The influence of the support on activity and selectivity of nickel sulfide catalysts in the decarbonylation of stearic acid to heptadecenes

Article abstract of DOI:10.1007/s11172-017-1755-2

The influence of the support nature on the performance of nickel sulfide catalyst in the decarbonylation of stearic acid to heptadecenes was investigated. The catalyst supported on silica demonstrated higher activity and selectivity in comparison with the catalyst on γ-Al₂O₃ used as a reference. The reaction schemes over these catalysts are nearly the same; however, the contributions from the side reactions of hydrogenation and oligomerization are reasonably different. Introduction of the products of decarbonylation (CO and water vapor) decreases the stearic acid conversion; and in the case of the catalyst supported on silica, the addition of CO strongly reduces the rate of hydrogenation of heptadecenes. The reasons for the observed differences were discussed. It was suggested that the dispersion of the nickel component as well as the nature of support acidity played a significant role.

Full text of DOI:10.1007/s11172-017-1755-2

Russian Chemical Bulletin, International Edition, Vol. 66, No.3, pp. 463—467, March, 2017
463
The influence of the support on activity and selectivity of nickel sulfide catalysts
in the decarbonylation of stearic acid to heptadecenes
V. Ya. Danyushevsky, P. S. Kuznetsov, E. A. Katsman, A. S. Kupriyanov, V. R. Flid, and A. S. Berenblyum
Moscow Technological University (Institute of Fine Chemical Technologies),
86 prosp. Vernadskogo, 119571 Moscow, Russian Federation.
Fax: +7 (495) 434 9287. Eꢀmail: berenblyum@gmail.com
The influence of the support nature on the performance of nickel sulfide catalyst in the
decarbonylation of stearic acid to heptadecenes was investigated. The catalyst supported on
silica demonstrated higher activity and selectivity in comparison with the catalyst on γꢀAl2O3
used as a reference. The reaction schemes over these catalysts are nearly the same; however,
the contributions from the side reactions of hydrogenation and oligomerization are reasonably
different. Introduction of the products of decarbonylation (CO and water vapor) decreases the
stearic acid conversion; and in the case of the catalyst supported on silica, the addition of CO
strongly reduces the rate of hydrogenation of heptadecenes. The reasons for the observed
differences were discussed. It was suggested that the dispersion of the nickel component as well
as the nature of support acidity played a significant role.
Key words: support, catalyst, activity, selectivity, decarbonylation, stearic acid, higher
olefins, dispersion, oligomerization.
Study of the selective synthesis of higher olefins (HO)
from higher fatty acids of renewable feedstock (inedible
vegetable oils and fats) is important for the development
of principles of the catalysis as well as for the developꢀ
ment of industrial processes based on these reactions.1
Recently, it has been published that sulfidized Niꢀ, Coꢀ,
and Моꢀbased supported catalysts favour decarbonylaꢀ
tion of fatty acids to olefins and are also active in hydroꢀ
deoxygenation of fatty acids to parafines.2 Moreover,
the ratio between contributions of these reactions varied
from 0.3 to 4.2 in the presence of different catalysts.2
We could establish that in the presence of nickel sulfide,
which was obtained by the hydrogen treatment of
nickel sulfate on alumina oxide surface, deoxygenation of
stearic acid (St) proceeds only via decarbonylation and
the initial selectivity to the olefin production may be
over 90%.1,3—5
The choice of a catalyst support has a crucial role for
the design of active and selective catalysts as it exerts a
significant influence on composition and dispersion of acꢀ
tive phase as well as can directly participate in the catalytꢀ
ic process owing to surface functional groups.6 The effect
of the support was shown by using palladium catalysts7 on
St deoxygenation.
Experimental
Reagents, solvents, and materials. Nickel sulfate (NiSO4•
•7H2O, 99.5%, reagent grade, State Standart GOST 4465ꢀ74),
sodium hydroxide (98.0%, analytical grade, State standart GOST
4328ꢀ77), sulfuric acid (94%, reagent grade), dodecane (99%,
high purity grade, Khimmed, Russia), tridecane (high purity
grade), chloroform (reagent grade), ethyl alcohol (96%, rectifiꢀ
cate), methyl alcohol (analytical reagent for liquid chromatoꢀ
graphy), stearic acid (99.2%, analytical grade, State Standart,
GOST 9419ꢀ78), hydrogen (99.99%, A grade), nitrogen
(99.999%, special purity grade, 1 sort), and carbon monoxide
(99.8%, high purity grade,) were used as received. Silica (KSKG,
State Standart GOST 3956ꢀ76) was used as a support. The speꢀ
cific surface area of this material, which we earlier determined
by BET, was 205 m2 g–1).
Catalyst synthesis. The support with 0.08—0.20 mm particle
size was used. To prepare a 3.24% nickel catalyst, 1.6 g of
NiSO4•7H2O was diluted in 12 mL of distilled water and added
to a support sample (10 g) placed in a plastic container. The
mixture was slowly stirred for an hour, then transferred into
a porcelain dish and evaporated for 3 h at 80—90 °С. The obꢀ
tained sample was dried for 3 h at 120—130 °С with stirring at
regular intervals before loading into the reactor.
Procedure for catalytic experiments. The experiments were
carried out in a 50 mL autoclave equipped with a stirrer. (Autoꢀ
clave Engineers); the reactor was heated in a bath with Wood´s
alloy following the known protocol.8 The catalyst (0.25 or 0.5 g)
was loaded inside the reactor and reduced for 3 h in a hydrogen
flow (50—60 mL min–1) at 350 °С and pressure of 10 atm. The
specific surface area determined by BET was 195 m2 g–1. The
autoclave was cooled and 9 mL (6.5—7.0 g) of dodecane as
The purpose of the present study is to estimate the
effect of the support on the performance of nickel sulfide
catalyst in decarbonylation of St to form HO (heptaꢀ
decenes). Silica and γꢀalumina, which was studied earliꢀ
er,5 were selected as supports.
Published in Russian in Izvestiya Akademii Nauk. Seriya Khimicheskaya, No. 3, pp. 0463—0467, March, 2017.
1066ꢀ5285/17/6603ꢀ0463 © 2017 Springer Science+Business Media, Inc.

464
Russ.Chem.Bull., Int.Ed., Vol. 66, No. 3, March, 2017
Danyushevsky et al.
Table 1. The characteristics of stearic acid conversion over nickel sulphide catalyst supported on silica (NSC/SiO2)
Runa
τ/min
Т/°С
P(Н2)
/atm
Conversion
(%)
Selectivity (%)
CO2/(CO + CO2)
to Нd
to Нde
1
2
3
4
5b
6c
7d
8e
9
10
11
12
130
160
190
160
160
160
160
160
130
145
160
120
325
325
325
325
325
325
325
325
350
350
350
350
15
15
15
12
15
15
17
12
15
15
15
15
41.2
71.1
79.3
52.2
83.7
45.7
67.1
41.8
43.6
62.9
74.6
98.8
17.3
14.4
15.5
16.7
21.4
14.9
10.5
14.8
10.5
13.1
14.2
30.1
56.4
61.7
56.2
73.4
54.6
60.3
56.8
62.2
48.3
56.5
56.6
43.5
0
0.0073
0.0071
0.0095
0.0128
0.0104
0.0072
0.0533
0.0241
0.0360
0.0559
0.0202
a Standart conditions: catalyst loading is 0.5 g, 6 g of dodecane, and 2 g of St; Н2 pressure is 15 atm.
b St weight is 1 g.
c Catalyst weight is 0.25 g.
d СО pressure is 8 atm.
e Н2О/St = 2 : 1.
a solvent was introduced into the reactor through a special vessel
(under the nitrogen flow). This procedure protects the catalyst
from oxidation when the air contacts with the catalyst. Then the
autoclave was opened and 1 or 2 g of stearic acid was added.
Then the autoclave was closed and the content mixture was
doubly blown with hydrogen flow. The experiments were conꢀ
ducted under the conditions presented in Table 1. The reaction
time was read from the moment when a temperature reached
325 or 350 °С (see Table 1, runs 1—12). After the experiment
was completed, the reactor was rapidly cooled with water and
gaseous and liquid products were sampled. The experiments were
run under the conditions which nearly totally exclude diffusion
kinetics, and the catalyst activity was maintain constant.8 At
these conditions, the reagent conversion was independent of the
stirring rate and the catalyst particle size (Table 2). The experiꢀ
mental results summarized in Table 1 were obtained at the stirring
rate of 1000 rpm and the catalyst particle size of 0.08—0.2 mm.
When the experiment was carried out over the catalyst sample
already used, the yield to olefin remained constant (see Table 1,
run 12).
The catalysts were characterized by transmission electron
microscopy (TEM) on a LEO 912 AB OMEGA instrument at an
accelerating voltage of 100 kV. The specific surface area was
calculated by BET method using the data obtained on an ASAP
2020 instrument (Micromeritics). Nickel content in the catalyst
was measured using a SpectroscanꢀLF Xꢀray fluorescent analyzer.
The instrumental error was 0.5%.
Results and Discussion
Nickel sulfide catalyst on silica (NSC/SiO2) was preꢀ
pared by reduction of the supported nickel sulfate using
the same technique as in the case when γꢀalumina was
used as a support. General catalytic characteristics of
NSC/SiO2 are given in Table 1. The comparison of the
obtained data for NSC/SiO2 and those for its analogue,
NSC/Al2O3, which were published earlier5 leads to the
following conclusions.
Table 2. Dependences of stearic acid conversion (Х)
on the stirring rate (N) and the average catalyst partiꢀ
cle size (see the reaction conditions in Table 1, run 9)
N/rpm
Average size/mm
X (%)
The reaction schemes of St deoxygenation over these
catalysts are closely related. Decarbonylation of St is the
main course of the primary transformation of St, and
heptadecenes (Hde) are the predominant products (see
Table 1). Relatively high selectivity to Hde formation
(73.4%) was attained at low hydrogen pressure (2 atm)
(see Table 1, run 4), but St conversion was not high enough.
However, it is known that during a longꢀterm run at low
Н2 pressure, the catalysts deactivates due to the coke forꢀ
mation.9,10 As it can be seen from Table 1, as St converꢀ
sion increases and Hde accumulates in the reaction
mixture, the selectivity to Hde formation decreases. The
400
600
800
0.14
0.14
0.14
0.14
0.14
0.14
0.14
0.26
0.52
2.30
5.00
25.2
36.9
42.1
43.6
44.1
43.2
43.6
43.1
43.9
39.2
21.3
1000
1200
1400
1000
1000
1000
1000
1000

Efficiency of nickel sulfide catalyst
Russ.Chem.Bull., Int.Ed., Vol. 66, No. 3, March, 2017
465
reasons are the secondary reactions of hydrogenation and
oligomerization of olefines5 yielding heptadecane (Hd) and
Hde oligomers (mainly dimers), respectively. The experiꢀ
mental results show that St hydrodeoxygenation hardly
occures (only traces of octadecane were found) and the
contribution from St decarboxylation and CO steam
conversion is fairly low (a fraction of CO2 in the reaction
products was low, see Table 1). There was no methanization
of carbon oxide, which was produced via St decarbonylaꢀ
tion (methane was not detected in the reaction products).
Despite the resemblance of the reaction schemes in
the presence of NSC/SiO2 and NSC/Al2O3, considerable
differences in their activity and less noticeable in selectivꢀ
ity can be observed. To obtain nearly equal St conversion,
a significantly lower residence time and/or temperature is
needed in the presence of NSC/SiO2 than in the presence
of NSC/Al2O3 under the other identical conditions (temꢀ
perature, hydrogen pressure, reaction time, catalyst loading).
Thus in the presence of NSC/Al2O3, the reactant converꢀ
sion of ∼77% with the selectivity to Hde of ∼46% is reached
in 2 h at 350 °C and 15 atm.5 On the other hand, less than
1.5 h at 325 °C (see Table 1, run 3) or ∼1 h at 350 °C (see
Table 1, run 11) is needed to approach the conversion
level in the presence of NSC/SiO2. The selectivities to
Hde are equal at the above indicated temperatures (∼56%).
Obviously, in the case of the catalyst deposited on SiO2,
the gain in selectivity to Hde formation is nearly 10%.
Nevertheless, as was indicated earlier,1 selectivity depends
at least on two factors: the catalyst activity in olefin hydroꢀ
genation and the propensity of olefin to oligomerization.
Increasing contribution of these factors evidently decreases
the catalyst selectivity.
The detailed examination of the catalysts supported on
oxides reveals that a total gain from replacing NSC/Al2O3
by NSC/SiO2 (see Ref. 1) as the catalyst for Hde formaꢀ
tion can be estimated as 10%. This estimate includes the
selectivity gain (17—18%) obtained due to a lower degree
of Hde oligomerization offset by a 7—8% loss caused by
a higher extent of hydrogenation of Hde to Hd (see Table 1,
runs 3 and 11). These calculations are based on the mateꢀ
rial balance assessment of the studied process considering
that Hde converts only to Hd and oligomers. Thus, in
comparison with NSC/Al2O3, NSC/SiO2 increases the
rate of Hde hydrogenation to a higher extent and the rate
of Hde oligomerization to a lesser extent than the target St
conversion to the olefins.
time, a fraction of oligomers, which was determined from
the material balance, remained unchanged. Probably, the
reaction was inhibited by a substrate. In addition, a comꢀ
petition between the reactants that form adsorbed comꢀ
plexes with the catalyst active sites of different strength
can not be excluded. An idea of such competition is evident
from the increased rate of Hde hydrogenation as St conꢀ
version approaches a 100% level (see Table 1). A decrease
in the catalyst loading results in the reasonable decrease in
St conversion while the selectivity remains nearly equal
(see Table 1, runs 2 and 6).
An unusual effect of the products of decarbonylation
(CO and water vapor) on the reaction characteristics was
found in the presence of NSC/SiO2 (see Table 1, runs 7
and 8). This behaviour of this catalyst is different from
that of NSC/Al2O3. In both cases, St conversion decreasꢀ
es, however, CO decreases the rate of Hde hydrogenation
stronger in the presence of NSC/SiO2.
It is obvious that the difference in the selectivities and
activities of the above catalysts is related to the nature of
supports (silica and alumina oxide).
As it follows from the data of kinetic modelling pubꢀ
lished earlier,1 the main reason for the decrease in selecꢀ
tivity to Hde formed via St decarbonylation is the conseꢀ
quent oligomerization of the product. The latter reaction
may proceed via thermal or catalytic mechanism. When
a catalytically inert support is used, oligomerization is supꢀ
posed to follow the thermal course. In the case of supports
containing active functional groups or active sites, the reꢀ
action is likely to follow the catalytic pattern. It was shown
that HO oligomerization occurs over the catalysts and
supports containing acid sites, both Brönsted and Lewis.11
By comparing the supports used in the catalyst prepaꢀ
ration, i.e. silica and γꢀalumina,1,5 it is evident that the
surface of the latter support is characterized by relatively
strong Lewis acidity and weaker Brönsted acidity owing to
the presence of bridged ОНꢀ groups.6,12 That is why aluꢀ
mina is utilized as a catalyst in a number of similar proꢀ
cesses.6 In the case of silica, its surface contains only Brönꢀ
sted acids, which are mainly terminal OH groups characꢀ
terized by fairly low acidity (4.0 ≤ H0 ≤ 6.8).12 These data
may verify the fact that the observed reduced selectivity to
HO formation in the presence of the aluminaꢀbased cataꢀ
lyst is related to the involvement of olefin in the acid oligoꢀ
merization promoted by γꢀAl2O3.
From the results of TEM study of NSC/SiO2 the reaꢀ
sons for the difference in activities of the above indicated
catalysts can be inferred. Darken roundꢀshaped particles
of nickel sulfide are readily visible on silica support (Fig. 1).
The average size of these particles is 4.3 1.8 nm (Fig. 2).
We should mention that the average size of the particles of
nickel compounds on γꢀAl2O3 surface was approximately
4.7 0.7.5 However, the particleꢀsize distribution on these
two supports is different (see Fig. 2). A majority of the
particles on alumina were 4—6 nm in diameter, whereas
A lower hydrogen pressure at 325 °C decreases the
total conversion of St and increases the selectivity of its
conversion to Hde. Such behavior is generally explained
by the reduced fraction of Hd in the reaction products (see
Table 1, runs 2 and 4).
At 325 °C, a twofold reduction in the initial concenꢀ
tration of St improves its conversion and decreases the
selectivity to Hde owing to increased fraction of Hd in the
reaction mixture (see Table 1, runs 2 and 5). At the same

466
Russ.Chem.Bull., Int.Ed., Vol. 66, No. 3, March, 2017
Danyushevsky et al.
N (%)
50
a
40
30
20
10
<3
3—4
4—5
5—6
>6 d/nm
N (%)
b
100 nm
20
15
10
5
<1 1—2 2—3 3—4 4—5 5—6 6—7 7—8 9—10 >10 d/nm
Fig. 2. The hystogramms of particleꢀsize distribution (d) on Al2O3
(a) and SiO2 (b) surfaces; N is particle fraction.
The reasons for a limited number of interplanar spacꢀ
ings in these catalytic systems were discussed in our previꢀ
ous work.3
50 nm
Fig. 1. SEM images of NSC/SiO2.
One can suggest that the presence of small nanopartiꢀ
cles of nickel sulfide characterized by a high fraction of
surface atoms contribute to an enhanced activity of the
catalyst supported on silica in St decarbonylation. Similar
effects were earlier observed in a number of reactions.13,14
The influence of the support nature on the performance
of nickel sulfide deposited catalysts for selective decarboꢀ
nylation of fatty acids to HO has no straightforward exꢀ
planation. Probably, the support characteristics signifiꢀ
cantly affect the dispersion of sulfides formed through
nickel sulfate reduction by hydrogen. Using silica and
alumina oxides as reference supports, we showed that
a great number of small nanoparticles (1—3 nm) were
generated on the silica surface unlike those on alumina
surface. This finding can be the reason for higher activity
of NSC/SiO2 than NSC/Al2O3 for deoxygenation of fatty
acids. At the same time, as it follows from our experiꢀ
ments, the catalyst supported on silica demonstrated the
highest activity not only in St decarbonylation but in hyꢀ
nearly 30% of the particles deposited on silica were charꢀ
acterized by smaller diameters (1—3 nm). The specific
surface area of the examined silica was 205 m2 g–1 which is
slightly different from the surface area of alumina oxide,
which was determined in our previous work (254 m2 g–1).1,5
Examination of NSC/SiO2 catalyst by TEM in the
electron diffraction mode clearly detects six interplanar
spacings of the crystallites that is less than on the difractoꢀ
gramms typical of crystalline sulfide. Equal result was obꢀ
tained for alumina supported catalyst.3 In both catalysts,
Ni3S2 and NiS phases were identified. However, the interꢀ
planar spacings characterizing pure nickel were not found
which can be accounted for by small sizes of metal partiꢀ
cles. The interplanar spacings (Å) for the particles in the
catalyst of interest have the following values: Ni3S2, 3.906,
2.875, 2.407, 2.113, 1.785, JCPDS Card No. 44ꢀ1418;
NiS, 1.327, JCPDS Card No. 47ꢀ1739.

Efficiency of nickel sulfide catalyst
Russ.Chem.Bull., Int.Ed., Vol. 66, No. 3, March, 2017
467
(Zvenigorod, April 11—15, 2016), Zvenigorod, 2016, p. 47
(in Russian).
drogenation of the target product, i.e., Hde. This could
reduce the selectivity of the process but the observed gain
from the olefin consumption in oligomerization exceeds
the loss of olefin in hydrogenation.
In additition, the acidity of supports may play a crucial
role in the performance of the supported sulfide catalysts
in the studied reaction. There is a reason to suppose
that the ability of acid sites to increase the rate of
olefin oligomerization is revealed stronger in the case of
alumina than in the case of silica. Moreover, it can explain
a higher selectivity to olefin observed over NSC/SiO2
catalyst.
4. V. Ya. Danyushevsky, A. S. Berenblyum, P. S. Kuznetsov,
E. A. Katsman, Sb. Tez. dokl. of II Ros. congressa po katalizu
"ROSKATALIZ" [Abstracts of the II Russsian Congress in Cataꢀ
lysis "ROSKATALIZ"] (Samara, October 2—5, 2014), Samara,
2014, p. 130 (in Russian).
5. E. A. Katsman, V. Ya. Danyushevsky, P. S. Kuznetsov, R. S.
Shamsiev, A.S. Berenblyum, Kinet. Catal., 2017, 58, 147.
6. Concepts of Modern Catalysis and Kinetics, 2nd ed., Eds
I. Chorkendorff, J. W. Niemantsverdriet, WileyꢀVCH, Weinꢀ
heim, 2007, 477 pp.
7. A. S. Berenblyum, R. S. Shamsiev, T. A. Podoplelova, V. Ya.
Danyushevsky, Russ. J. Phys. Chem., 2012, 86, 1199.
8. A. S. Berenblyum, T. A. Podoplelova, E. A. Katsman, R. S.
Shamsiev, V. Ya. Danyushevsky, Kinet. Catal., 2012, 53, 595.
9. I. Simakova, O. Simakova, P. MäkiꢀArvela, D. Yu. Murzin,
Catal. Today, 2010, 150, 28.
10. J. G. Immer, H. H. Lamb, Energy & Fuels, 2010, 24, 5291.
11. V. Sh. Fel´dblyum, N. V. Obeshchalova, Russ. Chem Rev.,
1968, 37, 789—797.
This work was financially supported by the Ministry of
Education and Science of the Russian Federation (Rusꢀ
sian Federational State Programme No. 2014/114, Project
No. 564, Governmental Assignment No. 114120870179).
References
12. O. V. Krylov, Geterogennyi kataliz [Heterogeneous Catalysis],
Akademkniga, Moscow, 2004, 679 pp. (in Russian).
13. V. I. Bukhtiyarov, M. G. Slin´ko, Russ. Chem Rev., 2001, 70,
147—159.
14. V. A. Yakovlev, S. A. Khromova, V. I. Bukhtiyarov, Russ.
Chem. Rev., 2011, 80, 911—925.
1. A. S. Berenblyum, V. Ya. Danyushevsky, P. S. Kuznetsov,
E. A. Katsman, R. S. Shamsiev, Petrol. Chem., 2016, 56, 663.
2. S. Brillouet, E. Baltag, S. Brunet, F. Richard, Appl. Catal. B:
Environ., 148—149, 201.
3. V. Ya. Danyushevsky, A. S. Berenblyum, P. S. Kuznetsov,
E. A. Katsman, Sb. Tez. dokl. "Fizikoꢀkhimiya nanoꢀ
structurirovannykh katalizatorov" [Abstracts of the Conferꢀ
ence Physics and Chemistry of Nanostructured Catalysts]
Received October 31, 2016;
in revised form November 16, 2016