Oxidation of dibenzothiophene by hydrogen peroxide catalyzed by solid bases

Article abstract of DOI:10.1016/S0021-9517(02)93721-8

The removal of sulfur compounds from hydrocarbon-based distillate fuels is essential due to environmental concerns (i.e., acid rain and airborne particulate material production), and because a few parts per million of sulfur are enough to poison the catalysts used for the purification of the exhaust gases of diesel cars. The oxidation of dibenzothiophene (DBT) by H₂O₂ at 333 K was investigated using the catalysts (hydrotalcite (HT) and MgLa mixed oxide) and solvents (methanol, acetonitrile, benzonitrile, acrylonitrile, and 3-methoxypropionitrile). High activity was found only after calcinations followed by rehydration of HT. Acetonitrile was the best solvent, while much lower reaction rates were observed in methanol. The decomposition of H₂O₂ into oxygen was observed for HT, and was the major reaction at > 353 K. This reaction did not occur in the absence of nitrile, and was faster over MgLa mixed oxide of higher basic strength. The activity increased with increasing Mg/Al ratio due to a lower rate of H₂O₂ decomposition attributed to a lower basicity of the solid.

Full text of DOI:10.1016/S0021-9517(02)93721-8

Journal of Catalysis 211, 103–108 (2002)
doi:10.1006/jcat.2002.3721
Oxidation of Dibenzothiophene by Hydrogen Peroxide Catalyzed
by Solid Bases
Jorge Palomeque, Jean-Marc Clacens, and Franc¸ois Figueras1
Institut de Recherches sur la Catalyse, 2 Avenue Albert Einstein, 69626 Villeurbanne Cedex, France
Received February 8, 2002; revised April 12, 2002; accepted May 28, 2002
sulting sulfone by means of a solvent. Using peracetic acid
at 356 K to treat a light atmospheric gas oil with an ini-
tial sulfur content of 0.42 wt%, Gore et al. reported 96%
conversion of sulfur compounds after 1 h (2, 3). The pro-
cess could be also applied using perboric, persulfuric, and
peracetic acids or ozone. This oxidation procedure is very
well suited to desulfuration, because the compounds, which
react faster in oxidation, are substituted thiophenes, which
show the lowest reactivity in HDS.
The oxidation of dibenzothiophene by hydrogen peroxide was
performed at 333 K using hydrotalcite and MgLa mixed oxide as
catalysts and nitriles or methanol as solvents. High activity is found
only after calcination followed by rehydration of hydrotalcite. The
kinetics involves competitive adsorption of the reactants. Of the
nitriles, acetonitrile is the best solvent, while much lower reaction
rates are observed in methanol. Decomposition of hydrogen perox-
ide into oxygen is also observed for hydrotalcites and is the major
reaction above 353 K. This reaction does not occur in the absence
of nitrile and is faster over the MgLa mixed oxide of higher basic
strength, in agreement with the decomposition of an iminoperacid
intermediate formed by the base-catalyzed addition of hydrogen
peroxide to the nitrile. The activity increases with increasing Mg/Al
ratio due to a lower rate of H2O2 decomposition attributed to a
c
Peracids do not require a catalyst but are more expensive
than hydrogen peroxide; a second possibility is to use H2O2
in the presence of a catalyst: methyltrioxorhenium(VII)
(CH3ReO3) (4), CF3COOH (5), formic acid (6), and re-
cently titano silicates (7) were reported. While TS-1 was not
very active due to the restricted access of reactants to the
porosity, Ti-beta and Ti-HMS proved to be effective in the
liquid phase, with Ti-beta reaching conversion above 90%
after 5 h at 333 K. On a real kerosene charge, 90% removal
of sulfur was observed after 2 h. Acetonitrile seems to be
the best solvent for this reaction, but methanol and ethanol
could also be used. When nitriles are used as solvents, hy-
drogen peroxide forms a peroxycarboximidic acid in the
presence of a solid base such as hydrotalcite (HT). Indeed,
HT acts as a catalyst to form a perhydroxyl anion species
(HOO−), which in turn nucleophilically attacks the nitrile
to generate peroxycarboximidic acid, which is an active in-
termediate oxidant (8). The addition of hydrogen peroxide
on nitriles was first performed in 1961 in organic chemistry
(9), and the reaction mechanism depicted in Scheme 1 is
widely accepted (10).
The peroxycarboximidic acid is too reactive to be isolated
but its existence was proved by vibrational spectroscopy
(11). It has been shown that step 1 is slow and that step 2,
which leads to the decomposition of hydrogen peroxide, is
fast. In the presence of an olefin, the iminoperacid reacts at
the double bond to give dioxolane, which decomposes into
the epoxide and the amide (8). With HTs, the best additives
are benzonitrile and dicyanobenzene (12) and the same ef-
fect can be obtained by adding an amide (13). This reaction
enables the epoxidation of a large spectrum of substrates
but leads to the consumption of the nitrile or amide addi-
tive. Selective oxidation can be extended to epoxidation of
lower basicity of the solid.
ꢀ 2002 Elsevier Science (USA)
Key Words: oxidative HDS assisted by nitriles; activation of hy-
drotalcite; basic catalysis; kinetics of oxidation; dibenzothiophene.
INTRODUCTION
Environmental concerns are the driving force behind
the goal of removing sulfur compounds from hydrocarbon-
based distillate fuels, initially because of the tendency of the
compounds to produce precursors to acid rain and airborne
particulate material. The main reason for S removal is, how-
ever, that a few parts per million of sulfur are enough to
poison the catalysts used for the purification of the exhaust
gases of diesel cars (1). Several processes were proposed
for solving the problem of removing these compounds from
fuels. The most common industrial process is the treatment
of the fuel with hydrogen at high temperature and high
pressure. This hydrotreating process, hydrodesulfurization
(HDS), is relatively expensive, especially for the fractions
corresponding to diesel fuels, because these feeds contain
the S compounds that are most refractory to HDS. There-
fore, other processes were investigated for the same pur-
pose, and recent research has focused on the oxidation of
sulfur compounds, coupled with the extraction of the re-
1 To whom correspondence should be addressed. Fax: 334 7244 5399.
E-mail: figueras@catalyse.univ-lyon1.fr.
103
0021-9517/02 $35.00
c
ꢀ 2002 Elsevier Science (USA)
All rights reserved.

104
PALOMEQUE, CLACENS, AND FIGUERAS
tivated at 723 K in N2 or air, then desorbed in situ at
473 K in vacuum. Thermal analysis of the solids was per-
formed on a SETARAM apparatus, and the intensity of the
mass 44 characteristic of CO2 was measured during thermal
analysis.
SCHEME 1. Formation and decomposition of peroxycarboximidic
acid.
Catalytic Measurements
unfunctionalized olefins (14) and has recently been applied
to the oxidation of thioethers to sulfones using calcined
HTs (15). The evaluation of the possibilities of basic cata-
lysis for oxidative desulfurization of benzothiophene
derivatives and the comparison of the results with those
of Ti-BEA were of interest. We report here the results
obtained with different solid bases including HT and a
La2O3/MgO mixed oxide of high basic strength.
The oxidation of dibenzothiophene (DBT) and diethyl-
sulfide by hydrogen peroxide was studied in a batch reactor
composed of a triple-necked 50-mL round-bottomed flask,
which was heated to 298–373 K in an oil bath. In a typi-
cal experiment 0.5 mmol of DBT (Aldrich), 10 mL of ace-
tonitrile as the solvent, and 0.51 mL of hydrogen peroxide
(30% in water, Prolabo) were introduced. The reactor was
swept with a flow of nitrogen to remove CO2 from the atmo-
sphere and stirred at 450 rpm. When thermal equilibrium
was reached, the catalyst (0.1 g), freshly activated at 723 K
and rehydrated at room temperature, was introduced; the
progress of the reaction was followed by gas chromatogra-
phy using a DB5 column. An external standard was used
for the chromatographic analysis: an aliquot of the reaction
medium was taken and diluted in methanol containing a
fixed amount of hexanol. With acetonitrile as the solvent
the conversion of DBT exclusively gave sulfones on HT,
such that the conversion of DBT corresponds to the sul-
fone yield, defined as the ratio of moles of DBT converted
to sulfone to moles of DBT introduced.
EXPERIMENTAL
Preparation of the Catalyst Precursors
HTs were prepared by co-precipitation at 338 K and pH 9
according to experimental procedures described in detail
elsewhere (16–19). Aqueous solutions containing the first
0.75 mol/L of MgCl2 · 6H2O and 0.25 mol/L of AlCl3 · 6H2O
(for a Mg/Al ratio = 3) and the second 2 mol/L of KOH
and 0.5 mol/L of K2CO3 were introduced by two electric
pumps to a 4-L flask and mixed vigorously. The mixture
was kept at 338 K for 18 h and was stirred continuously.
The precipitate was washed several times until the solution
was free of chloride ions (AgNO3 test) and then dried at
393 K.
It was also observed that some of the H2O2 was decom-
posed into oxygen:
1
The products of synthesis contain Cl− and CO32− as
charge-compensating anions. They were exchanged by
carbonates using a solution of Na2CO3: 2 g of HT for
100 ml of a 0.2 M solution of Na2CO3 stirred at 343 K for
2 h. A MgLa mixed oxide with a Mg/La atomic ratio of 3
was obtained by co-precipitation of Mg and La nitrates at
a constant pH of 10.
H2O2 → H2O + O2.
2
The material balance of the reaction was established by
analysis of the products of oxidation and by volumetric
measurement of the amount of oxygen evolved during the
reaction. From this experimental measurement, the selec-
tivity S(O2) of decomposition of H2O2 is expressed as a
ratio of the number of moles of oxygen evolved to half the
number of moles of H2O2 engaged in the reaction.
Activation of the Catalysts
Catalysts (about 0.15 g) were first heated to 723 K in a
flow of nitrogen at 10 K/min, and were kept at this temper-
ature for 12 h. In rehydration experiments, the HT was first
decarbonated, cooled to room temperature, and treated in a
flow of nitrogen saturated with the vapor pressure of water
at this temperature. HT-13 (18) represents a sample of HT
calcined at 723 K and treated for 18 h in a flow of nitrogen
saturated by water. The MgLa mixed oxide was activated
at 923 K and used without rehydration.
RESULTS AND DISCUSSION
Characterization of the Catalysts
Table 1 lists the chemical composition of the different
solids. The characterization of HTs was reported elsewhere
(20). The XRD patterns show that all HT samples are pure
and of good crystallinity. The structure of the HTs exists in
the range 1.5 < Mg/Al < 4.5 (17, 19); a sample with a Mg/Al
ratio of 6 must contain an excess of amorphous MgO, which
is undetectable by XRD. The calcination of HTs results in
Characterizations
X-ray powder diffraction patterns were recorded on a mesoporous mixed oxides with a high surface area, and it
Phillips instrument using CuKα 1 radiation. The isotherm was reported that rehydration leads to a considerable loss
for N2 adsorption at 77 K was determined on solids ac- of surface area. Rehydration restores the layered structure

DBT OXIDATION
105
TABLE 1
Chemical Compositions of the Hydrotalcites Used
Sample
Mg/Al solution
Mg/Al solid
Chemical composition
HT-6
2.02
2.99
4
5
6
1.82
2.93
3.90
4.85
5.79
Mg0.64Al0.36(OH)2(CO3)0.17(NO3)0.0004K0.0005, 1.2 H2O
Mg0.75Al0.25(OH)2(CO3)0.15(NO3)0.0003K0.0004, 0.47 H2O
Mg0.8Al0.2(OH)2(CO3)0.18(NO3)0.0067K<0.00002, 0.52 H2O
HT-13
HT-14
HT-15
HT-16
Mg0.83Al0.17(OH)2(CO3)0.19(NO3)0.017
Mg0.85Al0.15(OH)2(CO3)0.24(NO3)0.017
K
<0.00003, 0.24 H2O
K<0.0003, 0.5 H2O
and converts the solid to a low surface area (about 20 m2/g) when the sample was first decarbonated and then rehy-
of meixnerite-like material with OH− as the compensating drated. When the rehydration time is short, activation oc-
anions (21).
curs during the course of the reaction. Constant activity
XRD indicates that the MgLa mixed oxide is a bipha- is observed after about 10 h of rehydration, and the stan-
sic system containing both magnesia and La carbonates dard rehydration time was 18 h. With MgLa-3 as the cata-
(Fig. 1). The surface area of this mixed oxide is 45 m2/g lyst oxidized sulfur compounds were not observed and the
after calcination at 923 K. The solid probably consists of a only reaction product was oxygen produced from hydro-
layer of La oxide deposited on magnesia. The basic proper- gen peroxide with a selectivity close to 90–100%. During
ties can be evaluated from the thermal desorption of CO2 pretreatment the solids undergo decarbonation and dehy-
measured by mass spectrometry (Fig. 2) for two HTs and droxylation; the low activity of the calcined oxides suggests
the MgLa sample. This analysis showed that MgLa is decar- that basic hydroxyls are required for the oxidation reaction.
bonated at about 875 K and is a stronger base than either
Influence of the Reaction Conditions
HT-13 or HT-15, which are decarbonated below 700 K. The
intensity of the CO2 peak, however, is much more intense
on HT-13 than on HT-15, and shows a higher number of
sites on HT-13.
The catalyst used in this study was HT-13 (Mg/Al ratio =
2.9), rehydrated for 18 h, with a small surface area and no
microporosity (21). As reported earlier the reaction yields
the sulfone and oxygen only. Small amounts of acetamide
were detected (Scheme 1).
Effect of the Activation Conditions
The influence of hydrogen peroxide and DBT concen-
trations on the initial rate of the reaction is shown in Figs. 4
and 5. The reaction is initially first-order, and the rate goes
through a maximum as a function of the concentration of
DBT, suggesting competition between the two reactants
The effects of the activation conditions were investigated
under the following conditions: 0.05 M substrate, 10 g/L
catalyst, and 10 mol/mol hydrogen peroxide/substrate,
333 K reaction temperature (Fig. 3). Nondecarbonated HT
and the mixed oxide obtained after heat treatment in nitro-
gen or air show low activity. Catalytic activity increased
FIG. 2. Variations in the intensity of the mass 44 peak, normalized to
1 g of sample, as a function of temperature for MgLa-3 mixed oxide and
two HTs.
FIG. 1. Powder XRD pattern of MgLa = 3 mixed oxide.

106
PALOMEQUE, CLACENS, AND FIGUERAS
FIG. 3. Effect of time of rehydration on catalytic activity at 333 K of
HT-13 calcined at 723 K.
FIG. 6. Effect of reaction temperature on the initial rate of DBT ox-
idation on HT-13.
for adsorption. Fig. 6 shows the effect of the reaction tem-
perature on the initial rate, measured with a mixture of
DBT/H2O2 (0.5/2.5). The initial rate reaches a maximum
at about 333 K. The selectivity for the decomposition of hy-
drogen peroxide to oxygen increased with the reaction tem-
perature (Table 2). At the higher temperatures, the rate of
decomposition of hydrogen peroxide was fast and affected
the initial rate.
The decomposition of hydrogen peroxide was relatively
slow on the uncalcined sample, higher on the mixed oxides,
and decreased after rehydration. The LaMg mixed oxide
showed selectivity to O2 above 90%. The faster decompo-
sitiononabaseasstrongasLaMgandcalcinedHT(Table2)
is in agreement with the basic mechanism represented by
Scheme 1.
FIG. 4. Effect of DBT concentration on the rate of HT-13, at 333 K
and a constant concentration of hydrogen peroxide (0.25 M).
TABLE 2
Selectivity of Formation of Oxygen after 90-min Reactions
at Different Temperatures Using 0.5 mmol DBT
Reaction temperature
(K)
Oxygen selectivity
(%)
Catalyst
HT-13 as synthesized
HT-13 calcined
333
333
17
53
88 a
30
47
73
67
50
40
53
97
HT-13 (18)
HT-13 (18)
HT-6 (18)
HT-14 (18)
HT-15 (18)
HT-16 (18)
HT-13 (18)
HT-13 (18)
298
313
333
333
333
333
353
373
FIG. 5. Effect of concentration of hydrogen peroxide on the initial
rate of HT-13 at 333 K and a constant concentration of DBT (0.05 M).
a After 3 h.

DBT OXIDATION
107
O2 formation must then be related to the specific chemical
properties of acetonitrile, and step 2 of Scheme 1 is a well-
established path for this reaction. Since step 1 is slow, the
reactivity of the nitrile and the basic properties of the solid
would then control O2 formation.
Effect of the Basicity of the Catalyst
ThebasicpropertiesofMgAlHTchangewhentheMg/Al
ratio and the heat treatments are changed. The uncalcined
HT-13 sample showed low activity for oxidation of DBT
(Fig. 3), suggesting that basicity is involved in the reaction.
The rate of isomerization of isophorone enables the deter-
mination of the number of basic sites of HT in the liquid
phase after rehydration, in conditions close to those of the
reaction. The effects of the Mg/Al ratio on the rates of DBT
oxidation, the final yield in sulfone, and rate of isophorone
isomerization are given in Fig. 8. At a ratio Mg/Al of about
3, isophorone isomerization shows maximum activity as ob-
served for many other reactions such as Knoevenagel con-
densation (23, 24) or aldolization (25). In contrast, HT oxi-
dation activity increases steadily with the Mg content up to
Mg/Al = 6. The samples with Mg/Al ratios higher than 5 are
not single phases, and HT-16 (Mg/Al ratio = 5.7) is proba-
bly composed of a surface layer of amorphous magnesia
supported on HT, and probably shows the basic properties
of magnesia. Modification of catalyst composition affects
the final yield to a greater extent than the initial rate; this
is attributed to a lower selectivity for H2O2 decomposition
at high Mg/Al ratios.
FIG. 7. Effect of the nature of the solvent on the oxidation of DBT
at 333 K.
The effects of the solvent were investigated by comparing
methanol, acetonitrile, benzonitrile, acrylonitrile (ACN),
and 3-methoxypropionitrile (3MPN). The results are re-
ported in Fig. 7 and Table 3. No reaction was observed with
methanol as the solvent, even after partial rehydration for
3 h. Small amounts of 1-hexanol (1 mmol/10 mL of solvent
in acetonitrile as the solvent) also had a detrimental effect.
Since the mechanism involves the formation of perhydroxyl
anions from hydrogen peroxide, competition between the
alcohol and the hydrogen peroxide for adsorption at the
surface can account for this inhibition. A comparison of
the different nitriles shows that acetonitrile, benzonitrile, or
3-methoxypropionitrile give high final yields, with a higher
reaction rate with acetonitrile. Dumitriu et al. (15) reported
that the decomposition of H2O2 occurred in acetonitrile in
the absence of thioether. Moreover, the decomposition of
H2O2 was not observed during 3-cyclohexen-1-one epox-
idation performed on HT in toluene as the solvent (22).
TABLE 3
Influence of the Solvent on the Conversion of DBT on HT-13
at 333 K with H2O2/substrate ratio = 10
React
time
(h)
DBT
Sulfone
conv. O2 selectivity
Catalyst
Solvent selectivity (%)
(%)
HT-13 uncalcined
HT-13 calcined
HT-13 calcined a
HT-13 (18)
HT-13 (18)
HT-13 (18)
3
3
6
1.5
3
3
CH3CN
CH3CN
CH3OH
CH3CN
C6H5CN
3MPN
100
100
0
100
100
100
47.3
32
0
96
100
86
17
53
33
20
18
22
FIG. 8. Initial rates of oxidation and conversion after 90 min of DBT
compared to the rate of isomerization of isophorone on HTs of different
Al contents, which were calcined and then rehydrated for 18 h.
HT-13 (18)
3
ACN
17.6
a H2O2/substrate ratio = 5.

108
PALOMEQUE, CLACENS, AND FIGUERAS
Comparison of Ti-Catalysts with Hydrotalcites
5. Treiber, A., Dansette, P. M., El Amri, H., Girault, J. P., Ginderow,
D., Mornon, J. P., and Mansuy, D., J. Am. Chem. Soc. 119, 1565
(1997).
6. Otsuki, S., Nonaka, T., Takashima, N., Qian, W., Ishihara, A., Imai, T.,
and Kabe, T., Energy Fuels 14, 1232 (2000).
7. Hulea, V., Fajula, F., and Bousquet, J., J. Catal. 198, 179 (2001).
8. Schirmann, J.-P., and Delavarenne, S. Y., “Hydrogen Peroxide in
Organic Chemistry.” Edition et Documentation Industrielle, Paris,
1979.
9. Payne, G. B., Deming, P. H., and Williams, P. H., J. Org. Chem. 26, 659
(1961).
10. Sawaki, Y., and Ogata, Y., Bull. Chem. Soc. Jpn. 54, 793 (1981).
11. Vacque, V., Dupuy, N., Sombret, B., Huvenne, J.-P., and Legrand, P.,
J. Mol. Struct. 384, 165 (1996).
12. Aramendia, M. A., Borau, V., Jimenez, C., Luque, J. M., Marinas, J. M.,
Ruiz, J. R., and Urbano, F. J., Appl. Catal., A 216, 257 (2001).
13. Yamaguchi, K., Ebitani, K., and Kaneda, K., J. Org. Chem. 64, 2966
(1999).
14. Ueno, S., Yamaguchi, K., Yoshida, K., Ebitani, K., and Kaneda, K.,
Chem. Commun. 295 (1998).
15. Dumitriu, E., Guimon, C., Cordoneanu, A., Casenave, S., Hulea,
T., Chelaru, C., Martinez, H., and Hulea, V., Catal. Today 66, 529
(2001) .
The comparison can be made using the results reported
by Hulea et al. (7) at 333 K for Ti-BEA in acetonitrile as
the solvent; the DBT concentration was 0.0645 M, catalyst
concentration was 3.22 g/L, and H2O2/DBT was equal to
2.4 mol/mol. Since conversion reached 34% in 120 min,
the rate was 5.6 × 10−5 mol min−1 g−1 for Ti-BEA. For
HT, the first point (Fig. 5) was measured at 333 K in ace-
tonitrile solvent; the substrate concentration was 0.05 M,
catalyst concentration 10 g/L, and H2O2/DBT was equal to
2.5mol/mol. Theratewas1.2 × 10−5 molmin−1 g−1. Thereal
difference is smaller, because hydrotalcite loses about 50%
weight upon calcination. The specific activity per square
meter would favor HT (surface area 20 m2/g) compared to
Ti-BEA (470 m2/g).
In conclusion, the oxidation of DBT to sulfone is pos-
sible by a base-catalyzed mechanism on HT, using a solid
that was calcined and then rehydrated, suggesting that the
catalytic sites are probably hydroxyls. The best solvent is
acetonitrile, which facilitates the possibility that peroxycar-
boximidic acid is an intermediate. The activity reached by
this process is comparable to that obtained with Ti catalysts.
16. Miyata, S., Clays Clay Miner. 23, 369 (1975).
17. Cavani, F., Trifiro, F., and Vaccari, A.,Catal. Today 11, 173 (1991).
18. De Roy, A., Forano, C., El Malki, K., and Besse, J. P., in “Synthesis of
Microporous Materials” (M. L. Occelli and H. Robson, Eds.), Vol. 2,
p. 108. Van Nostrand-Reinhold, New York, 1992.
19. Miyata, S., Clays Clay Min. 28, 50 (1980).
20. Sanchez Valente, J., Figueras, F., Gravelle, M., Lopez, J., and Besse,
J.-P., J. Catal. 189, 370 (2000).
21. Lopez, J., Jacquot, R., and Figueras, F., Proc. Int. Cong. Catalysis,
Granada 2000 (A. Corma, F.V. Melo, S. Mendioroz, and J. L. G. Fierro,
Eds.); Stud. Surf. Sci. Catal. 130, 491 (2000).
ACKNOWLEDGMENT
J. Palomeque thanks the Mexican Institute of Petroleum for a Ph.D.
grant.
REFERENCES
22. Palomeque, J., Lopez, J., and Figueras, F., J. Catal., in press.
23. Corma, A., Forne´s, V., Martin-Aranda, R. M., Garcia, H., and Primo,
J., Appl. Catal. 59, 237 (1990).
24. Corma, A., Forne´s, V., Martin-Aranda, R. M., and Rey, F., J. Catal.
134, 58 (1992).
25. Figueras, F., and Lopez, J., in “Fine Chemicals through Heterogeneous
Catalysis” (R. A. Sheldon and H. Van Bekkum, Eds.), p. 327. Wiley-
VCH, Weinheim, 2000.
1. Guyon, M., Blejean, F., Bert, C., and Le Faou, Ph., Soc. Automot. Eng.,
SP-1399, 87 (1998).
2. Bonde, S. E., Chapados, D., and Gore, W. L., NPRA meeting, San
Antonio, TX, 26–28 March, 2000.
3. Gore, W., U.S. Patent No. 6160193 (1998).
4. Brown, K. N., and Espenson, J. H., Inorg. Chem. 35, 7211 (1996).