Catalyst-free gas-phase epoxidation of alkenes

Article abstract of DOI:10.1246/cl.2005.584

Butadiene, styrene, cyclohexene, allyl acetate, methyl methacrylate, and allyl alcohol were epoxidized in a gas-phase reaction in the absence of a catalyst. The applied oxidizing agent is ozone. With exception of allyl alcohol (selectivity to glycidol: 58%), the selectivity to the corresponding epoxide ranged from 88 to 97%. For acrylonitrile, there was no measureable conversion. Copyright

Full text of DOI:10.1246/cl.2005.584

584
Chemistry Letters Vol.34, No.4 (2005)
Catalyst-free Gas-phase Epoxidation of Alkenes
Torsten Berndtꢀ and Olaf Boge
Leibniz-Institut fur Tropospharenforschung e.V., Permoserstrasse 15, 04318 Leipzig, Germany
¨
¨
¨
(Received January 28, 2005; CL-050128)
Table 1. Experimental findings for a temperature of 250 ꢁC
Butadiene, styrene, cyclohexene, allyl acetate, methyl meth-
acrylate, and allyl alcohol were epoxidized in a gas-phase reac-
tion in the absence of a catalyst. The applied oxidizing agent is
ozone. With exception of allyl alcohol (selectivity to glycidol:
58%), the selectivity to the corresponding epoxide ranged from
88 to 97%. For acrylonitrile, there was no measureable conver-
sion.
Selectivity
Alkene feed Conversion
Alkene
to epoxide By-products
/%
/vol %
/%
Butadiene
Styrene
1.4
1.3
1.4
1.6
28
22
36
18
97
91
89
97
Acrolein, furan
Benzaldehyde
—
Cyclohexene
Allyl acetate
Methyl
—
Methyl
Epoxides are characterized by the highly reactive epoxy
group caused by polarity and ring strain. This opens the possibil-
ity for a series of reactions frequently used in organic synthesis.1
Generally, for the formation of epoxides several routes are appli-
cable using percarboxylic acids,2 hydroperoxides,3 or hydrogen
peroxide.4 For the economically valuable ethylene oxide, a cat-
alytic direct oxidation path exists with O2 or air feed.5 For pro-
pylene oxide, this approach does not work and at present com-
plex multistep manufacturing processes are applied.5
In first communications from our laboratory, a novel route
for the epoxidation of alkenes was presented proceeding in the
homogeneous gas phase under low-pressure conditions.6 In this
method, O3 reacts with NO2 and the resulting N(V)-oxides
(NO3, N2O5, N2O6) epoxidize the alkene forming finally the cor-
responding epoxide and NO2. To date, it is not clear what the
dominating N(V)-oxide is for the epoxidation process. In the
course of the overall reaction, NO2 is not consumed.
1.3
1.4
23
15
88
58
methacrylate
Allyl alcohol
pyruvate
Acrolein
In Table 1 a summary of the experimental finding is given.
For the epoxidation of butadiene, Figure 1 shows the con-
version of the alkene and the selectivity to butadiene monoxide
as a function of the reaction temperature for constant feed com-
position. Generally, a small temperature dependence of conver-
sion and selectivity was observed in the investigated range of
180–300 ꢁC. Butadiene monoxide was the only detected
epoxide. There was no indication for the occurrence of the
corresponding dioxide. The ratio of reacted butadiene/initial
O3 = 1.0 measured in the whole temperature range stands for
an efficient utilization of O3. Acrolein, furan (GC–MS), and
traces of HNO3 (FT-IR) were identified as by-products. The oc-
currence of acrolein indicates that in competition to pathway (2)
carbonyl formation takes place after cleavage of the double
bond. Carbonyl formation as a competing process with respect
to epoxidation was already observed in former studies.6 For
the other by-product furan, from the mechanistic point of view,
it can only be speculated how the ring closure occurs. In a few
runs the feed composition was changed. The O3/NO2 feed
was increased by a factor of 3.6 for a constant butadiene feed
of 1.5 vol % resulting in an increase of the butadiene conversion
from 28 to 76% at 250 ꢁC. The selectivity to butadiene monox-
ide, however, dropped from 97 to 78% and the ratio of reacted
butadiene/initial O3 was ꢂ0:8. A further increase of the O3/
NO2 feed by a factor of 1.7 (excess of O3 over butadiene!) yield-
ed a selectivity to butadiene monoxide of 67% for total conver-
sion of butadiene. Obviously, under conditions of raised buta-
diene conversion, other reactions than the desired pathway (2)
become more important. Probably, in a consecutive step the
N(V)-oxide attacks butadiene monoxide resulting in a lowering
of the selectivity to this product.
(NO2)
O3
+
NO2
N(V)-oxide + O2
O
(1)
(2)
R4
R1
R3
R2
R4
R3
+ xNO2
N(V)-oxide +
R1 R2
Using this synthesis route, ethylene, propylene, and butenes
were epoxidized very efficiently with a selectivity to the epoxide
of close to 100%.6 Subject of this work is to find out whether this
route is also useful for other alkenes (conjugated dienes, cyclic
compounds, alkenes with polar groups) allowing a more gener-
alized application for alkene’s epoxidation in gas phase. In solu-
tion, a mixture of O3 and NO2 leads to a nitration of organics
without epoxide formation.7
The experiments have been performed in a flow tube
(length, 100 mm; inner diameter, 5 mm; quartz glass) at a total
pressure of 10 mbar. O3 and NO2 were premixed and the result-
ing mixture was added to the alkene (diluted in He) at the en-
trance of the heated reaction zone (180–300 ꢁC). NO2 was chos-
en in a ꢂ5-fold excess over O3 for preventing the reaction of O3
with the alkene. The feed fraction of the alkenes was 1.3–
1.6 vol % and the residence time of the gas mixture in the reac-
tion zone was 8–16 ms. At the outlet of the flow tube a gas cell
was attached for FT-IR analysis. Additionally, online GC–MS
analysis of gas samples was carried out with special attention
to the identification of by-products.
For styrene, cyclohexene, allyl acetate, and methyl metha-
crylate, temperature dependent measurements were carried out
with an alkene conversion of 18–36%. The corresponding epox-
ides represented the main products with selectivities close to
100% under these conditions, cf. examples given in Table 1.
By-products, if any detected, arose from the cleavage of the for-
mer double bond, i.e. benzaldehyde from styrene and methyl
Copyright Ó 2005 The Chemical Society of Japan

Chemistry Letters Vol.34, No.4 (2005)
585
and water is also possible directly from the reaction of the N(V)-
oxides with allyl alcohol in competition to the desired glycidol
formation. The temperature dependence measurement, given in
Figure 2, shows an increase of the selectivity to glycidol with in-
creasing temperature. For acrolein, an inverse behavior was
measured. If thermal decomposition of glycidol is responsible
for the formation of acrolein, with increasing temperature a de-
crease of glycidol is expected contrary to the observation. That
makes the direct pathway (in competition to glycidol formation)
more likely than the consecutive reaction (thermal decomposi-
tion) of glycidol formed. The ratio of reacted allyl alcohol/initial
O3 was ꢂ0:55 pointing at N(V)-oxide consuming steps other
than the epoxidation such as the reaction with water producing
HNO3.
100
butadiene monoxide
80
60
40
20
0
conversion
175
200
225
250
275
300
In experiments with acrylonitrile, the amount of reacted al-
kene was too low for a reasonable determination of the conver-
sion even for an increased acrylonitrile feed of 2.9 vol %. As a
result of GC–MS analysis, two trace products were obtained.
No endeavors have been made for identification.
Temperature / °C
Figure 1. Temperature dependent data for butadiene conver-
sion and selectivity to butadiene monoxide for constant feed
composition.
In summary, the formation of epoxides from a series of al-
kenes was studied in the gas phase in the absence of a catalyst.
A N(V)-oxide, produced in the reaction of O3 with NO2, served
as the real oxidizing agent. For butadiene, styrene, cyclohexene,
allyl acetate, and methyl methacrylate, the selectivity to the cor-
responding epoxide was 88–97% with an alkene conversion of
18–36%. Detected by-products were mainly carbonyls derived
from the cleavage of the double bond attacked. For allyl alcohol,
the selectivity to glycidol was distinctively lower (58%) caused
by a competing/consecutive path leading to acrolein. The reac-
tivity of acrylonitrile was found to be insufficient for a measur-
able reactant conversion.
100
80
HNO3
60
glycidol
40
acrolein
20
conversion
References
0
1
2
3
J. G. Smith, Synthesis, 1984, 629.
175
200
225
250
275
300
N. Prileshajev, Chem. Ber., 42, 4811 (1909).
K. B. Sharpless and T. R. Verhoeven, Aldrichimica Acta,
12, 63 (1979).
M. G. Clerici and P. Ingallina, J. Catal., 129, 167 (1993).
K. Weissermel and H.-J. Arpe, ‘‘Industrielle Organische
Chemie,’’ 5th ed., VCH, Weinheim (1998).
Temperature / °C
Figure 2. Temperature dependent data for the conversion of
allyl alcohol and the selectivity to glycidol, acrolein, and HNO3
for constant feed composition.
4
5
pyruvate from methyl methacrylate.
6
7
8
a) T. Berndt, O. Boge, J. Heintzenberg, and P. Claus, Ind.
¨
In the case of the epoxidation of allyl alcohol, acrolein and
HNO3 were detected in large amounts beside the main product
glycidol, cf. Figure 2. It is noteworthy, that HNO3 was twice
as much as acrolein. Glycidol is a less stable product that tends
to polymerize at higher temperatures or it decomposes.8 Thermal
decomposition of glycidol can be assumed for describing the
formation of acrolein together with water. In a subsequent step,
water can react with N2O5 (a possible N(V)-oxide) via H2O +
N2O5 ! 2 HNO3. On the other hand, the production of acrolein
Eng. Chem. Res., 42, 2870 (2003). b) T. Berndt and O. Boge,
¨
Ind. Eng. Chem. Res., 44, 645 (2005). c) Eur. Patent 1299367
(2001); Chem. Abstr., 136, 232683 (2002).
a) T. Mori and H. Suzuki, Synlett, 1995, 383. b) H. Suzuki
and T. Mori, J. Org. Chem., 62, 6498 (1997). c) H. Suzuki
and N. Nonoyama, J. Chem. Soc., Perkin Trans. 1, 1997,
2965.
A. Kleemann and R. Wagner, ‘‘Glycidol,’’ Huthig Verlag,
¨
Heidelberg (1981).
Published on the web (Advance View) March 19, 2005; DOI 10.1246/cl.2005.584