High catalytic activity of silicalite in gas-phase ketonisation of propionic acid

Article abstract of DOI:10.1039/c3cc41161c

Amorphous silica and crystalline silicalite (MFI structure) are demonstrated to be active and environmentally benign catalysts for propionic acid ketonisation at 450-500 °C to form 3-pentanone. The silicalite is particularly efficient, and its ketonisation selectivity is increased by base modification probably through generation of silanol nests.

Full text of DOI:10.1039/c3cc41161c

ChemComm
View Article Online
View Journal | View Issue
COMMUNICATION
High catalytic activity of silicalite in gas-phase
ketonisation of propionic acid†
Cite this: Chem. Commun., 2013,
49, 3842
Hossein Bayahia, Elena Kozhevnikova and Ivan Kozhevnikov*
Received 12th February 2013,
Accepted 6th March 2013
DOI: 10.1039/c3cc41161c
www.rsc.org/chemcomm
Table 1 Gas-phase ketonisation of propionic acid over amorphous silica Aerosil-300a
Amorphous silica and crystalline silicalite (MFI structure) are demon-
strated to be active and environmentally benign catalysts for propio-
nic acid ketonisation at 450–500 8C to form 3-pentanone. The silicalite
is particularly efficient, and its ketonisation selectivity is increased by
base modification probably through generation of silanol nests.
Selectivity [mol%]
Temp. Conv.
C1–C3 C2–C3
Catalyst
None
[1C]
[%]
3-Pentanone alkanes alkenes Othersb
500
12
23
39
40
36
87
85
81
14
2
1
19
1
3
31
10
11
14
Aerosil-300 400
Aerosil-300 500
Aerosil-300c 500
Carboxylic acids readily available from natural resources are attrac-
tive as renewable raw materials for the production of value-added
chemicals and bio-fuel components.1 For fuel applications,
carboxylic acids require reduction in their oxygen content, i.e.
deoxygenation. Therefore, much current research is focused on
deoxygenation of carboxylic acids using heterogeneous catalysis.2
Ketonisation of carboxylic acids (eqn (1)) is widely used as a clean
method for the synthesis of ketones.3a It allows for partial deoxy-
genation of carboxylic acids to be achieved accompanied by their
carbon backbone upgrade. Ketonisation is catalysed by many basic
and acidic metal oxide and mixed-oxide catalysts in the temperature
range of 350–500 1C.3 But the nature of catalytically active sites is not
clear. It is generally thought that basic sites are favourable for
ketonisation.3a However, heteropoly acid H3PW12O40 possessing very
strong proton sites has also been found active in propionic acid
ketonisation.3h We now report that neutral metal-free silica materials,
namely amorphous silica and, in particular, crystalline silicalite
(MFI structure4), are active and environmentally benign catalysts
for the gas-phase ketonisation of propionic acid. Propionic acid is
chosen as a representative of carboxylic acids with the number of
carbon atoms n r 6 derived from carbohydrate feedstocks.3e
2
3
a
0.2 g catalyst, 2 vol% propionic acid in N2 flow, 20 mL minꢀ1 flow rate,
4.0 h g molꢀ1 space time, 6 h time-on-stream. Other products
b
included propionic anhydride, isopropanol, acetone together with
c
unidentified organic compounds; CO and CO2 not monitored. Aerosil-
300 modified by treatment with 3.7 M NH3(aq) + 0.7 M NH4NO3.
really surprising because silica is practically a neutral material. It
lacks basicity and possesses a very weak acidity due to its silanol
groups, SiOH. Table 1 shows representative results for the com-
mercial Aerosil-300 (Degussa), high-purity amorphous silica
consisting of nanosized nonporous spheres fumed in a high-
temperature flame (300 m2 gꢀ1 specific surface area). Its activity
increased with increasing the reaction temperature from 400 to
500 1C. The catalyst showed a respectable 3-pentanone selectivity
of 85% at 39% conversion at 500 1C, and no catalyst deactivation
was observed for at least 28 h on stream. Blank experiments in the
absence of silica showed a small contribution of homogeneous
reaction: at 500 1C about 12% of propionic acid underwent
pyrolysis, in agreement with the previous report.5
Next we tested crystalline materials, i.e. highly siliceous
zeolites. HZSM-5 zeolite (Si/Al = 180) possessing strong proton
sites exhibited a low ketonisation activity at 300 1C but no such
activity at 400–500 1C, mainly causing cracking of propionic
acid to form light hydrocarbons, predominantly ethylene
(Table 2). In contrast, purely siliceous silicalite (MFI structure,
see ESI†) showed high catalytic activity during ketonisation,
which peaked at 500 1C to afford 3-pentanone in 50% yield.
In search for better performance, silicalite was chemically
modified with aqueous acidic (0.01 M and 0.1 M HCl) or basic
(3.7 M NH3(aq) and 3.7 M NH3(aq) + 0.7 NH4NO3) solutions.6
2CH3CH2COOH - (CH3CH2)2CO + CO2 + H2O
(1)
First we tested a wide range of amorphous silicas for ketonisa-
tion of propionic acid (eqn (1)) using a fixed-bed continuous flow
reactor (see ESI†). These silicas were high purity powdered
materials employed as catalyst supports and stationary phases
in chromatography (see ESI†). At 400–500 1C, they exhibited a
moderate to good catalytic activity in the formation of 3-penta-
none, with some propionic anhydride also being formed. This is
Department of Chemistry, University of Liverpool, Liverpool L69 7ZD, UK.
E-mail: kozhev@liverpool.ac.uk
† Electronic supplementary information (ESI) available. See DOI: 10.1039/c3cc41161c Whereas the acid treatment had little effect, the basic one, with
c
3842 Chem. Commun., 2013, 49, 3842--3844
This journal is The Royal Society of Chemistry 2013

View Article Online
Communication
ChemComm
Table 2 Gas-phase ketonisation of propionic acid over silicalitea
by coke deposition on the catalyst, which amounted to 6.6 wt%. Coke
could be removed by aerobic gasification at 450–500 1C, allowing full
recovery of catalyst activity. The 3-pentanone product was quite stable
under the reaction conditions, undergoing 3 and 8% conversion over
base-treated silicalite at 450 and 500 1C, respectively, to form a
mixture of C1–C3 alkanes and alkenes as well as some heavier
products (see ESI†). Although silicalite works at a higher temperature
Selectivityb [mol%]
Temp. Conv.
C1–C3
C2–C3
Catalyst
[1C]
[%]
3-Pentanone alkanes alkenes Othersc
HZSM-5
HZSM-5
HZSM-5
300
400
500
43
98
99
23
42
95
57
75
10
46
84
92
30
0
0
1
0
0
3
5
15
2
6
2
2
6
39
97
0
18
28
2
18
1
1
64
61
3
21
5
4
1
1
1
4
5
3c
Silicalite 400
Silicalite 450
Silicalite 500
Silicalited 500
Silicalitee 500
Silicalitef 400
Silicalitef 450
Silicalitef 500
Silicalitef 550
76
72
53
95
75
96
93
92
65
than the best metal oxide catalysts, e.g. CeO2–Mn2O3 and
CeO2–ZrO2,3e it has the important advantage of being an easy
available non-toxic metal-free material, having perfect performance
stability in the ketonisation reaction and allowing easy regeneration.
The possibility of catalysis by metal impurities in silicalite
and Aerosil-300 can be ruled out because ICP analysis showed
only traces of metals such as Al, Fe, etc. in these materials. Also
the results did not change when bi-distilled water was used for
catalyst preparation and modification (see ESI†).
2
3
1
5
27
a
0.2 g catalyst, 2 vol% propionic acid in N2 flow, 20 mL minꢀ1 flow rate,
4.0 h g molꢀ1 space time, 6 h time-on-stream. Silicalite was prepared by the
literature procedure7 and calcined at 550 1C in air for 12 h. b The selectivity
defined as moles of product formed per 1 mole of propionic acid converted
and quoted in mole%; CO and CO2 not monitored. c Other products
included propionic anhydride, isopropanol, acetone together with uniden-
Characterisation of the silicalite catalysts using DRIFT spectro-
scopy8 revealed significant differences and gave important informa-
tion regarding the active sites in the ketonisation reaction. Fig. 2
shows the DRIFT spectra of our Aerosil-300 {(a) and (b)} and
silicalite {(c)–(g)} samples in the region of OH stretching modes
of silanol groups SiOH. These spectra are in agreement with those
reported in the literature.7,9 The sharp peak at 3744 cmꢀ1 is
attributed to the free terminal silanol groups located on external
and internal surfaces. The broader band around 3680 cmꢀ1 is
attributed to the hydrogen-bonded vicinal silanols. The very broad
band in the 3600–3100 cmꢀ1 region is generally ascribed to silanol
nests that consist of a number of silanol groups interacting through
extended hydrogen bonding. Such nests occur at silicon vacancies
(defects) created by removing a tetrahedral Si atom from the
framework and termination of the four loose O atoms by H atoms
(Scheme 1).9 DFT analysis shows10 that the OH groups in silanol
nests, due to the extended H-bonding, possess enhanced acidity
e
tified compounds. d Silicalite modified by 0.1 M HCl. Modified by 3.7 M
NH3(aq). f Modified by 3.7 M NH3(aq) + 0.7 M NH4NO3.
NH3(aq) + NH4NO3, significantly improved ketonisation selectivity
(Table 2). A similar effect has been found for the Beckmann
rearrangement of cyclohexanone oxime to e-caprolactam over silica-
lite, which has been attributed to the formation of framework defects
(silanol nests) in silicalite, acting as catalytically active sites.7 In con-
trast, the base treatment had practically no effect on the performance
of amorphous Aerosil-300 (Table 1). The conversion of propionic acid
over silicalite increased with the reaction temperature, although,
predictably, with some loss in 3-pentanone selectivity. At 500 1C
and a space time of 4.0 h g molꢀ1 the base-modified silicalite gave a
very good 3-pentanone yield of 77% with 92% selectivity at 84%
conversion (average for 6 h time-on-stream). Increasing the space
time to 10 h g molꢀ1 (i.e. a 2.5-fold residence time increase) increased
the yield to 87% which matches the performance of the best metal
oxide catalysts.3a The silicalite catalyst showed a stable performance
at this temperature for at least 28 h with 84–92% 3-pentanone
selectivity at 93–80% propionic acid conversion (Fig. 1). A small
decrease in catalyst activity over time-on-stream was probably caused
Fig. 2 DRIFT spectra of Aerosil-300 {(a) and (b)} and silicalite {(c)–(g)}. Aerosil-
300: (a) unmodified and (b) modified with 3.7 M NH3(aq) + 0.7 M NH4NO3, both
pretreated at 500 1C in N2 for 1 h. Silicalite unmodified, pretreated in N2 for 1 h
at: (c) 400 1C and (d) 500 1C; silicalite modified with 3.7 M NH3(aq) + 0.7 M
NH4NO3, pretreated in N2 for 1 h at: (e) 300 1C, (f) 400 1C, (g) 500 1C.
Fig. 1 Time course for propionic acid ketonisation: 0.2 g silicalite modified with
3.7 M NH3(aq) + 0.7 M NH4NO3, 500 1C, 2 vol% propionic acid, 20 mL minꢀ1 N2,
4.0 h g molꢀ1 space time.
c
This journal is The Royal Society of Chemistry 2013
Chem. Commun., 2013, 49, 3842--3844 3843

View Article Online
ChemComm
Communication
Ketonisation occurs with carboxylic acids possessing a-hydrogen
atoms.3a Several mechanisms have been proposed for this reaction.3
These include (i) decomposition of metal carboxylate, (ii) via an acid
anhydride intermediate route, (iii) via a b-keto acid intermediate
route and (iv) via a ketene intermediate route. The mechanism
may depend on many factors, especially the type of catalyst used.
Pestman et al.11 suggested ketene intermediacy for the ketonisation
of acetic acid over metal oxides (eqn (2)). Ketonisation of propionic
acid over heteropoly acid H3PW12O40 may also occur via the ketene
intermediate route.3h Silica has been patented as a catalyst for the
gas-phase production of ketenes from carboxylic acids.12 This
suggests that ketonisation of propionic acid over amorphous silica
and silicalite may occur via ketene intermediacy as well.
Scheme 1 Formation of silanol nest.
CH3CH2COOH ꢀ! CH3CHQCQO
ꢀH2O
(2)
þCH3CH2COOH
ðCH CH Þ CO þ CO
ƒƒƒƒƒƒƒƒƒ!
3
2
2
2
In conclusion, we have demonstrated that neutral materials –
amorphous silica and crystalline silicalite – are active and environ-
mentally benign catalysts for propionic acid ketonisation.
Silicalite is particularly efficient, and its ketonisation selectivity
is increased by base modification. DRIFT spectroscopy indicates
that silanol nests generated by base modification in silicalite are
responsible for its enhanced catalytic activity.
Fig. 3 XRD patterns for silicalite: (a) unmodified; (b)–(d) modified by: (b) 3.7 M
We thank Al-Baha University, Saudi Arabia, for PhD studentship
(H. Bayahia) and Dr Mshari Alotaibi for his help with experiments.
NH3(aq) + 0.7 M NH4NO3, (c) 3.7 M NH3(aq) and (d) 0.1 M HCl.
Notes and references
compared to that of an isolated silanol group. This effect leads to
increased reactivity of silanol nests, as documented for the
Beckmann rearrangement of cyclohexanone oxime over defective
silicalite.7 As can be seen from the band in the 3600–3100 cmꢀ1
region, our base-modified silicalite samples {(e)–(g)} have a much
higher density of silanol nests compared to that of the unmodified
silicalite {(c) and (d)} (Fig. 2). This is in agreement with the previous
report,7 which has demonstrated the formation of silanol nests in
silicalite upon base treatment. By contrast, the base treatment did
not create new silanol nests in amorphous Aerosil-300, cf. (a) and
(b), and the amount of silanol nests in it was significantly smaller
than in the base-treated silicalite. Therefore, our DRIFTS data
indicate that the silanol nests may be the active sites responsible
for the high selectivity of the base-modified silicalite in propionic
acid ketonisation. It should be noted that the density of silanol
nests decreases upon heating above 500 1C due to dehydration of
silanol groups.9a This may be the reason for the lower efficiency of
the unmodified silicalite (Table 2), which had been calcined at
550 1C for 12 h in the final step of its preparation (see ESI†).
The XRD pattern of our silicalite samples (Fig. 3) matched the
one of the authentic material.7 It did not change upon treatment
with base or acid, which indicates that the crystal structure of
silicalite was not affected by the chemical modification. Base-treated
samples (b) and (c), however, show small broad humps centred at a
2y of 151, which may be attributed to amorphous silica formed upon
generating silanol nests by base treatment (Scheme 1). Therefore the
XRD is in agreement with the formation of silanol nests upon base
treatment of silicalite. The texture of zeolite was not affected by the
chemical treatment, as its surface area, pore diameter and pore
volume practically did not change (see ESI†).
1 A. Corma, S. Iborra and A. Velty, Chem. Rev., 2007, 107, 2411;
E. L. Kunkes, D. A. Simonetti, R. M. West, J. C. Serrano-Ruiz,
C. A. Gaertner and J. A. Dumesic, Science, 2008, 322, 417.
2 H. Bernas, K. Eranen, I. Simakova, A.-R. Leino, K. Kordas,
J. Myllyoja, P. Maki-Arvela, T. Salmi and D. Yu. Murzin, Fuel,
2010, 89, 2033; M. Arend, T. Nonnen, W. F. Hoelderich, J. Fischer
and J. Groos, Appl. Catal., A, 2011, 399, 198 (and references therein).
3 (a) M. Renz, Eur. J. Org. Chem., 2005, 979; (b) T. Yokoyama and
N. Yamagata, Appl. Catal., A, 2001, 221, 227; (c) O. Nagashima, S. Sato,
R. Takahashi and T. Sodesawa, J. Mol. Catal. A, 2005, 227, 231; (d) A. D.
Murkute, J. E. Jackson and D. E. Miller, J. Catal., 2011, 278, 189;
(e) C. A. Gaertner, J. C. Serrano-Ruiz, D. J. Braden and J. A. Dumesic,
J. Catal., 2009, 266, 71; ( f ) H. Benaissa, P. N. Davey, Y. Z. Khimyak and
I. V. Kozhevnikov, J. Catal., 2008, 253, 244; (g) H. Benaissa, P. N. Davey,
Y. Z. Khimyak, E. F. Kozhevnikova and I. V. Kozhevnikov, Appl. Catal., A,
2008, 351, 88; (h) M. A. Alotaibi, E. F. Kozhevnikova and I. V.
Kozhevnikov, Appl. Catal., A, 2013, 447–448, 32.
4 Introduction to zeolite science and practice, ed. H. Van Bekkum, E. M.
Flanigen, P. A. Jacobs and J. C. Jansen, Elsevier, Amsterdam, 2nd edn, 2001.
5 P. G. Blake and K. J. Hole, J. Chem. Soc. B, 1966, 577.
6 Silicalite was treated with an aqueous HCl or NH3(aq) + NH4NO3
solution in a Teflon-lined autoclave at 90 1C for 1 h, filtered off,
washed thoroughly with distilled water and oven dried at 110 1C for
4 h.7 Prior to reaction the catalyst was heated in situ at reaction
temperature in N2 flow for 1 h to remove ammonia and nitrate.
7 G. P. Heitmann, G. Dahlhoff and W. F. Hoelderich, J. Catal., 1999, 186, 12.
8 The DRIFT spectra were recorded at room temperature and ambient
pressure under dry N2 using powdered catalyst mixtures with KBr.
The catalysts were pretreated at 300–500 1C for 1 h in N2 flow, as for
ketonisation reaction.
9 (a) S. Bordiga, P. Ugliengo, A. Damin, C. Lamberti, G. Spoto,
`
A. Zecchina, G. Spano, R. Buzzoni, L. Dalloro and F. Rivetti, Top.
Catal., 2001, 15, 43; (b) K. Barbera, F. Bonino, S. Bordiga, T. V. W.
Janssens and P. Beato, J. Catal., 2011, 280, 196.
10 T. Bucko, L. Benco and J. Hafner, Stud. Surf. Sci. Catal., 2005, 158, 601.
11 R. Pestman, A. Van Duijne, J. A. Z. Pieterse and V. Ponec, J. Mol.
Catal. A, 1995, 103, 175.
12 P. C. Watson, M. C. Libby and M. A. Barteau, U.S. Pat., 5475144, 1995.
c
3844 Chem. Commun., 2013, 49, 3842--3844
This journal is The Royal Society of Chemistry 2013