Mechanism of the hydrodenitrogenation of quinoline over NiMo(P)/Al2O3 catalysts

Article abstract of DOI:10.1006/jcat.1998.2181

The hydrodenitrogenation (HDN) of quinoline and its reaction intermediates was studied over NiMo(P)/Al₂O₃ catalysts. Phosphorus exhibited a promotional effect on the HDN of quinoline and ortho-propylaniline and a negative effect on t

Full text of DOI:10.1006/jcat.1998.2181

JOURNAL OF CATALYSIS 179, 18–27 (1998)
ARTICLE NO. CA982181
Mechanism of the Hydrodenitrogenation of Quinoline
over NiMo(P)/Al2O3 Catalysts
M. Jian and R. Prins
Laboratory for Technical Chemistry, Federal Institute of Technology (ETH), 8092 Zu¨rich, Switzerland
Received June 9, 1997; revised October 22, 1997; accepted October 29, 1997
sented in Fig. 1. Since the conversion of ortho-propylaniline
(OPA) is strongly inhibited by the presence of quinoline-
type molecules [quinoline, tetrahydroquinoline (THQ1
and THQ5), and decahydroquinoline (DHQ)] due to com-
petitive adsorption (6, 8, 10, 12), it has been concluded that
HDN of quinoline over NiMo/Al2O3 catalysts proceeds
exclusively via the fully saturated intermediates dec-
ahydroquinoline and 2-propylcyclohexylamine (PCHA)
(Q → DHQ → PCHA → hydrocarbons, see Fig. 1) (1–11).
A promotional effect of phosphorus on NiMo(P)/Al2O3
catalysts has been demonstrated in industrial HDN perfor-
mance and in studies with model compounds such as quino-
line (2, 13–16). It has been suggested that the positive ef-
fect of phosphorus originates from its accelerating effect on
C–N bond cleavage which is often supposed to be the most
difficult reaction step (2, 3, 17–19), although it has also been
concluded that C–N bond breaking and hydrogenation of
the aromaticringare about equallyslow(3, 7, 11). However,
from direct evidence of the C–N bond cleavage of piperi-
dine (20), it was concluded that phosphorus has a negative
influence on C–N bond cleavage over NiMo(P)/Al2O3 cata-
lysts, in contradiction to these suggestions in the literature.
To determine the role of phosphorus over NiMo(P)/
Al2O3 catalysts, we have investigated the HDN of quinoline
under various reaction conditions: in the presence as well as
absence of H2S, and at different temperatures. Important
reaction intermediates of the HDN network (DHQ and
OPA) were employed as reactants as well, to further study
the effect of phosphorus on the different HDN steps and
the reaction mechanisms. Simultaneous reactions of DHQ
and 1-methylcyclohexene (MCHE), of DHQ and ortho-
ethylaniline (OEA), and of Q and OEA were performed
to understand the product composition and inhibitive ad-
sorption effect in the HDN of quinoline.
The hydrodenitrogenation (HDN) of quinoline and its reaction
intermediates was studied over NiMo(P)/Al2O3 catalysts. Phospho-
rus exhibited a promotional effect on the HDN of quinoline and
ortho-propylaniline and a negative effect on the HDN of decahy-
droquinoline and on the hydrogenation of 1-methylcyclohexene in
its simultaneous reaction with quinoline. The HDN of quinoline
was found to take place via two pathways: via the saturated inter-
mediates decahydroquinoline and propylcyclohexylamine, and via
the aromatic intermediates 1,2,3,4-tetrahydroquinoline and ortho-
propylaniline. Although propylcyclohexane is the main HDN prod-
uct in both reaction paths, formation of propylbenzene is relatively
favored in the latter path. Thus, the relative contributions of the two
reaction paths in the quinoline HDN network could be estimated
from the HDN product composition. In contrast to previous sugges-
tions, the 1,2,3,4-tetrahydroquinoline–ortho-propylaniline reaction
path was found to be nonnegligible: About 40% of the quinoline
HDN takes place via this reaction path at 643 K, 3.0 MPa, and
PH S = 6.5 kPa, and about 35% at 623 K. H2S had an effect opposite
2
that of phosphorus on the HDN of decahydroquinoline and ortho-
propylaniline. H2S as well as phosphorus had a small positive effect
on the HDN of quinoline. This is due to a subtle balance between
their opposite effects on the HDN of decahydroquinoline as well as
ortho-propylaniline. Because of this, and of the opposite effect of
phosphorus as well as H2S on the HDNs of these two molecules, the
effect of phosphorus and H2S on the HDN of quinoline is a subtle
balance between the two HDN pathways, the catalyst composition,
c
and the reaction conditions.
ꢀ 1998 Academic Press
INTRODUCTION
Quinoline (Q) has many advantages over other nitrogen-
containing hydrocarbons as a model reactant in hydrode-
nitrogenation (HDN) (1, 2). Due to its molecular structure,
all reactions that take place in an industrial HDN process
also occur in the HDN network of quinoline, that is, C–N
bond cleavage, hydrogenation of aromatic heterocyclic
rings, and hydrogenation of benzenic rings. The reaction
mechanism of the HDN of quinoline has been studied by
EXPERIMENTAL
Catalysts used in this work had a composition of 3 wt%
several groups (1–8) and has been reviewed a few times nickel, 8 wt% molybdenum, and 0–2 wt% phosphorus. They
(9–11). It is generally accepted that the HDN reaction goes were prepared by incipient wetness impregnation, followed
through the intermediates indicated in the network pre- by drying at 393 K and calcination at 773 K (20). The
18
0021-9517/98 $25.00
c
Copyright ꢀ 1998 by Academic Press
All rights of reproduction in any form reserved.

MECHANISM OF HDN OF QUINOLINE OVER NiMo(P)/Al2O3
19
diates DHQ and PCHA: Thus, while the HDN activity of
quinoline increased, the activities for the hydrogenation
of alkenes, aromatics, and the aliphatic C–N bond cleav-
age decreased with the introduction of phosphorus into a
NiMo/Al2O3 catalyst (20, 21). This demonstrates that the
HDN ofquinoline cannot simplybyinterpreted byone rate-
limiting reaction step. Furthermore, since strong adsorption
of molecules is involved in the HDN process (16, 22), inhi-
bition effects due to the adsorption of the reactant, interme-
diates, or products must be taken into account in explaining
the behavior of the catalysts.
According to Satterfield and Gu¨ltekin (1), the presence
of H2S increases the rate of aliphatic C–N bond breaking
and decreases the rate of aromatic hydrogenation. Since
H2Spromoted the HDN ofquinoline (compare Figs. 2A and
2B), one could argue that the aliphatic C–N bond cleavage
limits this HDN reaction under the present conditions. This,
however, is in contradiction with the observation that phos-
phorus exhibits a negative influence on the aliphatic C–N
bond cleavage of piperidine (20). It suggests that either the
main HDN reaction of quinoline does not proceed through
the fully saturated amine intermediates (DHQ–PCHA) or
that the aliphatic C–N bond cleavage is not the only limiting
reaction.
Figures 1 and 3 show that there are two possible paths for
the HDN of quinoline: via DHQ–PCHA and via THQ1–
OPA. The concentration of the intermediate DHQ was
slightly lower, and that of OPA considerably higher, in the
presence than in the absence of H2S (Tables 1 and 2). The
former can be understood from the increased consumption
of DHQ through C–N bond cleavage to PCHA and hydro-
carbons in the presence of H2S, while the latter is due to the
fact that the transformation of THQ1 to OPA is enhanced
(an aliphatic C–N bond cleavage) and the further conver-
sion of OPA inhibited by the presence of H2S (see below).
The higher concentration of OPA thus indicates that HDN
via the THQ1–OPA path is limited by the transformation
of OPA in the presence of H2S.
FIG. 1. HDN reaction network of quinoline. Q, quinoline; THQ5,
5,6,7,8-tetrahydroquinoline; DHQ, decahydroquinoline; THQ1, 1,2,3,4-
tetrahydroquinoline; OPA, ortho-propylaniline; PCHA, 2-propylcyclo-
hexylamine; PCHE, propylcyclohexene; PCH, propylcyclohexane; PB,
propylbenzene.
reactions were carried out in a continuous-flow microreac-
tor. A sample of 0.5 g catalyst diluted with 9.5 g SiC was used
for each reaction. The catalyst was presulfided in situ with
a mixture of 10% (mol) H2S and H2 at 643 K and 1.5 MPa
for 4 h. The liquid reactant was fed to the reactor by means
of a high-pressure pump, with n-octane as the solvent. The
catalyst was stabilized at 643 K and 3.0 MPa for 18 h before
samples were taken. The initial reactant partial pressure
(P 0) of quinoline, DHQ, OPA, and OEA was 4.76 kPa in
all cases, and that of MCHE was 1.6 or 13.5 kPa. The partial
pressure of H2S was set at 0 or 6.5 kPa by adding dimethyl
disulfide (DMDS) to the liquid reactant. n-Nonane and
n-dodecane were used as internal standards.
RESULTS AND DISCUSSION
HDN of Quinoline
Tables 1 and 2 and Fig. 2 present the results of the HDN of
quinoline over the NiMo(P)/Al2O3 catalysts at 593, 623, and
643 K. The results demonstrate that in both the presence
and the absence of H2S, phosphorus promotes the HDN of
quinoline at all reaction temperatures.
This overall behavior of the HDN of quinoline is quite
different from that of each of the individual reaction steps
involved in the HDN path via the fully saturated interme-
Tables 1 and 2 also show that the hydrocarbon pro-
duct distribution was different for different catalysts and
TABLE 1
Results of the HDN of Quinoline in the Presence of H2S
Product composition (mol% )
T
(K)
PCH/PB
ratio
Catalyst
HC
PCHA
OPA
DHQ
THQ5
Q
THQ1
NiMo/Al2O3
NiMoP2/Al2O3
NiMo/Al2O3
NiMoP2/Al2O3
NiMo/Al2O3
NiMoP2/Al2O3
643
643
623
623
593
593
26.2
32.5
15.3
18.1
3.2
0
0
0
0
0.2
0.3
13.3
10.9
12.0
10.4
6.8
7.0
6.3
10.8
10.1
10.9
13.9
30.5
27.8
22.7
24.8
6.2
7.9
7.4
8.2
8.7
5.5
6.7
14.9
14.5
30.7
27.5
66.6
58.6
7.4
5.2
14
8.6
a
a
4.8
6.6
8.5
a PB was not detected.

20
JIAN AND PRINS
TABLE 2
Results of the HDN of Quinoline in the Absence of H2S
Product composition (mol% )
T
(K)
PCH/PB
ratio
Catalyst
HC
PCHA
OPA
DHQ
THQ5
Q
THQ1
NiMo/Al2O3
NiMoP2/Al2O3
NiMo/Al2O3
NiMoP2/Al2O3
NiMo/Al2O3
NiMoP2/Al2O3
643
643
623
623
593
593
9.2
20.8
4.7
9.0
0.7
0
0
0
0
0
0
3.6
2.3
2.4
2.3
1.0
0.9
9.7
12.3
14.1
12.3
14.3
13.9
40.8
35.8
24.8
22.5
7.1
12.4
10.6
11.7
10.5
7.0
22.8
19.9
41.3
42.3
69.6
70.3
3.2
2.8
3.6
2.4
a
a
1.7
6.5
6.5
a PB was not detected.
reaction conditions. Higher reaction temperatures, the compounds in the HDN product of DHQ also came close
presence of phosphorus, and the absence of H2S favored to those when quinoline was the reactant at higher temper-
the formation of PB, as shown by the lower PCH/PB ratio. atures. This would explain the positive effect of phosphorus
on the HDN of DHQ, because the fast conversions of DHQ
to Q, THQ1, and THQ5 at high temperatures caused this
reaction to resemble that of Q (Tables 3 and 1). The to-
HDN of Decahydroquinoline
DHQ isa keyintermediate in the HDN network ofquino-
line. Its molecular structure resembles that of piperidine
for which only C–N bond cleavage is needed to remove the
nitrogen atom. As in the HDN of piperidine (20), a nega-
tive effect of phosphorus prevailed in the HDN of DHQ at
593 K in the presence of H2S. However, the negative effect
became smaller at 623 K, and turned into a positive effect
at 643 K (Table 3, Fig. 4A).
The PCH/PB ratio in the HDN of DHQ decreased with
increasing temperature, and approached that of Q at the
highest reaction temperature. This suggests that the HDN
mechanism of Q and DHQ may be the same at higher reac-
tion temperatures, but different at lower temperatures. In
agreement with this, the relative yieldsofthe quinoline-type
tal HDN conversion of DHQ increased only slightly when
the reaction temperature was raised from 623 to 643 K,
much less than when the temperature was raised from 593
to 623 K. This is another indication that the HDN reaction
mechanism is different at high and low temperatures, or
that the rate-limiting reaction steps have changed.
The HDN conversion of DHQ was much lower in the ab-
sence of H2S (Table 4), and there was a strong promotional
effect of phosphorus at all reaction temperatures (Fig. 4B).
A comparison of the nitrogen-containing products shows
that the concentrations of THQ1 and Q are higher in the
absence than in the presence of H2S, due to the faster de-
hydrogenation. It indicates that in the absence of H2S, the
HDN of DHQ proceeds partly through the THQ1–OPA
FIG. 2. Influence of phosphorus on the HDN conversion of quinoline over NiMo(P)/Al2O3 in the presence (A) and absence (B) of H2S: ᭿, 643 K;
᭹, 623 K; ᭡, 593 K.

MECHANISM OF HDN OF QUINOLINE OVER NiMo(P)/Al2O3
21
FIG. 3. HDN reaction networks of decahydroquinoline (A) and ortho-propylaniline (B).
path. The dehydrogenation of DHQ to THQ1 is promoted ted. One mechanism functions mainly at low temperature
by the absence of H2S, and as a consequence, the DHQ net- and is favored by the presence of H2S, and the other func-
work as presented in Fig. 3A becomes coupled to that of tions at high temperature and is favored by the absence of
OPA (Fig. 3B). Furthermore, HDN via the DHQ–PCHA H2S. The latter might well be the HDN via the THQ1–OPA
path is slowed down in the absence of H2S. That a substan- path, since a high temperature and the absence of H2S fa-
tial part of the HDN proceeds via THQ1–OPA also agrees vor the dehydrogenation of DHQ to THQ1 and Q, which
with the fact that, although piperidine has a high C–N bond will be further converted to OPA and hydrocarbons. The
cleavage conversion in the presence and a very low one in two mechanisms are working in opposite directions with
the absence of H2S (20), the difference in the HDN conver- respect to phosphorus; in the first phosphorus has a nega-
sion of DHQ in the presence and absence of H2S is much tive effect and in the latter a positive effect. The positive
smaller, especially at high reaction temperatures (Tables 3 effect of phosphorus is associated with a low PCH/PB ratio,
and 4, Fig. 4).
and the negative one with a high PCH/PB ratio.
The PCH/PB ratiosin the HDN product were much lower
The inhibition effect alone cannot account for the func-
in the absence than in the presence of H2S. These results tion of phosphorus. Since DHQ and THQ5 form the ma-
resemble those of quinoline and again suggest that there jor part of the N-containing hydrocarbons in the products
might be a shift in the mechanism of the HDN of DHQ (Tables 3 and 4) and have the largest adsorption constants
when the temperature is increased or when H2S is omit- (22), one should expect that the major inhibition comes
TABLE 3
Results of the HDN of Decahydroquinoline in the Presence of H2S
Product composition (mol% )
T
(K)
PCH/PB
ratio
Catalyst
HC
PCHA
OPA
DHQ
THQ5
Q
THQ1
NiMo/Al2O3
NiMoP2/Al2O3
NiMo/Al2O3
NiMoP2/Al2O3
NiMo/Al2O3
NiMoP2/Al2O3
643
643
623
623
593
593
46.9
49.2
45.7
43.3
19.8
16.1
0
0
0.5
0.6
3.7
3.5
4.2
3.6
0.9
0.7
0
6.2
5.7
17.6
18.2
66.7
71.2
27.4
25.8
27.6
29.7
8.2
5.3
5.1
1.7
1.6
0
9.6
9.7
5.2
5.6
0.7
0.5
8.8
6.5
25
19
a
a
0
8.1
0
a PB was not detected.

22
JIAN AND PRINS
FIG. 4. Influence of phosphorus on the HDN conversion of decahydroquinoline over NiMo(P)/Al2O3 in the presence (A) and absence (B) of H2S:
᭿, 643 K; ᭹, 623 K; ᭡, 593 K.
from the adsorption of DHQ and THQ5. Tables 3 and 4 tures, in both the presence and the absence of H2S (Tables 5
show that at 623 K the sum of DHQ and THQ5 is much and 6).
higher in the absence than in the presence of H2S. Never-
The HDN product composition further supports the sug-
theless, phosphorus exhibits a positive effect in the former gestion about the role of the THQ1–OPA path. The lowest
and a negative effect in the latter case. On the other hand, a PCH/PB ratio is generally observed in the HDN of OPA,
positive effect of phosphorus prevails in both cases at 643 K the highest in the HDN of DHQ, while the PCH/PB ratio
where the sum of DHQ and THQ5 is still much higher in in the HDN of quinoline is inbetween (the PCH/PB ratio in
the absence of H2S.
the HDN ofpure OPA might be somewhat overestimated in
the absence ofH2Sdue to the hydrogenation ofPB to PCH).
Therefore, the HDN of quinoline might indeed be a com-
bination of HDN reactions via DHQ–PCHA and THQ1–
OPA. If the positive effect of phosphorus via the OPA path
can compensate for the negative effect via the DHQ path,
a promotional effect of phosphorus over NiMo(P)/Al2O3
catalysts will be observed, no matter which quinoline-type
compound is used as the reactant. The higher HDN conver-
sion of OPA in the absence of H2S demonstrates that the
relative low concentration of OPA in the HDN of DHQ and
Q is at least partly due to an increased transformation of
OPA to hydrocarbons under the corresponding conditions.
HDN of ortho-Propylaniline
OPA is the key intermediate in the HDN path via THQ1–
OPA (Figs. 1 and 3B). If, as suggested above, the promo-
tional effect of phosphorus in the HDN of Q and DHQ
arises from the reaction path via THQ1–OPA, especially
in the presence of H2S, phosphorus should have an accel-
erating effect on the conversion of OPA. To demonstrate
this suggestion, the HDN of pure OPA was studied under
the same conditions as that of Q and DHQ. Indeed, a pro-
motional effect of phosphorus was observed at all tempera-
TABLE 4
Results of the HDN of Decahydroquinoline in the Absence of H2S
Product composition (mol% )
T
(K)
PCH/PB
ratio
Catalyst
HC
PCHA
OPA
DHQ
THQ5
Q
THQ1
NiMo/Al2O3
NiMoP2/Al2O3
NiMo/Al2O3
NiMoP2/Al2O3
NiMo/Al2O3
NiMoP2/Al2O3
643
643
623
623
593
593
19.5
34.0
13.2
21.5
3.7
0
0
0
0
0
0
1.1
1.1
0
0
0
11.4
8.7
30.6
24.7
79.4
78.8
48.6
38.8
46.2
44.6
14.5
13.0
6.8
6.3
2.6
2.3
0.4
0.4
12.2
11.0
6.6
6.3
1.3
5.3
4.3
7.5
5.3
a
5.4
0
1.5
7.7
a PB was not detected.

MECHANISM OF HDN OF QUINOLINE OVER NiMo(P)/Al2O3
23
TABLE 5
TABLE 7
Results of the HDN of ortho-Propylaniline in the Presence of H2S
Hydrogenation of 1-Methylclohexene in the Presence of Quinoline
and H2S (P oMCHE = 1.6 kPa, P Qo = 4.76 kPa)
Product composition
(mol% )
Catalyst
T (K)
Conversion (% )
MCH/T
T
(K)
HDN
(% )
PCH/PB
ratio
Catalyst
PCH
PB
PCHE
NiMo/Al2O3
NiMoP2/Al2O3
NiMo/Al2O3
NiMoP2/Al2O3
NiMo/Al2O3
NiMoP2/Al2O3
643
643
623
623
593
593
94
85
79
64
51
36
11
9.0
25
NiMo/Al2O3
NiMoP2/Al2O3
NiMo/Al2O3
NiMoP2/Al2O3
NiMo/Al2O3
NiMoP2/Al2O3
643
643
623
623
593
593
84.1
99.7
64.3
90.9
23.5
34.6
68.6
76.5
53.8
75.3
17.8
23.7
14.9
21.5
6.1
12.2
0.9
1.4
0
3.8
2.7
4.8
9.1
4.8
3.6
8.8
6.1
19
21
a
a
1.8
13
a Toluene was not detected.
the ratios in the HDN of Q and OPA, while the PCH/PB
ratio in the HDN product of OPA was the furthest from the
MCH/T ratio.
Simultaneous Reaction of Quinoline
and 1-Methylcyclohexene
The HDN path via DHQ–PCHA (Fig. 3A) should result
in PCHE as the primary denitrogenation product which
undergoes further (de)hydrogenation to PCH and PB just
as MCHE. Therefore, the MCH/T ratio represents the ex-
treme case in which the aromatic product can be formed
only through dehydrogenation of the alkene intermediate.
On the other hand, the HDN path via THQ1–OPA (Fig. 3B)
enables the direct formation of PB by hydrogenolysis of
OPA, and accounts for the relatively high PB concentra-
tion in the product (24). It has been demonstrated that
PB can hardly be further hydrogenated in the presence of
quinoline-type compounds (21). Therefore, the hydrocar-
bon product composition can be used to estimate the rela-
tive contributions of the two HDN pathways in the quino-
line network. The high PCH/PB ratio in the case of DHQ
suggests that a relatively large part of the HDN reaction
takes place through PCHE.
Simultaneous hydrogenation of 1-methylcyclohexene
(MCHE) and HDN of quinoline were performed to investi-
gate the influence ofphosphoruson the elementaryreaction
steps and the HDN product composition (Table 7). MCHE
had no influence on the HDN of quinoline (the results are
not listed here), but quinoline strongly inhibited the con-
version of MCHE, from 100% in the absence of Q over all
catalysts to values below 94% (Table 7). An increase in the
initial partial pressure of MCHE by a factor of 8 had no
influence on the conversion of MCHE, indicating pseudo-
first-order hydrogenation behavior. Obviously, the catalyst
surface was almost completely covered by the quinoline-
type compounds and MCHE had little chance to compete
with quinoline on the surface.
Table 7 shows that phosphorus exhibited a negative effect
on the hydrogenation of MCHE over the NiMo(P)/Al2O3
catalysts, consistent with its negative effect on the hydro-
genation of cyclohexene in the simultaneous reaction with
piperidine (20). The MCH/T (MCH = methylcyclohexane,
T = toluene) ratios were higher than the PCH/PB ratios,
and were close to the equilibrium ratios of PCH/PB [11.5 at
643 K and 3.53 MPa and 24 at 623 K and 3.53 MPa (23)]. The
PCH/PB ratio in the HDN product of DHQ was closer to
the MCH/T ratio under the same reaction conditions than
Relative Contributions of the Two HDN Reaction Paths
The relative contributions of the two HDN pathways
were calculated under the assumptions that the reaction
path DHQ → PCHA → PCHE gives the same PCH/PB
ratio as MCH/T in the simultaneous reaction of Q and
MCHE (at 643 K: 11 for NiMo, 9.0 for NiMoP2; at 623
K: 25 for NiMo and 21 for NiMoP2), that the HC/PB ratio
from the reaction path THQ1 → OPA → HC is the same
as the HC/EB (EB = ethylbenzene) ratio of the HDN of
ortho-ethylaniline (OEA) in the simultaneous reaction of
OEA and Q, and that the amount of PCHE from each path
is proportional to the extent of that reaction path. In that
case one obtains
TABLE 6
Results of the HDN of ortho-Propylaniline in the Absence of H2S
Product composition
(mol% )
T
(K)
HDN
(% )
PCH/PB
ratio
Catalyst
PCH
PB
PCHE
NiMo/Al2O3
NiMoP2/Al2O3
NiMo/Al2O3
NiMoP2/Al2O3
NiMo/Al2O3
NiMoP2/Al2O3
643
643
623
623
593
593
90.5
100
76.5
96.0
37.2
52.8
66.2
72.1
61.7
75.3
28.6
40.3
21.8
26.0
12.6
19.2
5.0
1.0
0
1.8
1.0
3.7
5.0
3.0
2.8
4.9
3.9
5.7
5.3
HCDHQ + HCOPA = HCQ
PBDHQ + PBOPA = PBQ
PCHDHQ + PBDHQ + PCHEDHQ = HCDHQ
(PCH/PB)DHQ = MCH/T
[1]
[2]
[3]
[4]
7.6

24
JIAN AND PRINS
(PB/HC)OPA = (EB/HC)OEA
PCHEDHQ /PCHEQ = HCDHQ /HCQ
[5]
[6]
In the above equations, HCQ, PBQ, and PCHEQ are the
concentrations of total hydrocarbons (HC), PB, and PCHE
in the HDN product of quinoline, respectively. PCHDHQ
PBDHQ, and PCHEDHQ are the contributions of PCH, PB,
and PCHE in the HDN product of quinoline via the DHQ–
PCHA path, and HCOPA and PBOPA are the contributions of
HC and PB via the THQ1–OPA path, respectively. HCOEA
and EBOEA are the HC and EB in the HDN product of
OEA in the simultaneous reaction of OEA and Q under
the corresponding reaction conditions.
,
In this way, the contributions from the DHQ–PCHA
(path 1) and the THQ1–OPA (path 2) reaction paths in
the HDN network of quinoline can be calculated (Table 8).
The results show that under the present reaction conditions
about 40% of the quinoline HDN proceeds via the THQ1–
OPA path at 643 K, and 35% at 623 K. They confirm the
qualitative conclusions obtained above that both reaction
paths are important in the HDN of quinoline.
The (HC)OEA/EB ratio from the HDN of OEA in the
simultaneous reaction of OEA and Q is independent of
space time, while the (HC)OPA/PB ratio from the HDN of
pure OPA decreases with increasing space time (Fig. 5),
suggesting that the relative yield of the aromatic product is
influenced by an inhibition effect. The inhibition effect on
the HDN of OEA is so strong in the simultaneous reaction
of OEA and Q that a decrease in inhibitor concentration
by an increase in the space time has no influence on the
different HDN steps of OEA. On the other hand, in the
HDN of pure OPA the inhibition effect varies as the reac-
tion proceeds, because OPA is much more weakly adsorbed
than quinoline-type molecules. These results show that the
above method of calculating the relative contributions of
the two reaction paths can be used only when the conver-
sion of quinoline is not too high, when most of the surface
is still covered by strongly adsorbed species, otherwise dif-
ferent (HC)OPA/PB values have to be taken into account.
FIG. 5. (HC)OEA/EB ratio in the simultaneous HDN of OEA and Q
and the (HC)OPA/PB ratio in the HDN of pure OPA as a function of space
time at 643 K, 3.0 MPa, in the presence of H2S.
Simultaneous HDN of Q and OEA and of DHQ and
OEA was carried out to check the validity of the above cal-
culations. Table 9 shows that the HDN conversion of OEA
accounts for 30–40% of the total hydrocarbon production
in the simultaneous HDN of Q and OEA. The simultaneous
HDN of DHQ and OEA also demonstrates that 25–30%
of the total hydrocarbon product originates from OEA and
70–75% from DHQ at 643 K (Table 10). At lower reaction
temperature, however, the HDN contribution from OEA
was considerably lower. This is due to a stronger inhibition,
as the concentration of the inhibitors differed substantially
from that when starting from Q + OEA (cf. Tables 3 and 1).
The HDN conversion of OPA is much higher than that
of DHQ and Q, proving that pure OPA is much more reac-
tive (3, 25). Comparing Tables 9 and 10 with Table 5 shows
that the HDN of OEA is strongly inhibited by the pres-
ence of quinoline-type compounds. The stronger tempera-
ture dependence of the OEA conversion and the smaller
effect of phosphorus on the HDN of OEA in the simulta-
neous reaction compared to when it reacted alone indicate
TABLE 8
Relative Contributions of the DHQ–PCHA Path (Path 1) and the THQ1–OPA Path (Path 2) in the
HDN of Quinoline at 3.0 MPa and PH S = 6.5 kPa
2
Experimental result
Calculated result
Contribution (% )
Product composition (mol% )
T
(K)
HDN
(% )
Catalyst
PCH
PB
PCHE
Path 1
Path 2
NiMo/Al2O3
NiMoP2/Al2O3
NiMo/Al2O3
643
643
623
623
26.2
32.5
15.3
18.1
19.9
22.1
10.7
10.8
2.5
4.0
0.7
1.2
3.8
6.4
3.9
6.1
62
59
67
61
38
41
33
39
NiMoP2/Al2O3

MECHANISM OF HDN OF QUINOLINE OVER NiMo(P)/Al2O3
25
TABLE 9
Product Compositions in the Simultaneous HDN of Q and OEA in the Presence of H2S (P oOEA = P Qo = 4.76 kPa)
From OEA
HC (mol% )
From Q (mol% )
DHQ
T
(K)
Catalyst
HC/EB
HC
PCHA
OPA
THQ5
Q
THQ1
NiMo/Al2O3
NiMoP2/Al2O3
NiMo/Al2O3
643
643
623
623
15.7
17.5
8.9
7.5
6.0
12.7
8.3
20.6
28.0
12.8
16.8
0
0
0
0
12.4
10.4
10.8
8.7
8.4
9.5
14.7
19.1
30.6
27.8
21.3
20.3
9.5
8.0
9.2
8.1
17.6
15.2
30.9
25.7
NiMoP2/Al2O3
9.5
that it is mainly the inhibitive effect but not the intrinsic peratures more dehydrogenation of DHQ to THQ1 (and
kinetics that controls the HDN of OEA in the simultane- of THQ5 to Q) takes place. As a result of this coupling of
ous reactions. The negative effect of phosphorus on the the two networks shown in Figs. 3A and 3B, less HDN is
HDN of DHQ results in a higher concentration of uncon- possible via the quicker DHQ–PCHA path. In addition, de-
verted DHQ (THQ5) and, thus, in a stronger inhibition of pletion of DHQ due to its transformation to THQ5, which is
the HDN of OEA, which more than compensates for its thermodynamically favored at high temperature, is another
promotional effect. At the same time, OEA exhibits only limiting factor.
a weak inhibition effect (about 10% less total conversion)
on the HDN of DHQ (cf. Tables 3 and 10).
In the absence of H2S, the contribution from the DHQ–
PCHA path to HC production in the HDN of quinoline is
suppressed and that of the THQ1–OPA path may hardly be
changed, since decreased transformation of THQ1 to OPA
can be compensated byan increased transformation ofOPA
to HC. As a result, a low HDN conversion is obtained. The
negligible effect of phosphorus on the HDN of piperidine
in the absence of H2S (20) suggests that phosphorus should
have a small effect on the DHQ–PCHA path in the absence
of H2S as well. Thus, the positive effect of the THQ1–OPA
path in the HDN of Q will more than compensate for a
small negative effect of the DHQ–PCHA path, and a strong
promotional effect of phosphorus is observed.
The influence of phosphorus on the HDN of OPA is op-
posite that of H2S, in both the HDN conversion and the
PCH/PB ratio. Since the catalyst surface will be sulfur defi-
cient in the absence of H2S, this suggests that the introduc-
tion of phosphorus into a NiMo/Al2O3 catalyst may lead
to a more reduced environment of the catalytic sites, to
higher coordinative unsaturation, or to more sulfur vacan-
cies in the NiMoS phase. Phosphorus and H2S also exhibit
Influence of Phosphorus and H2S
The influence of phosphorus and H2S on the HDN con-
version and PCH/PB product ratio in the HDN of Q, DHQ,
and OPA is presented in Table 11. Phosphorus has a pro-
motional effect on the THQ1–OPA path and an inhibitive
effect on the DHQ–PCHA path in the presence of H2S. The
beneficial effect of H2S on the overall HDN conversion of
quinoline demonstrates that the majority of the HDN prod-
uct was formed via the DHQ–PCHA path. The higher for-
mation of HC via the DHQ–PCHA path in the presence
of H2S compensates for the lower formation of HC via the
THQ1–OPA path. At the same time, the promotional ef-
fect of phosphorus on the THQ1–OPA path compensates
for its negative effect on the DHQ–PCHA path. The result
is a higher total HDN conversion and a small positive effect
of phosphorus in the presence of H2S.
The small change in HDN conversion of DHQ between
623 and 643 K is then due to the fact that at the higher tem-
TABLE 10
Product Compositions in the Simultaneous HDN of DHQ and OEA in the Presence of H2S (P oOEA = P Do HQ = 4.76 kPa)
From OEA
HC (mol% )
From DHQ (mol% )
DHQ
T
(K)
Catalyst
HC/EB
HC
PCHA
OPA
THQ5
Q
THQ1
NiMo/Al2O3
NiMoP2/Al2O3
NiMo/Al2O3
NiMoP2/Al2O3
NiMo/Al2O3
NiMoP2/Al2O3
643
643
623
623
593
593
15.9
19.2
7.4
6.8
1.9
7.8
6.0
12
41.1
44.1
41.7
37.9
18.7
15.0
0
0
0.5
0.5
3.1
2.7
3.6
3.3
0.9
0.7
0
9.3
10.0
18.8
22.1
66.8
64.3
28.6
25.4
27.9
28.9
8.4
5.5
5.2
1.6
1.5
0
9.6
9.4
5.2
5.2
0.8
1.7
9.7
a
a
1.7
0
11.7
0.9
a EB was not detected.

26
JIAN AND PRINS
TABLE 11
Influence of Phosphorus and H2S on the Hydrocarbon Production and PCH/PB Product Ratio
in the HDN of Q, DHQ, and OPA
Hydrocarbon production
Effect of P Effect of H2S
No H2S With P No P
PCH/PB ratio
Effect of H2S
Effect of P
No H2S
H2S
H2S
With P
No P
Q
DHQ
OPA
+
ꢁ
++
+
+
+
ꢁ
+
++
ꢁ
ꢁ
ꢁ
ꢁ
ꢁ
+
+
+
+
+
+
ꢁ
ꢁ
++
++
opposite effects on the PCH/PB ratio in the HDN of DHQ the quinoline HDN takes place via the THQ1–OPA path,
and Q, which can be explained by the fact that the yield of while the proportion is 35% at 623 K.
PB is low in all cases and comes mainly from the reaction
Introduction of phosphorus into a sulfided NiMo/Al2O3
path via OPA. Phosphorus and H2S do not have an oppo- catalyst enhances the metallic properties of the catalyst,
site effect on the HDN conversion of Q. This is explained just like omitting H2S from the reactant. The HDN results
by the fact that although phosphorus and H2S have oppo- of quinoline and its reaction intermediates indicate that the
site effects on the HDN of OPA and DHQ (more or less), properties and numbers of the catalytic sites are altered by
the influence of phosphorus as well as H2S on the HDN of phosphorus and reaction conditions. To fully understand
DHQ is (more or less) opposite that on the HDN of OPA. the origin of the promotional effect of phosphorus and the
As a consequence, the influence of phosphorus and H2S is a HDN mechanism of quinoline, quantitative kinetic studies
subtle balance between the two HDN paths (DHQ–PCHA are necessary. These are described in another paper (27).
and THQ1–OPA), catalyst composition, and reaction
REFERENCES
conditions.
In a study of the HDN of quinoline over a Mo2N cata-
lyst, Lee et al. (26) reported that HDN took place exclu-
sively via the aromatic intermediates THQ1–OPA and that
onlyaromatichydrocarbonswere present in the HDN prod-
uct (benzene, toluene, ethylbenzene, and PB), suggesting
that only aromatic C–N bond cleavage occurred. These ex-
periments over Mo2N represent an extreme case, however,
because the sulfur-free Mo2N catalyst behaves more like
a metal than a sulfide, and the low pressure favors aro-
matic products. It does, however, indicate how strongly the
catalyst properties can influence the HDN reaction net-
work.
1. Satterfield, C. N., and Gu¨ltekin, S., Ind. Eng. Chem. Process Des. Dev.
20, 62 (1981).
2. Eijsbouts, S., van Gestel, J. N. M., van Veen, J. A. R., de Beer, V. H. J.,
and Prins, R., J. Catal. 131, 412 (1991).
3. Satterfield, C. N., and Yang, S. H., Ind. Eng. Chem. Process Des. Dev.
23, 11 (1984).
4. Yang, S. H., and Satterfield, C. N., Ind. Eng. Chem. Process Des. Dev.
23, 20 (1984).
5. Gioia, F., and Lee, V., Ind. Eng. Chem. Process Des. Dev. 25, 918
(1986).
6. Perot, G., Brunet, S., Canaff, C., and Toulhoat, H., Bull. Soc. Chim.
Belg. 96, 865 (1987).
7. Rico Cerda, J. L., and Prins, R., Bull. Soc. Chim. Belg. 100, 815
(1991).
8. Perot, G., Catal. Today 10, 447 (1991).
9. Schulz, H., Schon, M., and Rahman, N. M., Stud. Surf. Sci. Catal. 27,
201 (1986).
CONCLUSIONS
10. Ho, T. C., Catal. Rev. Sci. Eng. 30, 117 (1988).
11. Girgis, M. J., and Gates, B. C., Ind. Eng. Chem. Res. 30, 2021
(1991).
12. Gallant, M., Holder, K. A., Grange, P., and Delmon, B., Stud. Surf. Sci.
Catal. 100, 199 (1996).
13. Michelson, G. A., U.S. Patents 3,749,663–4; 3,755,148; 3,755,150; and
3,755,196 (1973).
14. Morales, A. L., Galiasso, R. E., Agudelo, M. M., Salazar, J. A., and
Carrasquel, A. L., U.S. Patent 4,520,128 (1985).
15. Tischer, R. E., Narain, N. K., Stiegel, G. J., and Cillo, D. L., Ind. Eng.
Chem. Res. 26, 422 (1987).
Phosphorus has a promotional effect on the HDN of
quinoline over NiMo(P)/Al2O3 catalysts, even though it has
a negative effect on aliphatic C–N bond cleavage. Two reac-
tion paths are responsible for this HDN reaction: The reac-
tion path via DHQ–PCHA is promoted by H2S and inhib-
ited by phosphorus, while the reaction path via THQ1-OPA
is favored by the absence of H2S and presence of phospho-
rus. The interpretation of the influence of reaction condi-
tions and catalysts should thus take both paths into account.
The former path results in a low PCH/PB ratio and the lat-
ter in a high PCH/PB ratio in the HDN products. At 643 K,
3.0 MPa, and H2S/H2 = 2.2 ꢂ 10ꢁ3 (mol/mol) about 40% of
16. Van Veen, J. A. R., Colijn, H. A., Hendriks, P. A. J. M., and van
Welsenes, A. J., Fuel Proc. Technol. 35, 137 (1993).
17. Fitz, C. W., and Rase, H. F., Ind. Eng. Chem. Prod. Res. Dev. 22, 40
(1983).

MECHANISM OF HDN OF QUINOLINE OVER NiMo(P)/Al2O3
27
18. Lewis, J. M., Kydd, R. A., Boorman, P. M., and van Rhyn, P. H., Appl.
Catal. 84, 103 (1992).
23. Cocchetto, J. F., and Satterfield, C. N., Ind. Eng. Chem. Process Des.
Dev. 20, 49 (1981).
19. Rajagopal, S., Grimm, T. L., Collins, D. J., and Miranda, R., J. Catal. 24. Jian, M., and Prins, R., Catal. Today 30, 127 (1996).
137, 453 (1992).
25. Eijsbouts, S., de Beer, V. H. J., and Prins, R., J. Catal. 127, 618
20. Jian, M., Rico Cerda, J. L., and Prins, R., Bull. Soc. Chem. Belg. 104,
225 (1995).
(1991).
26. Lee, K. S., Abe, H., Reimer, J. A., and Bell, A. T., J. Catal. 139, 34
21. Jian, M., Kapteijn, F., and Prins, R., J. Catal. 168, 491 (1997).
22. La Vopa, V., and Satterfield, C. N., J. Catal. 110, 375 (1988).
(1993).
27. Jian, M., and Prins, R., to be published.