J. Chem. SOC., Faraday Trans. I , 1986, 82, 3 5 4 3 Formation of an Epoxide Intermediate in the Photo-oxidation of Alkenes over Silica-supported Vanadium Oxide Tsunehiro Tanaka, Masaharu Ooe, Takuzo Funabiki and Satohiro Yoshida* Department of Hydrocarbon Chemistr?; and Division of Molecular Engineering, Kyoto University, Kyoto 606, Japan Photo-induced isomerization of propene oxide and photo-oxidation of ethene have shown that photo-oxidation of alkenes proceeds by a mechanism involving not only the oxidative cleavage of the olefinic double bond and oxidation via 71-ally1 intermediates, but also the formation of epoxides as a precursor. The product distribution in the photo-oxidation of butenes is well explained by the mechanism. A tracer study using 1 8 0 2 has indicated that oxygen involved in the epoxide intermediate comes from the lattice oxygen of the catalyst.Vanadium oxide supported on silica acts as a catalyst in the photo-oxidation of propene to produce aldehydes as the main products at a high level of conversion.l This is in marked contrast to the catalysis by TiO, which gives CO, as a main product even at a low conversion level (0.055%).2 In previous work we studied the mechanism of photo-oxidation of propene over V,O,/SiO, and proposed that the reaction proceeds along two pathway^.^ One is the double-bond fission of propene to form ethanal and the other is the formation of a n-ally1 intermediate by hydrogen abstraction from propene with adsorbed oxygen. The n-ally1 intermediate is fairly stable at room temperature and is decomposed thermally to form acrylaldehyde selectively.We have also shown that the oxygen atom in these aldehydes comes from lattice oxygen. In the proposed mechanism, the formation of propanal which was found as a minor product was neglected. The present work was carried out to elucidate the reaction path leading to propanal. A general mechanism for the photo- oxidation of lower alkenes over V,O,/SiO, will be discussed. Ex per imen t a1 V,O,/SiO, (V,O,, 5 wt%) was prepared conventionally by impregnating silica in an aqueous solution of ammonium metavanadate,4 and was treated with 60 T0rr.f. oxygen at 673 K for several hours before each run. The reactants were commercially supplied and purified by vacuum distillation at low temperature. lS0, was supplied from CEA-SEN Saclay (purity 99.88 %) and used without further purification.The reactions were performed in a conventional closed circulating system described el~ewhere.~ U.V. light was irradiated at room temperature from a 500 W Xe lamp through a glass filter transparent to wavelengths A > 300 nm. The temperature of the catalyst bed was elevated by ca. 10 K by the U.V. irradiation. Products were analysed by g.1.c. and mass spectrometry. Products were collected by the following steps: (A) after a reaction, products in the gas phase were frozen out in a liquid nitrogen trap in the dark at room temperature; (B) the catalyst was irradiated in vacuo to check photodesorption of the adsorbed species; and ( C ) the catalyst was heated in the dark to 473 K to collect the t 1 Torr z 133.3 Pa.3536 Photo-oxidation of Alkenes over Vanadium Oxide thermally desorbed products. Of these steps, A and B gave nearly the same distribution of products. Thus, except in specified cases, the amounts of products collected in steps A and B were summed and presented as the amount (A + B). Experiments under the same conditions were performed from twice or more in order to check the reproducibility of each result. The composition of products was highly reproducible, but the total yield of products was slightly affected by the different batches of the catalyst. However, the dispersion of data was not so large as to affect the points of the argument. The highest yields of products obtained under the same conditions are presented in the tables.Results Reaction of Propene Oxide As described previously, the photo-oxidation of propene over V,O,/SiO, formed CH,CHO, C,H,CHO and CH,=CHCHO with the mole ratio of 59: 10:39 at 41.5% conversion and in a 34% yie1d.l We have proposed a mechanism3 which explains the formulation of CH,CHO and CH,=CHCHO, but the formation of C,H,CHO was left unexplained. Since it is probable that C,H,CHO, is formed uia propene oxide, we investigated the reactivity of propene oxide over V,O,/SiO,. Results are given in table 1. When propene oxide was adsorbed on the catalyst in the absence of oxygen in the dark at room temperature and products were collected by heating the catalyst up to 473 K in the dark (step C), the desorbed gases were propene oxide, propanal, ethanal, acrylaldehyde and propene.The same products were formed in the photo-oxidation of propene,l while the product distribution was different. Propanone was not formed in either case. When the catalyst adsorbing propene oxide in the dark was irradiated for 30 min at room temperature and was then heated to 473 K in the dark, conversion of propene oxide and yield of products increased significantly. The percentage of C,H,CHO in the products decreased and that of CH,CHO increased in comparison with the results in the case without irradiation. Some portion of each product was photodesorbed (step B) and thermally desorbed (step C). Ethanal and propene were collected mainly in step B, propanal and acryl- aldehyde in step C. The former result suggests that the fission of the C-C bond and abstraction of oxygen to form ethanal and propene, respectively, occur in the photo- process but not in the thermal process.It was found that the presence of oxygen has little effect on the selectivity. These results indicate that propene oxide is converted predominantly to propanal and the U.V. irradiation greatly promotes the conversion of propene oxide, suggesting that propene oxide is one of intermediate in the photo- oxidation of propene. Photo-oxidation of Ethene The photo-oxidation of ethene was much more sluggish than that of propene and produced ethanal and carbon dioxide in the mole ratio of 39:61 at 5.8% conversion of ethene. 36.2% of produced ethanal and ca. 70% of carbon dioxide were collected in step A + B. Ethanal may be formed via ethene oxide.In order to clarify the source of oxygen incorporated into the products, oxygen tracer experiments were carried out in a series over the same catalyst. The first reaction (run 1) was performed with 1 6 0 2 , the second (run 2) and the third (run 3) with lSO, and the last (run 4) was performed with l60, again. Fig. 1 shows the mass spectra of products in runs 1 and 3 and table 2 summarizes the relative peak heights at selected m/e relating to the products. Pattern coefficients of fragment ions of ethanal were determined as 100.0 (m/e = 29, CHO+), 22.4 (m/e = 44, parent peak) and 16.2 (m/e = 43, CH,CO+) from the results for pure ethanal under the same analytical conditions. The ratio of peakT. Tanaka, M. Ooe, T. Funabiki and S . Yoshida 37 Table 1. Reaction of propene oxide over V,O,/SiO, U.V.products/pmol' /,umola /,urn01 tion* (% )" collectiond CH,CHO C,H,CHO CH,=CHCHO CH,=CHCH, PO 0, irradia- conv. product ~~~ -~ ~ ~~~ ~~ 32.6 0.0 dark 38.7 C 0.7 9.4 1 .o 1.4 30.6 0.0 U.V. 96.6 B,C 5.7(3.6) 18.6(1.1) 1.3(0.1) 3.8(2.3) 48.0 101.2 U.V. 90.9 A+B,C 9.9(7.8) 27.8(0.7) 2.4(0.1) 3. q2.2) ~ ~~~ ~ a Initial amount of propene oxide. * Irradiation time 30 min. " Based on propene oxide. collecting products. Steps A+B, B and C : see text. parentheses are the amount of products collected in step A + B or B. Procedure for A small amount of CO, was detected. The values in + 0 s? I 0 I 1 I I I I I I I I I I I I I I I I I I I I I I I I + + 0 '0 0 aD 0 I I 0 u u I I I + N 0 0 z I I I I I I I I I I I ] I ! I I I I 2 9 30 31 " 43 44 45 46 47 48 mle Fig.1. Mass spectra of products obtained from photo-oxidation of ethene. Full line, run 1 (table 2); dotted line, run 3 (table 2). heights of m/e = 29 and 43 (run 1) was in accordance with the mass pattern of ethanal. The peak of m/e = 44 is higher than that of the estimated value from the coefficients because of the contribution of carbon dioxide (CO;). One may conjecture that the small, but obvious peak of m/e = 30 is assigned to a parent peak of formaldehyde (H,CO+) formed via oxidative cleavage of ethene. However, analysis of products by g.1.c. did not indicate any trace of formaldehyde. This peak was always observed when any substance having m/e = 29 peak was measured, suggesting that the m/e = 30 peak is due to re- combination of fragment ions in the mass chamber.Therefore, this peak can be neglected. In runs 2 and 3, new peaks of m/e = 31 (CH180+) and 45 (CH,C180S) due to ethanal containing l80 appeared. The peak heights of these peaks increased from run 2 to 3, but they were much smaller than those of m/e = 29 (CH160+) and 43 (CH3C160+). In ad- dition, new peaks ofm/e = 46 and 48 due to C180160+ and C180z appeared after run 2 and the peak heights increased slightly from run 2 to 3. In run 4, m/e = 31 and 45 peaks still remained in spite of the reaction with 1 6 0 , and the peak heights of m/e = 29, 3 1,43 and38 Photo-oxidation of Alkenes over Vanadium Oxide Table 2. Mass pattern of products in the photo-oxidation of ethene relative peak height of mass number (rn/e)b runa 29 30 31 43 44 45 46 48 ~- ~~~~~ ~~~ 1 100.0 9.4 0.0 15.0 62.8 0.0 0.0 0.0 2 92.1 10.5 7.2 13.1 34.8 1.4 10.1 9.5 3 80.6 8.4 20.2 12.5 34.5 3.2 12.9 14.9 4 93.5 9.5 16.9 13.7 55.9 4.6 6.0 0.0 ~~~~ a Ethene and oxygen : 58.0 pmol.Photo-oxidation was performed by varying the atmosphere successively as follows: ( l ) , 1 6 0 2 ; (2), ‘*02; (3), 1 8 0 2 ; (4), 1 6 0 2 . Irradiation time was 30 min in each atmosphere. Peak heights relative to that of m/e = 29 in run 1. 0 0 * 5 1 .o 1.5 2 .o Po2 Fig. 2. Influence of partial pressure of oxygen on the yields of products. C,H,, 58.0pmol; 0, CH,CHO; A, CO,. 45, which correspond to ethanal, were not very different from those in run 3. On the other hand, the peak height of m / e = 46 decreased and the peak of m / e = 48 vanished with the increase in the peak height of m/e = 44.These peaks of m/e = 44,46 and 48 are due mainly to carbon dioxide. The formation of carbon dioxide containing l60 in runs 2 and 3 and containing l80 in run 4 suggests that lattice oxygen is also incorporated into carbon dioxide. The peaks due to fragments of ethanal were less dependent on whether lAOZ or l60, was used than those due to carbon dioxide. This result indicates that the oxygen atom of the ethanal comes from the lattice oxygen of the catalyst and infers the formation of an intermediate of epoxide type. The result that ethanal containing l80 was detected even when the reaction was performed with lSO, over the same catalyst (run 4) indicates that gaseous oxygen is consumed not only in the formation of carbon dioxide but alsoT. Tanaka, M.Ooe, T. Funabiki and S. Yoshida 39 in the reoxidation of the catalyst. As shown in fig. 2, the formation of ethanal depended upon partial pressure of oxygen, and ethanal was not formed in the absence of oxygen. The curves of ethanal and carbon dioxide in fig. 2 clearly indicate that conversion of ethanal to carbon dioxide is promoted at the higher partial pressure of oxygen. Photo-oxidation of Butenes To clarify the formation of the epoxide type intermediate, butenes are more suitable substances than ethene and propene because the variety of products gives more information than that from ethene and propene. The results are given in tables 3-5. It is evident that oxidation hardly proceeds without oxygen. Isomerization of butenes did not occur in the dark, however U.V.irradiation induced isomerization, although the extent was small in the 30 min irradiation and independent of the pressure of oxygen. Oxidation products were those formed by bond fissions (CH,CHO, CH,CH,CHO, CH,=CHCHO), saturated ketone and aldehyde (CH,COC,H,, n-C,H,CHO), unsaturated ketone and aldehyde (CH,COCH=CH,, CH,CH=CHCHO), and buta- 1,3-diene. The composition of those products was dependent on the structure of the butenes. The main product from but-1-ene was CH,CH,CHO, but the yield of n-C,H,CHO was fairly large (table 3). On the other hand, the main product from but-2-enes was CH,CHO (tables 4 and 5). In the case of cis-but-2-ene, yields of products other than CH,CHO and buta-l,3-diene were very low. In the case of trans-but-2-ene, yields of CH,COC,H, and buta-1,3-diene are higher than those of other products except for CH,CHO.It is noted that buta-1,3-diene is produced even in the absence of oxygen, but the yield of buta-1,3-diene is increased in the presence of oxygen. These results suggested that hydrogen abstraction from butenes is effected not only by the photoformed hole centre (Orattice) likewise from CH,,6 but also by the adsorbed oxygen.' It was also observed that the yield of buta-1,3-diene obtained by the step C increased when the reaction was performed in the presence of oxygen. This supports the suggestion that the n-ally1 intermediate formed by the abstrac- tion of hydrogen from butene by the adsorbed oxygen is thermally converted to Table 3. Composition of organic compounds in the reaction of but- 1-ene over U.V.irradiated V,O,/SiO,a ~ ~~ ~ yield/pmolb products with 0,' without 0, CH,=CH, 0.4 (0.0) 0.2 (0.1) CH,=CHCH, 0.3 (0.2) 0.5 (0.4) CH,=CHC,H, 19.8 (19.3) 37.6 (37.6) t-CH,,CH=CHCH, 0.2 (0.1) 1.8 (1.7) CH,=CHCH=CHI 2.3 (0.6) 1.5 (1.2) c-CH,CH=CHCH, 0.1 (0.0) 2.0 (2.0) CH,CHO 2.2 (1.6) 0.2 (0.0) C,H,CHO 10.5 (9.0) 0.2 (0.0) CH,=CHCHO 0.9 (0.8) tr" C H ,COC , H , 0.5 (tr) tr n-C,H,CHO 3.5 (1.4) - CH,=CHCOCH, 0.5 (0.2) - CH ,3C H =CH C H 0 1.3 (0.2) _ _ ~ ~~~~ ~~ ~~ ~~~ a Initial amount of but-I-ene: 55 pmol. Irradiation time: 30 min, at room temperature. Values in parentheses are the amounts of products collected in step A + B. 90 pmol. Trace,40 Photo-oxidation of Alkenes over Vanadium Oxide Table 4. Composition of organic compounds in the reaction of trans-but-2-ene over U.V.irradiated V,O,/SiO,a yield/pmol* products with 0,' without 0, - __ CH,=CH, 0.3 (0.3) 0.3 (0.3) CH,=CHCH, 0.2 (0.2) trd CH,=CHC,H, 0.3 (0.3) 0.6 (0.6) t-CH,CH=CHCH, 26.0 (25.7) 30.7 (30.6) c-CH,CH=CHCH, 2.8 (2.6) 3.0 (2.9) CH,=CHCH=CH, 2.6 (1.2) 1.6 (1.4) CH,CHO 15.9 (14.3) 0.6 (0.1) C,H,CHO 0.3 (0.3) tr CH,COC,H, 1.5 (0.1) - n-C,H,CHO 0.2 (tr) ~ CH,=CHCOCH, 0.2 (tr) - CH,CH=CHCHO 1.7 (0.6) - CH,=CHCHO 0.3 (0.2) - - a-d See footnotes in table 3. trans-But-2-ene (55 pmol) was used. Table 5. Composition of organic compounds in the reaction of cis-but-2-ene over U.V. irradiated V,O,/SiO," ~ - - ~ ~~~~~ yield/prnol" products with 0,' without 0, CH,=CH, CH,=CHCH, CH,=CHC,H, t-CH,CH=CHCH, c-CH,CH=CHCH, CH,=CHCH=CH, CH,CHO C,H,CHO CH,=CHCHO CH,COC,H, n-C,H,CHO CH,=CHCOCH,, CH,CH=CHCHO 0.4 (0.0) 1.0 (0.6) 0.5 (0.3) 5.3 (5.2) 20.0 (19.7) 25.4 (19.7) 0.3 (0.2) 0.3 (0.1 j 0.7 (0.1) 0.3 (0.1 j 5.3 (1.8) 0.1 (0.0) 0.2 (0.0) 0.3 (0.2) 0.1 (tr) 0.6 (0.5) 7.1 (7.0) 29.5 (28.7) 1.4 (1.3) 1.1 (1.0) trd a-d See footnotes in table 3.cis-But-2-ene (55 pmol) was used. buta-l,3-diene. The result that propanal from but-l-ene and ethanal from the but-2-enes are collected mainly in step A + B indicates that thermal energy is less important in the double-bond fission of butenes. Discussion Epoxide Route and Mechanism In the photo-oxidation of propene over V,O,/SiO,, propanal was formed as a minor product in addition to the main products, ethanal and acrylaldehyde. In a previousT.Tanaka, M. Ooe, T. Funabiki and S. Yoshida 41 paper,3 we neglected the reaction path for the formation of propanal. The present result that propanal was formed as a main product in the reaction of propene oxide suggests strongly that propene oxide is one intermediate in the photo-oxidation of propene. Acceleration of the reaction of propene oxide by U.V. irradiation can explain why the propene oxide is not detected during photo-oxidation of propene. Propene oxide as an intermediate in the photo-oxidation of propene was proposed by Pichat et aL2 They detected a small amount of propene oxide in the photo-oxidation of propene over TiO, and proposed that the oxygen species attacking propene to form the unstable primary product in question might be adsorbed atomic oxygen with neutral charge.8 Over U.V.irradiated TiO,, Djeghri and Teichnerg also reported the formation of alkene oxide in the photo-oxidation of pentanes. Previously, we have proposed that a 1 : 1 complex of propene and molecular oxygen is formed in the first stage of the rea~tion.~ The dependence of the reactivity of ethene photo-oxidation on partial pressure of oxygen as illustrated in fig. 2 is probably caused by the formation of a similar complex between ethene and oxygen, resulting in the very small degree of reaction in the absence of oxygen. The formation of such a complex was inspected by quantum chemical calculations.1° The calculations suggested that the ethene is adsorbed onto the lattice oxygen (V=O) to form species like ethene oxide. The photo-formation of species like alkene epoxide was also reported by Anpo and Kubokawa based on e.s.r.data." Subsequent adsorption of molecular oxygen on vanadium ions may cause the removal of the species from the site. Tracer studies using 1 8 0 , support the theory that lattice oxygen is transferred to the alkene. If isomerization of ethene oxide occurs, the formation of ethanal from ethene by photo-oxidation over V,O,/SiO, supports the presence of the epoxide route. The difference in reactivity between ethene and propene must be due to the different formation constants of the adsorbed species and the presence of the methyl hydrogen in propene which is withdrawn to form a n-ally1 intermediate. The photo-oxidation of propene over V,O,/SiO, is explained in scheme 1 .d ou b le-bond- fission CH3CHO H,C-CH-.-CII, CH*=CHCHO - CHsCH2CHO I epoxide formation H2C-CHCH3 '0' Scheme 1 The first route is the oxidative cleavage of the double bond to form ethanal, the second is the formation of a n-ally1 intermediate which is converted to acrylaldehyde, and the third is the formation of propene oxide which is isomerized to propanal. The second route is not possible in the case of ethene, but is predominant in the case of propene. We have concluded that thermal energy must not be required to complete the first route (double bond fission), but is required in the second route (n-ally1 intermediate oxidation).3 As for the third route, containing an epoxide intermediate, the following results have indicated that thermal energy also plays an important role in isomerization of the42 Photo-oxidation of Alkenes over Vanadium Oxide Table 6.Amount of ethanal collected after photo-oxidation of ethanal and adsorption of ethanal on V,O,/SiO, ________ ___ _ _ ~ _ _ _ ~ _ _ _ _ _ _ _ _ _ _ _ _ ~ _ _ ~~~ __ __ ___ ~ _ _ reaction total amount of amount of ethanal collected in each step (%) ethanal/pmol A B C oxidation of ethene adsorption of ethanal (58 pmol)a (3.2 pmol) 1.5 11 25 64 2.9 77 23 - Irradiation time: 30 min. O,, 60 pmol. activated epoxide to aldehyde or ketone. Thus, (i) propanal was formed by thermal treatment of adsorbed propene oxide; (ii) U.V. irradiation of the catalyst (for propene oxide) enhances the reactivity only; (iii) not only is isomerization photo-induced, the activation of the epoxide is also promoted by irradiation; (iv) as seen in table 6, most of the adsorbed ethanal could be recovered after irradiation, but ethanal formed in the photo-oxidation of ethene was mostly collected thermally in step C, indicating that thermal energy is essential to convert an activated epoxide intermediate to final products.Mechanism of Photo-oxidation of Butenes When the three routes are applied to the case of butenes, products other than CH,=CHCHO are explained as shown in scheme 2. (1) double-bond-fission route (0) CH2=CHC2H5 - C2H5CH0 (3) epoxide route CHO / Scheme 2 The oxidative cleavage of the C=C double bond will lead to C,H,CHO formation from but-1-ene and CH,CHO formation from but-2-enes as the major products. Formation of CH,CHO from but- 1 -ene indicates the isomerization of a part of but- 1 -ene to but-2-ene before the oxidative cleavage. The low yields of C,H,CHO from but-2-enes reflects the fact that the isomerization of but-2-enes to but-1-ene hardly occurs.T.Tanaka, M. Ooe, T. Funabiki and S. Yoshida 43 The folmation of CH,=CHCHO is rather difficult to explain by simple routes. The species might be formed by the oxidative cleavage of the n-ally1 intermediates or buta-1,3-diene, but we do not have any data for specification of the formation process. The saturated ketone and aldehyde, CH,COC,H, and n-C,H,CHO, are probably formed via epoxide intermediates. 1-Epoxybutane can form both products, but 2- epoxybutane forms only CH,COC,H,. Since the oxygen atom in all alkene epoxide prefers to be transferred to the terminal position, as found in the isomerization of propene oxide, the predominant formations of n-C,H,CHO from but-1 -ene and CH,COC,H, from but-2-enes are consistent with the 1- and 2-epoxybutane intermediates in each reaction.Unsaturated aldehyde and ketone (CH,CH=CHCHO and CH,=CHCOCH,) and buta- 1,3-diene may be formed via a n-methylallyl intermediate. Since but-1 -ene and trans-but-2-ene may be converted to a syn-n-methylallyl intermediate, the composition of the above three products is not so different. In the case of cis-but-2-ene, the yield of buta-l,3-diene becomes very high compared with other two products. This is probably because the anti-n-methylallyl intermediate is less stable than the syn isomer and the dehydrogenation proceeds before the oxygen addition. Products other than formed by the double bond fission, i.e. buta-1,3-diene9 CH3COC2Hs, n-C,H,CHO, CH,=CHCOCH, and CH,CH=CHCHO, are mostly obtained in step C. This result supports that the epoxide and n-ally1 routes which require thermal energy are involved in the photo-oxidation of butenes. This work was partially supported by a grant-in-aid for scientific research from the Japanese Ministry of Education. References 1 S. Yoshida, Y. Magatani, S. Noda and T. Funabiki, J. Chem. SOC., Chem. Commun., 1982, 601. 2 P. Pichat, J-M. Herman, J. Disdier and M-N. Mozzanega, J. Phys. Chem., 1979, 83, 3122. 3 S. Yoshida, T. Tanaka, M. Okada and T. Funabiki, J . Chem. SOC., Faraday Trans. 1, 1984, 80, 119. 4 S. Yoshida, T. Matsuzaki, T. Kashiwazaki, M. Mori and K. Tarama, Bull. Chem. SOC. Jpn, 1974, 47, 5 S. Yoshida, Y. Matsumura, S. Noda and T. Funabiki, J . Chem. SOC., Faraday Trans. 1, 1981,77, 2237. 6 S. L. Kaliaguine, B. N. Shelimov and V. B. Kazansky, J. Catal., 1978, 55, 384. 7 M. Iwamoto and J. H. Lunsford, J. Phys. Chem., 1980, 84, 3079. 8 J. M. Hermann, J. Disdier, M-N. Mozzanega and P. Pichat, J. Catal., 1979, 60, 369. 9 N. Djeghri and S. J. Teichner, J . Catal., 1980, 62, 99. 1564. 10 H. Kobayashi, M. Yamaguchi, T. Tanaka and S. Yoshida, J. Chem. SOC., Faraday Trans. I , in 11 M. Anpo and Y. Kubokawa, J. Catal., 1982, 75, 204. press. Paper 4/ 1968 ; Received 19th Nocember, 1984