L. Yang et al. / Journal of Catalysis 330 (2015) 330–343
341
(0.37 g min/mol) over WP, Ni2P, and MoP were all calculated
(Table S4 in the Supplementary Material). Over Ni2P, as generally
accepted, the TOF of 4,6-DMDBT in the absence of piperidine was
much lower than that of DBT. However, 4,6-DMDBT showed a
higher TOF than DBT at 0 kPa piperidine. These results suggest that
WP is particularly promising for deep HDS. Not only the hydro-
genation of the sulfur-containing compounds, but also that of
DM-BP, was accelerated by the methyl groups. The rate of hydro-
genation of DM-BP in the presence of 0.5 kPa DBT (0.36 mol/g min)
was five times as fast as that of BP at 0.5 kPa BT (0.07 mol/g min),
demonstrating that the hydrogenation of DM-BP cannot be
neglected in the HDS of 4,6-DMDBT. Also, over a Ni–MoS2/
of 4,6-DMDBT was completely inhibited by 0.2 kPa piperidine at
6 1.1 g min/mol. Only when piperidine was almost completely
converted at = 1.5 g min/mol (Fig. S4 in the Supplementary
Material) did 4,6-DMDBT start to react. Therefore, we did not cal-
culate the rate constants of its HYD and DDS pathways. The overall
active-site-based rate constant of the HDS of DBT in the absence of
s
s
piperidine over WP (0.28 mol/
over Ni2P (0.336 mol/ mol min) and higher than that over MoP
(0.22 mol/ mol min) (Table 2). For the HDS of 4,6-DMDBT, the
overall active-site-based rate constant over WP (0.44 mol/ mol min)
was 10 times higher than over Ni2P (0.045 mol/ mol min)
lmol min) was comparable to that
l
l
l
l
(Table 2), indicating that the intrinsic activity of WP in the HDS
of 4,6-DMDBT is much higher than that of Ni2P. However, the sit-
uation was reversed after the addition of piperidine. For the HDS
of DBT, WP became the least active catalyst (Table 2). Ni2P was
more nitrogen-tolerant than WP and MoP. Even in the presence
of 0.5 kPa piperidine, Ni2P was still the most active catalyst in
catalyzing the HDS of DBT (Table 2). The variations of the TOFs of
DBT and 4,6-DMDBT over Ni2P, MoP, and WP (Table S4 in
the Supplementary Material) agreed with these results. At
c-Al2O3 catalyst, the methyl groups promoted the hydrogenation
of 4,6-DMDBT, TH-4,6-DMDBT, and HH-4,6-DMDBT; this was
suggested to be the result of electron donation by the methyl
groups [13].
While the DDS of TH-DBT (0.33 mol/g min) was 2.6 times
slower than that of HH-DBT (0.85 mol/g min) (Scheme 1), the
DDS of TH-4,6-DMDBT (0.33 mol/g min) was faster than that of
HH-4,6-DMDBT (0.23 mol/g min) (Scheme 2). This is in agreement
with results reported for the HDS of DBT, 4,6-DMDBT, and their
hydrogenated intermediates over metal sulfides [13,32]. At
300 °C and total pressure 3 MPa, the rate constants of the DDS of
s
= 0.37 g min/mol, the TOF of 4,6-DMDBT over WP decreased to
zero after the addition of piperidine. Thus, WP is more sensitive
to piperidine than Ni2P and MoP. This may explain why, although
WP is superior to Ni2P in the HDS of 4,6-DMDBT in the absence of
piperidine, its activity in simultaneous HDS and HDN is lower than
that of Ni2P [5,6]. The inhibitory effect of nitrogen-containing com-
pounds on HDS is usually ascribed to their competitive adsorption
with sulfur-containing compounds. If the same holds true for WP,
the adsorption of piperidine onto WP should be much stronger
than that of DBT and 4,6-DMDBT. Since the HDS of 4,6-DMDBT
was more inhibited by piperidine than that of DBT, the adsorption
of 4,6-DMDBT over WP is expected to be weaker than that of DBT.
The degree of inhibition of piperidine on the DDS of 4,6-DMDBT
and its hydrogenated intermediates was in the order 4,6-DMDBT ꢃ
TH-4,6-DMDBT > HH-4,6-DMDBT. This must be due to the different
adsorption constants of these molecules. As discussed above, the
DDS of 4,6-DMDBT, TH-4,6-DMDBT, and HH-4,6-DMDBT may
occur mainly through hydrogenolysis over WP. A prerequisite for
the hydrogenolysis of the aromatic sulfides over metal sulfides or
metals is the perpendicular adsorption of these molecules onto
TH-4,6-DMDBT and HH-4,6-DMDBT over a Ni–MoS2/
lyst were 0.23 kPa mol/g min and 0.21 kPa mol/g min, respectively
[13]. Over a Co–MoS2/ -Al2O3 catalyst under identical conditions,
c-Al2O3 cata-
c
the obtained rate constants of DDS of TH-DBT and HH-DBT were
0.13 and 0.10 mol/g min, respectively [32]. Sun and Prins sug-
gested that the removal of the sulfur atom from these molecules
over metal sulfides proceeds by a hydrogenolysis reaction through
a late transition state, determined by formation of a carbon–metal
bond, rather than through an early transition state, determined by
breaking of the carbon–sulfur bond [32]. In other words, the hydro-
genation weakens one of the C–S bonds in the DBT molecule, but
aryl–metal bonds are more stable than alkyl–metal bonds [32].
Thus, the fact that the rate constant of the desulfurization of
TH-4,6-DMDBT to DM-CHB is higher than that of HH-4,6-DMDBT
to DM-CHB would suggest that the desulfurization of
TH-4,6-DMDBT and HH-4,6-DMDBT over WP does not proceed
through elimination, but through hydrogenolysis.
Our results thus suggest that the methyl groups in 4,6-DMDBT,
TH-4,6-DMDBT, and HH-4,6-DMDBT have several effects. They not
only provide strong steric hindrance for the direct removal of the
sulfur atom and substantially accelerate the hydrogenation/
dehydrogenation of the compounds, but also affect the modes of
cleavage of the C–S bonds in HH-DBT and HH-4,6-DMDBT. The
steric hindrance and few b-hydrogen atoms could explain why
hydrogenolysis is preferred for the cleavage of the cycloalkyl C–S
bond in HH-4,6-DMDBT over WP. To fully understand the role
the methyl groups, further experimental and theoretical work
and better insight into the nature of the active site(s) of WP are
required.
the active site through the sulfur atom
(r-adsorption).
Hydrogenation of a phenyl ring increases the electron density on
the sulfur atom, and thus increases the interaction with the active
sites of the catalyst [37]. Thus, the DDS of HH-4,6-DMDBT was less
inhibited by piperidine than the DDS of TH-4,6-DMDBT and
4,6-DMDBT over WP. However, as shown in Scheme 1, the inhibi-
tion of piperidine was much stronger on the DDS of HH-DBT than
on the DDS of DBT or TH-DBT. This might be another indication
that the DDS of HH-DBT occurs through a different mechanism.
According to our previous work [10], the desulfurization of
HH-DBT over WP most likely proceeds by b-elimination. Two types
of active sites are required for b-elimination: a basic site to abstract
the b-hydrogen atom and an acid or vacancy site to bind the sulfur
atom of the reactant [38]. The acid or vacancy sites can be severely
poisoned by piperidine, because it is a strong base.
4.3. The inhibitory effect of piperidine
All steps in the networks of the HDS of DBT and 4,6-DMDBT
were strongly inhibited by piperidine (Schemes 1 and 2). Over
the bulk WP catalyst, the DDS and HYD pathways of DBT were
about equally inhibited by piperidine. The rate constants of these
two pathways were both about six times lower in the presence
than in the absence of piperidine. This is different from the cases
over metal sulfides [34–36] as well as over bulk Ni2P [10] and bulk
MoP catalysts [11]. Over these catalysts, the HYD pathway of DBT
was more strongly inhibited by nitrogen-containing compounds
than the DDS pathway. The reaction of 4,6-DMDBT was much more
strongly inhibited by piperidine than that of DBT. The conversion
Piperidine had a different effect on the dehydrogenation of
HH-DBT or HH-4,6-DMDBT and the hydrogenation of TH-DBT or
TH-4,6-DMDBT (Schemes 1 and 2), which was also observed over
Ni2P [12]. The equilibrium between the tetrahydro and hexahydro
intermediates should not be affected by an inhibitor. An explanation
could be that different amounts of the four HH-4,6-DMDBT isomers
(4,4a-trans;4a,9b-cis, 4,4a-trans;4a,9b-trans, 4,4a-cis;4a,9b-cis,
and 4,4a-cis;4a,9b-trans) were present in the reaction products,
which may complicate the thermodynamics [12]. Once we
have prepared large enough amounts of the pure isomers of
HH-4,6-DMDBT and HH-DBT, this question can be studied further.