GModel
APCATA-14269; No. of Pages9
ARTICLE IN PRESS
2
S.A. D’Ippolito et al. / Applied Catalysis A: General xxx (2013) xxx–xxx
(
MCH) was taken as model molecule. The studied samples are
i.e., C –C6 products resulting from either deep hydrogenolysis or
1
those prepared in the first part of this paper [37], characterized
by measurement of the Pd and Rh contents, X-ray diffraction, H2
chemisorption, temperature-programmed desorption of pyridine,
isomerization of 3,3-dimethyl-1-butene (33DM1B), and evaluated
for the selective ring opening of decalin.
cracking; (ii) isomerization products, i.e., saturated cyclic alkanes
containing the same number of carbon atoms as MCH but with C5
ring; (iii) ring opening products, i.e., C7 alkanes resulting from the
opening of MCH and its isomers; (iv) dehydrogenation product, i.e.,
toluene. The yield in each kind of products was defined as the per-
centage of MCH converted into the given products. The selectivities
were calculated by dividing the yield in each kind of products by
the amount of MCH converted as reported previously [36].
2
. Experimental
Fig. 1 shows the conversion as a function of reaction temperature
for all the studied catalysts supported either on SIRAL 5, SIRAL 20
or SIRAL 40. All the catalysts display a high MCH conversion which
increases as the temperature increases. For the three series, the
Rh1 monometallic catalysts are more active than Pd1 monometallic
ones, putting in evidence the higher activity of Rh than Pd. For this
reason, bimetallic catalysts with higher Rh content are more active.
The activity follows the order: Rh1 > R2 > R1 = R0.5 = Pd1. So, the
addition of Rh to Pd allows working at lower temperatures to obtain
high conversion.
2
.1. Catalysts preparation
Three commercial SiO –Al O supports provided by Sasol
2
2
3
(
SIRAL 5, SIRAL 20 and SIRAL 40) were used as support. Previously,
◦
◦
−1
3
−1
they were calcined at 450 C for 4 h (10 C min , air, 60 cm min ).
Rh and/or Pd were added by a common impregnation method. An
aqueous solution of HCl (0.2 mol L ) was added to the support and
the system was left unstirred at room temperature for 1 h. Then an
−1
∼
∼
aqueous solution of RhCl and/or PdCl (Sigma–Aldrich) was added
3
2
in order to have a 1 wt% of total metal charge. For the bimetal-
lic catalyst, the Rh/Pd atomic ratio was x = 0.5, 1 and 2. The slurry
was gently stirred for 1 h at room temperature and then it was
put in a thermostated bath at 70 C until a dry solid was obtained.
The drying was completed in a stove at 120 C overnight. Finally,
Fig. 2 shows the yield and the selectivity to RO products as a
function of the reaction temperature. All the catalysts have a max-
imum of yield to RO products with the reaction temperature, with
the sole exception of the Pd1 catalysts that do not yield RO prod-
ucts at all. It can be seen that the temperature of maximum yield is
shifted to lower temperatures when the Rh content increases in the
catalysts. Moreover, as the Rh content in the bimetallic catalysts is
increased, higher yields to RO products are obtained, the optimum
being obtained for the R2 bimetallic samples, which also present
higher yield in RO products than the monometallic Rh catalyst.
Likewise, at higher conversions (>70%), the R2 bimetallic catalysts
are more selective to RO than Rh1 catalysts. The effect of Rh addi-
tion to the Pd metal function is stronger in the case of the SIRAL5
catalyst; only with this support the bimetallic R0.5 catalyst was
more selective to RO than the Pd1 catalyst. On the S20 and S40
series, the increase in selectivity to RO is noticeable especially for
R2 catalysts. The maximum yield to RO on the S5 series increases
in the following order: R2 > R1 > Rh1 > R0.5 ꢀ Pd1, while for the
S20 series the order is: R2 > Rh1 ꢀ R1 > R0.5 > Pd1, and for the R40
series: R2 > Rh1 ꢀ R0.5 > R1 > Pd1. At conversions higher than 80%,
the selectivity to RO products decreases dramatically because the
catalysts produce cracking and dehydrogenated products, which
are thermodynamically favored at higher reaction temperature as
it can be seen in Figs. 3 and 6. Again, the higher hydrogenolytic char-
acter of Rh compared to Pd leads to an increase of the RO reaction.
However, the maximum yields to RO products (Fig. 2) are obtained
on the bimetallic R2 catalyst. Both catalysts (R2 and Rh1) supported
on SIRAL 40 have similar metallic dispersion [37], but the Rh1 cat-
alyst presents the highest activity, which could be attributed to a
higher accessibility of rhodium in the monometallic catalyst than in
the bimetallic R2 catalyst. The lower yield in RO product obtained
with the Rh1 catalysts compared to the R2 ones can be explained by
the higher yields in cracking products obtained with these catalysts
as shown in Fig. 3. Thus, the presence of rhodium at the bimetallic
particle surface favors the activity in hydrogenolysis but the pres-
ence of Pd limits deep hydrogenolysis thus favoring the formation
of RO products.
◦
◦
3
−1
◦
the samples were calcined in flowing air (60 cm min ) at 300 C
3
−1
◦
for 4 h and reduced under flowing H2 (60 cm min , 500 C, 4 h).
The monometallic catalysts are named Pd1/Sy or Rh1/Sy, while the
bimetallic are named Rx/Sy, where Sy is the support (SIRAL) and y
is the weight percentage of SiO . In the case of the bimetallic cata-
lysts, R corresponds to the Rh/Pd atomic ratio and x is the value of
2
this ratio.
2.2. Ring opening of methylcyclohexane
MCH hydrogenolysis was performed in a fixed-bed continuous
reactor (stainless steel tube of 1.3 cm inner diameter) with 1.5 g
of catalyst, under a total pressure of 39 atm, with a molar ratio
H /MCH = 8 (weighted hourly space velocity WHSV = 2 h ) and
in a temperature range from 250 to 425 C. Previously, the cata-
lysts were reduced in situ at the reaction pressure at 500 C for
−
1
2
◦
◦
3
−1
1
h using 60 cm min
of H . The methylcyclohexane flow was
2
controlled using a HPLC pump (Gilson). Effluent products were ana-
lyzed by an on-line chromatograph (Varian 3900) using a FID and
equipped with a PONA capillary column. The conversion was varied
by changing the reaction temperature, by steps of 25 C. The initial
reaction temperature was chosen according to the metal loading
and the support of each catalyst. After reaching a conversion near
◦
1
00%, the temperature was not increased more. Four measure-
ments were performed at each temperature; the values given in
the results correspond to the average of these four measurements.
The error associated with the data is at maximum of 5%. It was
checked on some catalysts that the deactivation occurring during
the reaction is negligible. For that purpose, after having reached the
highest reaction temperature and then the maximum of conver-
◦
sion, the temperature was decreased by steps of 25 C to the lowest
temperature. The MCH conversion as well as the products distri-
bution obtained was considered as similar. The absence of intra
and extragranular diffusion limitations in these experimental con-
ditions was checked by varying independently the particle size (in
the 0.1–0.6 mm range), the amount of catalyst (from 500 mg to 2 g)
In fact, as shown in Fig. 3, the yield and selectivity to cracking
products increase with the reaction temperature, however with
quite different behaviors according to the nature of the catalyst
and the support. For the catalysts supported on SIRAL 5, all the sys-
tems containing Rh lead to cracking products, Rh1 and R2 being the
most selective ones. On the other hand, on SIRAL 20 and SIRAL 40,
the R0.5 and R1 catalysts (besides Pd1 systems) have little selec-
character of Rh is more evident on the S5 (Table 1). The differ-
ences in the behavior of catalysts supported on SIRAL 5 and those
−
1
and the reactant flow rates (from 0.0325 to 0.1300 mL min ).
3
. Results and discussion
The MCH conversion was determined as a function of tempera-
ture. The reaction products are classified in: (i) cracking products,