首页   按字顺浏览 期刊浏览 卷期浏览 Tin oxide surfaces. Part 15.—Infrared study of the adsorption of propene on tin(I...
Tin oxide surfaces. Part 15.—Infrared study of the adsorption of propene on tin(IV) oxide, tin(IV) oxide–silica and tin(IV) oxide–palladium oxide

 

作者: Philip G. Harrison,  

 

期刊: Journal of the Chemical Society, Faraday Transactions 1: Physical Chemistry in Condensed Phases  (RSC Available online 1985)
卷期: Volume 81, issue 6  

页码: 1329-1343

 

ISSN:0300-9599

 

年代: 1985

 

DOI:10.1039/F19858101329

 

出版商: RSC

 

数据来源: RSC

 

摘要:

J. Chem. Soc., Faruduy Trans. I, 1985, 81, 1329-1343 Tin Oxide Surfaces Part 15.-Infrared Study of the Adsorption of Propene on Tin(1v) Oxide, Tin(1v) Oxide-Silica and Tin(rv) Oxide-Palladium Oxide BY PHILIP G . HARRISON* AND BARRY MAUNDERS Department of Chemistry, University of Nottingham, University Park, Nottingham NG7 2RD Received 4th June, 1984 The adsorption of propene on tin(1v) oxide, tin(rv) oxide-silica and tin(1v) oxide-palladium oxide has been studied by infrared spectroscopy. Tin(1v) oxide-silica were found to chemisorb propene as a surface acetate species when the oxides were outgassed at ambient temperatures, while only tin(1v) oxide-silica reacted with propene after outgassing at 716 K, again to form a surface acetate species. Tin(1v) oxide-palladium oxide outgassed at ambient temperatures was found to chemisorb propene both as a surface acetate and acrylate, while after outgassing at 570 K reaction with propene leads only to a surface acetate species.The most probable mechanism for the formation of the surface acetate involves initial electrophilic addition of acidic hydroxy groups to the C=C double bond, giving a surface isopropoxide which under- goes oxidation to the acetate via an intermediate, coordinated acetone. Consistent with this hypothesis, only tin(rv) oxide-silica is active at higher temperatures of pretreatment, where Bronsted acidity for this sample remains high but is very low for the other two oxides. The surface acrylate is formed by a similar process, although the initial process appears to involve a palladium-induced C-H bond fission of the methyl group of propene, generating a surface allyloxide species which undergoes oxidation as before.The reaction of propene over metal oxides has been the subject of several studies and a variety of behaviour has been observed. Only physisorption occurs on silica gel,l but where chemisorption does take place both dissociative and associative mechanisms can occur, leading to the formation of surface n-allylic and isopropoxide species, respectively. The majority of studies are concerned solely with the composition of the product vapour phase, with relatively few diagnosing the nature of the surface-adsorbed species and intermediates. Infrared data for adsorption on y-alumina have been interpreted as coordination of propene via the C=C double bond to a surface A13+ ion and a methyl-group hydrogen to a surface oxide.This initial species was then reported to convert into a surface n-allylic species.2 Reversible dissociative chemisorption to surface n-allylic species occurs on zinc oxide, which has been shown to undergo oxidation by an 0; ion to ac~olein,~-~ although in another study oxidation to acetate plus fonnate was reported.6 In a preceding paper of this series we have studied the reaction of ethane and ethene with tin(1v) oxide, tin(1v) oxide-silica and tin(1v) oxide-palladium oxide, and although surface acetate was the ultimate product in each case, different chemisorption mechanisms operated.’ Here we report the results of a similar study of the adsorption of propene on the same oxides. 13291330 TIN OXIDE SURFACES 1 I I I I I I I wavenumber/cm-' Fig.1. Infrared spectra of tin(1v) oxide: (1) evacuated 320 K, 18 h, < 1.33 x N m-2, (2) exposed to propene, 320 K, 18.75 h, 0.27 k n mP2, (3) evacuation, 320 K, 2 h, < 1.33 x N mW2, (4) exposed to propene, 473 K, 2.5 h, 0.37 kN mP2; subsequent evacuation (5) 320 K, 2.5 h, < 1.33 x N m-2 and (7) 563 K, N m-2, (6) 523 K, 2.75 h, < 1.33 x 15.5 h, < 1.33 x N m-2. EXPERIMENTAL The preparation of tin(1v) oxide, tin(rv) oxide-silica and tin(rv) oxide-palladium, the manufacture of infrared-transmitting discs therefrom and the general techniques employed have been described previously.' Infrared spectra were recorded using a Perkin-Elmer 577 spectrometer. RESULTS TIN(IV) OXIDE N m-2) (fig.1) was exposed to propene vapour (320 K, 0.27 kN m-2) for 18.75 h. Only weak bands due to vapour-phase propene were observed on top of the background spectrum of the disc, and these were removed by evacuation (320 K, < 1.33 x lo-* n m-2). The disc was subsequently heated at increasing temperatures in the presence of propene vapour but no new bands were seen in the infrared spectrum until the sample was heated at 473 K. An evacuated tin(1v) oxide disc (320 K, < 1.33 x42 55 % 68 86 h riJ * c .- E - g 10 Y o h Y - 61 57 P. G . HARRISON AND B. MAUNDERS 1331 ~~~ 1800 1600 1400 1200 wavenumber/cm -' Fig. 2. Infrared spectra of tin(1v) oxide-palladium oxide: ( 1 ) evacuated 320 K, 2.5 h, <-1.33 x N m-2; exposed to propene, 320 K, 15.5 h, 0.27 kN m-2 and subsequently evacuated (2) 320 K, 2.5 h, < 1.33 x lo-* N m-2, (3) 445 K, 15 h, < 1.33 x lop4 N m-2, (4) 485 K, 19.5 h, < 1.33 x lop4 N m-2, ( 5 ) 564 K, 3.75 h, < 1.33 x N m-2 and (6) 61 1 K, 3.5 h, < 1.33 x N m-a.Under these conditions a broad band was observed centred at 1530 cm-l. Evacuating the cell (320 K, < 1.33 x N mA2) left stable absorption bands at 1515, 1425 and 1345 cm-l, the latter being very weak. The 1515 and 1425 cm-l bands increased in intensity upon heating under vacuum at 523 K. A tin(rv) oxide disc that had been pretreated by evacuation and then oxygen treatment at 605 K did not exhibit new absorption bands with propene at any reaction temperature, nor did a tin(1v) oxide disc that had been evacuated and oxygen treated at 738 K and then exposed to pro- pene + water-vapour mixtures; no rehydroxylation was observed in the latter case.TIN(1V) OXIDE-PALLADIUM OXIDE A tin(rv) oxide-palladium oxide disc that had been evacuated (320 K, < 1.33 x lob4 N mP2) (fig. 2), exposed to propene (320 K, 0.27 kN m-2) for 15.5 h and subsequently re-evacuated (320 K, < 1.33 x N m-2) exhibited new adsorption bands at 1630, 1515, 1428, 1365, 1270 and 980cm-l. The 1515cm-l band was unsymmetrical with a shoulder on the high-wavenumber side. In addition, an appreciable decrease in the intensity of the strong, broad band between 3600 and 2000 cm-l was observed. Increasing the evacuation temperature to 445 K increased1332 TIN OXIDE SURFACES I I I I I I I 1800 1600 1400 1200 wavenumber/cm-' Fig.3. Infrared spectra of tin(1v) oxide-palladium oxide: (1) evacuated 477 K, 15 h, < 1.33 x N m-2, (2) exposed to propene, 320 K, 0.75 h, 0.19 kN m-2, then evacuated 484 K, 3 h, < 1.33 x N m-2; exposed to propene 487 K, 2.5 h, 0.27 kN m-2 and then subsequently evacuated (3) 320 K, 43 h, < 1.33 x N m-2, (4) 579 K, 15.25 h, c 1.33 x N m-2 and (5) 666 K, 15 h, < 1.33 x lo-* N m-2. the intensity of the 15 15 and 1428 cm-l bands with respect to the other bands and exposed the presence of two other weak bands at 1065 and 820 cm-l, while at higher evacuation temperatures (485,521 and 564 K) the 1630,1365,1270,1065 and 820 cm-l bands decreased in intensity and were effectively removed after evacuation at 61 1 K. At this latter temperature a very weak band at 1350 cm-l replaced the 1365 cm-l band with the 1428 cm-l band shifted to 1420 cm-l.Evacuation at 659 K brought about almost complete removal of the 151 5 cm-l band, removal of the 1420 cm-l band and appearance of two new bands at 1585 and 1395 cm-l. A tin(rv) oxide-palladium oxide disc that had been evacuated, treated with oxygen at 570 K and re-evacuated at 477 K (1 5 h, < 1.33 x N m-2) was exposed to propene (320 K, 0.19 kN mP2) for 2 h, evacuated and heated to 484 K (fig. 3 and 4). No new bands appeared in the spectrum, although the broad band centred at 31 50 cm-l was removed. The disc was subsequently heated in propene (487 K, 0.27 kN rn-2), cooled and evacuated. New absorption bands were observed at 1630-1 540 (broad and weak), 1505 and 1420 cm-l, the latter band having a shoulder on the low-wavenumber side.In addition, a broad band centred at 3200 cm-l appeared. Evacuation at increasing temperatures had little effect until 579 K was reached. At this temperature the 1420 cm-l band was shifted to 1390 cm-l. Evacuation at 666 KP. G . HARRISON AND B. MAUNDERS 1333 wavenumber/cm-' Fig. 4. Infrared spectra of tin(1v) oxide-palladium oxide: (1) evacuated 477 K, 15 h, < 1.33 x 10-4 N m-2, (2) exposed to propene, 320 K, 0.75 h, 0.19 kN m-2 and then evacuated 484 K, 3 h, < 1.33 x N mP2, (3) exposed to propene, 2.5 h, 0.27 kN m-2 and then evac- uated 320 K, 43 h, < 1.33 x lo-* N m-2. greatly reduced the intensity of the 1505 cm-l band. Two bands at 1585 and 1390 cm-l were now present and were very similar in nature to those observed on the low-temperatdre pretreated tin(1v) oxide-palladium oxide after evacuation at 659 K.A tin(1v) oxide-palladium oxide disc that had been pretreated by calcination and oxygen treatment at 704 K exhibited no new absorption bands with propene under the conditions employed. TIN(IV) OXIDE-SILICA A tin(1v) oxide-silica disc that had been evacuated at 320 K was exposed to propene vapour (320 K, 0.43 kN mP2) for 1.5 h, and subsequent evacuation left a broad band between 1670 and 1550 cm-l (fig. 5). The disc was re-exposed to propene at 390 K and after pumping off the vapour a slightly broader band remained at 1670-1470 cm-l; upon subsequent evacuation at 443 K absorption bands were observed at 1520, 1425 and 1355 cm-l. The infrared spectrum of a tin(1v) oxide-silica disc that had been evacuated and treated with oxygen at 716 K and then exposed to propene vapour (0.16 kN m-2) at 320 K exhibited weak absorption bands attributable to propene (fig.6 and 7). In addition, the sharp band at 3720 cm-l was removed and a broad band centred at 3350 cm-l appeared. Pumping off the vapour left a weak, broad band centred at1334 64. 65 8 68. 69 69 U .- 2 77 - 1 0 - TIN OXIDE SURFACES 1800 1600 1400 1200 wavenurnber/cm -' Fig. 5. Infrared spectra of tin(rv) oxide-silica: (1) evacuated 320 K, 17 h, < 1.33 x lop4 N mP2, (2) exposed to propene, 320 K, 2 h, 0.43 kNm-2, (3) evacuated 320 K, 0.5 h, < 1.33 x N m--2 , (4) exposed to propene, 390 K, 1.5 h, 0.33 kN m-2 and subsequent evacuation (5) 320 K, 0.5 h, < 0.5 h, < 1.33 x N m-2 and (6) 443 K, 18.5 h, < 1.33 x lop4 N m-2.1580 cm-l. The 3720 cm-l band was partially restored but shifted to 3700 cm-l and the broad band at 3350cm-l slightly reduced in intensity. Raising the evacuation temperature to 470 K produced broad absorption bands with maxima at 1530 and 1420 cm-l. Raising the evacuation temperature to 565 K significantly increased the intensity of the bands, the 1530 cm-l band maximum shifting to 1520 cm-1 with a shoulder on the high-wavenumber side at 1580 cm-l. On heating at 661 K, the 1520 cm-l band was greatly reduced in intensity, the 1420 cm-l band was removed, and two new bands were seen at 1585 and 1395 cm-l. The infrared spectrum of a tin(1v) oxide-silica disc that had been evacuated, treated with oxygen at 708 K and then treated with D,O vapour (633 K, 60 h) exhibited a sharp absor tion band at 2740 cm-l and a weak broad band centred at 1345 cm-l.Exposure to propene (0.13 kN mP2, 320 K, 4.5 h) reduced the intensity of the 2740 cm-1 band with slight broadening on the low-wavenumber side while not greatly altering the intensity of the 2530 cm-l band. In addition, a weak, broad band with its maximum at 3400 cm-l appeared. No new bands, or increase in intensity of 2530 cm-l (fig. 8). Wea ., bands were also present at 1585, 1510, 1425, 1390 andP. G . HARRISON AND B. MAUNDERS 1335 I I I I I I I 1800 1600 1400 1200 wavenum ber/cm-' Fig. 6. Infrared spectra of tin(1v) oxide-silica: (1) evacuated and treated with oxygen, 716 K, 3 h, < 1.33 x N m-2, (2) exposed to propene, 320 K, 0.5 h, 0.16 kN m-2, then evacuated 320 K, 0.25 h, < 1.33 x N mP2, N m-2 and subsequent evacuation (3) 470 K, 3 h, < 1.33 x (4) 565 K, 2.75 h, < 1.33 x N m-2 and (5) 661 K, 3.5 h, .c 1.33 x N m-2.the existing bands, in the 1600-1300 cm-l region were observed. Pumping off the propene and evacuating for 5 min at 320 K did not fully restore the 2740 cm-l band but did largely remove the 3400 cm-l band, although a shallow band remained. Evacuation at 470 K (9.5 h, < 1.33 x lop4 N mp2) caused a decrease in intensity of the 2530 and 2740 cm-l bands with the latter becoming broader on the low-wavenumber side. At the same time the broad band at 3400 cm-l increased in intensity, with a weak shoulder being present at 3695 cm-l.Treatment with propene at this temperature removed the 2740 and 2530 cm-l bands almost entirely, while the 3400 cm-l band increased further in intensity with a definite shoulder at 3695 cm-l. No new bands were seen in the 1600-1300 cm-l region of the spectrum. Evacuating the cell had relatively little effect on the spectrum, causing only a slight restoration of the 2740 cm-l band. Finally, heating at 470 K in oxygen and then evacuating the cell resulted in the appearance of strong absorption bands at 1520 and 1428 cm-l, along with a weak band at 1345 cm-l. At the same time the 3400 cm-l band became a little more intense but narrower, with a definite band at 3710 cm-l (relatively intense). The 2740 and 2530 cm-l bands were partially restored, but not to their original intensity. In a separate experiment using a Perkin-Elmer 598 spectrometer with data-handling facilities, a tin(1v) oxide-silica disc was evacuated and treated with oxygen at 753 K1336 TIN OXIDE SURFACES I ~ ~ l o l l l l l r l 4 000 3500 3000 wave nu m ber/cm -' Fig.7. Infrared spectra of tin@) oxide-silica: (1) evacuated and oxygen-treated, 716 K, 3 h, < 1.33 x N m-2, (2) exposed to propene, 320 K, 0.5 h, 0.16 kN m-2 and subsequent evacuation, (3) 320 K, 0.25 h, < 1.33 x N m-2. N rnp2 and (4) 470 K, 3 h, < 1.33 x and a background spectrum recorded. Propene was admitted to the cell at 320 K for 0.5 h and then pumped off before a second spectrum was recorded. After normalising and smoothing the spectra, the latter was subtracted from the former to reveal an absorption band centred at 1180 cm-l.Infrared bands together with assignments are summarised in table 1. DISCUSSION The infrared absorption bands observed on low-temperature pretreated tin(1v) oxide (at 1515, 1425 and 1345 cm-l) and tin@) oxide-silica (at 1520, 1425 and 1355 cm-l) can in both cases be assigned to the antisymmetric and symmetric v(C00) stretching modes and the symmetric 6,(CH,) deformation mode, respectively, of surface-adsorbed acetate species. The absorption bands observed on the 716 K pretreated tin(rv) oxide-silica and the 570 pretreated tin(rv) oxide-palladium oxide discs can also be ascribed to the acetate structure. Tin(rv) oxide-palladium oxide, pretreated at low temperature, exhibited a different spectrum, with the absorption bands observed being attributable to two surface species : an acetate, antisymmetric v ( C 0 0 ) stretch, 15 15 cm-l, symmetric v(C00) stretch, 141 8 crn-l, with the symmetric d(C-H,) deformation mode combined with the band at 1365 cm-l, and a surface acrylate, v(C=C) stretch, 1630 cm-l, 6(C-H) deformation, 1270 cm-l, CH, rock, 1065 cm-l, CH, twist, 980 cm-l, and CH bend, 820 cm-l.The unsymmetrical nature72 68 78 h 5 78 2 80 c! 4- 79 .- E 82 - 10% - 3, P. G. HARRISON AND B. MAUNDERS 1337 83 f / 4 2 , I I I I I I I 10 3600 3300 - 2800 2500 Fig. 8. Infrared spectra of tin(1v) oxide-silica: ( I ) evacuated and treated with oxygen, 798 K, 3 h, < 1.33 x N m-2, exposed to D20, 633 K, 60 h, then evacuated and cooled to 320 K, 3 h, < 1.33 x N m-2, (2) exposed to propene, 320 K, 4.5 h, 0.13 kN m-2; subsequent evacuation (3) 320 K, 0.25 h, < 1.33 x N m-2, (5) exposed to propene, 470 K, 1.5 h, 0.13 kN m-2, (6) evacuated 320 K, 1.5 h, < 1.33 x N m-2 and (7) oxygen-treated 470 K, 14 h, then evacuated 320 K, 2 h, < 1.33 x N mP2.N mP2, (4) 470 K, 9.5 h, c 1.33 x of the 1515 cm-l band is due to its being composed of the antisymmetric v(C00) stretch of the acrylate as well as the acetate, the symmetric v(C00) stretch being obscured by the 1418 cm-l band. The lack of absorption bands assignable to surface-adsorbed oxidation products on the high-temperature-pretreated tin(1v) oxide and tin(rv) oxide-palladium oxide suggests that surface hydroxy groups play an important role in the oxidation process.Reactions of propene with many metal oxides have been reported in the literature. Water vapour is usually present under the reaction conditions studied. In general, the various authors report one of two initial reaction intermediates, a n-allylic-type species reported for chromia,s zinc 4 7 cuprous ~ x i d e , ~ bismuth molybdatelO and mixed tin-antimony oxide,ll or an isopropoxy species reported for mixed tin oxide- molybdenum oxide and cobalt oxide-molybdenum oxide,l27 l3 chromium oxide,l39 l4 and for mixtures of nickel oxide,14 titania13 and ferric oxidel33 l4 with molybdenum oxide. With the n-allylic intermediate the reaction products tend to be acrolein, although Kubokawa et aL6 have reported surface-absorbed acetate plus formate groups on zinc oxide, while the isopropoxy intermediate tends to lead to acetone or acetic acid as the main product.In the present case, the interactions of propene with tin(1v) oxide and tin(1v)c-' w w 00 Table 1. Infrared absorption bands observed for the adsorption of propene on tin(1v) oxide, tin(1v) oxide-silica and tin(1v) oxide-palladium oxide pretreatment evacuation temperature temperature band position/cm-l oxide /K /K SnO, 320 SnO, . PdO 320 SnO, . PdO 570 SbO, . SiO, 3 20 SnO, . SiO, 716 surface acetate - surface acrylate - surface carbonate - 473 320 445 61 1 487 1630-1 540(br)e 579 320 1670-1 550(br)e 443 - 320 1580(br)" 470 565 - - - - - - - - - 1515 1425 1630 - 1515 1428 1630 - 1515 1428 - 1515 1420 - 1505 1420 - - - 1580(sh) 1505 1 390d - 1520 1425 - - - 1530(br) 1420 - 1580(sh) 1520 1420 assignments v(C=C) - vas(COO) v,(COO)S(CH) - "as(CO0) V,(COO) - - Vas(C0) - - a Also due to dS(CH,) of acetate.Also observed: band at 980 cm-l (CH, twist). Also observed: bands at 1240 cm-, (epoxide-ring vibration of Also due to carbonate decomposition product. v(C=O) glycidaldehyde), 1065 cm-l (CH, rock), 980 cm-l (CH, twist), and 820 cm-l (CH bend). of coordinated ketone.P. G. HARRISON AND B. MAUNDERS 1339 oxide-silica to give a surface isopropoxide intermediate leading to surface-bound acetone, which is known to oxidise rapidly on tin@) oxide to form a surface-adsorbed acetate,15 would explain the observed results. However, if this type of mechanism is operating, an initial reduction of the hydroxy stretching bands in the 4000-2000 cm-l region of the spectrum would be expected.With the tin(1v) oxide disc such a reduction in intensity could not be observed because of the very intense nature of the hydroxy stretching band. On the 716 K pretreated tin(1v) oxide-silica disc, however, the sharp band at 3720 cm-l, assigned to the isolated SiOH groups, was removed in the presence of propene vapour while a broad band appeared at 3350 cm-l. These phenomena can be interpreted in terms of hydroxy-group addition across the carbon-carbon double bond together with a certain amount of hydrogen-bonding interaction. Some reversibility of this hydroxy addition was observed on evacuating the cell with the partial restoration of the 3720 cm-l band, though its shifted position (to 3700 cm-l) and the remaining broad band at 3350 cm-i are indicative of a significant amount of hydrogen bonding still being present.The fact that this occurred under conditions where the acetate structure was not observed, together with the appearance of the acetate structure on raising the evacuation temperature, is corroboration of there being an adsorbed intermediate. In an attempt to clarify the situation, the hydrogens of the hydroxy groups of a high-temperature-pretreated tin(1v) oxide-silica disc were exchanged by treatment with D,O. The absorption band of the isolated SiOD groups, at 2740 cm-l, was again observed to reduce in intensity in the presence of propene, and the new weak, broad band at 3400 cm-l could be due to hydroxy groups formed by hydrogen4euterium exchange between the surface and the propene (fig.8). H-D exchange has been shown by Buiten16 to occur readily for five of the propene hydrogens, but not for the hydrogen on carbon-2, consistent with the formation of an isopropoxide hydroxy addition to propene. Evacuation almost entirely removed the 3400 cm-l band, but only partially restored the 2740 cm-l band, suggesting the presence of an adsorbed species which arises from the 0-D interaction with propene. The decreased intensity of the 0-D bands and increased intensity of the 3400 cm-l band on evacuating at 470 K is further evidence for the H-D exchange mechanism. The eventual appearance of the acetate bands after further treatment with propene at 470 K followed by treatment with oxygen at this temperature again showed that the propene must be fairly strongly held to the surface as an intermediate.Infrared characterisation of the surface isopropoxide species formed by surface hydroxy addition to propene is difficult to obtain since the most intense absorption of this species, the v(C-0) stretching mode, is expected to fall at < ca. 1200 cm-l in a region where bulk oxide absorptions also occur. Nevertheless, a difference spectrum obtained using a data station of propene adsorbed at 320 K onto 753 K treated tin oxide-silica exhibited a band at 1 180 cm-l, in the region (1 185-1 130 cm-l) characteristic of molecular metal isopropoxides. No corresponding C-H deforma- tion modes were apparent, most probably because they were too weak to be observed (cf. the weak nature of the C-H deformation modes of the surface acetate species).As a surface alkoxide species, the isopropoxide is expected15 to undergo facile oxidation via coordinated acetate to the observed acetate species. The overall mechanism for the adsorption is shown in scheme 1, where M is Si or Sn. The oxidative clearage of ketones to give surface acetate and hydrocarbon has been previously well documented by us.17 In the present study, bands in the region 1670-1 550 cm-l may well correspond to the surface-coordinated acetone species [cf. the v(C=O) stretching mode of surface coordinated ketones which fall in the range 1680-1 585 ~ m - l ] . ~ ~1340 TIN OXIDE SURFACES D/HCH*-CH -CH3 H I '0 0 0- I l l - - o,Sn, ,M, /Sn 0 0 "0 Scheme 1. The hydroxy addition to the C=C double bond may be considered a classical electrophilic addition by the proton.Hence tin(rv) oxide-silica, which has a greater Bronsted acidity than tin(1v) oxide a10ne,18 exhibits a higher reactivity towards propene. The weak Bronsted-acidic sites on tin(1v) oxide are readily removed on heating, consistent with the lack of reactivity of the high-temperature-pretreated tin(rv) oxide. Scheme 2. Tin(1v) oxide-silica therefore resembles tin@) oxide-molybdenum(1v) oxide in its behaviour towards propene. In the latter case, acetone and acetic acid are formed, presumably via a similar reaction 2o The small amounts of acetaldehyde which are also formed may readily be incorporated into the same reaction scheme by elimination of methane, rather than hydrogen, from the surface isopropoxide (scheme 2).The formation of absorption bands due to both acetate and acrylate on the low-P. G . HARRISON AND B. MAUNDERS 1341 temperature-pretreated tin(1v) oxide-palladium oxide sample implies that at least two competing mechanisms are occurring. The identity of the second species as an acrylate was confirmed by comparison with samples exposed acrylic acid vapour.21 The acetate formation can be explained in an analogous manner to that put forward for the reaction of propene with tin(rv) oxide and tin(1v) oxide-silica. Since ethane is observed to react with tin(1v) oxide-palladium oxide under very mild conditions to afford surface acetate,' the most likely mechanism for the formation of surface acrylate is the formation of surface 1-0-CH,-CH-CH, species, followed by conversion into surface-coordinated acrolein and surface acetate in a manner similar to the formation of surface acetate (scheme 3).Indeed, separate studies have confirmed the formation of surface acrylate by adsorption of acrolein, identical to that obtained by adsorption of acrylic acid.21 The CH,=CH, /H c Scheme 3. iterature is unanimous that the formation of acrolein from propene over meta oxide catalysts involves a surface n-allylic species.4' 5 9 11* l9 In the present case, the palladium component of the oxide certainly plays an essential role in the initial chemisorption process, which leads to the formation of surface allyloxide (or other coordinated acrolein precursor). The driving force for the initial interaction may be the formation of surface Pd-H bonds (scheme 4), since palladium-hydrogen bonds Scheme 4.are known to be very stable in molecular complexes. Alternatively, a surface palladium--n-allylic species may be formed, which may subsequently be transferred 1.0 a neighbouring tin oxide surface site (scheme 5). The proposed mechanism for the oxidation of propene over zinc oxide is actually rather complicated and involves the interaction of a surface ally1 species with a surface1342 TIN OXIDE SURFACES H, H2C=CH-C<, O I I 0 I I I *,S*', 2% ,Sn, 0 0 Scheme 5. 0; species to give an intermediate allylperoxy species which is subsequently transformed to surface glycidol and gly~idaldehyde.~. The latter species was characterised by an infrared absorption at 1237 cm-l, ascribed to the epoxide ring vibration.In this study, a small band is observed at ca. 1240 cm-l, which disappears upon evacuation at higher temperatures before the disappearance of the acrylate absorption bands. However, the participation of a glycidaldehyde in the present study remains equivocal. The decomposition of a n-allylic intermediate to surface acetate plus formate as have been reported in one study of propene over zinc oxide6 can be definitely excluded by comparison of authentic infrared spectra obtained from formic acid adsorption.21 Tin@) oxide-palladium oxide, pretreated at 570 K, was found to react only weakly with propene to give a species attributable to a surface acetate only. With pretreatment at even higher temperatures, no surface species were observed, which is understandable in view of the loss of the weak Bronsted-acidic sites at these temperatures.With all three oxides when the evacuation temperature was raised high enough, decomposition of the adsorbed species occurred in the same manner as with the ethene and ethane reactions with these oxides. In summary, the reaction of propene over all three oxides evacuated at ambient temperatures gives surface acetate species. The most likely mechanism for the formation involves initial electrophilic addition of acidic hydroxy groups to their carbon-carbon double bond. At higher pretreatment temperatures tin(1v) oxide- silica is active in the formation of a surface acetate, reflecting the Bronsted-acidic nature of its surface, whilst underlying the very weak nature of the acid hydroxy groups of the other two oxides. Tin(rv) oxide-palladium oxide evacuated at ambient temperature, in addition to exhibiting absorption bands attributable to a surface acetate, also displays bands assignable to a surface acrylate. We thank the S.E.R.C. and the International Tin Research Institute for support in the form of a CASE award to (B.M.).P. G. HARRISON AND B. MAUNDERS 1343 L. Kubelkova and F. Trifiro, J. Catal., 1972, 26, 242. T. A. Gordymora and A. A. Davydov, Kinet. Catal. (Engl. Transl.), 1979, 20, 604. A. L. Dent and R. J. Kokes. J. Am. Chem. SOC., 1970,92, 1092; 6709; 6718. B. L. Kugler and R. J. Kokes, J. Catal., 1974, 32, 170. B. L. Kugler and J. W. Gryder, J. Catal., 1976, 44, 126. Y. Kubokawa, H. Miyata, T. Uno and S. Kawasaki, J . Chem. SOC., Chem. Commun., 1974,655. R. L. Burwell, G. L. Haller, K. C. Taylor and J. F. Read, Ado. Catal., 1969, 20, 1 . H. H. Voge, C. D. Wagner and D. P. Stevenson, J. Catal., 1963, 2, 58. 398. ' P. G. Harrison and B. M. Maunders, J. Chem. SOC., Faraday Trans. I , 1985,81, 1309. lo J. M. Peacock, A. J. Parker, P. G. Ashmore and J. A. Hockey, J. Catal., 1969,15, 15; 373; 379; 387; l 1 J. R. Christie, D. Taylor and C. C. McCain, J. Chem. SOC., Faraday Trans. 1, 1976,72, 334. l2 S. Tan, Y. Moro-Oka and A. Ozaki, J. Catal., 1970, 17, 132. l3 Y. Moro-Oka, Y. Takita and A. Ozaki, J. Catal., 1971, 23, 183. l4 Y. Moro-Oka, Y. Takita, S. Tan and A, Ozaki, Bull. Chem. SOC. Jpn, 1968, 41, 2820. l5 E. W. Thorton and P. G. Harrison, J. Chem. SOC., Faraday Trans. I , 1975, 71, 2468. l6 J. Buiten, J . Catal., 1969, 13, 373. P. G. Harrison and B. Maunders, J. Chem. SOC., Faraday Trans. I , 1984, 80, 1329. l 8 P. G. Harrison and B. Maunders, J. Chem. SOC., Faraday Trans. I , 1984, 80, 1341. Y. Moro-Oka, Y. Takita and A. Ozaki, J. Catal., 1972, 27, 177. *O J. Buiten, J. Catal., 1968, 10, 188. 21 P. G. Harrison and B. Maunders, unpublished data. (PAPER 4/908)

 

点击下载:  PDF (859KB)



返 回