Iron-catalysed propylene epoxidation by nitrous oxide: Dramatic shift of allylic oxidation to epoxidation by the modification with alkali metal salts

Article abstract of DOI:10.1039/b402839b

A dramatic shift of allylic oxidation to epoxidation has been observed during the oxidation of propylene by N₂O when the FeO _x/SBA-15 catalyst is modified with alkali metal salts, and the roles of alkali metal salts are to suppress the reactivity of lattice oxygen and to induce an iron coordination structure effective for epoxidation with N ₂O.

Full text of DOI:10.1039/b402839b

Iron-catalysed propylene epoxidation by nitrous oxide: dramatic shift of
allylic oxidation to epoxidation by the modification with alkali metal salts
Xiaoxing Wang, Qinghong Zhang, Qian Guo, Yinchuan Lou, Lüjuan Yang and Ye Wang*
State Key Laboratory of Physical Chemistry of Solid Surfaces, Department of Chemistry, Xiamen University,
Xiamen 361005, P. R. China. E-mail: yewang@jingxian.xmu.edu.cn; Fax: (+86)-592-2183047;
Tel: (+86)-592-2187470
Received (in Cambridge, UK) 24th February 2004, Accepted 15th April 2004
First published as an Advance Article on the web 14th May 2004
A dramatic shift of allylic oxidation to epoxidation has been
observed during the oxidation of propylene by N2O when the
FeOx/SBA-15 catalyst is modified with alkali metal salts, and
the roles of alkali metal salts are to suppress the reactivity of
lattice oxygen and to induce an iron coordination structure
effective for epoxidation with N2O.
reactions were carried out using a fixed-bed flow reactor. The
products were analyzed by two on-line gas chromatographs
equipped with FID and TCD detectors, respectively. A Porapak T
column connected to an FID detector was used for the analysis of
C3H6, PO and other oxygenated products.
Table 1 shows the results for the oxidation of C3H6 with N2O
over the FeOx/SBA-15 with and without modification by KCl.
SBA-15 with or without modification was always inactive in the
absence of iron, suggesting that iron was the active site. Over the
FeOx/SBA-15 without KCl, acrolein was mainly formed in addition
to COx (CO and CO2), indicating the occurrence of the allylic
oxidation of C3H6. After the modification by KCl, PO became the
main product over the samples with iron content lower than 3 wt%.
PO selectivity decreased remarkably with an increase in iron
content up to 5 wt% and only COx was observed over the KCl–
Fe2O3, strongly suggesting that highly dispersed iron sites are
responsible for the epoxidation of C3H6 with N2O. The sample with
iron content of 1 wt% exhibited the highest performance for PO
formation. This iron content is much higher than that in the Na+–
Fe2O3/SiO2 catalyst (100–1000 ppm).7c We have obtained better
catalytic performances with such a relatively higher iron content
probably because iron can be dispersed in a better way in the
mesopores of SBA-15.
Fig. 1 shows the effect of KCl content in the 1 wt% FeOx/SBA-
15 on catalytic performances. The increase in the content of the
modifier dramatically increased the selectivity to PO and decreased
those to allylic oxidation and deep oxidation products. In other
words, a dramatic shift from allylic oxidation to epoxidation took
place by the modification of FeOx/SBA-15 with KCl. Moreover, it
is interesting that, with increasing K/Fe ratio, C3H6 conversion first
decreased slightly but then underwent a remarkable rise and
increased for more than four times as compared with the sample
without modification.
The epoxidation of propylene to propylene oxide (PO) is one of the
most challenging chemical reactions. Although TS-1 and Pd/TS-1
can catalyze this reaction efficiently in liquid phase with H2O2 and
O2–H2 gas mixture, respectively,1,2 only a few studies have
succeeded in achieving high efficiency for the gas phase epoxida-
tion of C3H6.3 Unlike C2H4, because C3H6 possesses allylic
hydrogen, the easy abstraction of this hydrogen in the gas phase
reaction usually results in allylic intermediates leading to allylic
oxidation and further deep oxidation. There have been many
attempts to use O2 as an oxidant for the epoxidation of C3H6, but
high PO selectivity can only be obtained over a few catalysts.3 PO
selectivity of ca. 90% has been reported when H2 is co-fed with O2
over Au–Ti-based catalysts, but the conversion of C3H6 and the
efficiency of H2 are still not satisfactory.4
On the other hand, there are incentives to exploit N2O as an
oxidant for selective oxidation reactions since the by-product of an
oxidation reaction would be N2 and the process would be
environmentally benign. The selective oxidation of benzene to
phenol with N2O has been intensively studied over iron-containing
ZSM-5.5 One of the present authors has found that CH4 and C2H6
can be oxidized to the corresponding alcohols with N2O over an
iron phosphate catalyst.6 Reports on the use of N2O for epoxidation
of alkenes are, however, scarce.7
Recently, we investigated the catalytic property of an iron-
containing SBA-15 catalyst for the gas phase oxidation of C3H6
with N2O, and found an interesting shift of reaction pathway from
allylic oxidation to epoxidation by the modification of the catalyst
with an alkali metal salt such as KCl. Although a Na+–Fe2O3/SiO2
has been reported for the epoxidation of C3H6, the role of the alkali
metal ion is not clear.7c Through our work, we have gained
significant information about the factors controlling the reaction
route to allylic oxidation or to epoxidation. In this communication,
we report the dramatic shift of reaction route by the modification of
the catalyst and the elucidation of the roles of the modifier.
SBA-15, a typical mesoporous silica, was synthesized with the
procedures reported previously.8 The FeOx/SBA-15 sample was
prepared by an impregnation method using an ethanolic solution of
[FeIII(acac)3] (acac = acetylacetonate). After calcination at 823 K,
the sample was further modified with alkali salt by the impregna-
tion followed by drying and calcination at 823 K for 6 h. XPS
measurements showed that the oxidation state of iron was III in the
calcined FeOx/SBA-15 and did not change after the modification
with KCl. XRD and TEM studies confirmed that the ordered
mesoporous structure of SBA-15 was sustained after the introduc-
tion of FeOx (0–5 wt%) and KCl (K/Si = 0–0.12), but the surface
area and pore volume determined by N2-sorption decreased
gradually with an increase in the content of either iron or KCl,
suggesting the filling of the mesoporous channels with FeOx and
KCl. That any crystalline phases of iron oxide and KCl were not
observed for the KCl–FeOx/SBA-15 (Fe, 0–5 wt%; K/Si, 0–0.12)
indicated the high dispersions of both FeOx and KCl. Catalytic
Table 1 Selective oxidation of propylene with N2O over FeOx/SBA-15 and
KCl-modified FeOx/SBA-15
Conversion
Catalyst
(%)
Selectivity (%)
PO
ALa AAb Others COx
0.5 wt% FeOx/SBA-15 0.65
0
0
0
57
42
36
10
1.8 1.1
1.6 1.1
1.8 0.5
2.1 0.6
1.7 0.5
2.6 1.4
7.8 1.2
3.7
4.5
2.6
0
7.3
7.5
7.4
5.0
6.1
7.3
4.7
5.3
6.8
32
46
54
85
15
18
12
12
14
16
40
100
1 wt% FeOx/SBA-15
5 wt% FeOx/SBA-15
Fe2O3
1.17
1.32
0.52
0
K+–0.5 wt% FeOx/SBAc 2.59
K+–1 wt% FeOx/SBAc 4.51
K+–1 wt% FeOx/SBAd 1.87
K+–1 wt% FeOx/SBAe 1.21
K+–1 wt% FeOx/SBAf 3.77
K+–3 wt% FeOx/SBAc 4.09
76
72
81
80
77
68
14
12
37
K+–5 wt% FeOx/SBAc 5.80
g
K+–Fe2O3
trace
Reaction conditions: P(C3H6) = 2.5 kPa, P(N2O) = 25.3 kPa; T = 598 K;
catalyst, 0.20 g; total flow rate, 60 ml min21; TOS, 30 min.a AL = acrolein.
b AA = allyl alcohol. c K/Si = 0.04. d TOS, 110 min. e TOS, 190 min.
f Regenerating the catalyst after 240 min of reaction, followed by reaction
for 30 min. g K/Fe = 1.
1396
C h e m . C o m m u n . , 2 0 0 4 , 1 3 9 6 – 1 3 9 7
T h i s j o u r n a l i s © T h e R o y a l S o c i e t y o f C h e m i s t r y 2 0 0 4

with K+. This corresponds to the inhibition of the allylic oxidation
and is consistent with the consensus that the nucleophilic lattice
oxygen species generally attack the allylic hydrogen and induce
allylic oxidation. In other words, one role of K+ is to inhibit the
reactivity of lattice oxygen and thus to suppress the allylic
oxidation.
Because not only PO selectivity but C3H6 conversion also
increased remarkably after the addition of K+, the presence of K+
must also accelerate the activation of N2O for C3H6 epoxidation.
Our investigations have suggested that the change in the coordina-
tion structure of iron brought by K+ may account for such an
accelerating effect. A distinct Raman band at ca. 1000 cm21 was
observed after the modification of the FeOx/SBA-15 with an alkali
metal salt in UV-Raman studies, which have been shown to be
useful in determining the coordination structure of iron.9 In Fig. 2B,
the FeOx/SBA-15 showed Raman bands at 981, 1074 and 1138
cm21 in the region of 900–1200 cm21. The latter two bands
probably arise from the tetrahedral iron site in the framework of
SBA-15. A part of iron may be incorporated into the framework of
SBA-15 in the FeOx/SBA-15. After the modification with K+, these
bands became unnoticeable, and a band at 1000 cm21 appeared
distinctly. Another band at 808 cm21 also appeared after the
modification with K+, but this band could also be observed for a K+/
SBA-15 without iron (curve g of Fig. 2B). Thus, the band at 808
cm21 may not be related to iron sites. The intensity of the band at
1000 cm21 increased with an increase in K/Fe ratio from 0 to 5.0,
and slightly decreased with a further increase in K/Fe ratio. Since
this band shifts slightly to a lower wavenumber as compared with
those assigned to the framework Fe–O–Si, we tentatively speculate
that it arises from the tetrahedral FeO4 on the surface, which is
probably stabilized by alkali ions. Such an iron structure should be
effective for the activation of N2O for the epoxidation of C3H6.
In conclusion, we have found a dramatic shift of reaction route
from allylic oxidation to epoxidation during the oxidation of
propylene with N2O by modifying the FeOx/SBA-15 with K+. The
roles of K+ are to suppress the allylic oxidation by decreasing the
reactivity of lattice oxygen and to induce a coordination structure of
iron effective for epoxidation.
Fig. 1 Changes of catalytic performances with a change in K/Fe atomic ratio
in the modified 1 wt% FeOx/SBA-15. Symbols of selectivity: (5) propylene
oxide, (8) allyl alcohol, (:) acrolein, (2) COx.
Further investigations revealed that, similar shift of allylic
oxidation to epoxidation likewise occurred when other potassium
salts such as KNO3, KAc or KBr were used as the modifier instead
of KCl. Thus, it was K+ but not Cl2 played an important role. Other
alkali metal salt such as LiCl, NaCl, RbCl or CsCl also worked as
useful modifiers for PO formation. PO selectivity and yield
decreased with the following sequence, KCl > NaCl > RbCl >
CsCl > LiCl. By changing reaction conditions, PO selectivity of
80% can be obtained at 4% C3H6 conversion over a K+–FeOx/SBA-
15 catalyst, higher than that reported over the Na+–Fe2O3/SiO2
catalyst (70% PO selectivity at 2.5% C3H6 conversion).7c The
turnover frequency for PO formation over our catalyst was 5.2 mol
(mol-Fe)21 h21. Similar to many catalysts reported so far,3,4,7c the
activity of our catalyst also decreased with time on stream (TOS) as
shown in Table 1. It has been clarified that the decrease in C3H6
conversion with TOS is caused by the carbon deposition, and the
activity can almost be recovered by regenerating the catalyst with a
gas flow containing He and O2 at 823 K.
We have clarified that the redox property is an important factor
in determining the reaction route. The H2–TPR studies indicated
that the reactivity of the lattice oxygen of FeOx in the FeOx/SBA-15
was remarkably inhibited by the modification with K+ (Fig. 2A). A
reduction peak corresponding to the reduction of Fe3+ to Fe2+ was
observed at 682 K for the sample without modification. The main
reduction peak was shifted to 888–918 K after the modification
This work was supported by the NSF of China (Nos. 20273054
and 20021002) and the National Basic Research Program of China
(No. 2003CB615803).
Notes and references
1 M. G. Clerici, G. Bellussi and U. Romano, J. Catal., 1991, 129, 159.
2 R. Meiers, U. Dingerdissen and W. F. Hölderich, J. Catal., 1998, 176,
376.
3 (a) K. Murata and Y. Kiyozumi, Chem. Commun., 2001, 1356; (b) J. D.
Jewson, B. Cooker and A. M. Gaffney, US 5780657, 1998.
4 (a) T. A. Nijhuis, B. J. Huizinga, M. Makkee and J. A. Moulijin, Ind. Eng.
Chem. Res., 1999, 38, 884; (b) B. S. Uphade, T. Akita, T. Nakamura and
M. Haruta, J. Catal., 2002, 209, 331; (c) M. P. Kapoor, A. K. Sinha, S.
Seelan, S. Inagaki, S. Tsubota, H. Yoshida and M. Haruta, Chem.
Commun., 2002, 2902.
5 G. I. Panov, Cattech, 2000, 4, 18 and references therein.
6 Y. Wang and K. Otsuka, J. Chem. Soc., Faraday Trans., 1995, 91,
3953.
7 (a) R. Ben-Daniel, L. Weiner and R. Neumann, J. Am. Chem. Soc., 2002,
124, 8788; (b) T. Yamada, K. Hashimoto, Y. Kitaichi, K. Suzuki and T.
Ikeno, Chem. Lett., 2001, 268; (c) V. Duma and D. Hönicke, J. Catal.,
2000, 191, 93.
8 (a) D. Zhao, J. Feng, Q. Huo, N. Melosh, G. H. Fredrickson, B. F.
Chmelka and G. D. Stucky, Science, 1998, 279, 548; (b) W. Yang, X.
Wang, Q. Guo, Q. Zhang and Y. Wang, New J. Chem., 2003, 27,
1301.
9 (a) Y. Yu, G. Xiong, C. Li and F.-S. Xiao, J. Catal., 2000, 194, 487; (b)
X. Wang, Y. Wang, Q. Tang, Q. Guo, Q. Zhang and H. Wan, J. Catal.,
2003, 217, 457.
Fig. 2 TPR profiles (A) and UV-Raman spectra (B). (a) 1 wt% FeOx/SBA-
15, (b) K+–1 wt% FeOx/SBA-15 (K/Fe = 2.5), (c) K+–1 wt% FeOx/SBA-15
(K/Fe = 5), (d) K+–1 wt% FeOx/SBA-15 (K/Fe = 10), (e) K+–1 wt% FeOx/
SBA-15 (K/Fe = 15), (f) SBA-15, (g) K+/SBA-15 (K/Si = 0.04).
C h e m . C o m m u n . , 2 0 0 4 , 1 3 9 6 – 1 3 9 7
1397