Ring opening of decalin and methylcyclohexane over alumina-based monofunctional WO3/Al2O3 and Ir/Al 2O3 catalysts

Article abstract of DOI:10.1016/j.jcat.2011.10.014

Ring-opening reactions of decalin and MCH were studied over monofunctional acid (WO₃/Al₂O₃) and metal (Ir/Al ₂O₃) catalysts containing, respectively, up to 5.3 at. W/nm² and 1.8 wt% Ir. The catalysts were characterized by X-ray diffraction, Raman spectroscopy, low-temperature CO adsorption followed by infrared spectroscopy, and H₂ chemisorption. A reaction network was proposed for both molecules and used to determine the kinetic parameters. Kinetic modeling allowed relating characterization results and catalytic performance. For WO₃/Al₂O₃ catalysts, ring contraction precedes ring opening of both molecules. The evolution of ring contraction activity was consistent with the development of relatively strong Bronsted acid sites. Ring opening occurs according to a classic acid mechanism. For Ir/Al₂O₃ catalysts, only direct ring opening was observed. Ring opening proceeds mostly via dicarbene mechanism. Analysis of products indicated that monofunctional metal catalysts are better suited than acid solids for upgrading LCO.

Full text of DOI:10.1016/j.jcat.2011.10.014

Journal of Catalysis 286 (2012) 62–77
Contents lists available at SciVerse ScienceDirect
Journal of Catalysis
journal homepage: www.elsevier.com/locate/jcat
Ring opening of decalin and methylcyclohexane over alumina-based
monofunctional WO3/Al2O3 and Ir/Al2O3 catalysts
Rodrigo Moraes a, Karine Thomas a, Sébastien Thomas a, Sander Van Donk b, Giacomo Grasso b,
Jean-Pierre Gilson a, Marwan Houalla a,
⇑
a Laboratoire Catalyse et Spectrochimie, ENSICAEN – Université de Caen – CNRS, 6 Bd. du Maréchal Juin, 14050 Caen, France
b Total Petrochemicals Research S.A., Zone industrielle CB-7181, Feluy, Belgium
a r t i c l e i n f o
a b s t r a c t
Article history:
Ring-opening reactions of decalin and MCH were studied over monofunctional acid (WO3/Al2O3) and
metal (Ir/Al2O3) catalysts containing, respectively, up to 5.3 at. W/nm2 and 1.8 wt% Ir. The catalysts were
characterized by X-ray diffraction, Raman spectroscopy, low-temperature CO adsorption followed by
infrared spectroscopy, and H2 chemisorption. A reaction network was proposed for both molecules and
used to determine the kinetic parameters. Kinetic modeling allowed relating characterization results
and catalytic performance. For WO3/Al2O3 catalysts, ring contraction precedes ring opening of both mol-
ecules. The evolution of ring contraction activity was consistent with the development of relatively
strong Brønsted acid sites. Ring opening occurs according to a classic acid mechanism. For Ir/Al2O3 cata-
lysts, only direct ring opening was observed. Ring opening proceeds mostly via dicarbene mechanism.
Analysis of products indicated that monofunctional metal catalysts are better suited than acid solids
for upgrading LCO.
Received 27 August 2011
Revised 17 October 2011
Accepted 18 October 2011
Available online 7 December 2011
Keywords:
Alumina
Tungsten
Iridium
Decalin
Methylcyclohexane
Ring-opening
Ó 2011 Elsevier Inc. All rights reserved.
1. Introduction
carbon atom ring with one alkyl ramification. This limits the num-
ber of possible reaction products and facilitates the establishment
In order to meet an increasing demand for diesel, especially in
Europe, oil refiners have the option to blend some of the Light Cy-
cle Oil (LCO) fraction, produced in the Fluid Catalytic Cracking
(FCC) process, with the diesel pool. However, with the exception
of the carbon number, the LCO does not meet current diesel spec-
ifications, especially in Europe. It contains high amounts of poly-
aromatics (ca. 48–69%) and low cetane number (ca. 18–25) when
compared to the European norms (polyaromatics: upper limit of
11% – cetane number: minimum of 51). Hence, the LCO must be
deeply hydrotreated. The polycyclic aromatic fraction can be low-
ered by hydrogenation, which also improves the cetane index.
However, this increase is not sufficient, and a ring opening of naph-
thenic compounds appears to be required for LCO to significantly
contribute to the diesel pool. As illustrated in Table 1, the cetane
number significantly increases when a molecule such as decalin
with two naphthenic rings is converted into linear or mono-
branched paraffins. Ring opening must, thus, be highly selective to-
ward these products. Naphthene ring opening can be achieved over
metal, acid and bifunctional catalysts. One of the simplest mono-
naphthenic molecule, and as a consequence the most studied, is
methylcyclohexane (MCH). This model molecule consists of a six-
of a reaction network. As a result, hydroconversion of MCH and
other similar single-ring naphthenes has been the subject of many
studies [1–24]. However, MCH cannot be considered as representa-
tive of the LCO mixture, and thus, its use as a probe molecule to
study the LCO upgrading reaction can be questioned. The LCO
product distribution shows that it consists mainly of two-ring aro-
matic systems [25], making two-ring naphthenes such as decalin,
in hydrotreated feeds, a logical choice as probe molecules for the
study of LCO upgrading. In contrast to studies involving mono C4,
C5, and C6 naphthenic rings, literature concerning selective ring
opening of model molecules containing two fused rings for LCO
upgrading, such as decalin, tetralin, or naphthalene, is more limited
[16,26–35]. The study of such molecules is very often hindered by
the analytical difficulties, due to the complexity of ring opening
and other concomitant reactions (e.g., isomerization and cracking).
Detailed identification of the intermediates and products obtained
in these reactions is required for a better understanding of the
mechanism of ring opening reaction and an appropriate selection
of the most suitable catalytic system [36].
Acid-catalyzed ring opening of naphthenes has been mainly
studied on zeolites [3–5,19,28,29,33,37,38]. The majority of these
studies involve single-ring molecules [3–5,19,37,38]. Recently, sev-
eral studies have attempted to gain insight into the acid-catalyzed
ring opening of decalin [28,29,33]. Kubicka et al. [28] and
⇑
Corresponding author. Fax: +33 231452822.
E-mail address: marwan.houalla@ensicaen.fr (M. Houalla).
0021-9517/$ - see front matter Ó 2011 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2011.10.014

R. Moraes et al. / Journal of Catalysis 286 (2012) 62–77
63
Table 1
Cetane number of some relevant decalin ring-opening products [35].
Santikunaporn et al. [29] showed that decalin ring opening over zeo-
lites is preceded by ring contraction (e.g., isomerization from a
cyclohexyl to a cyclopentyl ring). Kubicka et al. [28] also reported
that Brønsted acidity influences the ring-opening yield. The best
performance was obtained for catalysts with moderate Brønsted
acidity (Hb-25, Hb-75 and HY-12 zeolites) with, virtually, no activity
observed for catalysts with low Brønsted acidity (MCM-41) and fast
deactivation due to coke deposits with solids exhibiting strong
Brønsted acid sites (H-mordenite-20). It is believed that ring open-
ing of naphthenes over acid catalysts proceeds on Brønsted acid sites
and is initiated by protolytic dehydrogenation followed by chain
reactions of the carbenium ions formed (e.g., ring contraction,
b-scission, alkylations and hydride transfer) [4,28,33].
McVicker et al. [16] investigated selective ring opening of naph-
thenic molecules over several metals (Ru, Rh, Ni, Ir and Pt). It was
observed that Ru, Rh and Ni exhibit similar preference for cleaving
unsubstituted C–C bonds, but are generally less selective than Ir. Pt
is much less active but can break substituted C–C bonds, leading to
an approximately statistical product distribution [8]. Paál et al. [39]
reported that metals such as Rh, Pd, Ir and Pt promote single
hydrogenolysis of 3-methylpentane and methylcyclohexane,
whereas multiple fragmentation (i.e., leading to the formation of
compounds with less than 7 carbon atoms) was observed for Co,
Ni, Cu, Ru, Ag, Re and Os. Overall, the literature shows that irid-
ium-based catalysts exhibit high activity and selectivity for ring
opening of C5 and C6 monocyclic molecules [13,16,40].
With respect to ring opening mechanism over metal catalysts,
Gault [8] reported that hydrogenolysis of substituted naphthenes
can take place over Pt/Al2O3 according to three distinct mechanisms:
(1) non-selective mechanism (occurring on highly dispersed plati-
num) corresponding to an equal chance of breaking any C–C bond
of the ring (multiplet mechanism); (2) selective mechanism allow-
ing only the rupture of unsubstituted C–C bonds (dicarbene mecha-
nism); and (3) a ‘‘partially selective’’ mechanism, competing with
the dicarbene mechanism (on poorly dispersed Pt catalysts), involv-
ing metallocyclobutane species as an intermediate (metallocyclob-
utane mechanism). In recent studies, the metallocyclobutane
mechanism has been frequently invoked to interpret results, involv-
ing the cleavage of methyl-substituted C–C bonds that could not be
explained by the multiplet or the dicarbene mechanisms [6,16,41].
In addition to the behavior of monofunctional metal and acid
catalysts, it was observed that C6 ring-opening reaction of naphth-
enes is enhanced by the use of bifunctional catalysts
[2,13,16,21,26,29,32,42–46]. This has been interpreted by C6 to
C5 ring contraction over the acid function and a greater ease of
opening C5 rings over the metal function [16,42,43].
Zirconia was proved to be an attractive support for bifunctional
catalysts because of the possibility of fine tuning its acidity by con-
trolled tungsten deposition [47]. We have recently shown that by
using tungstated zirconia as supports, after Ir deposition, high per-
formance bifunctional Ir–W catalysts for selective ring opening of
methylcyclohexane can be developed [13,45]. However, zirconia
is not a typical industrial support and lacks the textural and
mechanical properties of more conventional supports such as alu-
mina. Moreover, the shaped industrial catalyst for such a reaction
will be produced by extrusion, and alumina is the binder of choice
for such a process. Thus, it was of interest to develop an alternative
alumina-supported system and compare its performance with the
corresponding zirconia-based system. The results showed that
identical performance (products distribution, and selectivity) for
MCH ring opening can be obtained when alumina is used as a sup-
port instead of zirconia [48].
The present work extends previous studies of MCH ring opening
over alumina-based catalysts to decalin, a molecule that is more
representative of the LCO cut. It involves a detailed investigation
of the performance of WOx-based monofunctional acid catalysts
and Ir-based monofunctional metal catalysts for decalin conver-
sion. Specifically, it includes a systematic study of the effect of W
surface density (WO3/Al2O3) and iridium loading (Ir/Al2O3) on the
structure of the catalysts and on their activity and selectivity for
decalin ring-opening reaction and a comparison of the results with
those obtained for MCH reaction. A reaction network is proposed
for both naphthenes and a kinetic model is developed to fit the
reactivity results, thus enabling the investigation of structure-cat-
alytic performance relationship. Finally, potential improvement of
the cetane number will be evaluated based on the nature and com-
position of decalin ring-opened products. The present work will
serve as a basis for future study aimed at achieving a better under-
standing of the behavior of bifunctional catalysts.
2. Experimental
2.1. Materials
The
c-Al2O3 support (AX 300) was supplied by Criterion. It
exhibited, after calcination at 773 K for 2 h, a specific surface area

64
R. Moraes et al. / Journal of Catalysis 286 (2012) 62–77
of 256 m2 gꢀ1 and a pore volume of 0.69 cm3 gꢀ1. The W phase (4–
32 wt% W) was deposited by impregnation of the support with
ammonium metatungstate solution followed by calcination at
1023 K for 3 h. The W loadings used correspond to surface densi-
ties ranging from 0.5 to 5.3 at. W/nm2. The latter were calculated
assuming that the surface area is only due to the alumina. Tung-
stated alumina catalysts were designated as AWx where x is the
W surface density expressed as atoms W/nm2. The series of yIr/
AW0 (y = 0.6, 1.2 and 1.8 wt.%) catalysts was prepared by incipient
wetness impregnation with a 50 g/l hydrogen hexachloroiridate
(IV) hexahydrate solution (Acros Organics) followed by drying at
393 K for 12 h and calcination at 723 K for 3 h with a (20% v/v O2
consumed amount observed was attributed to chemisorbed and
physisorbed hydrogen. After 10 min of purging under pure Ar, a
new series of pulses were injected over the sample, to determine
the reversible part of the adsorbed hydrogen. The irreversible frac-
tion (chemisorbed H2) was taken as the difference between the two
values. Dispersion data were obtained assuming a H/surface Ir stoi-
chiometry of adsorption of 1. The stoichiometry of H/Ir = 1 adopted
in the present study for Ir dispersions lower than 60% was based on
previous results for the same system where the Ir particle size esti-
mated from H2 chemisorption was compared with that obtained by
TEM measurements [49,50].
with N2 as carrier) flow of 25 ml minꢀ1
.
2.7. Methylcyclohexane ring opening
2.2. BET surface area
Methylcyclohexane (MCH) ring opening was carried out at a to-
tal pressure of 5 MPa, for a temperature range of 523–623 K and for
contact times between 10 and 20 g h molꢀ1, in a continuous flow
fixed-bed microreactor containing 1 g of catalyst, diluted in 1 g of
silicon carbide. The partial pressures of hydrogen and MCH were
set at 4.8 and 0.2 MPa, respectively. MCH was dried over activated
3A zeolite and delivered through a liquid chromatography pump
(Gilson 302). Before testing, each catalyst was activated in situ un-
der atmospheric pressure at 673 K with flowing dry air for 2 h and
cooled to 623 K under flowing dry He. The sample was then treated
at 623 K in a dry hydrogen flow for 1 h and then cooled to 573 K
prior to increasing the pressure. Hydrogen (Air Liquide, Grade I)
and helium (Air Liquide, Grade U) gases were further purified from
water and oxygen contaminants with 3A zeolite and BTS (FLUKA)
traps. The product gas mixture was collected at atmospheric pres-
sure and analyzed online with a Varian 3400 gas chromatograph
equipped with a flame ionization detector and a Chrompack CP-
SIL 5 CB WCOT fused silica capillary column.
The specific surface area of the catalyst was determined from
physical adsorption of N2 at T = 77 K, by applying the BET equation
on the part of the adsorption isotherm with 0.05 6 p/p° 6 0.35. The
adsorption isotherms were measured with a Micromeritics ASAP
2000 apparatus. The samples were first outgassed at 573 K in a dy-
namic vacuum for 1 h.
2.3. X-ray diffraction
Powder X-ray diffraction patterns were recorded with a Philips
X’pert diffractometer (PW1750 spectrometer), using the Cu K
radiation. The angle (2h) was varied from 20° to 80° in 0.02° steps.
The crystalline structure of the phases present was identified by
comparison with the database of the Joint Committee on Powder
Diffraction Standards (JCPDS).
a
2.4. Raman spectroscopy
The products analysis, attribution, and calculation of MCH con-
version are described elsewhere [13].
Raman studies were carried out under ambient conditions using
samples in a powder form. The Raman spectrometer consisted of a
confocal microscope Labram 300 (Jobin Yvon), equipped with a Nd/
YAG laser (frequency doubled, 532 nm) and a CCD detector. The
power on the sample was about 13 mW.
A wide range of MCH conversion was obtained by varying the
contact time and the reaction temperature.
2.8. Decalin ring opening
Decalin ring opening was carried out at a total pressure of
5 MPa, in a temperature range of 523–623 K and at contact times
between 170 and 320 g h molꢀ1, in a continuous flow fixed-bed
microreactor containing 1 g of catalyst, diluted in 1 g of silicon car-
bide. The feed, consisting of a mixture of 89% (v/v) n-heptane and
11% (v/v) decalin, (cis/trans molar ratio of 40:60), was delivered
through a liquid chromatography pump (Gilson 302). The mixture
decalin/n-heptane was dried over activated 3A zeolite. The partial
pressure of hydrogen, n-heptane, and decalin were set at 4.874,
0.113, and 0.013 MPa, respectively. The product gas mixture was
collected at atmospheric pressure and analyzed online with a
Varian 3800 gas chromatograph, fitted with a capillary column
(HP-PONA, length 50 m, internal diameter 0.2 mm) and a flame
ionization detector (FID) detector. Before the activity test, each cat-
alyst was activated in situ at atmospheric pressure at 673 K with
flowing dry air for 2 h, cooled to 623 K under dry He flow. As noted
for the methylcyclohexane reaction, the sample was treated at
623 K in a dry hydrogen flow for 1 h and then cooled to 573 K prior
to increasing the pressure. Hydrogen (Air Liquide, Grade I) and he-
lium (Air Liquide, Grade U) gases were further purified from water
and oxygen contaminants with 3A zeolite and BTS (FLUKA) traps.
Decalin conversion leads to a complex mixture of about 250
compounds. The products were identified by GC–MS, GC ꢂ GC-
TOFMS at Total Petrochemicals Research Centers. GC–MS analyses
were performed using a HP-PONA column identical to that men-
tioned above. The GC ꢂ GC-TOFMS method used as primary col-
umn a HP-5MS (non-polar, 30 m, internal diameter 0.32 mm, film
2.5. Infrared spectroscopy
The infrared spectra were recorded at low temperature with a
Nicolet Nexus FTIR spectrometer equipped with a MCT detector.
The powder was pressed into a self-supported wafer (ꢁ15 mg).
The activation of the wafer was carried out in situ in a flow infrared
cell under dry air at 673 K for 2 h, followed by treatment under H2
for 1 h at 623 K. The sample was then cooled to 473 K, evacuated
under vacuum for 2 h, and cooled to room temperature. The infra-
red cell was then evacuated for 1 h and cooled to 100 K. The infra-
red spectra were acquired following introduction of increments of
CO and finally in the presence of 133 Pa of CO at equilibrium. All
spectra were normalized to 15 mg of the solid.
2.6. H2 chemisorption measurements
Hydrogen chemisorption measurements were obtained with an
in house apparatus using a TCD detector. The samples were previ-
ously pre-treated at 573 K for 1 h, under 5 MPa of hydrogen, in or-
der to duplicate the conditions of the catalytic tests. Before the
measurements, the sample (300 mg) was reduced in flowing
hydrogen (50 cm3 minꢀ1) at 573 K for 1 h, purged in ultra pure Ar
(50 cm3 minꢀ1) for 1 h, and then cooled down to room tempera-
ture. The metallic function was characterized by two series of H2
micropulses. After activation of the catalyst, pulses of H2 were in-
jected at room temperature every minute up to saturation. The

R. Moraes et al. / Journal of Catalysis 286 (2012) 62–77
65
thickness 0.25
1.5 m, internal diameter 0.10 mm, film thickness 0.20
l
m), and as secondary column a SolGelWax (polar,
m).
l
To facilitate the evaluation of the results, the products were
grouped. All the alkyl-substituted monocyclic C10 were referred
to as one-ring-opening products (1ROP); linear or branched C10
were denoted two-ring-opening products (2ROP); bicyclical C10
where one or both rings consist of less than 6 carbon atoms were
named ring contraction products (RCP); compounds with lower
molecular weight than that of decalin were denoted as cracking
products (CP). Under some experimental conditions (high temper-
atures, active catalysts), the n-heptane solvent in the decalin feed
can react and form isomers and cracking products. This was taken
into consideration in estimating decalin conversion and products
yield.
Decalin conversion was varied by changing the contact time and
the reaction temperature.
3. Results
Fig. 1. X-ray diffraction patterns for the tungstated alumina catalysts (AWx).
3.1. Monofunctional acid catalysts
3.1.1. Texture and composition of the AWx solids
The BET surface area, corrected surface area of the Al2O3 sup-
port (assuming that the surface area is only due to the alumina
support), W loading (determined by ICP) and the corresponding
W surface density (at. W/nm2), over the WOx/Al2O3 catalysts, are
listed in Table 2. As shown in Table 2, the corrected surface area re-
mains essentially constant in the W surface density range exam-
ined. The surface area of the alumina support was lower than
that of the tungstated solids, reflecting the stabilization effect of
the W phase.
3.1.2. X-ray diffraction (XRD)
The XRD patterns of the AWx solids for various W surface den-
sities are shown in Fig. 1. All solids exhibit the main reflections
lines of the gamma alumina support (2h = 37.6°, 45.8° and 66.7°).
In addition, the main lines characteristic of WO3 (2h = 23.1°,
23.6°, 24.2°, 33.6° and 34.0°) were detected for W surface densities
P4.1 at. W/nm2. A weak and an ill-defined band, at 2h = 23.1–24.2°
which can be attributed to the most intense lines of WO3, was also
evidenced for the AW2.9 solid. This suggests that up to 4.1 at. W/
nm2, the tungsten phase is essentially present in the form of a sur-
face phase.
Fig. 2. Raman spectra of the AWx catalysts under ambient conditions.
3.1.4. Low-temperature CO adsorption monitored by infrared
spectroscopy
Several probe molecules, including carbon monoxide, pyridine
and lutidine, can be used to characterize the acidity of oxide sur-
faces. Low-temperature CO adsorption monitored by infrared spec-
troscopy, adopted in the present study, offers the advantage of
being sensitive to the strength of Brønsted acid sites thus titrated
[54,55]. The interaction between CO and the hydroxyl groups leads
3.1.3. Raman spectroscopy
The Raman spectra of the AWx samples under ambient condi-
tions are shown in Fig. 2. For all W surface densities, the catalysts
exhibit a Raman band at 952–976 cmꢀ1 which was assigned to W
surface interaction species [51–53]. Raman bands due to WO3 crys-
talline particles (ꢁ808, 717 and 273 cmꢀ1) were observed for W
surface densities P2.9 at. W/nm2.
to a shift to lower wavenumbers of the m(OH) vibration and a shift
to higher wavenumbers of the (CO) vibration. The position of (CO)
vibration is thus characteristic of the strength of the acid site
probed [54,56,57].
After activation, changes in the infrared spectra of different cat-
alysts were observed in the OH region (not shown). The intensity of
the area associated with the OH stretching region decreased grad-
ually with increasing W surface density. This was attributed to a
progressive replacement of the OH groups by W species [58,59].
Similar phenomena have been observed for other metal oxides
supported on alumina [60,61].
Table 2
Texture and composition of the AWx solids.
a
Sample
W loading
(wt.%)
SBET
Scorrected
W surface density
(at. W/nm2)
(m2 gꢀ1
)
(m2 gꢀ1
)
AW0
–
222
230
220
201
182
158
222
243
249
263
256
246
0
AW0.5
AW1.4
AW2.9
AW4.1
AW5.3
3.8
9.4
18.9
24.0
28.5
0.5
1.4
2.9
4.1
5.3
Infrared spectra, following introduction of small increments of
CO, were acquired to facilitate the detection of strong acid sites
which may be present in minor amounts, since these sites will re-
act first. Fig. 3 shows the infrared spectra for the AWx series in the
a
Smeasured
Scorrected
¼
.
1ꢀW ðwt:%Þꢃ1:216
100

66
R. Moraes et al. / Journal of Catalysis 286 (2012) 62–77
intensity and can be attributed to a preferential consumption of
2172
0.01 u.a
the most basic OH groups with initial tungsten deposition. For W
surface density P2.9 at. W/nm2, as seen in Fig. 3, a new band at
higher frequency appears (2172 cmꢀ1) and the position of
m(CO) in-
AW5.3
AW4.1
creases up to 2166 cmꢀ1 for AW5.3 catalyst.
The evolution of Brønsted and Lewis acid sites with W surface
density was then determined using the infrared spectra of the cat-
alysts following the introduction of 133 Pa of CO at equilibrium in
the 2060–2240 cmꢀ1 region. The decomposition of the envelope
was carried out according to the parameters determined in a pre-
vious work [63].
AW2.9
AW1.4
Fig. 4b shows an example of the decomposition of the envelope
for the solid AW4.1. The spectra were curve fitted using 7 Gaussian
bands: The three bands located in the 2180–2205 cmꢀ1 region
were assigned to Lewis acid sites. The band at 2172 cmꢀ1 was
attributed to relatively strong Brønsted acid sites and that at
2155–2164 cmꢀ1 was ascribed to weak Brønsted acid sites. The
band at 2143 cmꢀ1 and the shoulder at 2130–2132 cmꢀ1 are asso-
ciated with physisorbed CO. The variation of the normalized area of
the bands attributed to Brønsted and Lewis acid sites (Fig. 5a and
b) reflects the evolution of the abundance of the corresponding
acid sites. A decrease in the abundance of Lewis acid sites and
weak Brønsted acid sites with increasing W surface density is ob-
served (Fig. 5a). This was attributed to an increased coverage of
the alumina support on W deposition [63,64]. In contrast, a differ-
ent evolution of relatively strong Brønsted acid sites was observed
(Fig. 5b). It can be clearly seen that for W surface density lower
than 1.4 at. W/nm2, no strong Brønsted acid sites (band at
2172 cmꢀ1) were detected. For W surface density >1.4 at. W/nm2,
the abundance of the strong Brønsted acid sites increased almost
linearly with increasing W surface density.
AW0.5
2165
AW0
2120
2220
2200
2180
2160
2140
Wavenumber (cm-1)
Fig. 3. Infrared spectra in the
m(CO) region for AWx series after introduction of
3.5 mol CO following subtraction of the spectrum of the activated catalysts.
l
(CO) region after introduction of 3.5 lmol of CO and subtraction of
the spectrum of activated catalysts. The band located in the 2180–
2220 cmꢀ1 region is attributed to CO coordinated to Al3+ sites [62]
and that at 2165–2172 cmꢀ1 is ascribed to Brønsted acid sites [63].
For W surface density P2.9, the latter band is centered at ca. 2172–
2173 cmꢀ1, indicating the presence of relatively strong Brønsted
acid sites [63]. The development of these sites was associated with
the formation of polymeric W surface species.
Fig. 4a shows the infrared spectra in the m(CO) region obtained
after the introduction of 133 Pa CO. For the alumina support, CO
adsorption at 100 K gives rise to a strong band at 2155 cmꢀ1 as-
signed to CO weakly bonded to OH groups and a shoulder at
2143 cmꢀ1 ascribed to physisorbed CO on the alumina surface.
3.1.5. Decalin and MCH conversion: influence of W surface density
3.1.5.1. Activity. Fig. 6 shows the evolution of the composition of
cis- and trans-isomers in the unconverted decalin as a function
of conversion for the AW4.1 catalyst. The results indicate that
the cis/trans ratio gradually decreases with increasing conversion.
This is consistent with the results reported by Lai et al. [65], in the
case of HY zeolites, and may simply reflect the higher reactivity of
the cis form [29,66].
With increasing W surface density, the position of
m(CO) band,
associated with CO interaction with Brønsted acid sites, shifts up-
ward from 2155 for AW0 catalysts to 2162 cmꢀ1 for AW2.9 cata-
lysts. This shift was accompanied by
a strong decrease in
Fig. 4. Infrared spectra of the m(CO) region for the (a) series of AWx catalysts with different W surface density after the introduction of 133 Pa CO at equilibrium and following
subtraction of the spectrum of the activated catalysts and (b) curve fitting results for the AW4.1 catalyst. Solid line, experimental spectrum; dotted line, reconstituted
spectrum after curve fitting.

R. Moraes et al. / Journal of Catalysis 286 (2012) 62–77
67
Fig. 5. Evolution of the abundance of acid sites for tungsten alumina catalysts (AWx) after introduction of 133 Pa CO at equilibrium at 100 K: (a) relatively strong Brønsted
acid sites (N); weak Brønsted acid sites (j); and (b) Lewis sites (d).
Fig. 7 shows the variation of decalin and MCH conversion for the
AWx catalysts as a function of W surface density, for a reaction
temperature of 623 K and a contact time of 20 g h molꢀ1. The re-
sults for decalin were obtained by extrapolation of the conversion
obtained at a higher contact time assuming first order kinetics.
Similar conversions were observed for both naphthenes. No signif-
icant activity was observed for W surface densities lower or equal
to 1.4 at. W/nm2. For higher W surface densities, the conversion in-
creased with increasing W surface density up to 4.1 at. W/nm2 and
leveled off for the AW5.3 solid. The activity observed for these cat-
alysts appears to be associated with the formation of relatively
strong Brønsted acid sites for W surface densities higher than 1.4
at. W/nm2. A similar behavior was reported by Lecarpentier et al.
[13] for MCH conversion over WO3/ZrO2–SiO2.
3.1.5.2. Selectivity. The performance of the catalysts, for decalin and
MCH ring opening, was monitored. The evolution of ring opening,
ring contraction, and cracking products selectivities with conver-
sion for the three active catalysts is shown in Fig. 8. The results
clearly show that, for both reactions at a given conversion, the
three active catalysts exhibit similar selectivities. This suggests
that the active sites are associated with Brønsted acid sites of sim-
ilar strengths present in these solids.
Fig. 6. Cis- and trans-decalin internal composition as a function of conversion for
the AW4.1 catalyst.
Fig. 7. (a) Decalin and (b) MCH conversion obtained with the AWx catalysts as a function of the W surface density at 623 K and a contact time of 20 g h molꢀ1
.

68
R. Moraes et al. / Journal of Catalysis 286 (2012) 62–77
Fig. 8. Product selectivities obtained with the AWx catalysts as a function of: (a) decalin and (b) MCH conversion (open symbols: x = 2.9; half-filled symbols: x = 4.1; and filled
symbols: x = 5.3).
For low conversions, the selectivity for RCP tends to ca. 100%,
whereas those for 1ROP and CP tend to ca. 0% with decreasing con-
version. This indicates that cracking and ring opening of both
naphthenic molecules are preceded by ring contraction. The sec-
ondary nature of the ring opening products has also been reported
by Kubicka et al. [28] and Santikunaporn et al. [29] for decalin over
zeolites and by Lecarpentier et al. [13] for the conversion of meth-
ylcyclohexane over WO3/ZrO2–SiO2 catalysts. The observed de-
crease in RCP selectivity with increasing conversion can be
attributed to the cracking and ring opening of ring-contracted
products. At low conversion, the major cracking products of deca-
lin were isobutane and C6 (namely methylcyclopentane), as re-
ported by Kubicka et al. [28] and Santikunaporn et al. [29]. This
suggests that the cracking of 1ROP (butylcycloalkane intermediate)
to form isobutane and a methylcycloalkane is favored over that of
RCP (methyl-substituted binaphthene [36]), yielding typically by
cracking C1 and C9 products.
increasing conversion. However, the decrease in selectivity is less
pronounced for the MCH reaction, suggesting that the RCP formed
during the MCH reaction are less reactive than those of decalin.
Furthermore, a much lower CP selectivity was observed for the
MCH reaction when compared to that of decalin.
For both reactions, traces of aromatics (e.g., toluene and tetra-
lin) and olefins were observed. This is consistent with previous
studies of the ring opening of naphthenic molecules over acid cat-
alysts [28,29,69].
Fig. 9a depicts the evolution of the molar selectivities of C10H20
ROPs within the one-ring opening group of products, obtained for
the three active AWx catalysts, as a function of decalin conversion.
Analysis of the 1ROPs internal distribution shows that ethyl-dim-
ethylcyclohexanes (EDMCH) are the most abundant compounds.
For simplification purposes, less abundant products with three or
more alkyl substituents, such as tetramethylcyclohexanes and di-
methyl-propylcyclopentanes, were grouped and identified as ‘‘Oth-
ers’’. Lower amounts of methyl-propylcyclohexanes (MPCH),
butylcyclohexanes (BCH), pentylcyclopentanes (C5CP), and ‘‘Oth-
ers’’ were detected. It should be noted that under the examined
conditions, iso-butylcyclohexane accounts for over 80% of the
BCH group of product.
The high CP selectivity observed at high conversion of decalin
can be attributed to the cracking of RCP and 1ROP. Note that only
a small amount of 2ROP (not shown) was detected (up to 2% of
selectivity) at high conversion. This is consistent with the reported
difficulty in breaking endocyclic C–C bonds compared to exocyclic
ones [67,68] during hydrocracking of C10 naphthenes.
Fig. 9b depicts the evolution of the molar selectivities of C7H14
ROP within the ring opening group of products, obtained for the
three active AWx catalysts, as a function of MCH conversion.
As shown for decalin, RCP selectivity for the MCH reaction
(Fig. 8b) also tends to 100% at low conversion and decreases with
Fig. 9. Molar selectivities of: (a) C10H20 1ROPs and (b) C7H14 ROPs, within the one-ring opening group of products, obtained with the AWx catalysts, as a function of the
corresponding naphthene conversion (open symbols: x = 2.9; half-filled symbols: x = 4.1; and filled symbols: x = 5.3).

R. Moraes et al. / Journal of Catalysis 286 (2012) 62–77
69
1,0
0,8
0,6
0,4
0,2
0,0
version, the cis/trans ratio is very close to the thermodynamic
equilibrium composition (ca. 10:90). This suggests that, in contrast
to what was observed for the W monofunctional acid catalysts, the
cis–trans equilibrium is rapidly reached over Ir metal catalysts.
Similar observations were reported by Lai et al. [65] when compar-
ing proton-form and Pt or Pd loaded zeolites and where ascribed to
metal-catalyzed dehydrogenation/hydrogenation reactions.
trans-decalin
Fig. 11 shows the variation of decalin conversion for the yIr/
AW0 catalysts as a function of Ir loading for a reaction temperature
of 623 K and a contact time of 20 g h molꢀ1. As reported for the
monofunctional acid catalysts, decalin conversion was extrapo-
lated to a contact time of 20 g h molꢀ1, assuming first order kinet-
ics. The results show that conversion, for both decalin and MCH,
increases with iridium content. For these catalysts, increasing irid-
ium content brings about a slight decrease in the dispersion (55–
40%). As a result, the overall number of available surface metal
sites increases with iridium content. Hence, the observed increase
in conversion reflects the increase in the number of surface metal
sites. It is also observed that under the same conditions, for a given
Ir loading, MCH conversion is higher than that of decalin. Higher
reactivity of MCH as compared to decalin can be inferred from
the work of McVicker et al. [16]. As noted previously, over these
catalysts, the cis/trans thermodynamic equilibrium composition
(10–90 vs. 40–60 in the starting decalin) is rapidly reached, result-
ing in the enrichment of the reactant mixture in trans-decalin.
Trans-decalin is reportedly less reactive than cis-decalin [29,66],
which was attributed, for the acid-catalyzed reaction, to a more
hindered tertiary C–H bond, in trans-decalin, making protonation
and formation of adsorbed carbocations more difficult than in
the case of cis-decalin. In the present case, a lower reactivity of
the trans form is consistent with its lower ring strain as compared
with the cis-form (ca. ꢀ7.9 kJ molꢀ1 vs. 5.0 kJ molꢀ1, respectively
[70]). In addition, trans-decalin exhibits lower ring strain than that
of MCH (ca. 4.2 kJ molꢀ1 [16]). Thus, one can tentatively attribute
the lower activity of decalin to the higher stability of trans-decalin
when compared to that of MCH. It should be noted, however, that
because of the use of n-heptane as a solvent in the decalin feed, one
cannot rule out possible inhibition of decalin conversion because of
competitive adsorption of n-heptane.
cis-decalin
0%
5%
10%
15%
20%
25%
Decalin conversion (%)
Fig. 10. Cis- and trans-decalin internal composition as a function of conversion for
the yIr/AW0 catalysts.
Analysis of the internal distribution of ROP, for the MCH reaction,
shows that methylhexanes (2MHx and 3MHx) are the most abun-
dant compounds followed by dimethylpentanes (23DMP and
24DMP) and heptane (Hp) in comparable amounts. No trim-
ethylbutanes or ethylpentanes were detected. The results in
Fig. 9a and b show that for the three active catalysts, at a given con-
version, essentially the same product selectivity was obtained.
3.2. Monofunctional metal catalysts
3.2.1. H2. chemisorption measurements
A series of Ir/Al2O3 catalysts (noted yIr/AW0) containing up to
1.8 wt% Ir were prepared. Dispersion values of 56, 46, and 40%
were obtained for the catalysts containing 0.6, 1.2, and 1.8 wt%,
respectively.
3.2.2.2. Selectivity. The performance of the monofunctional metal
catalysts, for decalin ring opening, was monitored and compared
with that of MCH. Fig. 12 reports the selectivities for 1ROP, CP,
and 2ROP, for all the monofunctional metal catalysts, as a function
of decalin and MCH conversion. The results show that, for a given
3.2.2. Decalin and MCH conversion: influence of Ir content
3.2.2.1. Activity. Fig. 10 depicts the evolution of the internal compo-
sition of cis- and trans-isomers as a function of conversion for the
yIr/AW0 catalyst. The results show that for very low decalin con-
Fig. 11. (a) Decalin and (b) MCH conversion obtained with the yIr/AW0 catalysts as a function of the Ir loading at 623 K and a contact time of 20 g h molꢀ1
.

70
R. Moraes et al. / Journal of Catalysis 286 (2012) 62–77
Fig. 12. Product selectivities obtained with the yIr/AW0 catalysts as a function of: (a) decalin and (b) MCH conversion (filled symbols: y = 0.6; half-filled symbols: y = 1.2;
open symbols: y = 1.8).
conversion, essentially the same selectivities are obtained for the
three catalysts. This suggests that the Ir loading has a minor influ-
ence on the product selectivity of both naphthenic molecules. For
low decalin conversion, the selectivity for one-ring opening was
around 100% and decreased with increasing conversion. CP selec-
tivity was low (ca. 0%) for low conversion and increased with con-
version. The high 1ROP selectivity observed at low decalin
conversions indicates that ring opening products are primary in
nature and typical of direct ring opening. Similar results have been
reported by McVicker et al. [16] for the conversion of decalin over
Ir/Al2O3 catalysts containing 0.9 wt% Ir. The decrease in 1ROP
selectivity with increasing conversion can be attributed to cracking
and second-ring-opening of 1ROPs. Fig. 12a also shows that 2ROP
selectivity slightly increases with conversion up to a maximum
of ca. 10% for the 1.8Ir/AW0 catalyst. Negligible amounts of aro-
matic compounds (e.g., tetralin) were detected.
Analysis of the products indicated that this could not be solely
explained by the dealkylation of MCH and that a second breaking
of the C–C bond appears to occur without desorption of methyl-
hexane intermediates (MHx), leading to the formation of 2 and
3 methylpentanes, as well as of methane. Fig. 13 depicts the mo-
lar selectivities of the C10H20 1ROPs, within the ring-opening
group of products, obtained for the yIr/AW0 catalysts, as a func-
tion of decalin conversion (Fig. 13a) and the corresponding
C7H14 ROPs selectivity as
a function of MCH conversion
(Fig. 13b). For the decalin reaction, methyl-propylcyclohexanes
(MPCH) and diethylcyclohexanes (DECH) were the major products
formed. It should be noted that, over these catalysts, MPCH con-
sisted essentially of 1-methyl-2-propylcyclohexane (shown below
in scheme 4), whereas over monofunctional acid catalysts, the
largest fraction of MPCH corresponds to more ramified methyl-
propylcyclohexanes (e.g., methyl-isopropylcyclohexane). Rela-
tively high amounts (ca. 10–20%) of n-butylcyclohexane (BCH)
and products with three or more alkyl ramifications (Others)
were also detected. Analysis of the internal distribution of ROPs,
for the MCH reaction, shows that 3-methylhexane (3MHx) is
the most abundant compound followed by 2-methylhexane
(MHx). Smaller amounts of n-heptane (Hp) were also detected.
For the MCH ring-opening reaction, in variance with the re-
sults reported for decalin, ROP and CP selectivities extrapolate,
respectively, to 80% and 20% for zero conversion (instead of
100% and 0%). Thus, cracking products are partially primary in
nature. Similar findings were reported by Lecarpentier et al.
[13] for the ring opening of MCH over Ir/ZrO2–SiO2 catalysts.
Fig. 13. Molar selectivities of: (a) C10H20 1ROPs and (b) C7H14 ROPs, within the one-ring opening group of products, obtained with the yIr/AW0 catalysts, as a function of
decalin and MCH conversion (filled symbols: y = 0.6; half-filled symbols: y = 1.2; open symbols: y = 1.8).

R. Moraes et al. / Journal of Catalysis 286 (2012) 62–77
71
(Scheme 3), respectively. The presence of such ring-contracted
products has been authenticated by GC–MS and GC ꢂ GC analyses
(not shown). In addition, they have also been observed by Kubicka
et al. [28] and Santikunaporn et al. [29] for the ring opening of dec-
alin over zeolites. Iso-butylcyclohexane can be formed as a second-
ary product by indirect opening of methyl-bicyclononanes [29]. It
should be noted that one cannot exclude the formation of n-butyl-
cyclohexane by direct ring opening of decalin (protolytic cracking
mechanism) [28,33,69]. However, the results in Fig. 8, showing a
near 100% RCP selectivity for low conversion, point to a minor con-
tribution of this pathway.
Analysis of the internal distribution of ROPs, for the MCH reac-
tion (Fig. 9b), shows that methylhexanes (2MHx and 3MHx) are
the most abundant compounds followed by dimethylpentanes
(23DMP and 24DMP) and heptane (Hp) in comparable amounts.
The presence of 23DMP and 24DMP indicates that ring contraction
precedes ring opening, since these compounds cannot be formed
by direct ring opening of MCH. The presence of n-heptane can be
explained by direct ring opening via b-scission of the methylcyclo-
hexane carbenium ion, as suggested by Cerqueira et al. [4] for the
transformation of MCH over zeolites.
4. Discussion
4.1. Reaction network
Scheme 1 (a and b) depicts the proposed reaction network for
the ring opening reaction of decalin and MCH, respectively.
The results obtained for the yIr/AW0 solids illustrate the ability
of Ir to open directly a six-carbon-atom ring. Thus, a direct path for
ring opening (rdo) must be considered in the proposed reaction net-
work. It has been shown for the AWx solids that ring contraction
reaction precedes that of ring opening. Such an indirect path is
illustrated by two consecutive reactions: ring contraction reaction
(rrcont) followed by indirect ring opening (rio). A second ring-open-
ing step (r2ro) was also considered for decalin, as it was observed
that some catalysts yield non-negligible amounts of paraffins. In
the proposed reaction network, three cracking pathways were con-
sidered: cracking of ROPs (rcrack1), cracking of RCPs (rcrack2), and
cracking of 2ROPs (rcrack3). The direct cracking of MCH (rcrack0
)
was considered over monofunctional metal catalysts since, con-
trary to decalin, MCH can react by cleavage of an exocyclic bond
to form methane and a cycloalkane. The proposed reaction net-
work differs from that reported in [26] by the cracking step of ring
contraction products, the second ring opening and the cracking of
the 2ROPs.
4.3. Monofunctional metal catalysts
As shown in Fig. 13, the most abundant products of decalin con-
version over monofunctional metal catalysts are MPCH and DECH.
The formation of these compounds can be readily explained by di-
rect ring opening. Their occurrence in a 2:1 ratio at low conversion
is consistent with the dicarbene mechanism [8,16,71], which in-
volves selective cleavage of non-substituted C–C bond (Scheme 4).
However, the formation of BCH (cleavage of decalin C–C bonds at
substituted positions) is not in line with this mechanism [16,71].
Its presence can be accounted for by a ‘‘partially selective’’ mecha-
nism (via a metallocyclobutane intermediate) (Scheme 5), compet-
ing with the dicarbene mechanism. Such a mechanism has been
proposed by Gault [8] to explain the results observed for the ring
opening of methylcyclopentane over low dispersed Pt/Al2O3 cata-
lysts and more recently by Do et al. [6] in the case of the direct ring
4.2. Monofunctional acid catalysts
As noted earlier, analyses of the 1ROPs internal distribution
(Fig. 9a) for the ring opening of decalin shows that ethyl-dimethyl-
cyclohexanes (EDMCH) were the most abundant compounds fol-
lowed by methyl-propylcyclohexanes (MPCH), butylcyclohexanes
(BCH), pentylcyclopentanes (C5CP), and ‘‘Others’’ (products with
three or more alkyl groups, except EDMCH).
Assuming that ring contraction precedes ring opening and that
over purely acid catalysts, the ring opening mechanism proceeds
via b-scission the presence of EDMCH and MPCH can be explained
by ring opening of ring-contracted products, such as trimethyl-
bicycloheptanes (Scheme 2) and methyl-bicyclononanes
a
b
Scheme 1. General reaction scheme for the ring-opening reaction of: (a) decalin and (b) MCH.
Scheme 2. Formation of EDMCH by b-scission of a trimethyl-bicyclo[2.2.1]heptane intermediate.

72
R. Moraes et al. / Journal of Catalysis 286 (2012) 62–77
Scheme 3. Formation of MPCH by b-scission of a methyl-bicyclo[3.4.0]nonane intermediate (adapted from [28]).
Scheme 4. Reaction pathway for the selective breaking of a non-substituted carbon–carbon bond of decalin over iridium, via the dicarbene mechanism [35].
Scheme 5. Formation of butylcyclohexane from decalin via the metallocyclobutane mechanism.
opening of 1,3-dimethylcyclohexane and 1,2-dimethylcyclohexane
over Ir/Al2O3 catalysts. Similarly, for the MCH reaction [16], the for-
mation of n-heptane (cleavage of substituted C–C bonds) is consis-
tent with ring opening via a metallocyclobutane mechanism. Note
that the minor formation of the products noted as ‘‘Others’’
(Fig. 13) cannot be explained by direct ring opening. The presence
of such products can be tentatively explained by the reported abil-
ity of iridium to promote isomerization [71–73].
4.4.1. Monofunctional acid catalysts (AWx): influence of W surface
densities
As noted earlier (Fig. 8), for these solids, ring-opening products
were secondary in nature. Direct ring-opening path (rꢄdo) was thus
deleted from the original general reaction network (Scheme 1). The
resulting reaction networks are presented in Scheme 6.
For the decalin reaction, cracking of ring-opened products
(crack 1) was observed, whereas this pathway was minor for the
MCH reaction as a result of the low yield of 1ROP.
4.4. Kinetic modeling – correlation between surface structure, acidity
and catalytic performance
Table 3 and Table 4 report the initial rates and activation ener-
gies calculated for the decalin and MCH reaction, respectively, over
the active solids (AW2.9, AW4.1, and AW5.3). The activation ener-
gies obtained for a given pathway appear to be little affected by the
W surface density, suggesting a similar nature of the active sites
(i.e., Brønsted acid sites of similar strength). In addition, the activa-
tion energies for ring contraction of decalin (ca. 80 kJ molꢀ1) and
MCH (ca. 95 kJ molꢀ1) are relatively close and can be attributed,
for both molecules, to the contraction of a C6 ring via an acid
mechanism.
For kinetic modeling (see Supplementary materials), all the
reactions shown in Scheme 1 were considered to be first order. Ini-
tial rates, determined as rꢄi ¼ kiꢄ ꢃ Pꢄ , where kꢄi is the rate constant at
523 K, were calculated using the reference conditions (P = 1 bar).
For the optimization of rate constants (ki) and activation energies
(Eai), the experimental data were treated as a function of the con-
tact time and the reaction temperature. The reaction network has
been used to model the performances obtained with various cata-
lysts. A very good agreement was obtained between the calculated
weight fractions and experimental results.
Fig. 14 shows the evolution of the initial rates for ring contrac-
tion (rꢄrcont), indirect ring opening (riꢄo) and cracking of mono-
naphthenes (rcꢄrack1 for decalin and rcꢄrack2 for MCH) calculated for

R. Moraes et al. / Journal of Catalysis 286 (2012) 62–77
73
a
b
Scheme 6. Schematic reaction used to determine the kinetic parameters from the experimental data for the monofunctional acid catalysts. (a) Decalin and (b) MCH.
Table 3
Initial rates for the decalin reaction at 523 K and activation energies expressed respectively in mol hꢀ1 kgꢀ1 and kJ molꢀ1 obtained with AWx catalysts (x in at. W/nm2).
rꢄrcont
rꢄio
rꢄcrack1
rꢄcrack2
rꢄ2ro
rꢄcrack3
x
Eacont
Eaio
Eacrack1
Eacrack2
Ea2ro
Eacrack3
2.9
4.1
5.3
0.20
0.40
0.44
85
77
78
1.94
3.55
3.59
48
46
48
1.50
2.39
2.67
50
47
48
0.25
0.35
0.47
66
67
64
0.09
0.13
0.13
67
62
62
0.46
0.70
0.72
89
85
83
Table 4
of 523 K as a function of the area of the infrared band at
2172 cmꢀ1 attributed to the interaction of CO with relatively
strong Brønsted acid sites [63]. A steady increase in the ring con-
traction rate for both decalin and MCH reactions with the area of
infrared band at 2172 cmꢀ1 was obtained, whereas no such rela-
tionship was evidenced with the bands associated with Lewis acid
sites. These results are consistent with those reported by Lecarpen-
tier et al. [13], for MCH reaction over tungstated zirconia catalysts.
The results in Fig. 14 also show that, for a given W surface density,
the rꢄrcont is similar for both naphthenes. However, the initial rate
observed for indirect ring opening of decalin is higher than that
of MCH. This can be tentatively attributed to higher strains of
ring-contracted bicyclonaphthenes when compared to cyclopen-
tanes [70] and is in line with the lower activation energy obtained
for the indirect ring opening of decalin. Higher indirect ring-open-
ing rate of decalin is also consistent with the reported increase in
indirect ring opening rate with the number of tertiary carbons in
the molecule [74].
Initial rates for the MCH reaction at 523 K and activation energies expressed
respectively in mol hꢀ1 kgꢀ1 and kJ molꢀ1 obtained with AWx catalysts (x in at. W/
nm2).
rꢄrcont
rꢄio
rꢄcrack2
x
Eacont
Eaio
Eacrack2
2.9
4.1
5.3
0.09
0.30
0.32
92
97
93
0.32
0.88
0.85
56
55
51
0.06
0.12
0.11
97
100
98
the AWx as a function of W surface density. In agreement with ear-
lier results (Fig. 7), negligible ring contraction activity (rrcont) was
detected for W surface densities 61.4 at. W/nm2, which is consis-
tent with the absence of relatively strong Brønsted acid sites in
these solids. For higher W surface densities, rrcont increases with
W surface density, reflecting the concomitant increase in the abun-
dance of relatively strong Brønsted acid sites (Fig. 5a). This rela-
tionship is clearly illustrated in Fig. 15, which shows the initial
ring contraction rate (rrcont) calculated for a reaction temperature
4,0
1,0
a
r°io
Decalin
MCH
b
r°io
0,8
0,6
0,4
0,2
0,0
3,0
2,0
1,0
0,0
r°crack1
r°rcont
r°rcont
r°crack2
0
1
2
3
4
5
6
0
1
2
3
4
5
6
W surface density (at./nm2)
W surface density (at./nm2)
Fig. 14. Initial rates for ring contraction, indirect ring opening, and cracking obtained with the AWx catalysts as a function of W surface density. (a) decalin: rꢄrcont (h), riꢄo
rcꢄrack1 (s); (b) MCH: rꢄrcont (h), riꢄo ), rꢄcrack2 (s).
(r),
(r

74
R. Moraes et al. / Journal of Catalysis 286 (2012) 62–77
Table 6
Initial rates for the MCH reaction at 523 K and activation energies expressed
respectively in mol hꢀ1 kgꢀ1 and kJ molꢀ1 obtained with the yIr/AW0 catalysts (y in
wt.%).
rꢄdo
rꢄcrack1
rꢄcrack0
y
Dispersion (%)
Eado
Eacrack1
Eacrack0
0.6
1.2
1.8
55
46
40
0.016
0.063
0.231
158
155
152
0.032
0.085
0.201
167
159
152
0.003
0.006
0.042
184
176
167
the dicarbene mechanism [6]. This is consistent with the results
described above indicating that the dicarbene mechanism is the
predominant pathway for direct ring opening of decalin and
MCH probe molecules.
Fig. 16 shows the evolution of the initial rates for direct ring
opening (rꢄdo) and cracking of ring-opened products (rꢄcrack1), calcu-
lated for the yIr/AW0 catalysts as a function of Ir loading, for both
naphthenic molecules. It is clear that, for both reactions, the initial
rate of each reaction step increases with iridium content. The same
tendency was observed for the remaining pathways. Similar find-
ings were obtained when the initial rates were plotted as a func-
tion of Ir surface atoms (titrated by H2 chemisorption) (Fig. 17).
However, the evolution is not linear with intercept at the origin,
suggesting an apparent structure sensitivity of the reaction. This
is clearly illustrated in Fig. 18, which depicts the turnover fre-
quency, TOF (expressed as initial rate rꢄdo per surface Ir atom), as
a function of Ir particle size. The results show that, for both mole-
cules, the TOF increases with increasing particle size. This is in
agreement with the results of Walter et al. [22] for ring opening
of MCH over Ir/Al2O3 catalysts and those reported by Angel et al.
[75] and Fuentes et al. [7], for ring opening of methylcyclopentane
and cyclohexane, respectively, over Rh/Al2O3 catalysts [75]. Inter-
estingly, in the case of the related Rh system [75], the most signif-
icant variations in the TOF for methylcyclopentane conversion
were observed for intermediate Ir dispersion range (between 30%
and 50%). This is consistent with the results of the present study
and can be interpreted, as proposed by the authors for the Rh sys-
tem, in terms of changes in the surface structure of Ir particles
occurring in this dispersion range. It should be noted that within
the range of dispersion examined, no change in product selectivity
was observed (Fig. 12 and Fig. 13).
Fig. 15. Initial ring contraction rate (rꢄrcont) at 523 K for AWx catalysts as a function
of the areas of the infrared CO vibration band at 2172 cmꢀ1
.
4.4.2. Monofunctional metal catalysts (yIr/AW0): influence of Ir
content
As noted previously (Fig. 12), negligible activity for ring con-
traction was observed over these catalysts. Hence, the indirect
ring-opening (rꢄio) pathway via ring contraction of MCH was not
considered in the modeling of the results. The reaction network
was reduced to the following reaction network (Scheme 7).
Table 5 and 6 report the calculated direct ring opening initial
rates and activation energies for the monofunctional metal cata-
lysts obtained for the decalin and MCH reactions, respectively.
The results show that, for both molecules, within the range of Ir
loadings examined, Ir content seems to have no significant effect
on the activation energy of a given pathway. Table 5 and 6 also
indicate that similar activation energies for direct ring opening of
both molecules (ca. 150 kJ molꢀ1) were obtained. The calculated
values are typical of those reported for direct ring opening via
a
b
Scheme 7. Schematic reaction used to determine the kinetic parameters from the experimental data for monofunctional metal catalysts. (a) Decalin and (b) MCH.
Table 5
Initial rates for the decalin reaction at 523 K and activation energies expressed respectively in mol hꢀ1 kgꢀ1 and kJ molꢀ1 obtained with the yIr/AW0 catalysts (y in wt.%).
rꢄdo
rꢄcrack1
rꢄ2ro
rꢄcrack3
y
Dispersion (%)
Eado
Eacrack1
Ea2ro
Eacrack3
0.6
1.2
1.8
55
46
40
0.003
0.014
0.027
150
146
142
0.004
0.007
0.010
167
159
159
0.004
0.008
0.014
142
138
138
0.006
0.013
0.018
167
163
163

R. Moraes et al. / Journal of Catalysis 286 (2012) 62–77
75
0,25
0,03
0,02
0,01
0,00
MCH
b
a
Decalin
0,20
r°do
0,15
0,10
r°crack1
r°crack1
0,05
r°do
0,00
0,0
0,5
1,0
1,5
2,0
0,0
0,5
1,0
1,5
2,0
Ir loading (wt%)
Ir loading (wt%)
Fig. 16. Initial rates for direct ring opening (rꢄdo) and cracking of ring opening products (rcꢄrack1) for the yIr/AW0 catalysts as a function of Ir loading. (a) Decalin and (b) MCH.
0,3
0,2
0,1
0,0
0,03
0,02
0,01
0,00
MCH
a
Decalin
b
0
10
20
30
40
0
10
20
30
40
Surface Ir atoms (µmol/g)
Surface Ir atoms (µmol/g)
Fig. 17. Initial rate of direct ring opening (rꢄdo) at 523 K for yIr/AW0 catalysts as a function of the surface iridium atoms titrated by H2 chemisorption. (a) Decalin and (b) MCH.
1,0
0,8
0,6
0,4
0,2
0,0
8,0
6,0
4,0
2,0
0,0
a
Decalin
MCH
b
1,5
2,0
2,5
3,0
1,5
2,0
2,5
3,0
Particle size (nm)
Particle size (nm)
Fig. 18. Turnover frequencies (TOF) as a function of Ir particle size for direct ring-opening reaction of decalin and MCH. (a) Decalin and (b) MCH.
It should be noted that the results reported in the present study
were acquired under S-free conditions. The extreme sensitivity of
Ir to S (ppb levels) is well documented and is a major issue in oil
refining [76–79]. In commercial practice, when Ir catalysts are used
(e.g., in naphtha reforming), suitable guard beds prevent S from
reaching the reactor. Thus, in principle, the sulfur content in the

76
R. Moraes et al. / Journal of Catalysis 286 (2012) 62–77
feed can be extensively reduced before contact with the catalysts.
However, although the topic is outside the scope of this work, a
systematic study of its influence on the catalytic performance is
of obvious interest.
the monofunctional metal catalysts for 2ROPs, for which the CN
can reach 77. This clearly indicates that monofunctional metal sol-
ids are better suited for the LCO upgrading than monofunctional
acid catalysts. With respect to the performance of bifunctional sys-
tems, to date few studies have examined the potential improve-
ment of cetane number based on the nature of the most
abundant ring-opening products of decalin over these catalysts.
Santikunaporn et al. [29] noted that over bifunctional Pt/HY cata-
lysts, the products formed do not exhibit cetane numbers signifi-
cantly higher than those of the saturated aromatics. This is
consistent with the study of Santana et al. [35], which shows that
only preferential cleavage at substituted C–C bonds can yield prod-
ucts with CN substantially higher than that of the decalin feed.
4.5. Evaluation of possible improvement of the cetane number on
decalin conversion
An assessment of possible improvement of the cetane number
following ring opening was made for decalin by comparing the
CN of decalin (CN = 36 [35]) with that of the most abundant 1ROPs
formed. The CN of representative products resulting from opening
of the first decalin ring are presented in Table 7. The internal
distribution of C10H20 one-ring-opening products, over monofunc-
tional acid catalysts (Fig. 9a), showed that ethyl-dimethylcyclohex-
anes (EDMCH) were the most abundant compounds followed
by methyl-propylcyclohexanes (MPCH), iso-butylcyclohexanes
(BCH), pentylcyclopentanes (C5CP), and ‘‘Others’’ (products, other
than EDMCH, with three or more alkyl groups).
5. Conclusions
The activity and selectivity of monofunctional acid (WO3/Al2O3)
and metal catalysts (Ir/Al2O3) for ring opening reaction of decalin
and methylcyclohexane (MCH) were examined.
As shown in Table 7, EDMCH (CN = 30), MPCH (CN = 24) and
BCH (CN = 36) exhibit a lower or equal CN than that of decalin
(CN = 36) whereas the CN for the products C5CP and ‘‘Others’’
range between 15 and 29. Moreover, over these catalysts, a very
high cracking selectivity was observed. This strongly suggests that
no CN improvement can be obtained for the ring opening of decalin
over monofunctional acid catalysts. These observations are in
agreement with the predictions of cetane number improvement
reported by Santana et al. [35], for acid-catalyzed ring opening of
decalin.
In the case of monofunctional metal catalysts, the major prod-
ucts obtained were MPCH (1-methyl-2-propylcyclohexane) and
DECH (cleavage of the C–C bonds at non-substituted positions),
which exhibit higher cetane number than that of decalin
(CN = 39). Note that the 1ROP with the highest CN (n-butylcyclo-
hexane; CN = 47), formed via cleavage at substituted positions,
was detected with ca. 10% selectivity at high conversion. In addi-
tion, ca. 12% of selectivity at high conversion was observed over
For both molecules, the activity of monofunctional acid cata-
lysts was only observed for W surface density higher than 1.4 at.
W/nm2, which is consistent with the development of relatively
strong Brønsted acid sites. For these catalysts, ring contraction
product (RCP) selectivity was high at low conversion and de-
creased with increasing conversion, indicating that ring contrac-
tion is the first step for ring opening of decalin and MCH. The
decrease in RCP selectivity was less pronounced for the MCH reac-
tion when compared to that of decalin, suggesting that the RCPs
formed during the MCH reaction are less reactive than those of
decalin. Naphthenes ring opening occurs according to an acid
mechanism, which involves protolytic dehydrogenation, skeletal
isomerization, b-scission, and hydride transfer. For the active cata-
lysts, the low selectivity for one-ring opening and the low CN of the
1ROP formed suggest that these are not suitable catalysts for the
upgrading of LCO.
The performance of monofunctional metal catalysts yIr/AW0
was also compared for both naphthenic molecules. Conversion
Table 7
Major products and respective cetane number obtained for the ring opening of decalin, over monofunctional acid and metal catalysts [35].

R. Moraes et al. / Journal of Catalysis 286 (2012) 62–77
77
[29] M. Santikunaporn, J.E. Herrera, S. Jongpatiwut, D.E. Resasco, W.E. Alvarez, E.L.
Sughrue, J. Catal. 228 (2004) 100.
[30] K.C. Mouli, V. Sundaramurthy, A.K. Dalai, Z. Ring, Appl. Catal. A 321 (2007) 17.
[31] K.C. Mouli, V. Sundaramurthy, A.K. Dalai, J. Mol. Catal. A: Chem. 304 (2009) 77.
[32] D.Y. Murzin, D. Kubicka, I.L. Simakova, N. Kumar, A. Lazuen, P. Maki-Arvela, M.
Tiitta, T. Salmi, Pet. Chem. 49 (2009) 90.
[33] A. Corma, V. Gonzalez-Alfaro, A.V. Orchilles, J. Catal. 200 (2001) 34.
[34] S. Dokjampa, T. Rirksomboon, S. Osuwan, S. Jongpatiwut, D.E. Resasco, Catal.
Today 123 (2007) 218.
[35] R.C. Santana, P.T. Do, M. Santikunaporn, W.E. Alvarez, J.D. Taylor, E.L. Sughrue,
D.E. Resasco, Fuel 85 (2006) 643.
[36] C. Flego, N. Gigantiello, W.O. Parker Jr, V. Calemma, J. Chromatogr. A 1216
(2009) 2891.
[37] A. Slagtern, I.M. Dahl, K.J. Jens, T. Myrstad, Appl. Catal. A 375 (2010) 213.
[38] C. Berger, A. Raichle, R.A. Rakoczy, Y. Traa, J. Weitkamp, Micropor. Mesopor.
Mater. 59 (2003) 1.
[39] Z. Paál, P. Tetenyi, Nature 267 (1977) 234.
[40] P. Samoila, M. Boutzeloit, C. Especel, F. Epron, P. Marécot, Appl. Catal. A 369
(2009) 104.
was relatively low for the lowest Ir loading and increases signifi-
cantly with increasing iridium content. Ir loading has a minor
influence on the product selectivity for both naphthenic molecules.
For low conversion, the results for monofunctional metal catalysts
showed that ROP selectivity is high and decreases with increasing
conversion.
The internal distribution of C10H20 one-ring-opening products
evidenced that, over Ir/Al2O3, hydrogenolysis via the dicarbene
mechanism is the preferred path to ring open both naphthenes.
However, a competitive ‘‘partially selective’’ mechanism (via a
metallocyclobutane intermediate), leading to the formation of high
CN products, is also operative. The CN and the ring opening prod-
ucts selectivity (1ROP and 2ROP) obtained from ring opening over
monofunctional metal catalysts suggest that these catalysts are
better suited for the LCO upgrading than monofunctional acid
catalysts.
[41] B. Coq, E. Crabb, F. Figuéras, J. Mol. Catal. A: Chem. 96 (1995) 35.
[42] H. Schulz, J. Weitkemp, H. Eberth, in: J.W. Hightower (Ed.), Proc. 5th ICC,
North-Holland, Amsterdam, 1973, p. 1229.
[43] G. Onyestyak, G. Pál-Borbély, H.K. Beyer, Appl. Catal. A 229 (2002) 65.
[44] W.C. Baird, D.P. Klein, J.G. Chen, G.B. McVicker, US Patent 6683020, 2004 (to
Exxon-Mobil).
Appendix A. Supplementary material
[45] S. Lecarpentier, J. Van Gestel, J.-P. Dath, C. Collet, J.-P. Gilson, WO Patent
Application 2007/006924, 2007 (to Total-France).
[46] M.A. Arribas, P. Concepción, A. Martínez, Appl. Catal. A 267 (2004) 111.
[47] T.N. Vu, J. van Gestel, J.-P. Gilson, C. Collet, J.-P. Dath, J.C. Duchet, J. Catal. 231
(2005) 453.
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.jcat.2011.10.014.
[48] J.-W. Park, K. Thomas, J. van Gestel, J.-P. Gilson, C. Collet, J.-P. Dath, M. Houalla,
Appl. Catal. A 388 (2010) 37.
References
[49] D.S. Cunha, G.M. Cruz, Appl. Catal. A 236 (2002) 55.
[50] P.N. Da Silva, M. Guenin, C. Leclercq, R. Frety, Appl. Catal. 54 (1989) 203.
[51] T. Onfroy, G. Clet, M. Houalla, J. Phys. Chem. B (2005).
[52] S.S. Chan, I.E. Wachs, L.L. Murrell, N.C. Dispenziere, in: S. Kaliaguine, A. Mahay
(Eds.), Stud. Surf. Sci. Catal., Elsevier, 1984, p. 259.
[53] I.E. Wachs, Catal. Today 27 (1996) 437.
[54] O. Cairon, T. Chevreau, J.C. Lavalley, J. Chem. Soc. Faraday Trans. 94 (1998)
3039.
[55] O. Cairon, K. Thomas, A. Chambellan, T. Chevreau, Appl. Catal. A 238 (2003)
167.
[56] L.M. Kustov, V.B. Kazanskii, S. Beran, L. Kubelkova, P. Jiru, J. Phys. Chem. 91
(1987) 5247.
[1] H. Belatel, H. Al-Kandari, F. Al-Khorafi, A. Katrib, F. Garin, Appl. Catal. A 275
(2004) 141.
[2] V. Calemma, A. Carati, C. Flego, R. Giardino, R. Millini, Abstr. Pap. Am. Chem.
Soc. 229 (2005) 51.
[3] P. Castaño, A.G. Gayubo, B. Pawelec, J.L.G. Fierro, J.M. Arandes, Chem. Eng. J.
140 (2008) 287.
[4] H.S. Cerqueira, P.C. Mihindou-Koumba, P. Magnoux, M. Guisnet, Ind. Eng.
Chem. Res. 40 (2001) 1032.
[5] A. Corma, F. Mocholi, V. Orchilles, G.S. Koermer, R.J. Madon, Appl. Catal. 67
(1991) 307.
[6] P.T. Do, W.E. Alvarez, D.E. Resasco, J. Catal. 238 (2006) 477.
[7] S. Fuentes, F. Figueras, J. Catal. 61 (1980) 443.
[57] L. Kubelkova, S. Beran, J.A. Lercher, Zeolites 9 (1989) 539.
[58] C. Pfaff, M.J.P. Zurita, C. Scott, P. Patino, M.R. Goldwasser, J. Goldwasser, F.M.
Mulcahy, M. Houalla, D.M. Hercules, Catal. Lett. 49 (1997) 13.
[59] L. Karakonstantis, C. Kordulis, A. Lycourghiotis, Langmuir 8 (1992) 1318.
[60] W.S. Millman, M. Crespin, A.C. Cirillo, S. Abdo, W.K. Hall, J. Catal. 60 (1979) 404.
[61] A.M. Turek, I.E. Wachs, E. Decanio, J. Phys. Chem. 96 (1992) 5000.
[62] C. Morterra, G. Magnacca, Catal. Today 27 (1996) 497.
[63] X. Chen, G. Clet, K. Thomas, M. Houalla, J. Catal. 273 (2010) 236.
[64] S.L. Soled, G.B. McVicker, L.L. Murrell, L.G. Sherman, N.C. Dispenziere, S.L. Hsu,
D. Waldman, J. Catal. 111 (1988) 286.
[65] W.-C. Lai, C. Song, Catal. Today 31 (1996) 171.
[66] E.F. Sousa-Aguiar, C.J.A. Mota, M.L. Valle Murta, M. Pinhel da Silva, D. Forte da
Silva, J. Mol. Catal. A: Chem. 104 (1996) 267.
[67] C.J. Egan, G.E. Langlois, R.J. White, J. Am. Chem. Soc. 84 (1962) 1204.
[68] D.M. Brouwer, H. Hogeveen, Recl. Trav. Chim. Pays-Bas 89 (1970) 211.
[69] H.B. Mostad, T.U. Riis, O.H. Ellestad, Appl. Catal. 63 (1990) 345.
[70] K.B. Wiberg, Angew. Chem.-Int. 25 (1986) 312.
[71] F. Weisang, F.G. Gault, J. Chem. Soc. Chem. Commun. 11 (1979) 519.
[72] D. Kubicka, N. Kumar, P. Mäki-Arvela, T. Venäläinen, M. Tiitta, T. Salmi, D.Y.
Murzin, Stud. Surf. Sci. Catal. 158 (2005) 1669.
[8] F.G. Gault, Adv. Catal. 30 (1981) 1.
[9] H. Glassl, K. Hayek, R. Kramer, J. Catal. 68 (1981) 397.
[10] S.L. Gonzalez-Cortes, S. Dorkjampa, P.T. Do, Z. Li, J.M. Ramallo-Lopez, F.G.
Requejo, Chem. Eng. J. 139 (2008) 147.
[11] R. Kramer, H. Zuegg, J. Catal. 80 (1983) 446.
[12] R. Kramer, H. Zuegg, J. Catal. 85 (1984) 530.
[13] S. Lecarpentier, J. Van Gestel, K. Thomas, J.-P. Gilson, M. Houalla, J. Catal. 254
(2008) 49.
[14] F. Locatelli, D. Uzio, G. Niccolai, J.M. Basset, J.P. Candy, Catal. Commun. 4 (2003)
189.
[15] G. Maire, G. Plouidy, J.C. Prudhomme, F.G. Gault, J. Catal. 4 (1965) 556.
[16] G.B. McVicker, M. Daage, M.S. Touvelle, C.W. Hudson, D.P. Klein, W.C.J. Baird,
B.R. Cook, J.G. Chen, S. Hantzer, D.E.W. Vaughan, E.S. Ellis, O.C. Feeley, J. Catal.
210 (2002) 137.
[17] G.B. McVicker, O.C. Feeley, J.J. Ziemiak, D.E. Vaughan, K.C. Strohmaier, W.R.
Kliewer, D.P. Leta, J. Phys. Chem. B 109 (2005) 2222.
[18] G. Moretti, W.M.H. Sachtler, J. Catal. 116 (1989) 350.
[19] A. Raichle, Y. Traa, F. Fuder, M. Rupp, J. Weitkamp, Angew. Chem. Int. Ed. 40
(2001) 1243.
[20] G. Rodriguez-Gattorno, L.O. Aleman-Vazquez, X. Angeles-Franco, J.L. Cano-
Dominguez, R. Villagomez-Ibarra, Energy Fuels 21 (2007) 1122.
[21] T. Sugii, Y. Kamiya, T. Okuhara, Appl. Catal. A 312 (2006) 45.
[22] C.G. Walter, B. Coq, F. Figuéras, Appl. Catal. A 133 (1995) 95.
[23] Z. Wang, A. Nelson, Catal. Lett. 123 (2008) 226.
[24] Y. Zhuang, A. Frennet, Appl. Catal. A 134 (1996) 37.
[25] B.H. Cooper, B.B.L. Donnis, Appl. Catal. A 137 (1996) 203.
[26] D. Kubicka, N. Kumar, P. Mäki-Arvela, M. Tiitta, V. Niemi, H. Karhu, T. Salmi,
D.Y. Murzin, J. Catal. 227 (2004) 313.
[73] F.G. Gault, V. Amirebrahimi, F. Garin, P. Parayre, F. Weisang, Bull. Soc. Chim.
Belg. 88 (1979) 475.
[74] C. Marcilly, Catalyse acido-basique, Application au Raffinage et
Pétrochimie, Editions Technip, 2003.
à la
[75] G.D. Angel, B. Coq, R. Dutartre, F. Figuéras, J. Catal. 87 (1984) 27.
[76] J. Barbier, E. Lamy-Pitara, P. Marécot, J.P. Boitiaux, J. Cosyns, F. Verna, Adv.
Catal. 37 (1990) 279.
[77] R. Frety, P.N. Da Silva, M. Guenin, Appl. Catal. 57 (1990) 99.
[78] S. Nassreddine, L. Massin, M. Aouine, C. Geantet, L. Piccolo, J. Catal. 278 (2011)
253.
[27] D. Kubicka, T. Salmi, M. Tiitta, D.Y. Murzin, Fuel 88 (2009) 366.
[28] D. Kubicka, N. Kumar, P. Mäki-Arvela, M. Tiitta, V. Niemi, T. Salmi, D.Y. Murzin,
J. Catal. 222 (2004) 65.
[79] R.W. Rice, D.C. Keptner, Appl. Catal. A 262 (2004) 233.