpendicular adsorption of two
adjacent carbon atoms onto two adjacent metal atoms is predom-
8
6
4
0
0
0
inant onto Ir/SiO , whereas the metallocyclobutane mechanism,
2
requiring the flat adsorption of three neighboring carbon atoms
may also occur onto Ir/Al O . Knowing that the dicarbene mode
2
3
results in the cleavage of unsubstituted secondary–secondary C
C
bonds, and the metallocyclobutane mode in the cleavage of sub-
stituted C C bonds, this difference in reaction mechanism may
explain the difference in the formation of n-hexane observed in
the present study. The higher yield in C1–C5 products observed for
the Ir/SiO catalyst at high conversion compared to Ir/Al O may be
explained by the difference in the temperature of reaction needed
to reach high conversions, the deep hydrogenolysis being favored
when the temperature increases.
20
0
2
2
3
225
250
275
300
325
350
Temperature(°C)
Fig. 6. Conversion of MCH vs. temperature over 0.6Ir(0.22Cl)/Al2O3 (–),
0
.6Ir(0.68Cl)/Al2O3 (×), 0.6Ir(1.17Cl)/Al2O3 (ꢀ) and 0.6Ir(0.10Cl)/SiO2 (♦)
−1
(
P = 3.95 MPa, WHSV = 2.1 h , H2/MCP = 8 mol/mol).
3.4. Influence of the acidity on the MCH ring-opening
The acidity of the support and the isomerization ability of the
(3EP), dimethylpentanes (2,2DMP, 2,3DMP, 2,4DMP, 3,3DMP),
resulting catalysts have been reported as important parameters for
the ring-opening of MCH, the isomerization of MCH into more reac-
tive C5 rings favoring high activities in MCH ring-opening [4,30]. In
this part, the behavior of Ir, Rh and Pt supported on alumina with
various amounts of chlorine and on silica will be discussed. On these
catalysts, the conversion of methylcyclohexane yields a wide range
of products, which can be classified in four groups:
and, in smaller amounts, 2,2,3-trimethylbutane (2,2,3TMB).
3) Dehydrogenation product: toluene.
4) Deep hydrogenolysis or cracking products: C1–C6 products.
(
(
3.4.1. MCH ring-opening on Ir-supported catalysts
The evolution of MCH conversion vs. temperature obtained
in the presence of 0.6Ir(0.22Cl)/Al O , 0.6Ir(0.68Cl)/Al O ,
2
3
2
3
0
.6Ir(1.17Cl)/Al O3 and 0.6Ir(0.10Cl)/SiO2 is reported in Fig. 6. For
2
(
1) Isomers (Isom): ethylcyclopentane (ECP) and dimethylcy-
clopentanes (1,1DMCP, 1,2DMCP and 1,3DMCP).
2) Ring-opening products (RO): n-heptane (n-C7), 2-
methylhexane (2MH), 3-methylhexane (3MH), 3-ethylpentane
Ir supported on alumina, the change in the chlorine content has
not a strong effect on the conversion. However, the catalyst with
the highest chlorine content shows the highest MCH conversions
(
◦
◦
at 275 C and 300 C compared to the other catalysts supported on
1
00
100
(
a )
(b1)
1
8
6
4
2
0
0
0
0
0
80
60
40
20
0
2
25
250
275
300
325
350
0
20
40
60
80
100
Conversion(%)
Temperature (°C)
1
00
100
80
60
40
20
0
(
a )
(b2)
8
6
4
2
0
0
0
0
0
2
2
25
250
275
300
325
350
0
20
40
60
80
100
Temperature (°C)
Conversion(%)
Fig. 7. C1–C6 products and major ring-opening products (RO) obtained during MCH transformation on various Ir catalysts supported on alumina or silica: 0.6Ir(0.22Cl)/Al2O3
–), 0.6Ir(0.68Cl)/Al2O3 (×), 0.6Ir(1.17Cl)/Al2O3 (ꢀ) and 0.6Ir(0.10Cl)/SiO2 (♦). (ax) Yield as function of temperature; (bx) selectivity as function of MCH conversion (P = 3.95 MPa,
(
−
1
WHSV = 2.1 h , H2/MCP = 8 mol/mol).

P. Samoila et al. / Applied Catalysis A: General 462–463 (2013) 207–219
215
Table 7
◦
conversion at 300 C, whereas at this temperature a maximum
TOF of Ir, Pt and Rh catalysts supported on alumina or silica for MCH conversion
of 50% of conversion is reached with Ir/Al O . In fact, the Ir/SiO
−1
2
3
2
(P = 3.95 MPa, WHSV = 2.1 h , H2/MCP = 8 mol/mol).
catalyst exhibits the lowest H/M value, which may be attributed
to the presence of particles larger than on alumina. As the H/M
value decreases with the chlorine content on the alumina support,
and the lowest H/M ratio is obtained on silica, it can be inferred
that the MCH conversion increases with the fall of the iridium
dispersion in accordance with the results reported in the literature
◦
TOFa (h−1
)
Catalyst
T ( C)
Conversion (%)
0.6Ir(0.22Cl)/Al2O3
0.6Ir(0.68Cl)/Al2O3
0.6Ir(1.17Cl)/Al2O3
0.6Ir(0.10Cl)/SiO2
275
275
275
275
13.6
14.4
20.5
31.5
93
108
157
321
0.6Pt(0.46Cl)/Al2O3
0.6Pt(0.62Cl)/Al2O3
0.6Pt(1.36Cl)/Al2O3
0.6Pt(0.13Cl)/SiO2
350
350
350
350
7.1
8.9
14.9
2.0
54
74
127
24
for Ir/Al O3 catalysts [66].
2
The only reaction products (Fig. 7) are the ring-opening and
C1–C6 products. Neither isomers nor toluene are observed, what-
ever the catalyst and MCH conversion, as already described in Ref.
0.6Rh(0.34Cl)/Al2O3
0.6Rh(0.59Cl)/Al2O3
0.6Rh(1.02Cl)/Al2O3
0.6Rh(0.10Cl)/SiO2
300
300
300
300
14.5
13.1
15.1
16.6
84
80
94
[
21] for Ir(1.17Cl)/Al O . The only ring-opening products are 2MH,
2 3
3
MH, and n-C7. These three products can be obtained by direct
175
opening of MCH, according to Scheme 1. In order to discriminate
the primary products from the others formed by successive reac-
tion, the selectivity to the various ring-opening products is reported
in Fig. 8 as a function of MCH conversion. When extrapolating to
zero conversion, the non-primary products intercept the y-axis at
zero, whereas the primary products exhibit a non-negligible selec-
tivity [12,67]. Fig. 8a and b shows that 2MH and 3MH are primary
a
TOF determined from H/M ratio.
◦
alumina. The values of TOF measured at 275 C (Table 7) show that
the activity of the alumina-supported Ir catalysts increases slightly
with the chlorine content. Surprisingly, the 0.6Ir(0.10Cl)/SiO2
catalyst is by far the most active of the series, with 95% of MCH
60
40
20
0
60
(
a)
(b)
40
20
0
0
20
40
60
80
100
0
20
40
60
80
100
Conversion (%)
Conversion (%)
6
4
0
0
60
40
20
0
(
c)
(d)
20
0
0
20
40
60
80
100
0
20
40
60
80
100
Conversion (%)
Conversion (%)
20
15
10
5
0
20
15
10
5
(
e)
(f)
0
0
20
40
60
80
100
0
20
40
60
80
100
Conversion (%)
Conversion (%)
Fig. 8. Selectivity for the formation of the various ring-opening products vs. MCH conversion over (a) 0.6Ir(1.17Cl)/Al2O3, (b) 0.6Ir(0.10Cl)/SiO2, (c) 0.6Rh(0.59Cl)/Al2O3, (d)
0
(
.6Rh(0.10Cl)/SiO2, (e) 0.6Pt(1.36Cl)/Al2O3 and (f) 0.6Pt(0.13Cl)/SiO2. 2-methylhexane (ꢀ), 3-methylhexane (ꢂ), n-heptane (ꢁ), 2,2-dimethylpentane (×), 3,3-dimethylpentane
−1
), 2,3-dimethylpentane (᭹), 2,4-dimethylpentane (+), 3-ethylpentane (
) (P = 3.95 MPa, WHSV = 2.1 h , H2/MCP = 8 mol/mol).

2
16
P. Samoila et al. / Applied Catalysis A: General 462–463 (2013) 207–219
1
00
100
8
0
0
0
0
80
60
40
20
0
6
4
2
0
2
75
300
325
350
375
400
325
350
375
400
425
450
Temperature (°C)
Temperature (°C)
Fig. 9. Conversion of MCH vs. temperature over 0.6Rh(0.34Cl)/Al2O3 (–),
.6Rh(0.59Cl)/Al2O3 (×), 0.6Rh(1.02Cl)/Al2O3 (ꢁ) and 0.6Rh(0.10Cl)/SiO2 (ꢀ)
Fig. 11. Conversion of MCH vs. temperature over 0.6Pt(0.46Cl)/Al2O3 (–),
0.6Pt(0.62Cl)/Al2O3 (×), 0.6Pt(1.36Cl)/Al2O3 (ꢂ) and 0.6Pt(0.13Cl)/SiO2 (ꢃ)
0
−1
−1
(P = 3.95 MPa, WHSV = 2.1 h , H2/MCP = 8 mol/mol).
(P = 3.95 MPa, WHSV = 2.1 h , H2/MCP = 8 mol/mol).
products on 0.6Ir(1.17Cl)/Al O3 and 0.6Ir/SiO . The favored forma-
tion of 2MH and 3MH is indicative of a dicarbene mechanism as
conversion (Fig. 7). This confirms that the change in the chlorine
content does not modify the behavior of the Ir/Al2O3 catalysts.
Thus, the increase in the acidity with the chlorine content is not
sufficient to favor the isomerization pathway according to a bifunc-
tionnal mechanism. Compared to the alumina supported catalysts,
Ir/SiO2 is able to maintain a high selectivity to RO products even at
high conversion. This may be explained by the higher activity of the
silica-supported catalyst, 95% of MCH conversion being obtained at
2
2
already demonstrated on Ir/SiO2 [5] and Ir/Al O3 [4]. Thus, iridium
2
supported on alumina or silica is very active for the direct opening
of MCH and a bifunctional mechanism involving the acidity of the
support is not favored, in accordance with the results obtained by
Resasco et al. on Ir/SiO2 [5]. As a dicarbene mode needs the pres-
ence of two adjacent Ir atoms, large metal ensembles, and then
large particle sizes, will favor MCH ring-opening, which explains
the increase in the TOF values with the decrease in the H/M ratio.
A similar evolution of the selectivity to RO products and C1–C6
products as a function of conversion is obtained on Ir/Al O , what-
◦
◦
300 C instead of 325 C for the Ir/Al2O3 catalysts.
3.4.2. MCH ring-opening on Rh-supported catalysts
Results obtained on Rh supported on alumina with various chlo-
rine contents and on silica are reported in Figs. 8–10.
2
3
ever the chlorine content, with a high selectivity to C1–C6 at high
1
00
100
(
a )
(b1)
1
8
6
4
2
0
0
0
0
0
80
60
40
20
0
2
75
300
325
350
375
400
0
20
40
60
80
100
Temperature (°C)
Conversion(%)
1
00
100
80
60
40
20
0
(
a )
(b2)
2
8
6
4
2
0
0
0
0
0
2
75
300
325
350
375
400
0
20
40
60
80
100
Temperature (°C)
Conversion(%)
Fig. 10. C1–C6 products and major ring-opening products (RO) obtained during MCH transformation on various Rh catalysts supported on alumina or silica:
.6Rh(0.34Cl)/Al2O3 (–), 0.6Rh(0.59Cl)/Al2O3 (×), 0.6Rh(1.02Cl)/Al2O3 (ꢁ) and 0.6Rh(0.10Cl)/SiO2 (ꢀ). (ax) Yield as function of temperature; (bx) selectivity as function
0
−
1
of MCH conversion (P = 3.95 MPa, WHSV = 2.1 h , H2/MCP = 8 mol/mol).

P. Samoila et al. / Applied Catalysis A: General 462–463 (2013) 207–219
217
6
4
2
0
0
0
0
100
(
a )
(b1)
1
8
6
4
2
0
0
0
0
0
3
25
350
375
400
425
450
0
20
40
60
80
100
Temperature (°C)
Conversion(%)
6
0
100
(
a )
(b2)
(b3)
(b4)
2
8
6
4
2
0
0
0
0
0
4
2
0
0
0
3
25
350
375
400
425
450
0
20
40
60
80
100
Temperature (°C)
Conversion(%)
6
0
100
(
a )
3
8
6
4
2
0
0
0
0
0
4
2
0
0
0
3
25
350
375
400
425
450
0
20
40
60
80
100
Temperature (°C)
Conversion(%)
6
0
0
0
0
100
(
a )
4
8
6
4
2
0
0
0
0
0
4
2
3
25
350
375
400
425
450
0
20
40
60
80
100
Temperature (°C)
Conversion(%)
Fig. 12. C1–C6 products, ring-opening products (RO), isomerization products (Isom) and dehydrogenation product (toluene) obtained during MCH transformation on various
Pt catalysts supported on alumina or silica: 0.6Pt(0.46Cl)/Al2O3 (–), 0.6Pt(0.62Cl)/Al2O3 (×), 0.6Pt(1.36Cl)/Al2O3 (ꢂ) and 0.6Pt(0.13Cl)/SiO2 (ꢃ). (ax) Yield as function of
−
1
temperature; (bx) selectivity as function of MCH conversion (P = 3.95 MPa, WHSV = 2.1 h , H2/MCP = 8 mol/mol).
The evolution of the conversion as a function of temperature
Fig. 9) is similar for all the Rh catalysts supported on alumina,
increases (Fig. 10). The analysis of the selectivity to RO products
(
vs. conversion (Fig. 8c and d) shows that the only RO products are
2MH and 3MH, which are primary products, whatever the support.
The selectivity to isomers is around 10% for the alumina-supported
catalysts for conversions lower than 60%, whereas it is negligible
for the silica supported catalyst (Fig. 10). On Rh/Al O , a part of
whatever the chlorine content, whereas Rh/SiO2 shows lower con-
versions at the highest temperatures. The rhodium catalysts are
much less active than the iridium ones, the TOF values varying from
−
1
◦
−1
8
0 to 175 h at 300 C, whereas a TOF of 321 h is reached over
2
3
◦
Ir/SiO2 at 275 C (Table 7).
MCH is converted into isomers that are then opened, whereas on
Rh/SiO2 only the direct opening on the metal function is possible.
The major secondary products are 2,3DMP and 2,4DMP (Fig. 8c
Whatever the catalyst, a decrease in the RO selectivity, due to
an increase in C1–C6 selectivity, is observed when the conversion

2
18
P. Samoila et al. / Applied Catalysis A: General 462–463 (2013) 207–219
and d), resulting from the breaking of secondary C C bonds of
,2DMCP and 1,3DMCP, respectively, via dicarbene intermediates.
in the acidity of alumina with the chlorine content is not suffi-
cient to favor the isomerization pathway and then a bifunctional
mechanism. Consequently, the difference in activity observed on
1
On Rh/SiO , the main RO products are 2MH and 3MH, confirming
2
◦
that the conversion of MCH occurs by direct ring-opening on this
catalyst.
silica (31.5% of MCH conversion at 275 C) and alumina (13.6% of
◦
conversion at 225 C for 0.6Ir(0.22Cl)/Al O ) with various chlorine
2
3
contents is explained in term of metal dispersion (H/M = 0.64 for
Ir/SiO and H/M = 0.95 for 0.6Ir(0.22Cl)/Al O ). For Ir, whatever the
support, MCH is directly opened via a dicarbene intermediate yield-
ing 2MH and 3MH. Contrary to Ir, the acidity of the alumina support
combined with Rh favors the production of the isomers, dimethyl-
cyclopentanes, which are then opened to dimethylpentanes. On
2
2
3
3
.4.3. MCH ring-opening on Pt-supported catalysts
Results obtained on Pt supported on alumina with various chlo-
rine contents and on silica are reported in Figs. 11 and 12. Compared
to iridium and rhodium, the effects on the conversion of the chlo-
rine content and of the support, and consequently of the acidity,
are more pronounced for the platinum catalysts. The conversion
increases in the following order: 0.6Pt(0.13Cl)/SiO < 0.6Pt(0.46Cl)/
Al O < 0.6Pt(0.62Cl)/Al O < 0.6Pt(1.36Cl)/Al O . On platinum-
supported catalysts, whatever the support, the ring-opening
products are not primary products (Fig. 8e and f) and consequently
are obtained by opening isomers. This is in accordance with the
results of Calemma et al. [12] and confirms that MCH ring-opening
occurs in two steps on platinum catalysts, the first step consisting
in isomerizing the C6 cycle into C5 cycles, which are then opened.
Rh/SiO , in the absence of any acidity, the direct opening of MCH
2
is possible. A third type of behavior is observed with Pt, which is
not at all able to directly open C6-rings, since the activity of the Pt
catalysts depends strongly on the acidity and the chlorine content
of the support.
To conclude, depending on the metal and the supports, three
types of behavior were observed for MCH ring-opening: (i) a direct
ring-opening on the metal function whatever the support for irid-
ium, (ii) a first step of isomerization, and then a need of a sufficiently
acidic support, for Pt and (iii) an intermediate behavior for Rh,
which is able to either directly convert MCH in absence of acidic
support or favor a bifunctional mechanism on chlorinated alumina.
2
2
3
2
3
2
3
On 0.6Pt(0.13Cl)/SiO , the main reaction is dehydrogenation yield-
2
ing toluene, and no isomer is observed (Fig. 12). This demonstrates
that platinum is not able to directly open MCH and the acidity
of the silica support is not sufficient to produce enough isomers
for the ring-opening reaction. Thus, isomerization is the limiting
step on 0.6Pt(0.13Cl)/SiO . On Pt/Al O catalysts, dehydrogenation
also occurs but the production of isomers, and consequently of RO
products, is not negligible and increases as the chlorine content
increases.
Acknowledgment
2
2
3
P. Samoila acknowledges the financial support of European
Union’s Seventh Framework Program (FP7/2007-2013) under grant
agreement no. 264115 – STREAM.
References
4
. Conclusion
[
1] D. Kubi cˇ ka, N. Kumar, P. Mäki-Arvela, M. Tiitta, V. Niemi, T. Salmi, D.Y. Murzin,
J. Catal. 222 (2004) 65–79.
2] M. Arribas, P. Concepcion, A. Martinez, Appl. Catal. A 267 (2004) 111–119.
3] A. Corma, V. Gonzalez-Alfaro, A. Orchilles, J. Catal. 200 (2001) 34–44.
4] G.B. McVicker, M. Daage, M.S. Touvelle, C.W. Hudson, D.P. Klein, W.C. Baird Jr.,
B.R. Cook, J.G. Chen, S. Hantzer, D.E.W. Vaughan, E.S. Ellis, O.C. Feeley, J. Catal.
Pt, Ir and Rh catalysts were prepared on alumina and silica in the
[
[
[
presence of various amounts of chloride. On alumina, the Brønsted
and Lewis acidities increase with the chlorine content in the cata-
lyst. The Brønsted acidity of the chlorinated alumina is not strong
enough to protonate pyridine, used as probe molecules, but its
strength is sufficient for the isomerization of 33DMB1, a test reac-
tion. Catalysts supported on silica, which does not retain chlorine
species, present few Lewis acid sites and no Brønsted acid sites.
For MCP ring-opening, the acidity of the support has no sensi-
tive effect on the activity of the catalysts. Similar conversions are
obtained for Rh and Ir, in the same range of temperatures, lower
2
10 (2002) 137–148.
5] M. Santikunaporn, W.E. Alvarez, D.E. Resasco, Appl. Catal.
75–187.
[6] S. Lecarpentier, J. Van Gestel, K. Thomas, J.-P. Gilson, M. Houalla, J. Catal. 254
2008) 49–63.
[
A 325 (2007)
1
(
[
7] L.M. Kustov, T.V. Vasina, O.V. Hasloboishchikova, E.G. Khelkovskaya, P. Zeuthen,
Stud. Surf. Sci. Catal. 130 (2000) 227–232.
[8] F. Locatelli, J.P. Candy, B. Didillon, G. Niccolai, D. Uzio, J.M. Basset, J. Am. Chem.
Soc. 123 (2001) 1658–1663.
9] J.B.F. Anderson, R. Burch, J.A. Cairns, J. Catal. 107 (1987) 351–363.
10] F. Locatelli, D. Uzio, G. Niccolai, J.M. Basset, J.P. Candy, Catal. Commun. 4 (2003)
189–194.
11] T. Sugii, Y. Kamiya, T. Okuhara, Appl. Catal. A 312 (2006) 45–51.
12] V. Calemma, A. Carati, C. Flego, R. Giardino, F. Gagliardi, R. Millini, G. Bellussi,
ChemSusChem 1 (2008) 548–557.
[13] A. Djeddi, I. Fechete, F. Garin, Top. Catal. 55 (2012) 700–709.
[14] S. Rabl, A. Haas, D. Santi, C. Flego, M. Ferrari, V. Calemma, J. Weitkamp, Appl.
Catal. A 400 (2011) 131–141.
[
◦
◦
than 300 C, whereas for Pt, temperatures higher than 350 C are
needed to start to convert MCP. Pt and Rh behave similarly in term
of conversion when supported on chlorinated alumina or silica,
whereas Ir is much less active when supported on silica. Thus, the
[
[
[
−
1
−1
g
metal
activity of Ir/Al O (1.9 mol h
that of Ir/SiO2 (0.5 mol h
) is 3.8 times higher than
2
3
−1
−1
◦
) at 225 C. This difference in
metal
g
activity was explained by different reaction mechanisms, with only
one possible way on silica, via a dicarbene intermediate, whereas
both dicarbene and metallocyclobutane intermediates are possible
on alumina. As far as the selectivity to RO products is concerned,
on alumina support, Ir is the most selective to the formation of
RO products, with 98.4% of RO products at 94% of MCP conver-
sion. No ring-enlargement (formation of cyclohexane or benzene)
and hydrocracking products are observed, even at high conversion.
Interestingly, Rh/SiO2 is highly selective toward RO products, as
Ir/Al O .
The effects of the support and the chlorine content of alumina
are more pronounced on MCH conversion than on MCP conversion,
but depend strongly on the metal. For Ir catalysts, the activity for
MCH ring-opening increases with the chlorine content of alumina,
and consequently with its acidity. However, surprisingly, the less
[
[
15] A. Djeddi, I. Fechete, F. Garin, Appl. Catal. A 413–414 (2012) 340–349.
16] P.T. Do, W.E. Alvarez, D.E. Resasco, J. Catal. 238 (2006) 477–488.
[17] S. Hantzer, M.S. Touvelle, J.G. Chen, G.B. McVicker, WO Patent 02:07881 (2002).
[
[
18] W.C. Baird, D.P. Klein, J.G. Chen, G.B. McVicker, WO Patent 02:07881 (2002).
19] P. Samoila, M. Boutzeloit, C. Especel, F. Epron, P. Marécot, Appl. Catal. A 369
(
2009) 104–112.
[20] P. Samoila, M. Boutzeloit, C. Especel, F. Epron, P. Marécot, J. Catal. 276 (2010)
37–248.
2
[
21] P. Samoila, M. Boutzeloit, I. Salem, D. Uzio, G. Mabilon, F. Epron, P. Marécot, C.
Especel, Appl. Catal. A 415–416 (2012) 80–88.
[22] F. Garin, P. Girard, G. Maire, G. Lu, L. Gugzi, Appl. Catal. A 239 (2003) 297–304.
[
[
23] D. Teschner, Z. Paal, React. Kinet. Catal. Lett. 68 (1999) 25–33.
24] S.G. Brandenberger, W.L. Callender, W.K. Meerbott, J. Catal. 42 (1976)
2
3
282–287.
[25] E.G. Christoffel, K.H. Röbschläger, Ind. Eng. Prod. Res. Dev. 17 (1978) 331–334.
[
[
26] A. Djeddi, I. Fechete, F. Garin, Catal. Commun. 17 (2012) 173–178.
27] A. Haas, S. Rabl, M. Ferrari, V. Calemma, J. Weitkamp, Appl. Catal. A 425–426
(
2012) 97–109.
[28] S. Lecarpentier, J. van Gestel, K. Thomas, J.-P. Gilson, M. Houalla, Appl. Catal. A
374 (2010) 137–141.
acidic support, SiO , leads to the most active Ir catalyst. The increase
2

P. Samoila et al. / Applied Catalysis A: General 462–463 (2013) 207–219
219
[
29] S. Lecarpentier, J. van Gestel, K. Thomas, M. Houalla, J. Catal. 245 (2007)
[49] N.B. Muddada, U. Olsbye, T. Fuglerud, S. Vidotto, A. Marsella, S. Bordiga, D.
Gianolio, G. Leofanti, C. Lamberti, J. Catal. 284 (2011) 236–246.
[50] J.P. Boitiaux, J.M. Devès, B. Didillon, Ch. Marcilly, in: G.J. Antos, A.M. Aitani, J.M.
Parera (Eds.), Catalytic Naphtha Reforming, Science and Technology, Part II,
Marcel Dekker, New York, 1995, p. 79 (Chapter 4).
[51] J.A. Cusumano, G.W. Dembinski, J.H. Sinfelt, J. Catal. 5 (1966) 471–475.
[52] D.W. Blakely, G.A. Somorjai, J. Catal. 42 (1976) 181–196.
[53] M. Guenin, M. Breysse, R. Frety, K. Tifouti, P. Marécot, J. Barbier, J. Catal. 105
(1987) 144–154.
4
5–54.
[30] H. Du, C. Fairbridge, H. Yang, Z. Ring, Appl. Catal. A 294 (2005) 1–21.
[31] K. Hayek, R. Kramer, Z. Paal, Appl. Catal. A 162 (1997) 1–15.
[32] M.J. Dees, M.H.B. Bol, V. Ponec, Appl. Catal. 64 (1990) 279–295.
[33] L.B. Galperin, J.C. Bricker, J.R. Holmgren, Appl. Catal. A 239 (2003) 297–304.
[34] M. Guisnet, P. Ayrault, C. Coutanceau, M.F. Alvarez, J. Datka, J. Chem. Soc. Faraday
Trans. 93 (1997) 1661–1665.
[
[
[
[
[
[
[
[
[
35] C. Kemball, H.F. Leach, B. Skundric, K.C. Taylor, J. Catal. 27 (1972) 416.
36] H.J. Pines, J. Catal. 78 (1982) 1.
37] J.L. Lemberton, G. Perot, M. Guisnet, J. Catal. 89 (1984) 69.
38] D. Martin, D. Duprez, J. Mol. Catal. A 118 (1997) 113.
39] J.P. Brunelle, Pure Appl. Chem. 50 (1978) 1211–1229.
40] J. Barbier, D. Bahloul, P. Marécot, Catal. Lett. 8 (1991) 327–333.
41] G. Leitz, H. Lieske, H. Spindler, W. Hanke, J. Völter, J. Catal. 81 (1983) 17–25.
42] D. Radivojevic, K. Seshan, L. Lefferts, Appl. Catal. A 301 (2006) 51–58.
43] T. Mang, B. Breitscheidel, P. Polanek, H. Knözinger, Appl. Catal. A 106 (1993)
[54] I. Rodriguez-Ramos, A. Guerrero-Ruiz, J. Catal. 135 (1992) 458–466.
[55] E. Rogemond, N. Essayem, R. Frety, V. Perrichon, M. Primet, F. Mathis, J. Catal.
166 (1997) 229–235.
[56] A.B. Gaspar, L.C. Dieguez, Appl. Catal. A 201 (2000) 241–251.
[57] G. Maire, G. Plouidy, J.C. Prudhomme, F.G. Gault, J. Catal. 4 (1965) 556–569.
[58] G. Maire, C. Corolleur, D. Juttard, F.G. Gault, J. Catal. 21 (1971) 250–253.
[59] C. Corolleur, F.G. Gault, D. Juttard, J.M. Muller, J. Catal. 27 (1972) 466–468.
[60] H. Glassl, K. Hayek, R. Kramer, J. Catal. 68 (1981) 397–405.
[61] G. Del Angel, B. Coq, R. Dutartre, F. Figueras, J. Catal. 87 (1984) 27–35.
[62] E.N. Rao, N. You, S. Yoon, D.P. Upare, Y.-L. Park, C.W. Lee, Catal. Lett. 141 (2011)
1047–1055.
[63] D.P. Upare, S. Yoon, C.W. Lee, Catal. Lett. 142 (2012) 744–752.
[64] D. Teschner, K. Matusek, Z. Paal, J. Catal. 192 (2000) 335–343.
[65] D. Teschner, D. Duprez, Z. Paal, J. Mol. Catal. A 179 (2002) 201–212.
[66] C.G. Walter, B. Coq, F. Figueras, Appl. Catal. A 133 (1995) 95–102.
[67] N.A. Bhore, M.T. Klein, K.B. Bishoff, Ind. Eng. Chem. Res. 29 (1990) 313–316.
2
39–258.
[
[
[
[
44] E.P. Parry, J. Catal. 2 (1963) 371–379.
45] H. Knözinger, Adv. Catal 25 (1976) 184–271.
46] H. Knözinger, P. Ratnasamy, Catal. Rev. Sci. Eng. 17 (1978) 31–70.
47] D. Guillaume, S. Gautier, I. Despujol, F. Alario, P. Beccat, Catal. Lett. 43 (1997)
2
13–218.
48] D. Guillaume, S. Gautier, F. Alario, J.-M. Devès, Oil Gas Sci. Technol. – Rev. IFP
4 (4) (1999) 537–545.
[
5