VOLUME 18, NUMBER 5 SEPTEMBER/OCTOBER 2004

© Copyright 2004 American Chemical Society

Articles
Search for an Efficient 4,6-DMDBT Hydrodesulfurization
Catalyst: A Review of Recent Studies
Shyamal K. Bej,*,† Samir K. Maity,‡ and Uday T. Turaga§
Department of Chemical Engineering, University of Michigan, 3230 H.H. Dow Building,
2300 Hayward Avenue, Ann Arbor, Michigan 48109-2136, Instituto Mexicano del Petroleo,
Eje Central Lazaro Cardenas 152, Mexico City, DF 07730, Mexico, and Fuel Science Program,
Department of Energy and Geo-Environmental Engineering,
The Pennsylvania State University, University Park, Pennsylvania 16802

Received November 19, 2003. Revised Manuscript Received May 11, 2004

Almost complete removal of 4,6-dimethyl dibenzothiophene (4,6-DMDBT) will perhaps be

inevitable for reducing the sulfur content of diesel to a level of 50 wppm and lower. The
hydrodesulfurization (HDS) of 4,6-DMDBT does not tend to occur through direct desulfurization,
a pathway typically followed by reactive sulfur compounds over conventional CoMo/Al2O3 catalysts.
Its reactivity can be enhanced either by increasing the rate of direct desulfurization or by
transforming it to a more activated molecule through hydrogenation, isomerization, demethylation,
and C-C bond scission. Attempts have been made to develop better catalysts using these concepts.
Different additives such as phosphorus, fluorine, and lanthanum have been added to the alumina
support for developing the required catalytic properties. Various other supports such as zeolite,
zirconia, titania, etc., by themselves or in admixtures with alumina have also been used to improve
the HDS activities of the catalysts. This article reviews the results of recent studies conducted
in this area and summarizes the advances that have taken place in this direction.

Introduction will also be quite similar.4 Sulfur present in diesel is

Stringent environmental regulations are exerting usually removed through hydrodesulfurization (HDS)
pressure to reduce the maximum allowable sulfur processes using alumina-supported CoMo- or NiMo-
content in diesel. In most advanced countries, the based catalysts. Using these currently operating HDS
allowable limit on diesel sulfur will be restricted to 50 technologies, the sulfur content in diesel is typically
wppm within another 5 years.1,2 For example, Europe reduced to a level of 300-500 wppm.
will tighten its diesel sulfur content to less than 50 The molecules, which need to be removed for bringing
wppm by 2005.3 The situation in the USA and Japan down the sulfur content from existing levels of 300-
500 wppm to a future level of <50 wppm, are mainly
* Corresponding author: Tel: 734-763-5748. Fax: 734-763-0459. the refractory dialkyl dibenzothiophenes.5-7 The posi-
E-mail: shyamal@engin.umich.edu.
† University of Michigan.
‡ Instituto Mexicano del Petroleo. (3) Bihan, L. L.; Yoshimura, Y. Fuel 2002, 81, 491.
§ The Pennsylvania State University. (4) Turaga, U. T.; Ma, X.; Song, C. Catal. Today 2003, 86 (1-4),
(1) Shin, S.; Yang, H.; Sakanishi, K.; Mochida, I.; Grudoski, D. A.; 265.
Shinn, J. H. Appl. Catal. A: General 2001, 205, 101. (5) Segawa, K.; Takahashi, K.; Satoh, S. Catal. Today 2000, 63, 123.
(2) Yoshimura, Y.; Yasuda, H.; Sato T.; Kijima, N.; Kameoka, T. (6) Meille, V.; Schulz, E.; Lemaire, M.; Vrinat, M. Appl. Catal. A:
Appl. Catal. A: General 2001, 207, 303. General 1999, 187, 179.

10.1021/ef030179+ CCC: $27.50 © 2004 American Chemical Society

Published on Web 07/24/2004
1228 Energy & Fuels, Vol. 18, No. 5, 2004 Bej et al.

tions of the alkyl groups in these dibenzothiophene efforts that have been made to develop better catalysts
derivatives play an important role in controlling reac- for increasing the reactivity of 4,6-DMDBT. The article
tivities of these molecules. For the dimethyl series, the will focus on only those studies which have been
ease of desulfurization over a CoMo/Al2O3 catalyst is conducted using 4,6-DMDBT as a model compound
reported to have the order8,9 under industrial conditions of pressure and tempera-
ture.
2,8-DMDBT > 3,7-DMDBT > 4,6-DMDBT
Approaches for Developing Better Catalysts
where 4,6-DMDBT has been found to exhibit the lowest
rate. The electronic effects of the alkyl groups are known The important routes through which the reactivity of
to be responsible for the relatively higher activity of 2,8- 4,6-DMDBT can be enhanced are shown in Figure 1.
DMDBT. The primary reason for the poor reactivity of Except for direct desulfurization, all other pathways
4,6-DMDBT has been attributed to the steric hindrance mainly center on removing the steric hindrance of the
of the methyl groups, which makes the sulfur atom methyl groups present at 4 and 6 positions. The first
inaccessible to the active sites of the catalyst.10-14 As a one, which has received considerable attention, is the
result, the HDS of 4,6-DMDBT does not tend to follow hydrogenation of one of the phenyl rings. The hydro-
the direct desulfurization route (shown below), typical genation of a phenyl ring imparts flexibility to the
methyl group resulting in the reduction of steric hin-
drance.14,21,22 This is represented in Figure 2 which
compares the flexibility of 4,6-DMDBT and its partially
hydrogenated derivative for approach to the active site
of the catalyst in head-on and side-on fashions.
Another way of reducing the steric hindrance of the
methyl groups is to shift these groups from 4,6 to 3,7 or
of other reactive sulfur compounds over CoMo/Al2O3 to 2,8 positions through an isomerization reaction.23-25
catalysts.15 This has, therefore, spurred intense interest The complete removal of one or both methyl groups
among researchers to explore alternative catalytic path- through a dealkylation reaction offers another possibil-
ways for activating this molecule. In addition, there has ity.21,21 The scission of the single C-C bond in the
been considerable interest in developing improved pro- thiophenic ring has also been attempted by various
cess configurations and novel reactors to facilitate the researchers.21 For this discussion, we will refer to these
efficient removal of these sulfur compounds. The pos- (isomerization, dealkyalation, and C-C bond scission)
sibilities as well as the challenges associated with these reactions as non-hydrogenative routes for desulfuriza-
engineering options for producing ultralow sulfur diesel tion.
have been recently reviewed.16,17 Ongoing research is The properties required in a catalyst for directing the
also focused on other non-HDS technologies for the HDS reaction through the above-mentioned pathways
efficient removal of these refractory sulfur molecules. can be grouped into two classes. The saturation of one
These include desulfurization via adsorption, desul- of the phenyl rings depends primarily on the hydro-
furization via extraction, oxidative desulfurization, and genation capability of the catalyst which can be en-
desulfurization via precipitation, etc. Recent advances hanced either by incorporating a suitable metal such
in these areas have been reviewed by Babich and as Ni, W, Pt, Pd, Ru, etc., and/or by providing a suitable
Moulijn18 and Song.19 Among these, desulfurization by support. All other routes such as isomerization, dealkyl-
adsorption deserves special mention and is in a very ation, and C-C bond scission depend mainly on the
advanced stage of development. acidic property of the catalyst.21 Support plays an
Advances up to the late 1990’s in HDS catalysis and important role in controlling the acidic property. Re-
other related aspects have been reviewed in depth by searchers have tried to use various mixed oxide supports
Whitehurst et al.20 In this article, we will review recent from a combination of Al2O3, TiO2, ZrO2, etc., to achieve
this goal.27 Besides the mixed oxide supports, other
(7) Schulz, H.; Bohringer, W.; Waller, P.; Ousmanov, F. Catal. Today acidic materials such as different types of zeolites and
1999, 49, 87. amorphous silica alumina (ASA) have also received
(8) Kilanowski, D. R.; Teeuwen, H.; Gates, B. C.; Beer, V. H. J. D.; considerable attention.28,29 For the purpose of discus-
Schuit, G. C. A.; Kwart, H. J. Catal. 1978, 55, 129.
(9) Houalla, M.; Broderick, D. H.; Sapre, A. V.; Nag, N. K.; Beer, V. sion, a general classification of the catalysts with
H. J. D. J. Catal. 1980, 61, 523. potential for the removal of 4,6-DMDBT is shown in
(10) Houalla, M.; Nag, N. K.; Sapre, A. V.; Broderick, B. H.; Gates, Figure 3.
B. C. AIChE J. 1978, 24 (6), 1015.
(11) Landau, M. V. Catal. Today 1997, 36, 393.
(12) Meille, V.; Schulz, E.; Lemaire, M.; Vrinat, M. J. Catal. 1997, (21) Landau, M. V.; Berger, D.; Herskowitz, J. Catal. 1996, 159, 236.
170, 29. (22) Ma, X.; Sakanishi, K.; Iosda, T.; Mochida, I. Ind. Eng. Chem.
(13) Kabe, T.; Ishihara, A.; Zhang, Q. Appl. Catal. A 1993, 97, L1. Res. 1995, 34, 748.
(14) Bataille, F.; Lemberton, J. L.; Michaud, P.; Perot, G.; Vrinat, (23) Michaud, P.; Lemberton, J. L.; Perot, G. Appl. Catal. A: General
M.; Lemaire, M.; Schulz, E.; Breysse, M.; Kasztelan, S. J. Catal. 2000, 1998, 169, 343.
191 (2), 409. (24) Lecrenay, E.; Sakanishi, K.; Mochida, I. Catal. Today 1997, 39,
(15) Girgis, M. J.; Gates, B. C. Ind. Eng. Chem. Res. 1991, 30 (9), 13.
2021. (25) Ozaki, H. Catalysis Surveys from Japan 1997, 1 (1), 143.
(16) Bej, S. K. Fuel Process. Technol. 2004, 85, 1503. (26) Lecrenay, E.; Mochida, I. Stud. Surf. Sci. Catal. 1997, 106, 333.
(17) Bej, S. K.; Dalai, A. K.; Maity, S. K. Rev. Process Chem. Eng. (27) Breysse, M.; Afanasiev, P.; Geantet, C.; Vrinat, M. Catal. Today,
2000, 3 (3), 203. in press .
(18) Babich, I. V.; Moulijn, J. A. Fuel 2003, 82, 607. (28) Isoda, T.; Takase, Y.; Kusakabe, K.; Morooka, S. Energy Fuels
(19) Song, C. Catal. Today 2003, 86 (1-4), 211. 2000, 14, 585.
(20) Whitehurst, D. D.; Isoda, T.; Mochida, I. Adv. Catal. 1998, 42, (29) Robinson, W. R. A. M.; van Veen, J. A. R.; de Beer, V. H. J.;
345. van Santen, R. A. Fuel. Process. Technol. 1999, 61, 103.
Efficient 4,6-DMDBT Hydrodesulfurization Catalyst Energy & Fuels, Vol. 18, No. 5, 2004 1229

Figure 1. Possible reaction pathways for enhancing the reactivity of 4,6-DMDBT.

methyl biphenyl) or simply BP (biphenyl) is generally

the result of the demethylation propensity of the
catalyst. Though desulfurized products formed through
the direct desulfurization route can also undergo phenyl
ring hydrogenation, the presence of relatively larger
amounts of MCHT compared to DMBP generally indi-
cates the preferential occurrence of HDS through hy-
drogenative desulfurization (i.e., prehydrogenation of
4,6-DMDBT). Similarly, the desulfurized products
(formed through the prehydrogenation route) can also
undergo isomerization or dealkylation. Specially de-
signed experiments are conducted to understand the
reactivities of various possible intermediates, which in
turn help to elucidate the reaction pathways in a
meaningful way.

Oxide-Supported Catalysts
Alumina alone or in admixtures with other compo-
Figure 2. Enhancement in flexibility of the partially hydro- nents has been used as supports for the active metals
genated 4,6-DMDBT molecule for approaching the active sites and promoters. In some studies other non-alumina-
of the catalyst.
based oxide supports have also been used. Hydro-
Studies aimed at developing better catalysts for 4,6- desulfurization catalysts are usually prepared by im-
DMDBT HDS have mostly been compared to either pregnation of active metals (Mo, W) and promoters (Co,
CoMo/Al2O3 or NiMo/Al2O3 catalysts. Conversion and/ Ni) onto the support surface following pore filling or
or reaction rate constants have generally been used as incipient wetness techniques. Impregnations of metals
a criterion for comparison. Analysis of intermediates are carried out either through a single step (co-
and final products has often been employed to elucidate impregnation) or a two-step (sequential impregnation)
reaction pathways. For example, the preferential forma- procedure. Finally, the impregnated sample is dried and
tion of dimethyl biphenyl (DMBP) compared to methyl calcined to disperse the active metals onto the support.
cyclohexyl toluene (MCHT) over a catalyst indicates the The dispersion of the active metals, a key property of
dominance of a non-hydrogenative desulfurization route the finished catalyst, depends on a number of param-
(refer Figure 1). The formation of DMBP could be the eters such as impregnation procedure, solute concentra-
result of an increase either in the intrinsic hydrogenoly- tion, pH, calcination temperature, etc. In the two-step
sis capability or in the isomerization capability of the impregnation procedure, the sequence of impregnating
catalyst. On the other hand, formation of MMBP (mono the active metal and the promoter sometime plays an
1230 Energy & Fuels, Vol. 18, No. 5, 2004 Bej et al.

Figure 3. A general classification of the catalysts holding potential for the removal of 4,6-DMDBT.

important role. The effects of these parameters on the Addition of phosphorus and fluorine has been claimed
properties of HDS catalysts have been discussed in to improve dispersion as well as to increase acidity of
several reviews.30-32 The oxides of the active metals are the alumina support.37,38 However, very little literature
then converted to their sulfide forms, which are gener- is available reporting the effects of these additives on
ally believed to be the active phases for the HDS the HDS of 4,6-DMDBT.39,40 A summary of recent
reaction. findings has been provided by Moon.41
Alumina-Supported Catalysts. Three combinations Lecrenay et al.42 compared the performances of CoMo/
of metals, viz., CoMo, NiMo, and NiW have generally Al2O3 and phosphorus-modified CoMo/Al2O3 catalysts.
been used as the active components over the alumina The Co and Mo contents of the catalysts were kept
support. constant, respectively, at 3 and 10 wt % for both cases,
Alumina-Supported CoMo Catalysts. As mentioned whereas the phosphorus content in the second catalyst
earlier, the rate of 4,6-DMDBT removal through a direct was 3 wt % (as P2O5). Upon introduction of phosphorus,
desulfurization route over CoMo/Al2O3 catalyst is slow the rate of 4,6-DMDBT HDS in decane increased 3-fold.
because of the steric hindrance of the methyl groups. They also compared cumene transalkylation and naph-
Early literature indicates that CoMo-based alumina thalene hydrogenation rates over both these materials.
catalysts have low activities for the prehydrogenation The phosphorus-containing catalyst increased the cumene
of the aromatic ring.13,33,34 transalkylation and the naphthalene hydrogenation
Various efforts have been made to increase the rates by approximately three and two times, respec-
activities of conventional CoMo-based alumina catalysts tively.
by incorporating more hydrogenation capabilities in The level in improvement depends on the amount of
these materials. These include loading of the active phosphorus added and the method of addition. Kwak
metals in greater amounts, improving dispersion of the et al.39 studied, in detail, the effect of phosphorus
active metals, and manipulating the acidity level of the addition on the behaviors of alumina-supported CoMoS-
alumina support. The first two objectives have been based catalysts. The reaction was conducted at 350 °C
achieved by increasing the surface area of the support under 4.0 MPa using 4,6-DMDBT in dodecane. The
and also by using better metal loading techniques.35 A phosphorus-modified alumina support was prepared by
number of hydrotreating catalyst manufacturers such impregnating γ-alumina with an aqueous solution of
as Akzo Nobel, Criterion, Haldor-Topsoe, and United H3PO4. With the addition of phosphorus, the conversion
Catalyst have improved the performances of their of 4,6-DMDBT initially increased and then decreased,
CoMo-based alumina catalysts using various novel showing a maximum corresponding to a P2O5 content
techniques, which have been recently summarized by of 0.5 wt %. They proposed that the enhancement in
Song and Ma.19,36 conversion with the addition of phosphorus was due to

(30) Topsoe, H.; Clausen, B. S.; Massoth, F. E. Hydrotreating (37) Atanasova, P.; Halachev, T.; Uchytil, J.; Kraus, M. Appl. Catal.
catalysis science and technology; Springer-Verlag: New York, 1996. 1988, 38, 235.
(31) Zdrazil, M. Catal. Today 1988, 3, 269. (38) Matralis, H. K.; Lycourghiotis, A.; Grange, P.; Delmon, B. Appl.
(32) Startsev, A. N. Catal Rev.sSci. Eng. 1995, 37, 353. Catal. 1988, 38, 273.
(33) Isoda, T.; Ma, X.; Mochida, I. Abstract of Papers of the ACS (39) Kwak, C.; Kim, M. Y.; Choi, K.; Moon, S. H. Appl. Catal. A 1999,
208: 59-PETR Part 2 1994, AUG 21, 1994. 185, 19.
(34) Ma, X.; Sakanishi, K.; Isoda, T.; Mochida, I. Abstract of Papers (40) Kwak, C.; Lee, J. J.; Bae, J. S.; Choi, K.; Moon, S. H. Appl.
of the ACS 208: 83-PETR Part 2 1994, AUG 21, 1994. Catal. A 2000, 200, 233.
(35) Mignard, S.; Kasztelan, S.; Dorbon, M.; Billon, A.; Sarrazin, P. (41) Moon, S. H. Catal. Surveys from Asia 2003, 7 (1), 11.
Stud. Surf. Sci. Catal. 1996, 100, 209. (42) Lecrenay, E.; Sakanishi, K.; Mochida, I.; Suzuka, T. Appl. Catal.
(36) Song, C.; Ma, X. Appl. Catal. B 2003, 41, 207. A 1998, 175, 237.
Efficient 4,6-DMDBT Hydrodesulfurization Catalyst Energy & Fuels, Vol. 18, No. 5, 2004 1231

the increase in molybdenum dispersion. This was also identical experimental conditions, the conversion of 4,6-
evident from nitric oxide chemisorption data. The DMDBT over the NiMo catalyst was 68%, whereas, that
decrease in HDS rate beyond this level of phosphorus over the CoMo catalyst was only 49%. A comparison of
content (0.5 wt % P2O5) was attributed to the formation the product distributions showed that hydrogenative
of relatively stable Co-Mo-P compounds. The amounts desulfurization was the predominant pathway over both
of both DMBP and MCHT increased with increasing these materials. However, some minor differences were
phosphorus content of the catalyst and attained a observed. For example, the extent of further hydrogena-
maximum value corresponding to a P2O5 content of 0.5 tion of another aromatic ring in MCHT was appreciable
wt %. However, it was interesting to note that the over the NiMo catalyst, whereas, it was negligible over
increase in the rate of DMBP formation was higher than the CoMo catalyst. Although DMBP, the product of
that of MCHT formation. This indicated that addition direct desulfurization, was formed in small quantities
of phosphorus enhanced the rate of non-hydrogenative over both the catalysts, its content was lower for the
desulfurization to a greater extent compared to the rate NiMo-based one.
of hydrogenative desulfurization. They proposed that Lecrenay et al.24 studied the HDS of 4,6-DMDBT (in
addition of phosphorus increased the Bronsted acidity decane solution) over both CoMo-based alumina and
of the catalyst and this in turn facilitated the migration NiMo-based alumina catalysts. The CoO content of the
of the methyl groups. The migration of methyl groups CoMo catalyst was 4.2 wt % while the NiO content of
was further supported by 2,2′-DMBP isomerization data. the NiMo-based catalyst was 3.1 wt %. Though both
The effect of fluorine addition to CoMo-based catalysts catalysts had nearly similar cracking activities (as
for the HDS of 4,6-DMDBT was studied by Kwak et al.40 estimated from their isopropylbenzene transalkylation
The method of successive impregnation was used to load rates), the hydrogenation activity (as estimated from its
the various components, where the order of impregna- naphthalene hydrogenation capability) of the NiMo-
tion was as follows: first fluorine, followed by molyb- based alumina catalyst was almost 2.5 times higher
denum , then finally nickel. NH4F was used as the than the CoMo-based catalyst. As a result of the higher
fluorine precursor. The HDS of 4,6-DMDBT (in dode- hydrogenation capability the NiMo-based material ex-
cane) was performed at a temperature of 350 °C, under hibited a higher HDS rate. The first-order rate constant
a hydrogen pressure of 4.0 MPa. It was observed that for 4,6-DMDBT HDS over the CoMo-based catalyst was
the conversion of 4,6-DMDBT increased with the in- ∼0.006 min-1 g-1 while that over the NiMo-based one
crease in fluorine content of the catalysts. DMBP and was ∼0.02 min-1 g-1. It was also interesting to note that
MCHT were observed as the two major products of 4,6- the ratio of the amount of products obtained through
DMDBT HDS over fluorinated CoMoS catalysts. The the hydrogenative desulfurization route to those ob-
concentrations of both these compounds increased with tained through the direct desulfurization route was very
increasing fluorine content of the catalysts. This sug- high (∼12 at 270 °C) for the Ni-Mo catalyst as
gested that both hydrogenative and non-hydrogenative compared to ∼4 at 270 °C for the CoMo-based catalyst.
desulfurization routes were favored by the addition of This further suggested that the enhancement in HDS
fluorine. The enhancement in hydrogenative and non- activity over the NiMo-based catalyst was possibly due
hydrogenative desulfurization activities with the in- to its increased hydrogenation capability.
creasing amount of fluorine content was also confirmed The relative performances of CoMo/Al2O3 and NiMo/
by BP hydrogenation and 2,2′-DMBP isomerization Al2O3 catalysts also depend on the aromatic content of
data, respectively. The improvement in hydrogenation the feedstock. Isoda et al.45 have compared the perfor-
activity was attributed to the increased dispersion of mances of these two catalysts using a mixture of 4,6-
the active phase CoMoS, whereas, the enhancement in DMDBT, decane, and naphthalene. Because of the
isomerization capability was attributed to the increased dominance of naphthalene hydrogenation over the
acidity of the catalysts. However, for any particular level NiMo-based catalyst, the performance of this material
of fluorine incorporation, the concentration of MCHT for the conversion of 4,6-DMDBT was strongly retarded
was higher than DMBP, indicating that hydrogenation due to the presence of naphthalene. Conversely, the
was the predominant pathway for the HDS of 4,6- CoMo-based catalyst was found to be less retarded by
DMDBT over these catalysts. Unlike the zeolite-modi- naphthalene, and as a result was found to be superior
fied catalysts, cracking reactions producing light hydro- to the NiMo-based one for the HDS of 4,6-DMDBT.
carbons were essentially absent over the fluorinated (as In another study, Lecreany et al.46 reported that the
well as phosphorus-modified) catalysts, indicating that CoMoS-based alumina catalyst showed a somewhat
the acidity generated was not high enough to allow higher activity for the transalkylation of isopropyl-
these undesirable side reactions. benzene than that of a NiMoS-based alumina catalyst.
Alumina-Supported NiMo Catalysts. It has been Each catalyst contained 3 wt % of promoter metal (Ni
reported that NiMo/Al2O3 catalysts normally have a or Co) and 10 wt % of molybdenum oxide. The NiMoS-
greater aromatic hydrogenation capability as compared based catalyst was also more active than the CoMoS-
to that of CoMo/Al2O3 catalysts.43 Isoda et al.44 carried based one for the hydrogenation of naphthalene. The
out a comparative study on the HDS of 4,6-DMDBT over pseudo-first-order rate constant for the HDS of 4,6-
alumina-supported commercial CoMo and NiMo cata- DMDBT (in decane) at 270 °C over the NiMo-based
lysts. Their studies were conducted at a temperature alumina catalyst was approximately double that of the
of 270 °C, and a hydrogen pressure of 3.0 MPa. Under CoMo-based alumina catalyst. This indicated that the
(43) Stanislaus, A.; Cooper, B. H. Catal. Rev. Sci. Eng. 1994, 36, (45) Isoda, T.; Nagao, S.; Ma, X.; Korai, Y.; Mochida, I. Appl. Catal.
75. A: General 1997, 150, 1.
(44) Isoda, T.; Nagao, S.; Ma, X.; Korai, Y.; Mochida, I. Energy Fuels (46) Lecrenay, E.; Sakanishi, K.; Nagamatsu, T.; Mochida, I.;
1996, 10, 1078. Suzuka, T. Appl. Catal. B 1998, 18, 325.
1232 Energy & Fuels, Vol. 18, No. 5, 2004 Bej et al.

hydrogenation property of the NiMo-based material had γ-Al2O3 catalyst were 3.0 and 9.5 wt %, respectively. The
more influence than the cracking property of the CoMo- NiMo/γ-Al2O3 catalyst contained 1.4 wt % of Ni and 7.9
based catalyst for the HDS of 4,6-DMDBT. wt % of Mo. The Ni and W contents of the NiW/γ-Al2O3
Knudsen et al.47 also reported that NiMo -based catalyst were 3.0 and 9.5 wt %, respectively. The HDS
alumina catalysts were more active than the CoMo- experiments were conducted using 4-E (4-ethyl), 6-M (6-
based ones for the removal of 4,6-DMDBT present in a methyl) DBT in hexadecane at 390 °C and a total
blend of 50% SRGO and 50% LCO. In this case, both pressure of 6.0 MPa. The NiW-based material was found
catalysts were tested at the same pressures, but LHSVs to be more active than the CoMo or NiMo catalyst
and reaction temperatures were adjusted to give the presumably due to its greater hydrogenation capability.
same level of HDS (∼97%). The higher activity of the NiW/Al2O3 catalyst as com-
The effects of phosphorus addition have also been pared to that of the CoMo/Al2O3 or NiMo/Al2O3 catalyst
reported to improve the performance of NiMo/Al2O3 has similarly been reported by Robinson et al.29
catalysts.42 Upon introduction of 3 wt % of P2O5 into a Catalysts Based on Zeolite Mixed Alumina Sup-
NiMo/Al2O3 catalyst containing 3 wt % Ni and 10 wt % ports. Various types of zeolites have been added to
Mo, the rate 4,6-DMDBT HDS could be increased by a alumina for increasing its acidity. The increase in
factor of about 3. This enhancement in HDS rate was acidity is expected to enhance the isomerization and
attributed to increase in both hydrogenation activity and dealkylation of the alkyl groups present in 4,6-DMDBT.
transalkylation activity. Fluorine has also been used as Isoda et al.44 studied the performance of a Y-zeolite
an additive to improve the activity of NiMo/Al2O3 mixed alumina-supported CoMo catalyst. The Co and
catalysts.48,49 Though addition of fluorine enhanced both Mo were loaded over the mixed support containing 5
non-hydrogenative desulfurization and hydrogenative wt % Y-zeolite and balance alumina following an
desulfurization pathways, the effect was more pro- impregnation procedure. Commercially available CoMo-
nounced in the former. Characterization of the catalysts and NiMo-based alumina catalysts were also tested for
revealed that both dispersion of the active metals (Ni comparison. The loadings of Co in the commercial CoMo
and Mo) and acidity of the support were improved due catalyst and the mixed supported CoMo catalyst were
to the addition of fluorine. XPS studies showed that 4.4 and 3.9 wt % (as CoO), respectively. The Ni loading
addition of fluorine had a negligible effect on the was 3.1 wt % (as NiO) in the commercial NiMo catalyst.
electronic properties of either Mo or Ni. The increase The MoO3 contents in the commercial CoMo, NiMo, and
in dispersion of the active metals caused an enhance- the mixed supported CoMo catalysts were 14.9, 14.9,
ment in the hydrogenative desulfurization route. The and 15.8 wt %, respectively. The HDS activity tests were
improvement in the non-hydrogenative desulfurization conducted using 4,6-DMDBT in decane at 270 °C under
route was the result of methyl migration presumably a hydrogen pressure of 3.0 MPa. The Y-zeolite mixed
due to the increased acidity of the catalysts. Pyridine alumina-supported CoMo catalyst showed the highest
IR and 2,2′-DMBP isomerization studies further sup- conversion. Under similar conditions, the conversion of
ported the notion of an increase in Bronsted acid sites 4,6-DMDBT obtained over the mixed-supported CoMo
due to the addition of fluorine. catalyst was 72%, whereas, the conversions over the
Ogawa et al.50 attempted to improve the catalytic commercial CoMo and NiMo catalysts were 49% and
activity of NiMo/Al2O3 catalyst by introducing lantha- 68%, respectively. The product distribution over the
num as an additive. The rate constant for the HDS of zeolite mixed alumina-supported CoMo catalyst was
4,6-DMDBT increased by about 80% when La loading significantly different from those obtained over the
was low (0.7 wt %). Increasing La loading beyond this alumina-supported CoMo and NiMo catalysts. The
level decreased the surface Ni concentration which product patterns obtained over the two alumina-based
resulted in a decrease in the value of the rate constant. materials indicated that the HDS of 4,6-DMDBT took
The structure of MoS2 crystallites was also affected by place mainly through the hydrogenation route. On the
the addition of La. Up to a La loading of 0.7 wt %, the other hand, isomerization of 4,6-DMDBT to 3,6-DMDBT
lateral size of the MoS2 crystallites increased. On the and dealkylation products were observed characteristi-
other hand, a higher La loading (more than 0.7 wt %) cally over the zeolite mixed alumina-supported catalyst.
inhibited the lateral growth of MoS2 crystallites. Significant amounts of various light hydrocarbons were
Alumina-Supported NiW Catalysts. NiW-based ma- produced over this catalyst presumably from the hydro-
terials are known to possess higher hydrogenating cracking reactions. The dominance of methyl migration
properties and hence hold potential for enhancing 4,6- and hydrocracking reactions were the results of the
DMDBT HDS rates.51 Reinhoudt et al.52 conducted a higher acidity of the zeolite mixed alumina-supported
comparative study on alumina-based CoMo, NiMo, and CoMo catalyst (23% more acid sites) as compared to that
NiW catalysts. The Co and Mo contents of the CoMo/ of the conventional alumina-supported CoMo and NiMo
catalysts. The absence of significant deactivation was
(47) Knudsen, K. G.; Cooper, B. H.; Topsoe, H. Appl. Catal. A 1999,
189, 205. also noticed over this zeolite mixed alumina-supported
(48) Kim, H.; Lee, J. J.; Moon, S. H. Appl. Catal. B 2003, 44, 287. material despite the presence of a higher number of acid
(49) Kim, H.; Lee, J. J.; Koh, J. H.; Moon, S. H. Stud. Surf. Sci. sites on it.
Catal. 2003, 145, 315.
(50) Ogawa, Y.; Toba, M.; Yoshimura, Y. Appl. Catal. A 2003, 246, Landau et al.21 also conducted a systematic study of
213.
(51) Gachet, G.; Breysse, M.; Cattenot, M.; Decamp, T.; Frety, R.;
4,6-DMDBT HDS over alumina-supported CoMo and
Lacroix, M.; de Mourges, L.; Portefaix, J. L.; Vrinat, M.; Duchet, J. C.; NiMo catalysts, a mixed HY-Al2O3-supported CoMo
Housni, S.; Lakhdar, M.; Tilliette, M. J.; Bachelier, J.; Cornet, D.; catalyst and a mixed HZSM-5-Al2O3-supported CoMo
Engelhard, P.; Gueguen, C.; Toulhoat, H. Catal. Today 1988, 4, 7.
(52) Reinhoudt, H. R.; Troost, R.; van Langeveld, A. D.; Sie, S. T.; catalyst. The MoO3 contents of these catalysts were 25.1,
van Veen, J. A. R.; Moulijn, J. A. Fuel Process. Technol. 1999, 61, 133. 24.0, 20.9, and 24.8 wt %, respectively. The surface
Efficient 4,6-DMDBT Hydrodesulfurization Catalyst Energy & Fuels, Vol. 18, No. 5, 2004 1233

Table 1. Performances of Various Oxide Mixed Alumina-Supported CoMo and NiMo Catalysts for the HDS of
4,6-DMDBT, Transalkylation of Cumene, and Hydrogenation of Naphthalene. The Co, Ni, and Mo Contents of All the
Catalysts are 3, 3, and 10 wt %, Respectively42,46
catalyst
8 wt % 25 wt % 20 wt % 15 wt %
Al2O3 TiO2-Al2O3 TiO2-Al2O3 ZrO2-Al2O3 B2O3- Al2O3
support support support support support
properties CoMo NiMo CoMo NiMo CoMo NiMo CoMo NiMo CoMo NiMo
(m2/g)
surface area 240 440 252 252 247 247 230 230 288 288
rate constant (× 10-3) for 5 10 14 14 18 14 11 15 19 32
HDS of 4,6-DMDBT
at 270 °C
rate constant (× 10-6) for 7 4 10 4 35 12 26 7 43 35
transalkylation of cumene
at 250 °C
rate constant (× 10-3) for 3 9 5 8 10 15 6 17 6 22
hydrogenation of naphthalene
at 250 °C

areas of the catalysts were 230, 140, 308, and 243 m2/ constants for the HDS of 4,6-DMDBT (in decane) for
g, respectively. The zeolite-alumina mixed supports these materials are shown in Figure 5. The value was
were prepared by coextrusion of aluminum hydroxide highest for the alumina-zeolite mixed supported CoMo
with zeolite powders. The HDS tests were carried out catalyst. It may be worth mentioning here that the
using 4,6-DMDBT dissolved in a mixed solvent contain- zeolite mixed catalyst also showed the maximum crack-
ing 29% n-decane, 50% n-octadecane, and 30% tetralin. ing activity (as determined from the isopropylbenzene
The first-order rate constants for the HDS of 4,6- transalkylation test). The cracking activities of the
DMDBT at 360 °C for all the Co-Mo loaded materials alumina-supported NiMo and CoMO catalysts were very
are compared in Figure 4. Upon introduction of HZSM-5 small. As expected, the alumina-supported NiMo cata-
into CoMo/Al2O3 catalyst, the HDS rate decreased by a lyst exhibited the highest activity for the hydrogenation
factor of about 1.3. There was no significant change in of naphthalene. They reported that as the acidic com-
the product distribution. A lower diffusion rate of the ponent was increased in the support, both cracking and
bulky 4,6-DMDBT molecule in the HZSM-5 channels hydrogenating capabilities of the catalysts were also
was thought to be responsible for the decrease in the enhanced. The desulfurization of 4,6-DMDBT over the
HDS activity. Because of this, the HZSM-5 component zeolite mixed catalyst occurred mainly through isomer-
remained inactive and acted as an inert diluent. Con- ization and cracking routes.
versely, the HDS rate increased by a factor of about 3 Catalysts Based on Other Oxide Mixed Alumina
for the case of HY-zeolite mixed alumina-supported Supports. The use of mixed oxide alumina-supported
CoMo catalyst. Analysis of the products revealed that CoMo- or NiMo-based catalysts have provided encour-
the reaction proceeded mainly through the scission of aging results for the HDS of thiophene and benzo-
the C-C bond connecting the two aromatic rings in the thiophene.53-57 This has prompted researchers to ex-
4,6-DMDBT molecule. plore the potentials of these materials for the HDS of
In another study, Lecrenay et al.24 compared the 4,6-DMDBT.
performance of a zeolite mixed alumina-supported CoMo The HDS of 4,6-DMDBT over alumina containing
catalyst with those of alumina-supported commercial mixed oxide-supported CoMo- and NiMo-based catalysts
NiMo and CoMo catalysts. The CoO (or NiO) and MoO3 were studied by Mochida and co-workers.42,46 The
contents of the catalysts were in a similar range. The results are summarized in Table 1. Two different TiO2-
surface areas of the CoMo/Al2O3, NiMo/Al2O3, and
zeolite mixed alumina-supported CoMo catalysts were
268, 273, and 220 m2/g, respectively. The first-order rate

Figure 4. Comparison of first-order rate constants for the Figure 5. Comparison of first-order rate constants for the
HDS of 4,6-DMDBT over alumina, HZSM-5 mixed alumina, HDS of 4,6-DMDBT over NiMo/alumina, CoMo/alumina, and
and HY mixed alumina-supported CoMo catalysts (adapted zeolite mixed alumina-supported CoMo catalysts (adapted
form ref 21). from ref 24).
1234 Energy & Fuels, Vol. 18, No. 5, 2004 Bej et al.

Figure 6. Effects of various additives on the properties of alumina-supported HDS catalysts.

Al2O3 commercial supports containing 8 and 25 wt % of contents in the mixed support were 20 and 15 wt %,
TiO2, respectively, were prepared by the co-hydrolysis respectively. The amounts of the active metals in these
of a mixture of alkoxides of Al and Ti. The active metals catalysts were the same as used in their earlier work
were loaded on these supports through the technique (as described in the previous section).46 The results are
of successive impregnation in which Mo was first shown in Table 1. The pseudo first-order rate constant
impregnated followed by either Co or Ni. Each catalyst at 270 °C over CoMo/Al2O3 was 5, and it increased to
contained 3 wt % of promoter metal (Co or Ni) and 10 11 and 19 over CoMo/ZrO2-Al2O3 and CoMo/B2O3-
wt % of molybdenum oxide. For the sake comparison, Al2O3 catalysts, respectively. Similarly, the NiMo/Al2O3
the metals were also loaded onto an alumina support. catalyst had a pseudo first-order rate constant of 10,
The HDS experiments were conducted using 4,6- and it became 15 for NiMo/ZrO2-Al2O3 and 32 for NiMo/
DMDBT dissolved in n-decane in the temperature range B2O3-Al2O3 catalyst. The acidity of these supported
of 270-360 °C and under a hydrogen pressure of 2.4- materials, as measured from the rate of isopropyl-
5.0 MPa. The activities of the mixed Al2O3-TiO2- benzene transalkylation, increased with the addition of
supported catalysts were higher than that of the alumina- the second oxide. In this case also, all the CoMo-based
supported ones. The highest activity was observed for catalysts showed higher acidity as compared to that of
a CoMo-based catalyst supported on Al2O3-TiO2 mix- the NiMo-based materials. The hydrogenation capabili-
ture containing 25 wt % of TiO2. The major product was ties of the catalysts, as evident from the naphthalene
MCHT, indicating the dominance of the hydrogenative hydrogenation activities, also increased. The enhance-
desulfurization route over the mixed supported materi- ment in 4,6-DMDBT HDS rate was presumably due to
als. The acidity of the catalysts increased due to the the increase in their hydrogenation capabilities in which
addition of TiO2, as evidenced by their higher capabili- acidity might also have played a role.
ties for the transalkylation of isopropylbenzene. The A mixture of alumina and amorphous silica-alumina
hydrogenation capability, as measured with the naph- was also used as a support for loading CoMo oxides.24
thalene hydrogenation reaction, also increased moder- The activity of the alumina and amorphous silica-
ately with the increasing TiO2 content of the catalysts. alumina mixed supported CoMo catalyst was higher
They have concluded that the increased acidity may also than the conventional CoMo/Al2O3 catalyst but it could
have caused enhancement in the hydrogenation capa- not outperform the NiMo/Al2O3 catalyst.
bility. Robinson et al.29 used amorphous silica alumina
The same group of researchers42 studied the activities (ASA) as a support for NiMo catalysts for the HDS of
of CoMo and NiMo catalysts supported on alumina 4-E,6-M-DBT in the hopes of taking advantage of the
mixed with other oxides such as ZrO2-Al2O3 and B2O3- strongly acidic properties of ASA. The HDS rate for the
Al2O3 for the HDS of 4,6-DMDBT. The ZrO2 and B2O3 ASA-supported NiMo catalyst was higher than that of
alumina-supported commercial CoMo as well as NiMo
(53) Rana, M. S.; Srinivas, B. N.; Maity, S. K.; Murali Dhar, G.; Rao, catalysts.
T. S. R. P. Stud. Surf. Sci. Catal. 1999, 127, 397. Based on the above information, Figure 6 summarizes
(54) Gutierrez-Alejandre, A.; Gonzalez-Cruz, M.; Trombetta, M.;
Busca, G.; Ramirez, J. Microporous Mesoporous Mater. 1998, 23 (5- the effects of various additives on the properties of
6), 265. alumina-supported HDS catalysts.
(55) Barrio, V. L.; Arias, P. L.; Cambra, J. F.; Guemez, M. B.;
Campos-Martin, J. M.; Pawelec, B.; Fierro, J. L. G. Appl. Catal. A 2003,
Non-Alumina-Based Oxide-Supported Catalysts.
248 (1-2), 211. Although γ-Al2O3, because of its excellent mechanical
(56) Grzechowiak, J. R.; Wereszczako-Zielinska, W.; Rynkowski, J.; as well as dispersing properties, has been widely used
Ziolek, M. Appl. Catal. A 2003, 250 (1), 95.
(57) Murali Dhar, G.; Srinivas, B. N.; Rana, M. S.; Kumar, M.; as a support for commercial HDS catalysts, various
Maity, S. K. Catal. Today 2003, 86 (1-4), 45. other oxides have also been used for supporting the
Efficient 4,6-DMDBT Hydrodesulfurization Catalyst Energy & Fuels, Vol. 18, No. 5, 2004 1235

active metals for the HDS of thiophene, benzothiophene, transalkyaltion were also found to take place over the
and dibenzothiophene.21,58-65 In fact, the non-alumina- zeolite-based materials. Both the Mo/HY and CoMo/HY
based other oxide-supported catalyst formulations have catalysts exhibited higher rates compared to the Mo/
opened up a new horizon in HDS catalysis research. Al2O3 and CoMo/Al2O3 catalysts. This enhancement in
Though there are a number of publications available HDS rate was reported to be due to the migration of
regarding thoiophene HDS over these catalysts, only the methyl groups.
very few studies have been reported involving the HDS Mesoporous materials such as MCM-41 possess po-
of 4,6-DMDBT. Landau et al.21 reported that the activity tential for acting as good supports. These materials also
of a silica-supported NiMo catalyst for the HDS of 4,6- offer less diffusional resistances to 4,6-DMDBT. Al-
DMDBT was higher than that of CoMo/Al2O3 and NiMo/ though some research68-71 has been conducted using
Al2O3 catalysts. Since only limited work has been done Mo- or CoMo-loaded MCM-41 for the HDS of DBT, only
on the HDS of 4,6-DMDBT over non-alumina-based a few definitive studies have been carried out for the
oxide-supported catalysts, there remain immense op- HDS of 4,6-DMDBT over these materials.4,66 Further
portunities for advances in this area. studies are required to understand and explore the full
potential of these novel materials for the HDS of 4,6-
Catalysts Supported on Non-Oxide-Based DMDBT.
Materials Carbon-Supported Catalysts. Recently, carbon has
Zeolite- and Mesoporous Materials-Supported received considerable attention as a support for HDS
Catalysts. Because of their higher surface areas, acidic catalysts. Carbon has a number of highly desirable
properties, and well-defined pore structures, zeolites properties such as high surface area, controllable pore
and mesoporous materials have attracted much atten- volume and pore size, and perhaps favorable support-
tion as supports for CoMo/NiMo-based HDS cata- metal interaction for HDS reactions.72
lysts.4,27,36,66 The higher surface areas allow loading of Farag et al.72,73 conducted a comparative study be-
higher levels of the active metals without affecting tween carbon and alumina as a support for CoMo
dispersion. catalysts for the HDS of 4,6-DMDBT. Cobalt and Mo
Isoda et al.28 studied the skeletal isomerization of 4,6- were loaded on two types of activated carbon using two
DMDBT over a Y-type zeolite-supported Ni catalyst at different techniques. One support had a surface area of
270 °C and a hydrogen pressure of 2.5 MPa. 4-MDBT 907 m2/g with an average pore size of 12.5 Å, while the
and 3,6-DMDBT, produced through demethylation and surface area and the average pore size of the other were
methyl group migration, respectively, were found to be 3213 m2/g and 9.9 Å, respectively. The Co and Mo
the major products of the reaction. Alkyl-DBTs contain- contents of the catalysts were 2 and 10 wt %, respec-
ing three methyl groups were also detected. No de- tively. A commercial CoMo/Al2O3 catalyst having 3.2 wt
sulfurized products were observed indicating that the % of Co and 13.7 wt % of Mo was also tested. The pseudo
Y-type zeolite-supported Ni catalyst induced no detect- first-order rate constant for the conversion of 4,6-
able desulfurization under the conditions of their in- DMDBT at 340 °C for the 907 m2/g carbon-supported
vestigation, however, enhanced the transalkylation of CoMo catalyst was about double that for the commercial
4,6-DMDBT. CoMo catalyst. On the other hand, the catalyst prepared
Bataille et al.67 studied the HDS of 4,6-DMDBT over on the 3213 m2/g surface area carbon support exhibited
dealuminated HY-zeolite-supported Co and CoMo cata- a rate that was almost equal to that of the commercial
lysts and compared the results with those of alumina- CoMo catalyst; however, the rate was lower than that
supported ones. The Mo/HY catalyst contained 8.3 wt of the lower surface area-supported material. This was
% Mo. The Mo and Co contents of the CoMo/HY catalyst attributed to the pore diffusional resistances of 4,6-
were 8.5 and 2.2 wt %, respectively. The reaction was DMDBT within the narrow pores of the high surface
carried out at a temperature of 330 °C and a hydrogen area carbon support. It was observed that the preferred
pressure of 3 MPa. Over all the materials, the HDS took pathway for HDS also depended on temperature. For
place through two common routes, i.e., the direct de- example, at 340 °C, the two routes (direct desulfuriza-
sulfurization route and the prehydrogenation one. In tion and hydrogenation) contributed equally to the
addition to these, methyl group isomerization and overall rate. At lower temperature (∼300 °C), the
hydrogenation route was the preferred one, while the
(58) Cinibulk, J.; Kooyman, P. J.; Vit, Z.; Zdrazil, M. Catal. Lett.
2003, 89 (1-2), 147.
direct desulfurization route became the dominant one
(59) Venezia, A. M.; La Parola, V.; Deganello, G.; Cauzzi, D.; at higher temperature (∼380 °C). Thermodynamic
Leonardi, G.; Predieri, G. Appl. Catal. A 2002, 229 (1-2), 261. limitations for the hydrogenation reaction at higher
(60) Maity, S. K.; Rana, M. S.; Bej, S. K.; Ancheyta-Juarez, J.; Murali
Dhar, G.; Rao, T. S. R. P. Appl. Catal. A: General 2001, 205, 215. temperature were probably responsible for such a shift
(61) Maity, S. K.; Rana, M. S.; Srinivas, B. N.; Bej, S. K.; Murali in reaction pathways.
Dhar, G.; Rao, T. S. R. P. J. Mol. Catal. A: Chemical 2000, 153, 121.
(62) Maity, S. K.; Rana, M. S.; Bej, S. K.; Ancheyta-Juarez, J.; Murali
Dhar, G.; Rao, T. S. R. P. Catal. Lett. 2001, 72 (1-2), 115. (68) Song, C.; Reddy, K. M. Appl. Catal., A 1999, 176, 1.
(63) Wang, D.; Qian, W.; Ishihara, A.; Kabe, T. J. Catal. 2002, 209, (69) Wang, A.; Wang, Y.; Kabe, T.; Chen, Y.; Yshihara, A.; Quian,
266. W. J. Catal. 2001, 199, 19.
(64) Dzwigaj, S.; Louis, C.; Breysee, M.; Cattenot, M.; Belliére, V.; (70) Wang, A.; Wang, Y.; Kabe, T.; Chen, Y.; Yshihara, A.; Qian,
Geantet, C.; Vrinat, M.; Blanchard, P.; Payen, E.; Inoue, S.; Kudo, H.; W.; Yao, P. J. Catal. 2002, 210, 319.
Yoshimura, Y. Appl. Catal. B 2003, 41, 181. (71) Klimova, T.; Calderon, M.; Ramirez, J. Appl. Catal. A 2003,
(65) Damyanova, S.; Andonova, S.; Stereva, I.; Vladov, C.; Petrov, 240, 29.
L.; Grange, P. React. Kinet. Catal. Lett. 2003, 79 (1), 35. (72) Farag, H.; Whitehurst, D. D.; Sakanishi, K.; Mochida, I. Catal.
(66) Turaga, U. T.; Song, C. Catal. Today 2003, 86 (1-4), 129. Today 1999, 50, 9.
(67) Bataille, F.; Lemberton, J. L.; Perot, G.; Leyrit, P.; Cseri, T.; (73) Farag, H.; Mochida, I.; Sakanishi, K. Appl. Catal. A 2000, 194-
Marchal, N.; Kasztelan, S. Appl. Catal. A 2001, 220, 191. 195, 147.
1236 Energy & Fuels, Vol. 18, No. 5, 2004 Bej et al.

proximately 1.6 times that of the commercial alumina-

based catalyst.

Other Catalysts

Carbide-, Nitride-, and Phosphide-Based Cata-

lysts. Recently, transition metal carbides and nitrides
have received considerable attention for HDS re-
actions.76-87 Carbides and nitrides are a class of inter-
stitial compounds of metal and, carbon or nitrogen,
respectively. These materials can be prepared in high
surface area form through temperature program reac-
Figure 7. Proposed reaction scheme for the HDS of 4,6-
DMDBT over carbon-supported CoMo catalyst (adapted from tion. Carbides and nitrides are also known to possess
ref 74). properties like those of the Pt-group materials.88
Furimsky87 has reviewed the potential of carbides and
Farag et al.74 carried out a kinetic analysis for the nitrides as catalysts for hydroprocessing reactions.
HDS of 4,6-DMDBT over a carbon-supported CoMo Though carbide-based catalysts were tested for the
catalyst containing 2 and 10 wt % of cobalt and HDS of various sulfur compounds, reports of their
molybdenum, respectively. The study was conducted performances for the HDS of hindered DBTs are rela-
with a view to elucidate the mechanism of 4,6-DMDBT tively few.78,79,86 Da Costa et al.78 investigated the 4,6-
conversion pathways over the carbon-supported cata- DMDBT HDS over alumina-supported molybdenum
lyst. They proposed a reaction network (refer to Figure carbide catalysts at a temperature of 340 °C and a
7) for the HDS of 4,6-DMDBT over the carbon-supported pressure of 4 MPa. Over the carbide-based catalyst, the
CoMo catalyst. The major products observed were HDS took place following both direct desulfurization and
3,3′-dimethylphenylcyclohexane (3,3′-DMPC) formed hydrogenative desulfurization routes. A comparison of
through a hydrogenative desulfurization route and 3,3′- the selectivities for various products indicated that the
dimethylbiphenyl (3,3′-DMBP) formed through the di- direct desulfurization route was favored over the alu-
rect desulfurization route. Partially hydrogenated 4,6- min-supported carbide catalyst in contrast to the con-
DMDBT was also detected in small quantities. Under ventional alumina-supported CoMo or NiMo catalysts,
the conditions of investigation, the formation of 3,3′- for which hydrogenative desulfurization was the pre-
DMPC through the hydrogenation of 3,3′-DMBP was ferred pathway. In another study,79 the same group of
found to be very slow. This confirmed the production of researchers compared the activity and stability of an
alumina-supported molybdenum carbide to that of an
3,3′-DMPC from the hydrodesulfurization of the par-
alumina-supported sulfided molybdenum oxide and to
tially hydrogenated 4,6-DMDBT. At a reaction temper-
that of a supported platinum catalyst for the HDS of
ature of 340 °C and a hydrogen pressure of 2.9 MPa,
4,6-DMDBT. The ranking of HDS activity was as
the ratio of the products generated from the direct
follows: MoS2/Al2O3 < Mo2C/Al2O3 < Pt/SiO2. The HDS
desulfurization and hydrogenation routes were about
reaction, as expected, mainly took place through the
41:59.
hydrogenative desulfurization route over both the sup-
Sakanishi et al.75 studied the kinetics and mechanism ported platinum and sulfided molybdenum oxide cata-
of 4,6-DMDBT HDS over carbon-supported NiMo cata- lysts. The Mo2C/Al2O3 catalyst favored the direct de-
lysts. Various types of carbon supports having a wide sulfurization route, despite its well-known hydrogena-
range of surface areas (480-3060 m2/g) were used. tion capabilities. The influence of phosphorus and nickel
Regardless of the nature of carbon, NiMo/C catalysts on the properties of alumina-supported molybdenum
exhibited higher activity than that of a commercial carbide for the HDS of 4,6-DMDBT was studied by
NiMo/alumina catalyst in the temperature range of Manoli et al.86 The introduction of phosphorus increased
340-380 °C. In this range, the direct desulfurization the Lewis acid sites, which in turn enhanced the non-
route was the dominant one. On the other hand, the hydrogenative desulfurization routes. On the other
hydrogenative desulfurization route was the major one
at lower temperatures (∼300 °C). In this range, the (76) Dhandapani, B.; St. Clair, T.; Oyama, S. T. Appl. Catal. A 1998,
selectivity for various products was dependent on the 168, 219.
(77) Sajkowski, D. J.; Oyama, S. T. Appl. Catal. A 1996, 134, 339.
level of 4,6-DMDBT conversion. (78) Da Costa, P.; Potvin, C.; Manoli, J. M.; Lemberton, J. L.; Perot,
Robinson et al.29 compared the performance of a G.; Mariadassou, G. D. J. Mol. Catal. A 2002, 184, 323.
(79) Da Costa, P.; Potvin, C.; Manoli, J. M.; Breysse, M.; Mariadas-
carbon-supported CoMo catalyst with that of an alumina- sou, G. D. Catal. Lett. 2003, 86 (1-3), 133.
supported commercial CoMo catalyst containing 3 wt (80) Brian, D.; Stephanie, J.; Denise, H. B.; Rebekah, M.; Diana, C.
% Co and 9.5 wt % Mo, respectively, for the HDS of P.; Scott, K.; Randy, S.; Mark, E. B. Catal. Today, in press.
(81) Liu, Y. Q.; Liu, C. G.; Oue, G. H. Energy Fuels 2002, 16 (3),
4-E,6-M-DBT. The carbon used for the support had a 531.
surface area of 1200 m2/g. The Co and Mo contents in (82) Trawczynnski, J. Catal. Today 2001, 65 (2-4), 343.
(83) Trawczynnski, J. Appl. Catal. A 2000, 197 (2), 289.
this catalyst were 1.6 wt % and 8 wt %, respectively. (84) Kim, D. W.; Lee, D. K.; Ihm, S. K. Catal. Lett. 1997, 43 (1-2),
The activity of the carbon-supported catalyst was ap- 91.
(85) Oyama, S. T.; Yu, C. C.; Ramanathan, S. J. Catal. 1999, 184,
535.
(74) Farag, H.; Sakanishi, K.; Mochida, I.; Whitehurst, D. D. Energy (86) Manoli, J. M.; Da Costa, P.; Brun, M.; Vrinat, M.; Mauge, F.;
Fuels 1999, 13, 449. Potvin, C. J. Catal. 2004, 221, 365.
(75) Sakanishi, K.; Nagamatsu, T.; Mochida, I.; Whitehurst, D. D. (87) Furimsky, E. Appl. Catal. A: General 2003, 240, 1.
J. Mol. Catal. 2000, 155, 101. (88) Levy, R. B.; Boudart, M. Science 1973, 181, 547.
Efficient 4,6-DMDBT Hydrodesulfurization Catalyst Energy & Fuels, Vol. 18, No. 5, 2004 1237

hand, upon introduction of nickel an increase in the catalyst. Both Pt-based catalysts were quite stable
hydrogenative desulfurization rate was observed. under the conditions of reaction. On the other hand, the
Molybdenum nitride alone or in combination with Pt catalyst supported on XVUSY, although showing a
other promoters has been reported to effectively catalyze very high initial activity (even higher than the Pt/ASA-
the HDS of thiophene, benzothiophene, and vacuum gas based catalyst), underwent a rapid deactivation within
oil.81-84 However, publications covering detailed studies a time period of about 4 h. Catalyst characterization
for the HDS of 4,6-DMDBT over nitrides are not studies indicated that an appropriate tuning of the
available in the literature. Similarly, only a few reports support acidity was very important for achieving high
are available regarding the use of phosphides for the activity and stability. According to them, the effect of
HDS of 4,6-DMDBT. Oyama et al.89 conducted the HDS acidic supports on the high activity for the HDS of 4-E,6-
of 4,6-DMDBT over nickel phosphide (Ni2P) catalysts M-DBT might be bipartite.
supported on silica, alumina, and potassium ion- The combined effect of Pt and Pd was also studied by
exchanged USY. The catalytic activity was measured Reihoudt et al.52 4-E,6-M-DBT was used as a model
at 340 °C and 3.1 MPa in a trickle bed reactor. The compound for the HDS reaction. The Pd-Pt supported
activity was found to be Ni2P/K-USY > Ni2P/SiO2 . on ASA (amorphous silica alumina) catalyst showed
Ni2P/Al2O3, on the basis of equal sites loaded in the higher HDS activity as compared to that of conventional
reactor. On the basis of the same concept (i.e., equal CoMo/A2O3, NiMo/Al2O3, and NiW/Al2O3 catalysts. The
sites loaded in the reactor), the conversion of 4,6- beneficial effect of the noble metal was more prominent
DMDBT over the Ni2P/K-USY catalysts was much on the ASA support than on the Al2O3. When both Pd
higher (98% HDS) than that of a commercial sulfided and Pt were present, the HDS activity was much higher
CoMo/Al2O3 catalyst (55% HDS). than that of a single metal-containing catalyst. It was
Noble Metal-Based Catalysts. Because of their high noted that a specific metal-support interaction was
hydrogenation capabilities, noble metal-based materials responsible for the higher activity of ASA-supported
have also been considered and studied for HDS re- catalysts.
actions.2,3,90-93 However, the use of these metals for A combination of various noble metals such as Re-
HDS is limited because of their very low sulfur toler- Pt, Ir-Pt, Sn-Pt, and Ge-Pt have also been used for
ances. Researchers2,3 have reported that a bimetallic the hydrogenation of LCO/SRLGO feedstock.91 However,
Pd-Pt catalyst supported on ytterbium-modified ultra- these combinations of noble metals did not show any
stable Y (USY)-zeolite is capable of effectively desul- positive effects on the hydrogenation reaction; however,
furizing 4,6-DMDBT. Pt or Pt-Pd-supported catalysts showed very high
Reinhoudt et al.90 studied the performances of several activities.91
Pt catalysts supported on various materials such as
amorphous silica alumina (ASA), γ-alumina, and sta-
Conclusions
bilized Y-zeolite (XVUSY) for the HDS of 4-E,6-M-DBT.
The study was conducted at a temperature of 360 °C Relatively less research has been conducted in the
and 6.0 MPa hydrogen pressure. They observed that area of 4,6-DMDBT HDS as compared to that for the
with a similar level of Pt loading (about 0.8-1.0 wt %), HDS of thiophene and benzothiophenes. Among all the
the Pt/ASA catalyst exhibited much higher activity as approaches studied for improving the activities of
compared to that of the Pt/γ-alumina-based catalyst. catalysts for the removal of 4,6-DMDBT, the inclusion
The activity of the commercial CoMo/γ-alumina catalyst of various additives such as phosphorus, fluorine, lan-
was also higher than that of the Pt/γ-alumina-based thanum, other oxides, and zeolites holds substantial
material; however, it was lower than that of the Pt/ASA promise. Most of these additives improve the catalytic
activity either by improving the dispersion of the active
(89) Oyama, T. S.; Lee, Y. K.; Chaturvedula, H. Accepted for
presentation at the 2003 annual AIChE meeting, San Francisco, metals (Mo and Co or Ni) and/or by generating acidic
November, 2003. properties in the catalysts. Carbon is also a potentially
(90) Reinhoudt, H. R.; Troost, R.; van Schalkwijk, S.; van Langeveld, useful support. Zeolites, when used alone as a support,
A. D.; Sie, S. T.; van Veen, J. A. R.; Moulijn, J. A. Fuel Process. Technol.
1999, 61, 117. do not provide very encouraging results. Much attention
(91) Takashi, F.; Kazuo, I.; Takeshi, E.; Hirofumi, M.; Kazushi, U. needs to be focused on mixed oxide-supported catalysts.
Appl. Catal. A 2000, 192, 253.
(92) Lourdes, I. Meriño; Centeno, Aristóbulo; Sonia, A. Giraldo. Appl. Noble metal- and carbide-based formulations also bring
Catal. A 2000, 197, 61. lots of promise for the application.
(93) Takashi, F.; Kazuo, I.; Katsuyoshi, O.; Hirofumi, M.; Kazushi,
U. Appl. Catal. A 2001, 205, 71. EF030179+