Influences of the hydrogen sulfide partial pressure and of a nitrogen compound on the hydrodeoxygenation activity of a CoMo/carbon catalyst

Article abstract of DOI:10.1006/jcat.2000.3103

The effect of H₂S partial pressure and of a basic nitrogen compound (diaminopropane) on the activity of a sulfided CoMo/carbon catalyst in hydrodeoxygenation were studied at 10-150 kPa H₂S, 7 MPa total pressure, and at 270° and 200°C in the transformation of a mixture of guaiacol (GUA), ethyldecanoate (ED), and 4-methylacetophenone (MA) in p-xylene. An increase in H₂S partial pressure decreased the average degree of reduction and uncoordination of the active sites and stopped the hydrogenolysis and hydrogenation pathways. H₂S did not influenced the CoMo/C catalyst activity for the overall conversion of GUA. ED conversion was not influenced by H₂S, and the selectivity in hydrogenated and decarboxylated products was slightly changed. MA was the only reactant molecule whose conversion was inhibited by H₂S.

Full text of DOI:10.1006/jcat.2000.3103

Journal of Catalysis 198, 47–55 (2001)
doi:10.1006/jcat.2000.3103, available online at http://www.idealibrary.com on
Influences of the Hydrogen Sulfide Partial Pressure and of a Nitrogen
Compound on the Hydrodeoxygenation Activity
of a CoMo/Carbon Catalyst
Maria Ferrari, Rosanna Maggi, Bernard Delmon, and Paul Grange1
Unite´ de catalyse et chimie des mate´riaux divise´s, Universite´ catholique de Louvain, Croix du Sud, 2 bte 17, B-1348 Louvain-la-Neuve, Belgium
Received June 13, 2000; revised October 2, 2000; accepted October 30, 2000; published online February 7, 2001
oxygenated compounds and water (6–8), or reduction of
the sulfided phase by hydrogen, depending on the operat-
ing conditions.
The influences of the hydrogen sulfide partial pressure and of
a basic nitrogen compound (diaminopropane) on the activity of a
sulfided CoMo/carbon catalyst in hydrodeoxygenation were stud-
ied from 10 to 150 kPa H2S, under a total pressure of 7 MPa, at 270
and 200 C in the transformation of a mixture of guaiacol, ethylde-
canoate, and 4-methylacetophenone in p-xylene. It is likely that
three different active sites are responsible for hydrogenolysis, hy-
drogenation, and acid-catalysed reactions. An increase in hydrogen
sulfide partial pressure decreases the average degree of reduction
and uncoordination of the active sites and inhibits the hydrogenol-
ysis and hydrogenation pathways, the first being affected to a larger
extent. This accounts for the decrease of guaiacol dehydroxylation
and the increase of 4-methylacetophenone conversion. The reac-
tions catalysed by acid sites located on the metal sulfide, namely
decarboxylation and alcohol dehydration, are promoted by hydro-
gen sulfide. The addition of diaminopropane strongly inhibits the
conversion of guaiacol and ethyl decanoate. The decarboxylation
c
HDO reactions are deeply influenced by hydrogen sul-
fide. In the conversion of 4-methylphenol the direct hy-
drogenolysis of the Caromatic–O bond is more strongly inhib-
ited than the hydrogenation of the aromatic ring followed
by hydrogenolysis (9–11). In the simultaneous hydrodesul-
furization(HDS) and HDO of phenol and mercaptophenol
(12, 13) the HDO of a hydroxyl group was suppressed by
the presence of a mercapto group either in the same or in
a separate molecule.
Depending on the origin of biomass, bio-oils are charac-
terised by a nitrogen content between 0.2 and 1 wt% (4, 5).
Nitrogen-containing molecules, like ammonia or amines,
are strong inhibitors of hydrogenation (HYD) (14–16) and
HDS (17, 18) reactions. They can interact with incom-
pletely coordinated molybdenum atoms, or with –SH acid
sites. HDO reactions are also affected by nitrogen com-
pounds. The HDO of phenols is inhibited by quinoline, o-
ethylaniline (19), and indole (20). Ammonia, generated by
decomposition of ammonium carbonate (10), or diamino-
propane (9) inhibits the HDO of 4-methylphenol.
selectivity is decreased as well.
2001 Academic Press
Key Words: hydrodeoxygenation; hydrogen sulfide; sulfided
catalyst; carbon support.
1. INTRODUCTION
A double objective will be pursued in this work: to im-
prove the understanding of HDO reactions and to inves-
tigate the modifications of the active sites occurring under
different reaction conditions. Our results will be supported
by comparisons with previous work performed in the same
field (9, 21, 22). The use of activated carbon instead of alu-
mina as support should make interpretation easier, because
there is no longer a contribution of the support acidity to
the conversion of the reactant. The role of the sulfhydryl
groups should be more easily understood.
Liquids derived from the pyrolysis of biomass, such as
bio-oils (1), should be upgraded to enable storage, trans-
portation, and use. Hydrodeoxygenation (HDO) could be
adequate for this purpose (2, 3). The reactions involved
are quite similar to those occurring in the hydrotreating of
petroleum fractions.
The present work deals with the influences of the hydro-
gen sulfide partial pressure and of the addition of a basic
nitrogen compound, i.e., diaminopropane, on the HDO ac-
tivity of a carbon-supported sulfided CoMo catalyst.
As biomass-derived liquids do not contain sulfur com-
pounds (4, 5), a partial pressure of hydrogen sulfide is
needed to preserve the sulfided catalyst from oxidation by
2. EXPERIMENTAL
2.1. The Catalyst
1 To whom correspondence should be addressed. Fax: +32-10-47.36.49.
The catalyst was prepared using a commercial activated
carbon, BKK-100 (particle size diameter between 0.3 and
E-mail: grange@cata.ucl.ac.be.
47
0021-9517/01 $35.00
c
Copyright
2001 by Academic Press
All rights of reproduction in any form reserved.

48
FERRARI ET AL.
0.5 mm, specific surface area 1100 m2 g 1, pore volume
Three series of experiments were done to study the in-
0.7 cm3 g 1). The support was impregnated with aqueous fluence of hydrogen sulfide. In the first series, four tests
solutions of ammonium heptamolybdate and cobalt nitrate, were performed at different hydrogen sulfide partial pres-
using the incipient wetness method. Molybdenum was im- sures (50, 75, 100, and 150 kPa), using a fresh catalyst for
pregnated first. After each impregnation, the sample was each test. In the second, called “PH S increase”, the test was
2
dried overnight at 130 C and then at 400 C for 3 h under an done at increasing hydrogen sulfide partial pressures (10,
argon flow. The composition of the catalyst was 15 wt% of 25, 35, 50, 75, and 150 kPa) without changing the catalyst
MoO3 and 3 wt% of CoO. This catalyst has been extensively and always waiting 24 h before modifying the operating
characterised by N2 physisorption, XPS, and SEM (23). The conditions (temperature and hydrogen sulfide partial pres-
textural characteristics of the catalyst were specific surface sure). Finally, after the activity was measured at a hydrogen
1
area, 677 m2 g 1, and pore volume, 0.4 cm3 g
.
sulfide partial pressure of 150 kPa, this pressure decreased
to 10 kPa and the activity was measured again. The last
point at which the initial conditions were restored is indi-
cated as 10 kPa bis. In the third series of experiments, called
2.2. The Catalytic Tests
The tests were carried out in a fixed-bed continuous-flow
tubular reactor (internal diameter, 1.75 cm; length, 70 cm).
A 1.5 g portion of catalyst (particle diameter between 0.3
and 0.5 mm) was diluted with glass spheres (diameter be-
tween 0.2 and 0.45 mm) to reach a volume of 18 cm3. The
catalytic bed (8 cm height) was loaded in the reactor be-
tween two plugs of glass wool. The rest of the reactor was
filled with 1 mm diameter glass spheres.
“PH S decrease,” the test was done at decreasing hydrogen
2
sulfide partial pressures (150, 100, 50, and 25 kPa) without
changing the catalyst and waiting 24 h before modifying the
operational conditions. Finally, after the activity was mea-
sured at 25 kPa, the hydrogen sulfide partial pressure was
increased back to 150 kPa (150 kPa bis) for a last series of
product analyses.
Two reaction temperatures were used for the test: 270 C
in order to evaluate the reactivity of GUA and ED and
200 C in order to evaluate the reactivity of MA.
Liquid samples, collected before the back pressure reg-
ulator every hour, were analysed by using a gas chromato-
graph equipped with an FID detector and a capillary col-
umn (stationary phase CP-Sil-8CB; length, 25 m).
Conversion, yield, and selectivity data reported here
were obtained in the period corresponding to 18 to 24 h
of reaction, where conversion and selectivities are nearly
constant. The error in the experimental values of the con-
version and selectivity was estimated at 10–15% .
The HDO pathways of the model compounds are re-
ported in Fig. 1.
Prior to the reaction, the catalyst was dried at 130 C
overnight, under nitrogen flow. The sulfidation mixture,
15 vol% H2S in H2, was then introduced, with the pres-
sure set at 0.3 MPa and the flow at 150 ml min 1; after a
30 min stabilisation period, the temperature was increased
to 400 C (3 C min 1). These conditions were held for 3 h.
The temperature was subsequently decreased to 270 C, the
gas switched to pure hydrogen, and the liquid feed intro-
duced. The pressure was progressively raised to 7 MPa in
1
1 h. The liquid feed rate was 45 ml h and the hydrogen
flow rate 24 l h 1. The tests were performed at a total pres-
sure of 7 MPa. This procedure was also followed for the
experiments performed without catalyst.
The decomposition of the hydrogen sulfide precursor,
dimethyl disulfide (DMDS), in the absence and in the
presence of the catalyst (thermal and catalytic reaction of
DMDS, respectively) was first investigated. We also studied
the reaction of the oxygenated model compound mixture
in the absence of the catalysts. For the study of the thermal
and catalytic reactions of DMDS, the liquid feed contained
DMDS (7.2 wt% , 0.67 mol l 1) and pentadecane (2 wt% ,
0.082 mol l 1) as internal standard for the chromatographic
analysis; p-xylene was used as solvent.
3. RESULTS
3.1. Preliminary Tests
The thermal decomposition of DMDS starts at a tem-
perature above 200 C and is completed at 300 C. DMS ap-
pears only at 270 C. The catalytic decomposition of DMDS
(Fig. 2) starts at a much lower temperature and at temper-
atures higher than 180 C the decomposition is complete.
Methylmercaptan is formed at 150 C.
The conversion of ED and MA at 270 C without the cata-
lyst is always lower than 1% , while the conversion of GUA
is 1.5% . The conversion of MA at 200 C is lower than 1% .
For the study of the HDO reactions, the composition of
the mixture of oxygenated model compounds was as fol-
lows: guaiacol (GUA), 3 wt% , 0.21 mol l 1; ethyl decanoate
(ED), 3 wt% , 0.13 mol l 1; 4-methylacetophenone (MA),
1
3 wt% , 0.20 mol l 1; and pentadecane, 2 wt% , 0.082 mol l
in p-xylene. DMDS was added as a precursor of H2S; its
concentration depended on the hydrogen sulfide partial
pressure required for the test. For the test done with di-
3.2. Influence of the Hydrogen Sulfide Partial Pressure
The GUA conversion was not affected by hydrogen
aminopropane (DAP), its concentration was 0.29 wt% , sulfide and almost similar values were obtained in the three
0.034 mol l 1, and the concentration of DMDS was 1.8 wt% , series of experiments. An important molar balance deficit
1
0.17 mol l
.
was observed; the sum of the product yields calculated on

INFLUENCE OF H2S PARTIAL PRESSURE ON HDO REACTIONS
49
FIG. 1. Hydrodeoxygenation pathways of 4-methylacetophenone, ethyl decanoate, and guaiacol.
the basis of gas chromatographic analyses corresponded
For the first series of tests performed with a fresh cat-
to 60–80% of the reactant conversion. Actually, when the alyst for each PH S, we report in Table 1 the conversion,
2
liquid samples were withdrawn from the reactor, a white molar balance, and product yields after 2 h of reaction, as
dense unidentified phase was observed. It is known that well as the mean values obtained from the analysis of three
phenolic compounds with two oxygenated substituents are
often involved in condensation reactions (24, 25). The gas
phase, which could contain aromatic and alicyclic com-
pounds, such as benzene, cyclohexene, and cyclohexane,
was not analysed.
TABLE 1
GUA Conversion, Molar Balance, and Yields of C6 Hydrocar-
bons (Benzene, Cyclohexene, Cyclohexane), Phenol, and Catechol,
for the First Series of Tests, Performed at Different P H S. (Total
2
Pressure, 7 MPa; Temperature, 270 C)
Yield (% )
Conv GUA Molar balance
PH
(% )
(% )
C6 Ph Cat
S
2
50 kPa 2 h
Nearly constant
state
75 kPa 2 h
20
19
82
67
5.0 4.6 7.0
4.5 3.9 4.3
22
19
76
67
4.5 4.4 6.8
4.8 3.8 4.0
Nearly constant
state
100 kPa 2 h
23
21
68
57
4.4 3.8 7.4
4.2 3.0 4.8
Nearly constant
state
150 kPa 2 h
20
17
88
83
4.2 3.9 9.6
3.7 2.9 7.4
Nearly constant
state
FIG. 2. Evolution of the DMDS and DMS concentrations in the liquid
phase in the catalytic decomposition of DMDS. The presence of MM is
also indicated (#), but a quantification of this product was not done. Total
pressure, 7 MPa.
Note. The values of conversion and yield were measured after 2 h of
reactions and at a nearly constant state.

50
FERRARI ET AL.
FIG. 3. Evolution of catechol and phenol + C6 yields as a function
of PH for the three series of experiments. Reaction temperature, 270 C.
Total pressure, 7 MPa.
FIG. 5. Evolution of C9 (nonane), C10 (decane + decene), and de-
canol yields as a function of PH for the three series of experiments.
Reaction temperature, 270 C. Total pressure, 7 MPa.
S
2
S
2
samples withdrawn in the period between 18 and 24 h
(nearly constant state).
The evolution of the yields in catechol and the sum of
phenol and C6 (which includes benzene, cyclohexene, and
at 150 kPa; in this case, the catechol yield is lower at 150 kPa
bis than at 150 kPa.
The conversion of ED was not affected by hydrogen sul-
fide. The sum of the product yields always corresponded
to 95% ( 5% ) of the ED conversion. The evolution of
the yields in nonane, decanol, and molecules containing
cyclohexane) as a function of PH S is shown in Fig. 3. For the
2
three series of experiments, phenol and C6 yields decrease
with increasing PH S, while catechol yields increase. As a
2
10 atoms of carbon (C10), as a function of PH S, is repre-
2
consequence, the phenol + C6-to-catechol ratio decreases
sented in Fig. 5. In the “PH S increase” series, nonane yields
2
with increasing PH S (Fig. 4). In the case of the “PH S in-
2
2
increase at low PH S; for PH S higher than 50 kPa they are
2
2
crease” series, higher ratios are obtained than in the other
constant. For the two other test series, the yields in decar-
boxylated products are almost constant with the increase of
hydrogen sulfide. C10 yields increase, while decanol yields
two test series. Only for the “PH S increase” series is this
2
ratio lower than 1. For the two other series, the phenol +
C6 amount is higher than that of catechol, except at high
PH S (150 kPa). In the “PH S increase” series, the phenol +
decrease with the increase of PH S. With the increase of
2
2
2
PH S the dehydration of the alcohol is favored. This behav-
2
C6-to-catechol ratio after 300 h of reaction (10 kPa bis) is
lower than the one observed at the beginning of the test
ior is more evident for the “PH S increase” series and at low
2
hydrogen sulfide partial pressure.
The product yields and the decarboxylation selectivity
(10 kPa). In contrast in the “PH S decrease” series, this ra-
2
tio, after 200 h of reaction (150 kPa bis), is higher than that
are reported in Table 2. In the “PH S increase” series, a
2
higher nonane yield is observed at 10 kPa bis than at
10 kPa; this results in a higher value for the decarboxylation
TABLE 2
Nonane (C9), Decane + Decene (C10), and Decanol Yields
and Decarboxylation Selectivity for the Conversion of ED for the
“P H Increase” and “P H Decrease” Series (Reaction Tempera-
2S
2S
ture, 270 C; Total Pressure, 7 MPa)
Yield (% )
H2S partial
pressure
Decane +
Decarboxylation S
Nonane
decene
Decanol
(% )
10 kPa
10 kPa bis
150 kPa
4.7
6.3
4.7
3.4
9.2
9.6
12.7
13.9
7.1
6.8
1.1
1.0
22
28
26
19
FIG. 4. Evolution of the phenol + C6-to-catechol ratio as a function
of PH for the three series of experiments. Reaction temperature, 270 C.
S
2
150 kPa bis
Total pressure, 7 MPa.

INFLUENCE OF H2S PARTIAL PRESSURE ON HDO REACTIONS
51
TABLE 3
TABLE 4
GUA Conversion, Molar Balance, Phenol + C6 and Catechol
Yields, and Phenol + C6/Cat Ratio in the Absence and in the
MA Conversion at Different P H (Reaction Temperature,
2S
200 C; Total Pressure, 7 MPa)
Presence of DAP (P H S, 100 kPa; Reaction Temperature, 270 C;
2
Total Pressure, 7 MPa)
H2S partial
pressure
Second series,
Third series,
First series
“PH increase”
“PH decrease”
S
S
2
2
Yield (% )
DAP
Conv
Molar
Ph /Cat
ratio
25 kPa
35 kPa
50 kPa
31% (2)
28% (3)
23% (4)
21% (5)
34% (4)
26% (3)
concentration GUA (% ) balance (% ) Phenol Catechol
22%
19%
19%
12%
1a
0 mol l
0 mol l
21
23
12
57
62
32
7.1
10.1
2.8
4.7
4.1
1.1
1.5
2.5
2.5
75 kPa
1
100 kPa
150 kPa
150 kPa bis
18% (2)
12% (1)
12% (5)
1
0.034 mol l
15% (6)
a Results obtained with a catalyst with the same composition, but pre-
pared at a different time and used in the series of tests for the study of the
influence of the hydrogen sulfide partial pressure.
selectivity. In the “PH S decrease” series a lower nonane
yield is observed at 150 kPa than at 150 kPa.
2
3.3. Influence of a Basic Nitrogen Compound:
Diaminopropane (DAP)
The results for the conversion of MA are reported
in Table 3. The hydrogenation of the carbonyl group
of MA was always inhibited by hydrogen sulfide. The
sum of the product yields always corresponded to 95–
100% ( 5% ) of the reactant conversion. The selectivities
in 4-ethylmethylbenzene and ,4-dimethylbenzyl alcohol,
as a function of the hydrogen sulfide partial pressure,
are represented in Fig. 6; they show different tendencies
for the three series of tests. In the tests performed with
a fresh catalyst for each hydrogen sulfide partial pres-
sure, the 4-ethylmethylbenzene selectivities are almost con-
stant, while the selectivity in the alcohol does not show
The GUA conversion and the molar balance deficit
sharply decreased (by about 50% ) when DAP was added
(Table 4). The phenol + C6 ratio did not seem to be af-
fected. ED conversion decreased when DAP was added
(Table 5). Decanoic acid yields were not affected, but the
decarboxylation selectivity strongly decreased. The conver-
sion of MA increased when DAP was added (Table 6), but
the molar balance was largely deficient and some unidenti-
fied peaks appeared in the chromatogram. However, even if
identified, they cannot account for the molar balance deficit
as their area is very small.
any clear trend. For the “PH S increase” series, the 4-
2
ethylmethylbenzene and the ,4-dimethylbenzyl alcohol
4. DISCUSSION
selectivities are constant. In the “PH S decrease” series the
2
4-ethylmethylbenzene selectivity increases, while the selec-
tivity in ,4-dimethylbenzyl alcohol decreases as a func-
tion of PH S. Comparing the “PH S increase” and the “PH S
decrease” series, higher 4-ethylmethylbenzene selectivities
are observed in the first case.
4.1. Thermal and Catalytic Reaction of DMDS
In the literature, it is reported that disulfides decom-
pose easily (26). DMDS decomposes more easily than DMS
in the presence of hydrogen, under atmospheric pressure
on hydrotreating catalysts, producing primarily methylmer-
captan at a temperature as low as 170 C; the maximum
yields reach 70% at 220 C while hydrogen sulfide produc-
tion starts at about 200 C. Under pressure, the decomposi-
tion of DMDS into Methylemercaptan can start even at
lower temperature (27). Our results confirm that in the
presence of a sulfided CoMo/C catalyst the decomposi-
tion of DMDS starts below 150 C and we calculated the
2
2
2
PH S values considering that DMDS completely decom-
2
poses into H2S.
4.2. The Influences of the Hydrogen Sulfide Partial
Pressure and of the Presence of Diaminopropane
on the HDO Activity: Identification of the Active
Sites Involved in HDO Reactions
FIG. 6. Evolution of 4-ethylmethylbenzene (EtMeB) and ,4-
We will compare our results with those obtained with
alumina-supported catalysts, presented previously (9, 21,
22).
dimethylbenzyl alcohol (MeBeAl) selectivities as a function of PH S for the
2
three series of experiments. Reaction temperature, 200 C. Total pressure,
7 MPa.

52
FERRARI ET AL.
TABLE 5
ED Conversion, Molar Balance, Nonane, Decane + Decene, Decanol, and Decanoic Acid Yields,
and Decarboxylation Selectivity in the Absence and in the Presence of DAP (P H S, 100 kPa; Reaction Tem-
2
perature, 270 C; Total Pressure, 7 MPa)
Yield (% )
Conv
ED
(% )
Molar
balance
(% )
DAP
concentration
Decane +
Decanoic
acid
Nonane
decene
Decanol
Decarboxylation S (% )
1a
0 mol l
0 mol l
23
27
8
96
102
88
5.9
5.4
0.3
12.5
16.1
3.7
2.5
3.6
1.0
2.5
2.1
2.2
28.3
21.4
5.8
1
1
0.034 mol l
a Results obtained with a catalyst with the same composition, but prepared at a different time and used in the series
of tests for the study of the influence of the hydrogen sulfide partial pressure.
4.2.1. Guaiacol. The conversion of GUA initially in- demethylation. Compared to alumina, carbon-supported
volves the breaking of the O–methyl bond (demethyla- catalysts are less active for the breaking of the O–methyl
tion) to give catechol. The reaction then proceeds through bond, but more active for the Caromatic–O bond hydrogenol-
hydrogenolysis of the Caromatic–O bond (dehydroxylation), ysis. With the CoMo/C catalyst, the total GUA conversion
first to phenol and then to benzene, cyclohexene, and cy- is constant, but the formation of dehydroxylated products,
clohexane (21, 28, 29). Phenol can also be directly formed phenol and C6, decreased with increasing the hydrogen sul-
from guaiacol by a demethoxylation reaction (hydrogenol- fide partial pressure. A molar balance deficit was always
ysis of the Caromatic–O bond) (30). In our experiments the observed and not influenced by the increase of the hydro-
main products were catechol, phenol or cyclohexene, and gen sulfide partial pressure. A previous work (24) showed
cyclohexene.
that phenolic-type molecules with two or more oxygenated
Previous results indicated that the HDO of 4-methyl- substituents are the major cause of coke formation during
phenol to toluene over sulfided CoMo/Al2O3 and HDO ofbio-oilson alumina-supported catalysts. The mech-
NiMo/Al2O3 catalysts is strongly inhibited by hydrogen sul- anism involved would be similar to that responsible for the
fide (9) whereas the demethylation of GUA was not af- synthesis of phenol–formaldehyde resins.
fected by hydrogen sulfide, nor by water. These results tend
As for the formation of dehydroxylated products, it is
to exclude the participation of Brønsted acid sites. Actually logical to consider possible analogies between the break-
it is reasonable to expect that Brønsted acidity increases by ing of the C–O and C–S bonds. Hydrogenolysis reactions
adsorption of water and especially hydrogen sulfide. In con- are always more inhibited than hydrogenation ones. The
trast, a strong inhibition by ammonia, combined with the cleavage of the Caromatic–O bond thus seems to correspond
fact that alumina alone had some activity for GUA conver- to a hydrogenolysis reaction and it would take place on
sion (about half that of the sulfided CoMo catalyst), giving strongly reduced molybdenum atoms with a high degree
catechol as single product, indicated that the acid sites of of coordinative unsaturation (B sites). Upon increasing the
the alumina support (CS sites) were mainly responsible for hydrogen sulfide partial pressure, the average degree of re-
duction and the coordinative unsaturation of the active sites
decrease and the hydrogenolysis of Caromatic–heteroatoms
bonds is consequently inhibited. Two different hypotheses
TABLE 6
MA Conversion, Molar Balance, and 4-Ethylmethylbenzene and
,4-Dimethylbenzyl Alcohol Yields in the Absence and in the Pres-
about the reaction and the active sites involved may be pro-
posed. If demethylation is considered as the first reaction
step and only a consecutive pathway is assumed for the
transformation of GUA to catechol and then to phenol, it
ence of DAP (P H S, 100 kPa; Reaction Temperature, 200 C; Total
2
Pressure, 7 MPa)
must be excluded that this reaction involves “classical” hy-
Yield (% )
DAP
concentration
Conv MA
(% )
Molar
balance (% )
drogenation or hydrogenolysis sites (A or B sites), or acid
sites on the metal sulfide (CM sites). In the first case the
conversion would be inhibited by an increase of hydrogen
sulfide, while in the second case it would be enhanced. In
the case of carbon-supported catalyst, as nonoxidized car-
bon contains very low amounts of phenolic and carboxylic
acid groups (31) and the the overall acidity in the support
(the total number of acid sites is less than 5.3 mequiv/g of
EtMeB
MeBeAl
1a
0 mol l
0 mol l
21
13
45
99
108
36
12.3
8.4
15.5
8.6
5.5
0.8
1
1
0.034 mol l
a Results obtained with a catalyst with the same composition, but pre-
pared at a different time and used in the series of tests for the study of the
influence of the hydrogen sulfide partial pressure.

INFLUENCE OF H2S PARTIAL PRESSURE ON HDO REACTIONS
53
which 90% correspond to weak acid sites) is very low, we tions confirm that sulfided catalysts are very flexible and
assume there is no contribution to the activity from the can adapt to different reaction conditions.
support. A first speculative interpretation would be that
4.2.2. Ethyl decanoate. The effect of hydrogen sulfide
the breaking of the O–methyl bond is a sort of hydrocrack-
on ED conversion is less marked than that observed on
ing reaction, involving H + and H species, generated by
alumina-supported catalysts (22). At low PH S the amount
of decarboxylated products slightly increases and the dehy-
dration of decanol to decene and decane is favoured. The
hydrogenation of the carboxylic group into alcohol is al-
most constant; the sum of decanol, decene, and decane is
unchanged.
2
a heterolytic scission of hydrogen on the sulfided phase.
The positively charged H + would attack the O atom of the
methoxy group, which carries a partial negative charge. The
H
would attack the carbon, which is partially positively
charged.
A second possible interpretation of the results is to con-
In principle, the de-esterification should be an acid-
catalysed reaction (CM sites), while the hydrogenation of
the carboxylic group to alcohol should involve hydrogena-
tion sites (A sites). It is astonishing that hydrogen sulfide
does not inhibit this reaction as does the conversion of
MA. A compensation between an increase of acid sites
(sulfhydryl species, CM) and a decrease of hydrogenation
sites (uncoordinated molybdenum atoms, A) due to hydro-
gen sulfide could result in an overall constant ED conver-
sion and hydrogenated products yield. The promotion ef-
fect of hydrogen sulfide on the transformation of decanol
into decene and decane and the weak increase of nonane
would indicate that acid sites, present on the metal sulfides
as–SH groups, are involved (CM) in these reactions.
The inhibition effect of diaminopropane also confirms
that acid sites are involved in the conversion of ED. The
decarboxylation reaction would take place on more elec-
trophilic sites than the hydrogenation one as it is more
strongly inhibited.
sider that parallel pathways are operating in the conversion
of GUA: the demethylation to catechol and the demethoxy-
lation to phenol occurring at the same rate on the carbon-
supported catalyst (30). Even if the breaking of the O–
methyl bond should be easier than the hydrogenolysis of
the Caromatic–O bond, several authors have supposed that
the direct demethoxylation can take place even on alumina-
supported catalysts (28, 29). However, in that case, Lewis
acid sites on the support would be much more numerous
than the demethylation and demethoxylation sites on the
sulfided phase. The dominant effect would be that of the
support acidity, with no notable influence of hydrogen sul-
fide on GUA conversion. When alumina is replaced by car-
bon, Lewis acidity is not present any more, and all the re-
actions take place on the sulfided phase. As the support
does not contribute to the activity, distinction between sites
responsible for the demethoxylation and dehydroxylation
would be possible. In this case the most coherent hypothe-
sis would be that demethylation involves H + and sulfhydryl
groups (CM), while demethoxylation, like dehydroxylation,
takes place on highly reduced, uncoordinated molybdenum
Using the same kind of arguments as for the reaction of
GUA, it could be speculated that an increase of the hy-
drogen sulfide partial pressure would relatively rapidly en-
hance the acidity of the sulfided phase by adsorption of
hydrogen sulfide and formation of sulfhydryl groups.
atoms. At increasing PH S, hydrogen sulfide would adsorb
2
on molybdenum vacancies, with the formation of H + and
sulfhydryl groups; the demethylation reaction would conse-
quently be enhanced, while the cleavage of the Caromatic–O
bonds is inhibited. The GUA conversion would thus be ap-
proximately constant, while the phenol + C6/catechol ratio
would decrease.
4.2.3. 4-Methylacetophenone. The inhibition by hydro-
gen sulfide could have been interpreted as the participation
of partially reduced sites, located on incompletely coordi-
nated molybdenum atoms. The fact that ammonia had al-
most no influence on this reaction suggested that the ke-
tonic group does not adsorb on electrophilic sites. It was
shown (9) that uncoordinated sites have an electrophilic
character and that they are even more strongly inhibited by
ammonia than by hydrogen sulfide. Two mechanisms were
therefore proposed by Laurent and Delmon (22), involving
the participation of nucleophilic species. According to the
first, the carbonyl group could adsorb through the carbon
atom on a nucleophilic sulfur atom, and the reaction would
proceed through the addition of a proton on the negative
oxygen atom and subsequently the addition of a hydrogen
atom on the carbon. This second hydrogen atom would be
nucleophilic like a sorbed hydrogen species or a hydrogen
atom bonded to molybdenum (hydride species). The sec-
ond possible mechanism supposed that the carbonyl group
The addition of DAP decreases both the demethylation
and the demethoxylation/dehydroxylation reactions. This
is logical, as this basic compound can poison sulfhydryl and
H
+ species as well as reduced, uncoordinated molybdenum
atoms. With increasing hydrogen sulfide partial pressure
an adsorption or reaction of hydrogen sulfide takes place
on the hydrogenolysis sites; uncoordinated sites would be
progressively transformed into blocked sites. When the hy-
drogen sulfide partial pressure is abruptly decreased, the
hydrogenolysis activity is only partially recovered. The for-
mation of the incompletely coordinated sites is a relatively
slow process. When the hydrogen sulfide partial pressure
is decreased stepwise from 150 to 25 kPa, hydrogenolysis
sites are progressively generated; the more reductive at-
mosphere could render them more stable. These observa-

54
FERRARI ET AL.
TABLE 7
Four Kinds of Active Sites have been Tentatively Identified: (A) Three-Fold Coordi-
natively Unsaturated Mo Atoms (Hydrogenation), (B) Three-Fold Coordinatively Unsatu-
rated Mo Atom with a Sulfhydryl Group (–SH) Neighbour (Hydrogenolysis Sites, Cleavage
of the Heteroatom–Carbon Bonds), (CS) Acid Sites of the Support, and (CM) Acid Sites on
the Metal Sulfide, Sulfhydryl Groups (–SH)
CoMo/Al2O3 –NiMo/Al2O3
CoMo/C
NH3
H2S
H2S
(0–200 kPa)
(0–400 kPa)
DAP
(25–150 kPa)
4-Methylphenol
hydrogenation
4-Methylphenol
hydrogenolysis
↓ (A)
↓ (A)
↓↓ (B)
↓↓ (B)
GUA Conv
Catechol
↓
=
↓
↓
↓
=
=
Not reported
Not reported
Not reported
↓ (CS )
↓↓ (B)
↓
↑ (CM
↓ (B)
↓
)
Phenol + C6
Phenol + C6/Cat ratio
DES or ED conv
Decarboxylation S
C-10
↓
↓
↑
↓
=
↑
↓
↑ =
↓
↑ (A)
↓
= (↑) (A + CM
)
C-9
↓↓
↑↑ (CM
)
↓↓
↑ = (CM
)
MA Conv
=
↓
(?)
↓ (A + CM
)
might react in a “pre-adsorbed” state with an activated hy-
dride species. The effect of hydrogen sulfide would have
been to modify the ability of the active phase to form hy-
dride species and thus to inhibit the conversion.
5. CONCLUSIONS
The activity of sulfided CoMo/C catalyst is deeply af-
fected by modifications of the hydrogen sulfide partial pres-
Decanol is more easily dehydrated into decene and de- sure; when the initial conditions are restored the catalyst is
cane at increasing hydrogen sulfide partial pressure. With able to regain, at least in part, its activity.
the increase of PH S, hydrogen sulfide would adsorb on
Hydrogen sulfide does not affect the activity of the
2
hydrogenation sites (partially reduced and uncoordinated CoMo/C catalyst for the overall conversion of GUA. With
molybdenum atoms):the first hydrogenation step ofthe car- regard to the products, hydrogen sulfide inhibits the direct
bonyl group into the alcohol would be inhibited (sites A), hydrogenolysis of the Caromatic–O bond, and the phenol +
with a decrease of MA conversion, while the dehydration of C6-to-catechol ratio is consequently decreased. Highly re-
the alcohol would be favored by the acidity of the sulfhydryl duced, coordinatively unsaturated sites, located on the
groups (sites CM) (32).
metal sulfides, are very likely responsible for the hy-
With regard to the influence of diaminopropane, the high drogenolysis of the Caromatic–O bond. The decrease of the
increase of the conversion and the large molar balance H2/H2S ratio decreases the average degree of reduction of
deficit, which is unusual for the reaction of MA, suggest that the active sites; phenol production is consequently inhib-
side phenomena take place. Diaminopropane could inter- ited. Acid species that can be located on the sulfide phase
act with the decomposition products of dimethyl disulfide as well as generated by heterolytic dissociation of hydro-
and form amine polysulfides (33).
gen would be responsible for the demethylation of GUA to
The results are summarised in Table 7 and allow us to catechol.
compare carbon and alumina support catalysts. Arrows in-
For ED, the conversion is not influenced by hydrogen
dicate whether the mentioned conversion or selectivity in- sulfide, and the selectivity in hydrogenated and decarboxy-
creases (↑) or decreases (↓) when a parameter is modified. lated products is slightly changed. Diaminopropane de-
A full agreement exists among the results of the different creasesboth the ED conversion and the decarboxylation se-
investigations concerning the inhibition by hydrogen sul- lectivity. Hydrogenation and acid sites located on the metal
fide of the reactions involving the cleavage of Caromatic–O sulfides are involved in these reactions; parallel and con-
bond and on the hydrogenation of the carbonyl group of secutive pathways are simultaneously possible, and thus the
MA. The role of coordinatively unsaturated molybdenum overall effect of hydrogen sulfide is not very marked.
atoms, probably differing in the degree of reduction and
uncoordination, has been confirmed and explained.
MA is the only reactant molecule whose conversion is
inhibited by hydrogen sulfide. Reduced and incompletely

INFLUENCE OF H2S PARTIAL PRESSURE ON HDO REACTIONS
55
coordinated molybdenum atoms are likely involved in the 14. Mignard, S., Marchal, N., and Kasztelan, S., Bull. Soc. Chim. Belg. 104
(4–5), 259 (1995).
hydrogenation of the carbonyl group, while sulfhydryl acid
species would be responsible for the dehydration of the
alcohol.
15. Marchal, N., Mignard, S., and Kasztelan, S., Catal. Today 29, 203
(1996).
16. Chadwick, D., Oen, A., and Siewe, C., Catal. Today 29, 229 (1996).
17. La Vopa, V., and Satterfield, C. N., J. Catal. 110, 375 (1988).
18. Satterfield, C. N., Modell, M., and Wilkens, J. A., Ind. Eng. Chem.
ACKNOWLEDGMENTS
Process Des. Dev. 19, 154 (1980).
19. Satterfield, C. N., and Yang, S. H., J. Catal. 67, 335 (1983).
20. Odebunmi, E. O., and Ollis, D. F., J. Catal. 80, 76 (1983).
21. Laurent, E., and Delmon, B., Appl. Catal. A: Gen. 109, 77 (1997).
22. Laurent, E., and Delmon, B., Appl. Catal. A: Gen. 109, 97 (1997).
This work was carried out as part of the UE project JOR3-CT95-0025.
We also acknowledge the FRNS (Belgium) for financial support.
REFERENCES
23. Ferrari, M., Lahousse, C., Centeno, A., Maggi, R., Grange, P., and
Delmon, B., in “Studies in Surface Science and Catalysis” (B. Delmon,
P. A. Jacobs, R. Maggi, J. A. Martens, P. Grange, and G. Poncelet,
Eds.), Vol. 118, p. 505. Elsevier, Amsterdam, 1998.
24. Laurent, E., Centeno, A., and Delmon, B., in “Studies in Surface
Science and Catalysis” (B. Delmon and G. Froment, Eds.), Vol. 88,
p. 573. Elsevier, Amsterdam, 1994.
25. Centeno, A., Laurent, E., and Delmon, B., J. Catal. 154, 288 (1995).
26. Weisser, O., and Landa, S., in “Sulphide Catalysts, their Properties and
Applications”. Pergamon, Oxford, 1975.
1. Bridgwater, A. V., Catal. Today 29, 285 (1996).
2. Furimsky, E., Appl. Catal. (A) 199, 147 (2000).
3. Girgis, M. J., and Gates, B. C., Ind. Eng. Chem. Res. 30, 2021 (1991).
4. Maggi, R., and Delmon, B., Fuel 73, 671 (1994).
5. Sipila, K., Kuoppala, E., Fagernas, L., and Oasmaa, A., in “Biomass
and Bioenergy”, Vol. 14, p. 103. Pergamon, Elmsford, NY, 1998.
6. Yoshimura, Y., Sato, T., Shimada, H., Matsubayashi, N., and Nishijima,
A., Appl. Catal. 73, 55 (1991).
7. Laurent, E., and Delmon, B., J. Catal. 146, 281 (1994).
8. Calafat, A., Laine, J., and Lopez-Agudo, A., Catal. Lett. 40, 229
(1996).
27. van Gestel, J., Leglise, J., and Duchet, J. C., J. Catal. 145, 429 (1994).
28. Hurff, S. J. M., and Klein, T., Ind. Eng. Chem. Fundam. 22, 426
(1983).
9. Laurent, E., and Delmon, B., Ind. Eng. Chem. Res. 32, 2516 (1993).
10. Gevert, B. S., Otterstedt, J. E., and Massoth, F. E., Appl. Catal. 31, 119
(1987).
11. Schulz, H., Rahman, N. M., Kordokuzis, G., Sharma, L. D., Coal Sci.
Technol. 11, 269 (1987).
12. Viljava, T. R., and Krause, A. O. I., Appl. Catal. A: Gen. 145, 237
(1996).
13. Viljava, T. R., and Krause, A. O. I., in “Studies in Surface Science and
29. Petrocelli, F. P., and Klein, M. T., Fuel Sci. Technol. Int. 5, 63 (1987).
30. Ferrari, M., Centeno, A., Lahousse, C., Maggi, R., Grange, P., and
Delmon, B., Prepr. Am. Chem. Soc. Div. Petrol. Chem. March 29–April
3, 43 (1), 94 (1998).
31. De la Puente, G., Pis, J. J., Menendez, J. A., and Grange, P., J. Anal.
Appl. Pyrol. 43, 12 (1997).
32. Durand, R., Geneste, P., Moreau, C., and Pirat, J. L., J. Catal. 90, 147
(1984).
Catalysis” (G. F. Froment, B. Delmon, and P. Grange, Eds.), Vol. 106, 33. Yang, M. H., Grange, P., and Delmon, B., Appl. Catal. A: Gen. 154, L7
p. 343. Elsevier, Amsterdam, 1997. (1997).