J. Chem. SOC., Faraday Trans. I, 1985, 81, 1311-1327 Tin Oxide Surfaces Part 14.-Infrared Study of the Adsorption of Ethane and Ethene on Tin(rv) Oxide, Tin(rv) 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 ethane and ethene onto tin(1v) oxide, tin(1v) oxide-silica, and tin@) oxide-palladium oxide at various pretreatment temperatures in the range 320-740 K has been studied by transmission infrared spectroscopy. In every case the ultimate surface product was a surface acetate, except in the case of ethane and tin(1v) oxide, where no evidence for adsorption could be obtained under any conditions. The proposed mechanism for the formation of acetate involves the oxidation of an initially formed surface ethoxide species, although this intermediate is thought to arise via dissociative chemisorption, with C-H bond fission in the case of ethane and an electrophilic addition of surface hydroxyl groups across the C=C double bond in the case of ethene.Decomposition of the surface acetate to a surface carbonate occurs at temperatures > 580 K. The partial oxidation of hydrocarbons to oxygen-containing organic products is a process of great industrial importance. Ethane in particular is an abundant raw material, yet the literature concerning ethane reactions over heterogeneous oxidation catalysts is rather sparse. Only physisorption appears to occur on nickel oxidel and zinc oxide2 at 293 K, and very little reaction occurs over nickel oxide at higher temperatures (573-6’73 K).3 However, ethane is oxidised at elevated temperatures in the absence of air over the oxides of vanadium (623-673 K), molybdenum (773-823 K) and tungsten (823-923 K) to give as oxidation products formaldehyde, acetic acid, carbon monoxide, carbon dioxide and, under certain conditions, ethene.* Conversion of ethane into ethene also occurs over mixed-oxide systems of molybdenum and vanadium combined with an oxide of Ti, Cr, Mn, Fe, Co, Ni, Nb, Ta or Ce at 473 K in the presence of ~ x y g e n .~ Ethene can undergo several different types of reaction over metal oxides. Polym- erisation occurs on oxides of nickel, copper and palladium supported on porous silica glass to give surface polymerised species whose nature depends upon the particular transition metal.In addition, the vapour phase over the nickel sample has been observed to contain trans-b~t-2-ene.~ Saturated surface species of the types -CH2-CH, and -CH2-CH2- have been demonstrated on alumina at low temperatures, whilst surface acetate is formed at temperatures > ca. 523-573 K.’-lo Ethene is both physisorbed and chemisorbed on zinc oxide pretreated at 773 K, the chemisorbed product being olefinic in character and bound to the surface by interaction of the TC e l e c t r o n ~ . ~ l - ~ ~ At temperatures > 573 K, ethene is oxidised by oxygen over zinc oxide to carbon dioxide and water, although again a surface n-olefin complex was p0stu1ated.l~ Oxidation to carbon dioxide, water and acetic acid occurs 131 11312 TIN OXIDE SURFACES over 10% Pd on titania.15 On titania itself, ethene is adsorbed as an alkoxide with cleavage of the C=C bond.16 No chemisorption of ethene occurs on ~i1ica.l~ There is, apparently, no description in the literature of reactions of ethene over tin(1v) oxide or mixed oxides containing tin(1v) oxide, and very few for ethene.Complete oxidation occurs over the neat oxide at temperatures > 623 K and the reactivity of the catalyst is enhanced by the incorporation of O.l-l% chromium in the surface.17 Oxidation to carbon dioxide occurs over tin@) oxide-molybdenum(v1) oxide, although acetaldehyde could also be detected at reaction temperatures -= 413 K.l* In this paper we report a detailed infrared study of the adsorption behaviour of ethane and ethene on tin(1v) oxide, tin@) oxide-silica and tin@) oxide-palladium oxide.EXPERIMENTAL The preparation of tin(rv) oxide, tin@) oxide-silica and tin(1v) oxide-palladium, the manufacture of infrared-transmitting discs therefrom and the general techniques employed have been described previo~sly.~~ Infrared spectra were recorded using a Perkin-Elmer 577 spectrometer. Spectral data are summarised in tables 1 and 2. RESULTS ETHANE ADSORPTION TIN@) OXIDE-SILICA A tin(1v) oxide-silica disc that had been evacuated at 320 K, exposed to ethane (0.33 kN m-2), re-evacuated and then subsequently heated to 488 K gave rise to new absorption bands at 1725, 1525 and 1431 cm-l, with weaker bands at 1380 and 1355 cm-l and a shoulder on the high-wavenumber side of the 1525 cm-l band (fig.1). Similar bands were observed when ethane (0.33 kN m-2) was adsorbed at 488 K on a tin(rv) oxide-silica disc pretreated by evacuation and oxygen treatment at 740 K (fig. 2). Additional weak bands were observed at 1630, 1330 and 1305 cm-l. The hydroxy stretching absorption bands of tin(rv) oxide-silica pretreated at 740 K exhibit a marked increase in intensity with the absorption of ethane, even before the appearance of the acetate structure (fig. 3). TIN(VI) OXIDE-PALLADIUM OXIDE A tin(1v) oxide-palladium oxide disc that had been evacuated at 320 K, exposed to ethane (0.21 kN mP2) and subsequently re-evacuated at 320 K gave rise to new absorption bands at 1720,1550-1 5 10 and 1422 cm-l, with weaker bands at 1370,1340 and 1290 cm-l (fig.4). Raising the evacuation temperature to 438 K left bands only at 1715, 1515 and 1422 cm-l, with a weak shoulder at 1380 cm-l. No new absorption bands were observed in the infrared spectrum of a tin(1v) oxide-palladium oxide disc pretreated by evacuation and oxygen treatment at 665 K then exposed to ethane vapour (0.27 kN m-2). Heating a similarly treated disc at 623 K resulted in the appearance of a band at 1590 cm-l, which like the tin(1v) oxide-silica system is most probably due to a surface carbonate decomposition product. In addition, the broad 1425-1 385 cm-l band shifts to 1390 cm-l and becomes much sharper. No surface adsorbed species were observed with pure tin(1v) oxide under any pretreatment conditions.Table 1.Infrared data for the adsorption on tin(1v) oxide-silica and tin(rv) oxide-palladium oxide 'd 9 pretreatment evacuation temperature temperature band position/cm-l /K /K 3: oxide Sn0,-SiO, 320/740 K 488 1725 1630 - 1525 1431 - 1380 1355 1330 1305 580 1725 - 1590 1525 1430 - - 1355 - 1305 8 Z 750 1720 - - 1510-1550 1422 - 1370 1340 - 1290 ' ? 320 K Z I M - - - 638 1705-1715 - 1590 1525 - 1390 - % - - - - - - - - 1590 - - - 438 1715 - - 1515 1422 - 1380 - assignments - 5 8 - W H J - - - - vaas(COO) V,(COO) - surface acetate - surface enolate - - v(C=C) - - - d(COH) - v(C-0) - - - - - - - - I1 I1 - I 320 K Sn0,-PdO I, physisorbed cyclic dimer of acetic acid; 11, carbonate formed by decomposition of surface acetate.e w P c Table 2. Infrared data for ethene adsorption on tin@) oxide, tin@) oxide-palladium oxide and tin(1v) oxide-silica pretreatment evacuation temperature temperature band position/cm-l oxide /K /K - 1535 1430 SnO, 320 320 - 1535 1425 445 1522 1420 475 - - - 1520 1422 Sn0,-PdO 320 320 513 - - 1520 1422 - 1520 1422 Sn0,-PdO 320/H, 320 Sn0,-SiO, 320 320 - - 1535 1435 1530 1430 566 632 1530 1428 - - 1530 1430 320 560 1720+1710 - - 1530 1430 660 - 1530 1430 320b - 1625 - 1550 1440 - 1530 1430 320 - - - - - - - - - - - - - - 1585 1530 - - - - - - Sn0,-SiO, 660 50P - - 1585-1590 1530 - - - - Sn0,-SiO, 660 506' - - assignments V,s(COO) Vs(CO0) - - - __ - surface acetate - - I11 I1 I 1340 - 1345 - 1350 - 1345 - - 1345 - 1345 - 1345 - 1350 - - - - - - - - - - 1395 1352 - 1352 - 1352 - 1395 - - 1352 - - 1352 - - - - - 1388 1305 - I, carbonate formed by decomposition of surface acetate; 11, hydrogen-bonded acetic acid; 111, physisorbed cyclic dimer of acetic acid.a In the presence of ethene vapour; in the presence of water vapour.P. G. HARRISON AND B. MAUNDERS 1315 1800 1600 ]LOO 1200 wavenumber/cm-' Fig. 1. Infrared spectra of tin(1v) oxide-silica: (1) evacuated 320 K, 18 h, 1.33 x lop4 N mp2, (2) exposed to ethane, 320 K, 2.5 h, 0.33 kN mp2; subsequent evacuation (3) 320 K, 20 h, 1.33 x N m+, (4) 440 K, 3 h, 1.33 x N mP2, (6) 580 K, 6 h, 1.33 x lo-* N m-2 and (7) 638 K, 22 h, 1.33 x lo-* N m-2. N m-2, (5) 488 K, 15 h, 1.33 x ETHENE ADSORPTION TIN(IV) OXIDE A tin(1v) oxide disc that had been evacuated, exposed to ethene vapour (0.13 kN m-2) and then re-evacuated, all at 320 K, for 23 h ( < 1.33 x lo-* N m-2) exhibited broad weak bands in the infrared spectrum centred at ca.1535 and 1430 cm-l and a weak shoulder at 1340 cm-l (fig. 5). Raising the evacuation temperature to 445 K increased the intensity of all three bands, with their maxima occurring at 1535, 1425 and 1345 cm-l. A further increase in the temperature to 475 K led to further increases in intensity with the maxima shifting to 1522, 1420 and 1350 cm-l. Evacuation at 538 K resulted in the removal of these bands. A tin@) oxide disc that had been evacuated and treated with oxygen at 563 K and subsequently exposed to ethene vapour did not exhibit any new absorption bands in the 1800-1 100 cm-l region. However, in the 4000-2000 cm-l region the initial hydroxy stretching envelope was seen to decrease in intensity during exposure to the ethene, but it was restored to its initial intensity upon subsequent evacuation (fig.6).1316 TIN OXIDE SURFACES r 1900 1700 1500 1300 1100 wavenurn ber/cm -' Fig. 2. Infrared spectra of tin(rv) oxide-silica: (1) evacuated 740 K, 1 1 h, 6.65 x lop5 N m-2, oxygen treated 740 K, re-evacuated; in the presence of ethane (0.33 kN mP2) (2) 320 K, 18 h, (3) 390 K, 5 h, (4) 488 K, 48 h; subsequent evacuation (5) 320 K, 16 h, 6.65 x N m-2, (6) 390 K, 3 h, 6.65 x N m-2, (7) 490 K, 6 h, 6.65 x N m-2, (8) 570 K, 2.75 h, 6.65 x N m-2, (9) 670 K, 4 h, 6.65 x N m-2 and (10) 750 K, 12 h, 6.65 x N m-2. TIN(IV) OXIDE-PALLADIUM OXIDE A tin(1v) oxide-palladium oxide disc that had been evacuated and then exposed to ethene vapour (0.23 kN m-2) for 1 h at 320 K led to the formation of new absorption bands at 1520, 1422 and 1345 cm-l, which grew in intensity after 3 h in the ethene vapour (fig.7). Subsequent evacuation at 320 K led to a further increase in the intensities, while raising the evacuation temperature to 513 K produced even more intense bands. Evacuation at 573 K produced a decrease in intensity, while the bands were virtually removed after evacuation at 645 K. A tin@) oxide-palladium oxide disc that had been evacuated and treated with oxygen at 673 K did not exhibit new absorption bands in the 1800-1 100 cm-l region upon exposure to ethene vapour (0.31 kN m-2) at 320 K. It did, however, exhibit a decrease in intensity of the band100 80 h s 2 60 e f 3 Y i .- E $ 4 0 L.Y 20 P. G . HARRISON AND B. MAUNDERS 1317 t t 01 I I I 4000 3500 3000 2 500 2000 waveiiumberlcm-' Fig. 3. Infrared spectra of tin(1v) oxide-silica: (1) evacuated 740 K, 11 h, 6.65 x N m-2, oxygen treated 740 K, re-evacuated; in the presence of ethane (0.33 kN m-2) (2) 320 K, 18 h, (3) 488 K, 48 h and (4) subsequent evacuation, 320 K, 16 h, 6.65 x low5 N m-2. in the hydroxy stretching region. When the disc was subsequently treated with hydrogen, evacuated and re-exposed to ethene vapour (0.23 kN mP2) at 320 K and then heated to 410 K, weak bands were observed in similar positions to the bands formed on the 320 K pretreated disc. A tin(1v) oxide-palladium oxide disc that had been evacuated, treated with hydrogen and then exposed to ethene vapour (0.37 kN m-2), all at 320 K, gave essentially the same spectra as observed in fig. 7 but with fewer bands of lower intensity.TIN(IV) OXIDE-SILICA A tin(rv) oxide-silica disc that had been evacuated, then exposed to ethene vapow (0.26 kN mP2) for 2.75 h and subsequently re-evacuated for 20 h, all at 320 K, gave rise to weak absorption bands at ca. 1535, 1435 and 1345 cm-l (fig. 8). The exact position of the 1535 cm-I band is obscured by the presence of the water bending mode of molecularly coordinated water, itself shifted slightly (ca. 10 cm-l) to 1600 cm-l. Increasing the evacuation temperature increased the intensity of these bands, with maximum intensity being observed after evacuation at 566 K. During this treatment, the 1535 and 1345 cm-l bands shifted slightly to 1530 and 1350 cm-l, respectively.The 1435 cm-l band, however, remained broad throughout and centred at ca. 1430 cm-l. The spectrum obtained after evacuation at 632 K indicated decomposition of the surface species, with the 1530 cm-l band greatly reduced and the 1430 cm-l band shifted to 1395 cm-l. A band at 1585 cm-l was also present. It is difficult to say at what temperature this band first appeared, as its position, coupled with its broad nature, is very close to that of the water binding mode. As such, it is hard to tell when the first was removed and the second formed. A tin(1v) oxide-silica disc that had been evacuated, treated with oxygen at 660K and then exposed to ethene vapour (0.32 kN mP2) at 320 K for 2.75 h exhibited no new absorption in the 1800-1 100 cm-ITIN OXIDE SURFACES 1318 1800 1600 1400 1200 wavenumber/crn-' Fig.4. 4 7 4 7 - 4 9 % 54 u m Y 'S 57 v) 3 6 0 - 10 O/O - I I I I I I 1000 1600 14 00 1200 wavenumber/cm-' Fig. 5. Fig. 4. Infrared spectra of tin(1v) oxide-palladium oxide: (1) evacuated 320 K, 19 h, 1.33 x N m-2, (2) exposed to ethane 320 K, 4 h, 0.21 kN m-2; subsequent evacuation (3) Fig. 5. Infrared spectra of tin(1v) oxide: (1) evacuated 320 K, 22 h, < 1.33 x N m-2, (2) exposed to ethene, 320 K, 25 h, 0.13 kN m-2; subsequent evacuation (3) 320 K, 3 h, < 1.33 x N m-2, (4) 320 K, 23 h < 1.33 x N m-2, (5) 445 K, 19.5 h, < 1.33 x 320 K, 17 h, 1.33 x N m-2 and (4) 438 K, 65 h, 1.33 x N mP2. N m-2 and (6) 475 K, 18.5 h, < 1.33 x N mF2. region of the infrared spectrum.However, after pumping off the ethene vapour, re-evacuating at 653 K with oxygen treatment and then exposing to ethene vapour (0.23 kN mb2) at 505 K, new bands were observed at 1530, 1428 and 1352 cm-l (fig. 9). Pumping off the ethene vapour had little effect on the spectra other than a small shift in the 1428 cm-l band to 1430 cm-l. Increasing the evacuation temperature to 560 K considerably increased the intensity of the three absorption bands, and a weak broad band, or doublet of bands, at 1720 and 1710 cm-l was also observed. The adsorbed species was observed to begin to decompose at an evacuation temperature of 615 K. In a similar manner to the disc treated at 320 K, the 1430 cm-l band broadened, then at 660 K a sharp band at 1395 cm-l was observed along with the 1585-1590 cm-l band.In the 4000-2000 cm-l region of the spectrum the effect ofP. G . HARRISON AND B. MAUNDERS 1319 100 8 0 h 5 8 60 5 € 2 2 40 Y c .- Y zoa 0 4000 3500 3000 2500 2000 wavenum ber1cm-I Fig. 6. Infrared spectra of tin(1v) oxide: (1) evacuated and oxygen treated at 563 K, re-evacuated; subsequent exposure to ethene, 320 K, 0.29 kN m-2 (2) 1.25 h, (3) 3 h and (4) re-evacuation, 320 K, < 1.33 x N m-2. ethene vapour on the hydroxy stretching bands can be observed. The most obvious effect of ethene adsorption is the specific reduction in intensity of the 3720 cm-l band, with a slight shift to 3700 cm-l, and the increase in the intensity of the broad band on the low-wavenumber side of the 3720 cm-l band. Pumping off the ethene vapour did not restore the original band.An evacuated tin@) oxide-silica disc (660 K with oxygen treatment) that had been treated with ethene to give the bands at 1530, 1430 and 1352 cm-l and then subsequently exposed to water vapour (0.13 kN m-2) exhibited absorption bands at 1625, 1550 (shoulder), 1440, 1388 (very weak) and 1305 ~ m - ~ (fig. 10). Subsequent evacuation restored the original bands. The shift in the acetate bands to 1550, 1440 and 1388 cm-l in the presence of water vapour in all three cases shows how the acetate structure is weakened by the water; indeed, some of it is protonated to acetic acid, which is bound to the surface by hydrogen bonding, some of the 1625cm-l band and the 1305cm-l band being attributable to the C=O stretching and 0-H deformation modes, respectively, of the hydrogen-bonded acetic acid.The remainder of the 1625 cm-l band can be assigned to the water bending mode. Pumping off the water vapour restores the positions of the acetate bands and removes most of the hydrogen-bonded acetic acid. DISCUSSION In all cases the observed absorption bands at ca. 1515-1535, 1420-1435 and 1345-1 355 cm-l can be assigned to the antisymmetric and symmetric v(C00) modes and the symmetric S(C-H) deformation mode, respectively, of a surface acetate species. We have previously reported the spectrum of surface acetate on tin(1v) oxide, whilst the spectra of surface acetate on tin(1v) oxide-silica on adsorption of acetic acid1320 TIN OXIDE SURFACES 1800 1600 1400 1200 wave nu rn ber/ c m - * Fig. 7.- I 1 I I I I I I 1800 1600 1400 1200 wavenumberlcm-' Fig. 8. Fig. 7. Infrared spectra of tin(rv) oxide-palladium oxide: (1) evacuated 320 K, 40 h, 1.33 x N m-2; subsequent exposure to ethene, 320 K, 0.23 kN mP2 (2) 1 h, (3) 3 h, re-evacuated, < 1.33 x 10-4 N m-2, (4) 320 K, 1 h, ( 5 ) 320 K, 21 h, (6) 513 K, 21 h and (7) 645 K, 17 h. Fig. 8. Infrared spectra of tin(1v) oxide-silica: (1) evacuated 320 K, 18 h, < 1.33 x N m-2, (2) exposed to ethene, 320 K, 2.75 h, 0.27 kN mP2; subsequent evacuation (3) 320 K, 20 h, < 1.33 x N m-2, (4) 445 K, 19 h, < 1.33 x loP4 N m-2, ( 5 ) 490 K, 3 h, < 1 . 3 3 ~ 10-4Nm-2.(6)535K,7h, < 1 . 3 3 ~ 1 0 - ~ N m - ~ , ( 7 ) 5 6 6 K , 2 0 h , < 1 . 3 3 ~ 1 O - ~ N m - ~ and (8) 632 K, 19 h, < 1.33 x N m-2. is shown in fig.1 1. No information could be gained from the 4000-2000 cm-l region of the spectra. However, spectra from the high-temperature-pretreated samples afforded indications as to the mechanism of the initial chemisorption reaction for both ethane and ethene. The increase in intensity of the hydroxy stretching band of tin(rv) oxide-silica that had been pretreated at 740 K indicates that new hydroxy groups are formed, either directly or from adsorption of water possibly produced in the reaction, as a result of carbon-hydrogen bond fission. The other species formed in this process would be expected to be an ethoxide group via dissociative chemisorption at two adjacent surface oxide sites (scheme 1, where M is tin or silicon). The mechanism of scheme IP. G. HARRISON AND B.MAUNDERS 1321 I 60 - 9 1 - 55- 6 5 - 6 9 - h aJ 0 m 4 4 Y .- : 2 Y 8 1- 8 5- - 10 "/o - I I I I I I 1800 1600 1400 1200 Fig. 9. wave number/ c m -' I I I I I I 1800 1600 1400 1200 wavenumber/cm -' Fig. 10. Fig. 9. Infrared spectra of tin(1v) oxide-silica: (1) evacuated and treated with oxygen at 660 K, re-evacuated < 1.33 x 10-4 N m-2, (2) exposed to ethene, 320 K, 2.75 h, 0.22 kN mP2, (3) evacuated and treated with oxygen at 653 K, exposed to ethene, 505 K, 1 h, 0.23 kN mP2; subsequent evacuation (4) 320 K, 16.5 h, < 1.33 x N m-2, (5) 560 K, 2 h, .< 1.33 x N m-2, (6) 615 K, 20 h, < 1.33 x N mP2 and (7) 660 K, 14 h, < 1.33 x lop4 N m-2. Fig. 10. Infrared spectra of tin@) oxide-silica: (1) evacuated and treated with oxygen at 653 K, exposed to ethene, 505 K, 1 h, 0.23 kN m-2, subsequent evacuation, 320 K, 16.5 h < 1.33 x N m-2, (2) exposed to water vapour, 320 K, 0.5 h, 0.13 kN mP2 and (3) re- evacuated, 320 K, 1.33 x N m-2.CH3 I I I I CHI i 7 i 0 0 0 6 6 0 0 0 0 CH3-CHz----H CH3-CH2-H I I I - I I I - I l l O/M\O/M\O/M\O O/M\O/M\O/M\O OOM\O/M\O/M\O 44 Scheme 1. FAR 11322 53 h 75 9 86 8 4 TIN OXIDE SURFACES 1900 1700 1500 1300 1100 wavenumber/cm-' Fig. 11. Infrared spectra of tin(rv) oxide-silica: (1) evacuated 750 K, 12 h, 6.65 x N m-2, oxygen treated 750 K, re-evacuated, (2) acetic acid vapour, 320 K, 1 h, 0.17 kN m-2; subse- quent evacuation (3) 320 K, 18 h, 6.65 x N m-2; (5) 560 K, 60 h, 6.65 x N m-2, (6) 658 K, 3.25 h, 6.65 x low5 N m-2 and (7) 750 K, 16 h, N m-2, (4) 470 K, 5 h, 6.65 x 6.65 x lod5 N m-2.is consistent with the nature of the oxide surface, but does not preclude the possible participation of surface 0,- and 0'- oxygen species, which have been proposed for other 21 Such species have been characterised by e.s.r. on tin@) oxide itself under similar pretreatment 23 Extensive interaction occurs between ethene vapour and the remaining surface hydroxy groups on high-temperature-pretreated samples of all three oxides. For both tin(1v) oxide and tin(1v) oxide-palladium oxide the hydroxy stretching bands initially shift to lower wavenumber and become broader, indicating a hydrogen-bonding interaction, then on prolonged exposure to ethene these bands decrease in intensity. This decrease can be rationalised in terms of addition of a surface hydroxy group across the carbon-carbon double bond of ethene.That the addition is reversible, under these conditions, is seen by the reappearance of the hydroxy band when the ethene vapour was pumped off in the case of tin(1v) oxide. With tin@) oxide-palladium oxide the hydroxy band did not reappear in its original form.P. G. HARRISON AND B. MAUNDERS 1323 However, as no other absorption bands were observed in the spectrum this could be caused by loss in transmission properties, the disc being orange-brown in colour before ethene absorption but much darker in colour afterwards, most probably because of oxygen depletion of the oxide. With the high-temperature-pretreated tin(1v) oxide-silica sample, the 3720 cm-l band was observed to decrease in intensity in the presence of ethene vapour and shifted to 3700-3710 cm-l with a slight increase in the broad OH band centred at 3360cm-l, indicating hydrogen bonding with surface Si-OH groups, presumably because of their higher Bronsted acidity.When the bands due to the surface acetate structure formed and ethene vapour was pumped off, the band due to isolated Si-OH groups was observed to increase again but did not return to its original value. Hence, although the initial surface adsorbate is formed reversibly, its transformation to surface acetate is irreversible. The initial interaction with ethene adsorption appears to be a reversible hydrogen- bonding interaction between the ethene and surface hydroxy groups, with subsequent formal addition of the 0-H band across the C=C double bond to generate a surface ethoxide group (scheme 2).In a previous study of the tin(v) oxide+ethene system, I CH3 I Scheme 2. Solymosi and B0Zs017 concluded that surface 0;- ions are the primary adsorption sites and that in the absence of gaseous oxygen surface lattice oxygen oxidises the hydrocarbon. However, no surface species were characterised and the only reaction products identified were carbon dioxide and water. Finklo has suggested that on alumina ethene possibly interacts with surface hydroxy groups. The greater reactivity of tin(v) oxide-silica towards ethene may be rationalised in terms of this mechanism. The rate-determining step in classical electrophilic addition to C=C double bonds is the addition of the positively charged part of the electrophile, i.e.: + I I > C = C < + f / W a >C-C-Y/w Hence, reaction with the more Bronsted-acidic surface silanol groups will be favoured compared with reaction with surface Sn-OH or Pd-OH groups, although these too undergo reaction. 44-21324 TIN OXIDE SURFACES Unfortunately, no direct evidence for a surface ethoxide intermediate could be detected from the spectra obtained in the present study. Bands that could be attributable to v(CH,) or v(CH,) stretches were obscured by the intensity and broadness of the hydroxy stretching modes, no deformation modes were observed and the region where a C-0 stretching band might have been observed was obscured by the strong absorptions of the oxide itself. Nevertheless, the occurrence of an intermediate surface ethoxide is highly probable and we have demonstrated previously the facile conversion of surface methoxy groups (generated by the chemisorption of either methanol or dimethyl carbonate) on tin(v) oxide into surface formate groups at temperatures as low as 320 K.24 Similar acetate formation from ethanol via an ethoxide intermediate has been reported by both G~eenler,~ and Kage126 on alumina and by Davydov et al.27 on chromia, whilst Fink2s has suggested that acetate formation from ethanol and alumina proceeds via a coordinated acetaldehyde species.Surface-bound acetaldehyde on tin(v) oxide also undergoes rapid oxidation at 320 K to produce surface acetate.24 Therefore the most probable mechanism for the formation of surface acetate from ethane and ethene on these oxides is by the further reaction of the ethoxide intermediate via the abstraction of a hydrogen atom from a surface ethoxide by a neighbouring oxide species giving an acetaldehyde molecule formally coordinated to a surface tin atom.This species then undergoes reaction with a surface hydroxy group to produce surface acetate and hydrogen (scheme 3). CH3 I \ 1: '\ I l l O/M\O/M\O/M\O CH3 H HC- H H HC,----? I \ I I l l O/M,O/M\o/y\o 0- d o d b 0 0 H CH, Scheme 3. No mass-spectroscopic analysis of the vapour phase was carried out in the present study. However, Kage126 has demonstrated that hydrogen is produced in the conversion of surface ethoxide to surface acetate on alumina. Note that in our previous study of the absorption of acetaldehyde onto tin(v) oxide we observed weak bands at 1620, 1380 and 1317 cm-l, which were assigned to the v(C=C), d(C0H) and v(C=O) modes, respectively, of a surface enolate, CH,=CH(OH),d,, species.24 Similar bands are also observed with ethane adsorption, but not with acetic acid adsorption, on tin(v) oxide-silica at 1630,1380 and 1330 cm-l, lending some support for the formation aldehyde-type intermediates.Ethane adsorption on the tin(v) oxide-silica disc that had been pretreated at 320 K was carried out at 320 K, after which it was pumped off, yet it was not until the discP. G. HARRISON AND B. MAUNDERS 1325 was heated at 488 K, still under vacuum, that the absorption band due to acetate appeared, because of further reaction of the initially surface-adsorbed intermediates. The two other weak bands, at ca.1720 and 1305 cm-l, are undoubtedly associated with the acetate structure, or acetic acid, for both these are observed with acetic acid adsorption as well as ethane adsorption. The 1720 cm-l band often appears split into two bands at ca. 1725 and 1710 cm-l; a similar band, and splitting, has been observed by Lorenzelli et ~ 1 . ~ ~ for acetic acid adsorbed on haematite, and it can be assigned to physisorbed cyclic dimers of acetic acid. The 1305 cm-l band may be associated with a C-H deformation mode or possibly the v(C0H) stretching mode of dimeric species. The nature of the acetate structure cannot be unambiguously determined from the infrared spectra. The separation of the v(C00) asymmetric and symmetric stretching modes for unidentate carboxylates [structure (I)] are greater than for the corresponding chelating bidentate carboxylates [structure (II)].For unidentate cyclopentadienyl titanium carboxylates [structure (111)] separations of up to ca. 300 cm-l have been observed, while for the bidentate chelate of structure (11) of carboxylate groups in Cp,Ti(CH,COO), and Cp,Ti(C,H,COO), separations of SO-60 cm-l were An alternative bidentate of structure (111), with the carboxylate groups bridging two titanium ions, in the dimers of CpTi(CH,COO), and CpTi(C,H,COO), exhibited a separation of ca. 170 cm-l. Lorenzelli et ~ 1 . ~ ~ observed a separation of 100 cm-l for an acetate structure on the surface of a-Fe,O,, which was assigned to the chelating bidentate configuration. In these studies the separation was observed to be 96 cm-l and consequently the chelating bidentate of structure (11) is proposed as the most likely structure, although the bidentate of structure (111) cannot be entirely ruled out.Decomposition of the surface acetate begins at 580 K, as observed by a reduction in the intensity of the acetate bands and the appearance of a band at ca. 1590 cm-l. A second band appears at 1390 cm-l, with the disappearance of the 143 1 cm-l band, on evacuation at 639 K. After evacuation at 750 K only the band at 1590 cm-l remains. The position of this band is characteristic of the v1(A1) mode of bidentate carbonate. The corresponding v,(B,) mode is expected to occur at ca. 1220 cm-l, but in the present case this region is marked by absorption of the bulk oxide material.,* Anshits et al.,l have reported that surface carboxylates on copper(1) oxide decompose to the formate at elevated temperatures over long periods. However, adsorption of formic acid on tin(1v) oxide-silica shows that this is not the case with this system.,, More likely is decomposition to a surface carbonate species.This type of decomposition occurs with acetate on magnesium oxide at 573-723 K, in which methane was also formed,,, and would be readily accounted for by reaction with neighbouring surface hydroxy groups (scheme 4). The results obtained in these systems demonstrate that tin(rv) oxide systems are very strongly oxidising towards hydrocarbons. In view of the poor activity of ethane over most metal oxides, the reactivity observed under the relatively mild conditions1326 TIN OXIDE SURFACES Scheme 4.in this study is quite surprising. A possible explanation for the reactivity of the tin(1v) oxide-silica is the net negative charge obtained on each condensed Sn-0-Si linkage. This would make these surface oxides more nucleophilic than on the pure tin@) oxide and thus more able to abstract a hydrogen from ethane. The activity of the tin@) oxide-palladium oxide is not likely to be caused by the same reason as the concentration ofpalladium oxide is much less than that of silicon oxide ; however, palladium-hydrogen bonds are well known in reactions involved with alkenes and alkynes, so it is possible that under those conditions the formation of a palladium-hydrogen bond may be the driving force.The reactions with ethene occur at much lower temperatures (320-445 K) than needed for surface acetate formation on alumina (523 K).l0 Again, the highly oxidising nature of the tin(1v) oxide surface, under such mild conditions, appears to be fairly unique amongst pure oxides, since only polymerisation of ethene was observed on the oxides of nickel, copper and palladium,6 while on zinc oxide ethene was observed to chemisorb by interaction of the n e l e c t r o n ~ . ~ ~ - ~ ~ Under more strongly oxidising conditions, in the presence of air or oxygen and at higher temperatures, ethene has been oxidised over zinc oxide14 and titania.l69 34 However, these studies did not observe the surface species formed, if any, but analysed the product gases to show exclusively the presence of carbon dioxide. The vapour phases of the present studies were not analysed, though it is possible that carbon dioxide was produced and that the surface acetate may either be an intermediate to CO, production or may form by a competing mechanism.Over mixed l8 quantities of acetic acid and acetaldehyde have been observed under certain conditions, as well as carbon dioxide, although when the conditions become harsher only carbon dioxide was observed, This may be because of oxidation of produced acetic acid or acetaldehyde, or it may just reflect that the harsh conditions favour carbon dioxide formation over acetic acid or acetaldehyde formation. The oxidising nature of tin(1v) oxide may well be caused by its ability to become non-stoichiometric by loss of oxygen, although, for formation of the surface acetate structure, the role of the surface hydroxy groups is also important.In reactions with ethene the tin@) oxide was always observed to discolour, which indicates the formation of non-stoichiometric tin oxide. Both tin(rv) oxide-palladium oxide and tin(1v) oxide-silica are strongly oxidising towards ethene under relatively mild conditions. The behaviour of tin(1v) oxide- palladium oxide is similar to that exhibited by palladium on titania, although in this case a temperature of 423 K was required for ethene conversion.15 Unlike the case for tin(rv) oxide itself and tin(xv) oxide-palladium oxide, a surface acetate is formed upon reaction of ethene with tin(1v) oxide-silica pretreated at a high temperature (660 K).This may be because of the retention of Bronsted-acidic silanol groups necessary for reaction under these conditions and/or the structural incompatibility between SnO, and SiO,, which prevents their heterolytic sintering and hence the propagation of oxygen deficiencies. Thus, the tin(1v) oxide retains its stoichiometry, as suggested by the lack of discoloration of the disc in this case, and still furnishes oxide ions for the oxidation of ethylene.P. G. HARRISON AND B. MAUNDERS 1327 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.). K. Kuchynku, Collect. Czech. Commun., 1968, 33, 3049. A. L. Dent and R. J. Kokes, J . Phys. Chem., 1970, 74, 3653. Y. F. Y. Yao and J. T. Kummer, J . Catal., 1973, 28, 124.H. Raudsepp and M. Mikkal, Tr. Tallin. Politekh. Inst., Ser. A, 1962, 198, 109; Chem Abs., 60, 150%. E. M. Thorsteinson, T. P. Wilson, F. G. Young and P. H. Kasai, J. Catal., 1978, 52, 116. L. H. Little, J . Phys. Chem., 1959, 63, 1616. P. J. Lucchesi, J. L. Carter and D. J. C. Yates, J. Phys. Chem., 1962, 66, 1451. F. Bozon-Veruraz and G. Pannetier, Bull. SOC. Chim. Fr., 1970, 3856. K. P. Zhdanova and N. I. Popova, Kinet. Catal. (Engl. Transl.), 1968, 9, 326. lo P. Fink, Rev. Roum. Chim., 1969, 14, 8 1 1 . l 1 A. L. Dent and R. L. Kokes, J . Phys. Chem., 1970, 24, 3653. l2 R. J. Kokes and A. L. Dent, Adv. Catal., 1972, 22, 1. l 3 A. G. Whitney and I. D. Gray, J. Catal., 1972, 25, 176. l4 Y. Kubekawa, T. Ono and N. Yano, J. Catal., 1973, 28,471. l5 A. Omar, G. Djega-Mariadassou, F. Bozon-Verduraz and G. Pannetier, Bull. SOC. Chim. Fr., 1974, l6 Y. M. Shchekochikhin, V. N. Filimonor, N. P. Keier and A. N. Terenin, Kinet. Catal. (Engl. Transl.), l7 F. Solymosi and F. BOZSO, Ber. Bunsenges. Phys. Chem., 1977, 81, 529. la S. Tan, Y. Moro-Oka and A. Ozaki, J. Cural., 1970, 17, 132. l9 P. G. Harrison and B. Maunders, J. Chem. Soc., Faraday Trans. I , 1975, 71, 461. 2o M. Iwamoto and J. H. Lunsford, J. Phys. Chem., 1980, 84, 3079. 21 S. L. Kaliguine, G. N. Shelimov and V. B. Kazansky, J. Catal., 1978, 55, 384. 22 J. C. H. Van Hooff, J. Catal., 1968, 11, 277. 23 M. Che, C. Naccache and B. Imelik, Bull. SOC. Chim. Fr., 1968, 12, 4791. 24 E. W. Thornton and P. G. Harrison, J . Chem. SOC., Faraday Trans. 1, 1975, 71, 2468. 25 R. G. Greenler, J . Chem. Phys., 1962, 37, 2094. 26 R. 0. Kagel, J. Phys. Chem., 1967, 71, 844. 27 A. A. Davydov, V. M. Shchekochikhin, P. M. Zaitsev, Yu. M. Shchekochikhin and N. P. Keir, Kinet. 28 P. Fink, Rev. Roum. Chim., 1969, 14, 81 1 . 29 V. Lorenzelli, G. Busca and N. Sheppard, J . Catal., 1980, 66, 28. 30 R. S. P. Coutts, R. L. Martin and P. C. Wailes, Aust. J . Chem., 1973, 26, 941. 31 A. G. Anshits, V. D. Sokolovskii, G. K. Boreskov, A. A. Davydov, A. A. Budneva, V. I. Ardeer and 32 P. G. Harrison and B. M. Maunders, unpublished data. 33 V. I. Yakerson and A. M. Rubinstein, Bull. Acad. Sci. USSR, Div. Chem. Sci., 1966, 1278. 34 B. Dmuchorsky, M. C. Freerks and F. B. Zienty, J . Catal., 1965, 4, 577. 12, 2740. 1964, 5, 113. Catal., 1971, 12, 611. I. I. Zakharov, Kinet. Catal., 1975, 16, 95. (PAPER 4/906)