Hydrodenitrogenation activities of methyl-substituted indoles

Article abstract of DOI:10.1006/jcat.1999.2685

The HDN reactivities of indole and several methyl-substituted indoles were obtained over a NiMo/Al₂O₃ catalyst and a CoMo/Al₂O₃ catalyst using a fixed-bed reactor at 340°C and 31 atm hydrogen pressure. A methyl

Full text of DOI:10.1006/jcat.1999.2685

Journal of Catalysis 189, 70–78 (2000)
Article ID jcat.1999.2685, available online at http://www.idealibrary.com on
Hydrodenitrogenation Activities of Methyl-Substituted Indoles
S. C. Kim and F. E. Massoth1
Department of Chemical and Fuels Engineering, University of Utah, Salt Lake City, Utah 84112
Received May 21, 1999; revised September 1, 1999; accepted September 13, 1999
group is located on a carbon atom adjacent to the sulfur
atom (1). Problems associated with these effects in achiev-
ing high HDS conversions have been recently addressed by
Whitehurst et al. (2).
The hydrodenitrogenation (HDN) reactivities of indole (IND)
and severalmethyl-substituted indoles(MIN)weredetermined over
a NiMo/Al2O3 catalyst and a CoMo/Al2O3 catalyst using a fixed-bed
reactor at 340 C and 31 atm hydrogen pressure. Appreciable dif-
ferences in reactivity were obtained, depending on the location of
the methyl group. Compared to IND, overall conversions and HDN
conversions were generally higher when methyl groups were on the
aromatic ring, and substantially lower for methyl groups on the N
ring. Two paths for initial conversion of IND (MIN) were deduced,
viz., a CNH path (C–N bond cleavage) and a HYD (aromatic hy-
drogenation) path. The NiMo catalyst favored the HYD path over
the CNH path, while the opposite occurred for the CoMo catalyst.
Reasonable trends were obtained between initial CNH rates and
the electrostatic potential on the N atom of the MINs, and between
c
A number of studies have been reported on the effect of
methyl groups on the hydrodesulfurization of model com-
pounds such as thiophene (3), benzothiophene (4, 5), and
dibenzothiphene (5–8). In these studies, it was found that
methyl groups on C atoms adjacent to the S atom resulted
in lower activity. It is believed that this is due to steric hin-
drance to adsorption through the S atom by the methyl
group.
Only limited studies have been reported on the effect
of alkyl substituents on hydrodenitrogenation (HDN) re-
activity of model N-heteroatom compounds. Cerny (9) re-
ported that the effect of methyl groups upon hydrogenoly-
sis of methyl-substituted pyridines followed in the series
pyridine > 2-methylpyridine > 2,6-dimethylpyridine >
2,4,6-trimethylpyridine, indicative of steric hindrance. Liaw
et al. (10) reported the order of reactivity as pyridine >
2-methylpyridine > 3-methylpyridine; the latter compound
should not exhibit steric hindrance if adsorption is through
the N atom. Also, Bhinde (11) reported that the HDN re-
activities of 2,6-, 2,7-, and 2,8-dimethylquinoline are com-
parable to that of quinoline. It is evident that steric effects
alone cannot explain the reactivity of six-membered het-
erocyclic compounds. The effect of alkyl substituents on
five-membered nitrogen compounds, such as indole, have
not been reported.
The HDN of heterocyclic nitrogen compounds generally
involves the following reactions: (1) hydrogenation of nitro-
gen heterorings, (2) hydrogenation of aromatic rings, and
(3) C–N bond cleavage (12). Several studies using indole
as a model compound for the HDN reaction have invoked
various reaction pathways involving a sequence of coupled
hydrogenation and hydrogenolysis steps (13–16). Recently,
Zhang and Ozkan (17) proposed a more complex network,
as shown in Scheme 1. A rapid hydrogenation of indole
(IND) to indoline (HIN) is followed by two different paths,
viz., direct hydrogenation of the aromatic ring of indoline
to form octahydroindoline (OHI), which proceeds rapidly
to another intermediate ethylcyclohexylamine (OECHA),
followed by C–N bond rupture to form the intermediate
initial HYD rates and the ionization potential.
2000 Academic Press
Key Words: indole; methyl-substituted indoles; hydrodenitro-
genation (HDN); NiMo catalyst; CoMo catalyst; computational
chemistry.
INTRODUCTION
Catalytic hydrotreating has become an important process
for removalofsulfur and nitrogen from petroleum due to in-
creasing environmental constraints. The heterocyclic com-
pounds of sulfur and nitrogen also have potential to cause
corrosion and poisoningofdownstream catalysts;therefore,
there is a need for hydrotreating. Heterocyclic compounds
containing sulfur and nitrogen are relatively stable struc-
tures because of their aromatic character. Even though the
hydrotreating process removes sulfur and nitrogen simulta-
neously, nitrogen compounds are more resistant than sulfur
compounds and require more severe reaction conditions.
Although the importance of hydrodenitrogenation is gain-
ing importance, less quantitative information is available on
the catalytic chemistry of hydrodenitrogenation than that
of hydrodesulfurization (HDS).
It is well-known that the difficulty in achieving high HDS
conversions with conventional feeds is due to certain “un-
reactive” sulfur compounds. These have been identified
as alkyl-substituted dibenzothiophenes in which the alkyl
1 To whom correspondence should be addressed.
70
0021-9517/00 $35.00
c
Copyright
2000 by Academic Press
All rights of reproduction in any form reserved.

HDN ACTIVITIES OF METHYL-SUBSTITUTED INDOLES
71
SCHEME 1. Suggested reaction pathways for the HDN of indole.
orthoethylaniline (OEA). The latter then reacts further varied proportionally so that a constant feed concentra-
by two paths, viz., direct denitrogenation leading to ethyl- tion was maintained throughout the run. After the cata-
benzene (EB), or aromatic ring hydrogenation to ethyl- lyst was aged for 2 days under reaction conditions, cata-
cyclohexylamine. The latter, forming from two paths, is lyst activityremained essentiallyconstant over 300h. Liquid
very reactive and rapidly forms ethylcyclohexene (ECHE) samples taken at various space times were analyzed by gas
and ethylcyclohexane (ECH). The difference between this chromatography (30-m 0.32-mm glass capillary column
scheme and previous ones is the proposal for reaction paths (Hewlett Packard, HP-1) containing dimethylpolysiloxane)
through the intermediates OHIN and OECHA, which were and use of a flame ionization detector and temperature pro-
detected under certain reaction conditions (18).
gramming of 10 C/min. The identity of individual products
The aim of the present study was to determine the effect was determined by comparison with pure reference sam-
of methyl substituents on the HDN of indole and attempt ples and GC/MS analysis. Molar GC factors were deter-
to explain the results. The numbering system for indole is mined using available samples of IND and its reaction prod-
shown below.
ucts. The same relative factors were assumed to apply to
the MINs.
The major compounds found in the samples from runs
with indole were indole (IND), o-ethylaniline (OEA), two
isomers of ethylcyclohexene (ECHE), ethylcyclohexane
(ECH), and ethyl benzene (EB); analogous compounds
containing a methyl group were found for runs with methyl
indoles (MIN). Small amounts of the dihydroindoles (HIN
or HMIN) were detected, and these were considered as
the reactants together with IND or MIN. Total conversion
(TOT) of IND or MIN was determined from the sum of
the mole fractions of products, while HDN conversion was
determined from the sum of the mole fractions of the non-
N-containing products. Space time, , was defined as the
weight of catalyst divided by the total gas flow (hydrogen
plus vaporized liquid feed).
Modeling of the various indoles by computational chem-
istry was carried out on a PC platform, while ab initio
calculations were done on an IBM-SP parallel supercom-
puter. The ground-state electronic properties of indole and
the methyl-substituted indoles and their partially hydro-
genated analogs were calculated using both Mopac (Chem-
3D) and Gaussian 98. At first, the semi-empirical PM3
method was employed for initial geometry optimization.
Subsequently, the geometryoptimization wasrepeated with
EXPERIMENTAL
The catalysts used were Topsøe TK-554, which consisted
of 4.1% CoO and 20.5% MoO3 supported on alumina
(220 m2/g), and Topsøe TK-555, which consisted of 3.8%
NiO, and 24% MoO3 supported on alumina containing
2% phosphorous (160 m2/g). The 1.27-mm extrudates were
crushed and sieved to 40- to 60-mesh particles. Reactions
were carried out in a fixed-bed reactor at 340 C and 35 atm
under vapor-phase conditions. Either a 0.25- or 0.1- g sam-
ple of catalyst, mixed with 5 cm3 of glass beads, was presul-
fided with a 10% H2S–90% H2 mixture by volume under
atmospheric pressure and 400 C for 2 h. The liquid feed
consisted of 0.5 wt% of indole (IND) or methyl-substituted
indole (MIN) and 1.0 wt% dimethyldisulfide, in n-heptane
solvent. The hydrogen flow rate and liquid flow rate were

72
KIM AND MASSOTH
Gaussian 98, using a Hartree-Fock SCF/6-31G(d) basis set.
Further details are provided elsewhere (19).
RESULTS
A. Reactivities of MINs
Reactions of IND and various MINs were carried out
with the NiMo catalyst at three space times using 0.25-g
catalyst. Product distribution data are given in Table 1, and
total (TOT) and HDN conversions of the various MINs
are plotted in Fig. 1 for one space time. It is evident that
the location of the methyl group has a significant effect on
reactivity. The order of reactivities is as follows:
TOT conversion: 6-MIN > 5-MIN 7-MIN IND
> 2-MIN 2,5-MIN > 3-MIN,
HDN conversion: 6-MIN > IND 5-MIN 7-MIN
2-MIN > 2,5-MIN > 3-MIN.
Thus, methyl groups on the aromatic ring slightly increase,
while those on the N ring decrease in TOT conversion com-
pared to IND. On the other hand, only the 6-MIN promotes
HDN activity, while the 5- and 7-MIN have no effect; again
methyls on the N ring decrease HDN. In the case of 2,5-
MIN, evidently the effect of the 2-position overrides that
of the 5-position. It is also evident from Fig. 1 that, for the
FIG. 1. Total (TOT) and HDN conversions of indole and methyl-
substituted indoles. Numbers refer to the carbon atom on which the methyl
group is attached.
TABLE 1
Product Distributions and Conversions of MINs for an NiMo
Catalyst (340 C and 0.25 g of Catalyst)
HDN of indole, steric hindrance due to methyl groups ad-
jacent to the N atom is not important in contrast to the rel-
atively large effects found in the HDS of sulfur compounds.
Changes in distribution of products for the various MINs
can give further insight into the reason for the differences
between TOT and HDN conversions. Figure 2 displays
product distributions in terms of hydrogenated hydrocar-
bon products, aromatic hydrocarbons, and intermediate
anilines. It is evident that the relatively low HDN responses
for the 5-, 6-, and 7-MINs, ascompared to their TOT conver-
sions, is reflected in their greater yield of the intermediate
aniline compound. That is, the aniline compound is formed
faster than it reacts. For methyls on the N ring, there is rel-
atively less anilines, and consequently their HDN conver-
sions are not greatly different from their TOT conversions.
Since the aromatic yields are small compared to the yield
of the hydrogenated products, the latter predominates the
HDN activity.
Conv.
a
Cpd.
IND
2-MIN
2,5-MIN
3-MIN
5-MIN
6-MIN
7-MIN
IND + HIN OEA ECH EB ECHE TOT HDN
0.74
1.11
1.48
0.593
0.735
0.749
0.815
0.568
0.467
0.582
0.122 0.129 0.015 0.141 0.407 0.285
0.009 0.147 0.007 0.102 0.265 0.256
0.050 0.072 0.021 0.108 0.251 0.201
0.031 0.072 0.026 0.056 0.185 0.154
0.169 0.056 0.016 0.192 0.432 0.264
0.197 0.076 0.024 0.237 0.533 0.337
0.160 0.075 0.018 0.165 0.418 0.258
IND
2-MIN
2,5-MIN
3-MIN
5-MIN
6-MIN
7-MIN
0.411
0.538
0.549
0.694
0.328
0.267
0.346
0.158 0.248 0.025 0.158 0.589 0.431
0.010 0.321 0.015 0.116 0.462 0.452
0.069 0.160 0.044 0.177 0.451 0.381
0.306 0.271
0.672 0.465
0.233 0.156 0.043 0.302 0.733 0.501
0.213 0.142 0.041 0.259 0.654 0.442
0.034 0.163 0.038 0.07
0.207 0.14 0.035 0.29
IND
2-MIN
2,5-MIN
3-MIN
5-MIN
6-MIN
7-MIN
0.276
0.375
0.420
0.615
0.172
0.144
0.255
0.17 0.371 0.034 0.149 0.724 0.554
0.008 0.515 0.024 0.077 0.625 0.616
0.078 0.239 0.046 0.216 0.580 0.501
0.034 0.230 0.044 0.078 0.385 0.352
0.189 0.233 0.053 0.353 0.828 0.639
0.233 0.229 0.061 0.334 0.856 0.624
0.227 0.183 0.049 0.286 0.745 0.518
Reactions of IND and various MINs were carried out
with the CoMo catalyst at several space times using 0.1 g of
catalyst. Product distribution data are listed in Table 2, and
total and HDN conversions of the various MINs are plotted
in Fig. 1 for one space time. The order of reactivities is as
a
in (kg min)/m3.

HDN ACTIVITIES OF METHYL-SUBSTITUTED INDOLES
73
TABLE 2
Product Distributions and Conversions of MINs for a CoMo
Catalyst (340 C and 0.1 g of Catalyst)
Conv.
a
Cpd.
IND
2-MIN
5-MIN
7-MIN
IND + HIN OEA ECH EB ECHE TOT HDN
0.277
0.875
0.961
0.816
0.837
0.077 0.019 0.001 0.029 0.125 0.049
0.005 0.021 0.000 0.014 0.039 0.035
0.133 0.010 0.002 0.040 0.184 0.052
0.128 0.008 0.003 0.023 0.163 0.034
0.341
0.444
0.634
IND
2-MIN
2,5-MIN
3-MIN
5-MIN
6-MIN
7-MIN
0.828
0.951
0.944
0.964
0.766
0.754
0.801
0.099 0.029 0.005 0.039 0.172 0.073
0.008 0.026 0.000 0.015 0.049 0.041
0.023 0.020 0.000 0.013 0.056 0.033
0.011 0.011 0.006 0.007 0.036 0.024
0.169 0.015 0.004 0.046 0.234 0.065
0.193 0.014 0.004 0.036 0.246 0.054
0.151 0.013 0.004 0.031 0.199 0.048
IND
2-MIN
2,5-MIN
3-MIN
5-MIN
6-MIN
7-MIN
0.780
0.915
0.891
0.925
0.694
0.661
0.735
0.125 0.046 0.003 0.046 0.220 0.095
0.008 0.046 0.001 0.030 0.085 0.077
0.036 0.042 0.001 0.030 0.109 0.073
0.03 0.025 0.005 0.015 0.075 0.045
0.213 0.024 0.007 0.061 0.306 0.092
0.248 0.025 0.007 0.059 0.339 0.091
0.188 0.019 0.006 0.053 0.265 0.078
IND
2-MIN
2,5-MIN
3-MIN
5-MIN
6-MIN
7-MIN
0.694
0.871
0.831
0.883
0.579
0.545
0.628
0.175 0.075 0.006 0.049 0.306 0.130
0.008 0.083 0.000 0.039 0.129 0.122
0.042 0.061 0.000 0.066 0.169 0.127
0.021 0.065 0.017 0.016 0.117 0.098
0.280 0.050 0.011 0.081 0.421 0.142
0.314 0.044 0.013 0.084 0.455 0.141
0.257 0.034 0.011 0.069 0.372 0.114
FIG. 2. Mole fractions of the reaction products vs the methyl posi-
tion in the indoles. Hydrogenated products, ECHE + ECH; intermediate
product, OEA; aromatic product, EB; also methyl analogs of these.
0.887 2,5-MIN
3-MIN
0.682
0.840
0.477
0.050 0.148 0.006 0.115 0.318 0.269
0.021 0.106 0.017 0.017 0.160 0.140
0.352 0.060 0.015 0.095 0.523 0.170
follows:
6-MIN
a
in (kg min)/m3.
TOT conversion: 6-MIN > 5-MIN > 7-MIN > IND
> 2,5-MIN > 2-MIN 3-MIN,
indicating that the hydrogenation of 2-MIN to 2-HMIN is
also not rate-limiting.
HDN conversion: 6-MIN 5-MIN > IND 7-MIN
2-MIN 2,5-MIN > 3-MIN.
B. Initial Rates
Again, methyl groups on the aromatic ring enhance, while
those on the N ring repress TOT conversion compared to
IND. In no case is HDN increased, although 5- and 6-MIN
are about the same as IND. Large differences are observed
between TOT conversion and HDN for methyl groups on
the aromatic ring, while only small differences are observed
on the N ring. Again, the major reason for the former lies
in the relatively high OEA yields, as seen in the product
distribution plot of Fig. 2.
Referring to Scheme 1, Scheme 2 was developed for the
purposes of analysis of initial rates. As the intermediates,
OHI and OECHA, were not detected under our reaction
conditions, they are assumed to react fast.
Reaction product data, obtained for each MIN at three
space times and one feed composition, are insufficient to
Table 3 presents data on the reactivity of HIN vs IND
under the same conditions. It is seen that there is no appre-
ciable difference in reactivity whether starting with IND or
with HIN. This indicates that the reaction between IND and
HIN is in equilibrium and the conversion of IND to HIN
is not a limiting step in the HDN of IND under our reac-
tion conditions. Also given in Table 3 are data for 2-MIN vs
2-HMIN. Again, no difference in conversions is observed,
SCHEME 2. Simplified scheme for kinetic analysis.

74
KIM AND MASSOTH
TABLE 3
Reactivity of HIN vs IND and 2-MIN vs 2-MHIN
IND vs HIN
Conversion
NiMo catalyst
Cpd
TOT
HDN
IND
HIN
IND
HIN
IND
HIN
0.444
0.444
0.341
0.341
0.222
0.222
0.257
0.249
0.198
0.206
0.122
0.126
0.173
0.172
0.133
0.140
0.078
0.084
FIG. 3. Mole fractions vs space time for indole HDN. x, TOT conver-
sion; B, OEA; C, ECHE + ECH (also methyl analogs).
2-MIN vs 2-MHIN
Conversion
CoMo catalyst
Cpd
TOT
HDN
The initial rates obtained for the various MINs are shown
in Fig. 4 for the NiMo catalyst, where the values are in
m3/(kg min). The initial HYD rate (rHYD)0 is faster than
the CNH rate (rCNH)0 for IND and methyl groups on the
N ring, while the opposite occurs for Me groups on the
aromatic ring. Compared to IND, the effect of the methyl
position on the N ring is to lower the initial rates of both
paths. However, for methyl groups on the aromatic ring, the
CNH path is enhanced, whereas the HYD path is slightly
decreased (5- and 7-positions), except for 6-MIN, which is
not much different from IND.
2-MIN
2-MHIN
2-MIN
2-MHIN
2-MIN
2-MHIN
0.634
0.634
0.444
0.444
0.277
0.277
0.129
0.126
0.085
0.083
0.050
0.063
0.121
0.118
0.077
0.075
0.041
0.055
perform a kinetic analysis on all reaction paths. However,
initialratesofconversion ofIND via the CNH path, (rCNH)0,
and the HYD path, (rHYD)0, can be estimated to determine
the effect of the methyl position on these two paths. Thus,
according to Scheme 2, as → 0,
(rCNH )0 = [d B/d ]0,
(rHYD )0 = [d P/d ]0,
(rTOT)0 = (rCNH )0 + (rHYD )0 = [dx/d ]0,
[1]
[2]
[3]
where (rTOT)0 is the total rate of indole disappearance, B
is the mole fraction of OEA, P is the mole fraction of
ECHE + ECH (the symbols also apply to the methylated
products), x is the total conversion of IND or MIN, and
is the space time. The rates in Eqs. [1]–[3] are in terms
of feed flows, i.e., m3/(kg min). Rates in terms of moles of
IND or MIN are proportional to these and are obtained
by multiplying by the constant feed concentration of 3.4
10 3 mol/m3.
Plots of B, P, and x were extrapolated to = 0 to obtain
initial slopes, both by taking the tangent of the curve and
by polynomial fitting. Although determining initial slopes
from the limited data available is subject to some error, both
methods were in reasonable agreement, with an estimated
relative error of 15% . As an example, the mole fraction vs
plots for IND are given in Fig. 3 for the CoMo catalyst.
It is significant that the initial slope of the P vs plot is
definitely not zero, confirming a direct route to P exists, i.e.,
the HYD pathway. If P were only generated from the CNH
pathway (via OEA), the initial slope should approach zero.
Similar results were obtained with the MINs.
FIG. 4. Initial CNH and HYD rates vs methyl position.

HDN ACTIVITIES OF METHYL-SUBSTITUTED INDOLES
75
The initial rates of the MINs for the CoMo catalyst are et al. (29) showed that both - and -adsorption modes are
displayed in Fig. 4. Now, it is seen that the CNH rates are possible for pyridine adsorbed on a model MoS cluster. In
faster than the HYD rates. For methyl groups on the aro- the case of indole, direct adsorption through the N atom
matic ring, the CNH rates are considerably enhanced, while may not be favorable because the extra pair of electrons is
the HYD rates are not, compared to IND. Again, both rates shared in the cloud of the N ring. A possible mode of ad-
are decreased when a methyl group is on the N ring.
sorption could be through simultaneous adsorption of the N
Catalyst activities cannot be compared directly because atom and the N ring on a dual site, leading to rapid hydro-
of different amounts of catalyst used. But initial rates can be genation of the C2–C3 double bond, freeing the electron
compared since the catalyst weight is accounted for in the pair for adsorption of the N atom on the CNH site.
space time. In a comparison of the effect of methyl groups
According to Scheme 1, the first step in conversion of
on the initial rates for the two catalysts, there are similarities IND is its hydrogenation to HIN. As demonstrated above,
in trends and differences in relative values. For both cata- conversion of IND to HIN is not rate-limiting for IND and
lysts, generally the CNH rate is enhanced for methyls on 2-MIN, and presumably for the other MINs. Subsequently,
the aromatic ring and depressed for methyls on the N ring. HIN reacts via the two pathways, viz., the CNH path and
On the other hand, the order in the initial rates of the CNH the HYD path. For convenience, the following discussion is
and HYD paths for the two catalysts are reversed. Thus, the in terms of IND, but is intended to apply also to the MINs.
NiMo catalyst exhibits higher HYD vs CNH rates (except
for 5-MIN), while the opposite pertains to the CoMo cata- The CNH Pathway
lyst. Second, compared to CoMo, the NiMo catalyst exhibits
Scheme 1 shows the steps involved in conversion of IND
higher HYD rates and lower CNH rates. This is in line with
to OEA via this pathway. The second step, path 1, involves
observations from many HDS studies (20), namely, that
C–N bond cleavage and is assumed to be the rate-limiting
CoMo catalysts are superior for HDS (essentially a C–S
step. The C–N cleavage reaction likely takes place by a
bond breaking step) and NiMo for HDN (predominantly
Hofmann-type elimination mechanism (30). This involves
a HYD step). This also explains the relatively lower OEA
proton transfer of an H + from a Bronsted site to the N
yield for the NiMo catalyst. Considering the effect of the
atom, followed by C–N cleavage and -elimination of an H
catalyst on the relative rates of the initial steps gives the
atom to form an olefin (vinylaniline), which is rapidly hy-
following:
drogenated to OEA. An alternate mechanism involving a
nucleophilic substitution by catalyst SH (after formation
Path
Reaction
Init, rate
of the charged N) leads to a similar result (12). Then, the
enhancement in reactivity by a methyl group on the aro-
matic ring could be due to an inductive effect. Thus, the
methyl group donates electrons to the aromatic ring, which
in turn makes the N atom more negative, increases the pro-
ton transfer rate prior to C–N cleavage, and consequently
increases the rate. For the methyl group on the N ring, once
HIN is formed and the C2–C3 double bond becomes a sin-
gle bond, methyls at the 2- or 3-position no longer exert an
inductive effect. The methyl at the 2-position could possibly
interfere with the transfer ofthe proton to the N atom. How-
ever, this fails to account for the low reactivity of 3-MIN.
For reaction via the CNH path, we propose adsorption
via the N atom. Then, we may expect, in a general way, that
reactivity should be related to the electrostatic potential
(EP) on the N atom. Politzer and Murray (31) discuss the
IND → OEA
IND → P
CNH
HYD
HYD
NiMo < CoMo
NiMo > CoMo
NiMo > CoMo
OEA → P
Consequently, the NiMo catalyst favors HYD over the
CoMo catalyst, producing less OEA by all three paths. Al-
though differences in catalyst formulations, e.g., loading,
presence, or absence of phosphorous, may affect catalyst
activities to some extent, it is believed that the remarkable
differences in CNH and HYD rates between the NiMo and
CoMo catalysts are predominantly due to the intrinsic na-
ture of the promoter (Ni vs Co).
DISCUSSION
There is evidence in the literature that two separate path- merits of using electrostatic potential for determining the
ways are involved in hydrotreating reactions of heterocyclic interaction of a molecule with its environment. This corre-
compounds, viz., one involving C–X (X is a heteroatom) lation is based on the assumption that the CNH reaction
bond cleavage and another hydrogenation (HYD) of the proceeds when electrons on the N atom are coordinated to
aromatic ring (13, 21–28). Many of these authors assumed the active site. According to Perot (12), the C2 atom must be
that the two pathways require different sites. We propose in sp3 hybridization (HMIN) for C–N cleavage. Therefore,
that IND is adsorbed via the N atom on a CNH site and EP calculations were made for the HMIN structures. A plot
adsorption of IND occurs via the aromatic system on a of initial CNH rates vs EP(HMIN) is shown in Fig. 5 for
HYD site, although separate sites are not necessary to the the two catalysts. Rough trends are observed in both cases.
following discussion. From theoretical calculations, Ruette The EPs do not vary much for all the HMINs, except for

76
KIM AND MASSOTH
to suppose that the RLS is the hydrogenation of the aro-
matic ring (path 2), as the resonance energy of the aromatic
system must be overcome.
The initial hydrogenation rate was somewhat lower for
5- and 7-MIN as compared to IND. It has been reported
(34) that o- and p-alkyl-substituted phenols showed lower
hydrogenation than phenol. It is noted that 7-MIN is ortho
and 5-MIN is para to the N atom in indole, suggesting that
a similar effect could be operative in the indole system.
Furthermore, 6-MIN is more active for HYD than 5- and
7-MIN. Since the methyl group in 6-MIN is meta to the N
atom, there appears to be a positive “meta effect”, as also
observed for the CNH path.
It has been reported (1) that the hydrogenation of the
C2–C3 double bond of methyl-substituted benzothiophenes
(MBT) is in the order BT > 2-MBT > 3-MBT, similar to our
results with the analogous MINs (Fig. 1). Ma et al. reported
(32) that the C2–C3 bond orders for the MBTs decrease in
the above order, suggesting that the reactivity is directly
related to the bond order of the C2–C3 bond. This implies
that hydrogenation of IND to HIN is rate-controlling. How-
ever, this was not found to be the case in our calculations;
bond orders of all MINs were found to be essentially the
same, and our results (Table 3) showed this conversion was
not rate-limiting under our conditions. Alternatively, once
HMIN is formed by a rapid hydrogenation of the C2–C3
double bond, the N ring becomes puckered out of plane of
the aromatic ring. Thus, part of the time, the approach of
FIG. 5. Initial CNH rates vs EP for HMIN structures.
2-HMIN (which also exhibits the lowest CNH activity);
thus, the effect of the methyl group on the N ring nearest
to the N atom has the greatest effect on the EP. Although
the correlation is reasonably good, the fit of the 3-MIN is
appreciably off the line, especially for the CoMo catalyst.
Actually, a better correlation was obtained using the EP
values for the MINs instead of the HMINs, as seen in Fig. 6.
This may be envisioned by supposing the adsorption mode
proposed earlier involves a concerted reaction, with simul-
taneous hydrogenation of the C2–C3 double bond and ad-
sorption through the N atom. Then, the controlling species
would be the activated state of the MIN, rather than the
HMIN.
In a study by Ma et al. (32), the desulfurization activity of
a variety of sulfur compounds was related to the electron
density of the HOMO on the sulfur atom, but the corre-
lation was not applicable to the alkyldibenzothiophenes.
Recently, Isoda et al. (33) reported a correlation between
the desulfurization rate and the electron density of the
HOMO-5 (molecular orbital five levels lower than the
HOMO), which represents S– bonding. Both HOMO and
HOMO-5, however, failed to give any correlation with the
MIN rate data.
The HYD Pathway
Scheme 1 presents the steps involved in the conversion
of IND to OHI and ultimately to ECHE. It is reasonable
FIG. 6. Initial CNH rates vs EP for MIN structures.

HDN ACTIVITIES OF METHYL-SUBSTITUTED INDOLES
77
constants for the different MINs could have a bearing on
their reactivity, inasmuch as this could lead to different con-
centrations of adsorbed MIN on the catalyst sites. Further
investigation into the energetics of adsorption and reaction
modes is clearly needed to shed more light on the role of
the catalyst on these reactions.
CONCLUSIONS
1. The effect of a methyl group attached to the aromatic
ringofindole isto increase the overallconversion compared
to indole, while the opposite occurs for a methyl group on
the N ring.
2. The initial conversion of indole and methyl indoles
occurs by two parallel paths, a CNH path leading to an
intermediate N compound and a HYD path leading to hy-
drogenated products.
3. The CNH path is enhanced by a methyl group on
the aromatic ring, whereas the HYD path is relatively un-
affected. Both paths are considerably lowered by methyl
groups on the N ring.
4. The NiMo catalyst is superior to the CoMo catalyst
for the HDN of indole by virtue of its relatively lower in-
termediate N-compound yield.
5. It may be difficult to achieve high HDN conversions
for real feeds because of the very low reactivity of methyl
groups on the N ring, a situation analogous to that of HDS
of real feeds, where certain sulfur-containing compounds
are resistant to HDS.
FIG. 7. Initial HYD rates vs IP for HMIN structures.
2- or 3-HMIN may be hindered from adsorption through
the aromatic ring, and the overall rate of hydrogenation
would be lowered. However, this fails to explain the lower
HYD rate for 3-MIN compared to that for 2-MIN.
For reaction via the HYD path, we propose adsorption
through the system of the aromatic ring. Thus, we would
expect that the reactivity should be related to the charac-
teristics of the complex. Nag (35) noted a correlation be-
tween hydrogenation activity of aromatic compounds and
the ionization potential (IP). Plots of initial HYD rates of
the MINs vs IP are given in Fig. 7 for the two catalysts. The
correlation for the MINs is reasonable; the 2- and 3-MIN
are in line. However, IND itself does not fit the correlation.
APPENDIX: NOMENCLATURE
B
CNH
EB
intermediate N compound
carbon–nitrogen hydrogenolysis
ethylbenzene
ECH
ethylcyclohexane
ECHE ethylcyclohexene
EP
HDN
HIN
electrostatic potential
hydrodenitrogenation
indoline (dihydroindole)
HMIN methyl-substituted dihydroindole
HYD
IND
IP
MIN
OEA
P
hydrogenation
indole
Final Comments
ionization potential
methyl-substituted indole
o-ethylaniline
hydrogenated products
total conversion of indole
space time
In the above attempts to rationalize methyl effects in
terms of chemical properties of the MINs, the role of the
catalyst, e.g., adsorption of the reactants and products, was
ignored. Especially the differences in reactivities between
NiMo and CoMo catalysts are not explicitly addressed
in these correlations. Preliminary kinetic analysis of IND
showed that IND is strongly adsorbed, while reaction prod-
ucts are relatively weakly adsorbed. Furthermore, analysis
showed that different values for the adsorption constant
of IND were obtained for the two pathways, indirect ev-
TOT
ACKNOWLEDGMENTS
Support of this research by Haldor Topsøe A/S is gratefully acknowl-
edged. We wish to thank Per Zeuthen for the GC/MS analyses. We also
thank Dr. Patrick Redington for helpful discussions on computational
idence for different sites. Therefore, different adsorption chemistry methods.

78
KIM AND MASSOTH
20. Topsøe, H., Clausen, B. S., and Massoth, F. E., “Catalysis-Science
REFERENCES
and Technology” (J. R. Anderson and M. Boudart, Eds.). Springer,
New York, 1996.
1. Ma, X., Sakanishi, K., and Mochida, I., Ind. Eng. Chem. Res. 35, 2487
(1996).
2. Whitehurst, D. D., Isoda, T., and Mochida, I., Adv. Catal. 42, 345
(1998).
3. Raje, A. P., Liaw, S. J., and Davis, B. H., Appl. Catal. A 150, 319
(1997).
4. Geneste, P., Bonnet, M., and Graffin, P., J. Catal. 61, 115 (1980).
5. Kwart, H., Schuit, G. C., and Gates, B. C., J. Catal. 61, 128 (1980).
6. Kabe, T., and Ishihara, A., Ind. Eng. Chem. Res. 31, 1577 (1992).
7. Isoda, T., Nagao, S., Karai, Y., and Mochida, I., Am. Chem. Soc., Div.
Pet. Chem. 212, 559 (1996).
8. Vanrysselberghe, V., and Froment, G. F., Ind. Eng. Chem. Res. 37, 4231
(1998).
9. Cerny, M., Colloid Czech. Chem. Commun. 44, 85 (1979).
10. Liaw, S.-J., Raje, A. P., Thomas, G. A., and Davis, B. H., Appl. Catal.
A 150, 343 (1997).
11. Bhinde, M. V., Ph.D. Dissertation, University of Delaware, 1979.
12. Perot, G., Catal. Today 10, 447 (1991).
13. Olive, J.-L., Biyoko, S., Moulinas, C., and Geneste, P., Appl. Catal. 19,
165 (1995).
14. Geneste, P., Moulinas, C., and Olive, J. L., J. Catal. 105, 254 (1987).
15. Massoth, F. E., Balusami, K., and Shabtai, J., J. Catal. 122, 256
(1990).
21. Bhinde, M. V., Shih, S., Zawadski, R., Katzer, J. R., and Kwart, H.,
“Proceedings Third International Conference on the Chemistry and
Uses of Molybdenum” (H. F. Barry and P. C. H. Mitchell, Eds.),
p. 184. Climax Molybdenum Co., Ann Arbor, MI, 1979.
22. Satterfield, C. N., and Gultekin, S., Ind. Eng. Chem. Process Des. Dev.
20, 62 (1981).
23. Odebunmi, E. O., and Ollis, D. F., J. Catal. 80, 56 (1983).
24. Nagai, M., and Kabe, T., J. Catal. 81, 440 (1983).
25. Gevert, B. S., Otterstedt, J.-E., and Massoth, F. E., Appl. Catal. 31, 119
(1987).
26. Moreau, C., Aubert C., Durand, R., Zmimita, N., and Geneste P., Catal.
Today 4, 117 (1988).
27. van Gestel, J., Leglise, J., and Duchet, J.-C., Appl. Catal. A 92, 143
(1992).
28. Vanryssel, V., and Froment, G. F., Ind. Eng. Chem. Res. 35, 3311
(1996).
29. Ruette, F., Poveda, F. M., Sierraalta, A., Sanchez, M., and Rodriguez-
Arias, E. N., J. Mol. Catal. 119, 335 (1997).
30. Nelson, N., and Levy, R. B., J. Catal. 58, 485 (1979).
31. Politzer, P., and Murray, J. S., in “Reviews in Computational Chem-
istry” (K. B. Lipkowitz and D. B. Boyd, Eds.) Vol. 2, Chap. 7. VCH
Publishers, New York, 1991.
32. Ma, X., Sakamishi, K., Isoda, T., and Mochida, I., Energy Fuels 9, 33
(1995).
16. Callant, M., Holder, K. A., Grange, P., and Delmon, B., Bull. Soc.
Chim. Belg. 104, 245 (1995).
33. Isoda, T., Takase, Y., Kusakabe, K., and Morooka, S., “Symposium on
Chemistry of Diesel Fuels,” 216th National Meeting, p. 575. American
Chemical Society, Washington, DC, 1998.
34. Aubert, C., Durand, R., Geneste, P., and Moreau, C., J. Catal. 112, 12
(1988).
17. Zhang, L., and Ozkan, U. S., Stud. Surf. Sci. Catal. 101, 1223 (1996).
18. Zhang, L., and Ozkan, U. S., in “Hydrotreatment and Hydrocracking
of Oil Fractions” (G. F. Froment, B. Delmon, and P. Grange, Eds.),
p. 69. Elsevier, New York, 1997.
19. Kim, S. C., Ph.D. Dissertation, University of Utah, Salt Lake City,
Utah, 1999.
35. Nag, N. K., Appl. Catal. 10, 53 (1984).