Adsorption and Polymerization of Acetylene on Oxide Surfaces A Raman Study BY JOHN HEAVISIDE AND PATRICK J. HENDRA* Department of Chemistry, The University, Southampton SO9 5NH PETER TSAI AND RALPH P. COONEY Department of Chemistry, University of Newcastle, N.S.W., Australia 2308 AND Received 20th October, 1977 Laser Raman spectra have been recorded of adsorbed poly(acety1ene) formed by the polymeriza- tion of acetylene on y-alumina and zeolite KX. Spectral evidence indicates that on both y-alumina and zeolite KX monomeric acetylene is physisorbed via its 7~ electron system in a ‘‘ side-on ” orienta- tion. The chain length of the polyene chains formed on y-alumina appear to increase with increasing activation temperature. The line intensities in the Raman spectra of the coloured polyenes are significantly enhanced by resonance effects.The adsorption of acetylene on alumina has been studied previously by infrared rnethod~.l-~ Two distinct modes of interaction have been observed by Yates and Lucce~i,~ and supported by Bhasin et aL5 A stronger “ end-on” interaction involving hydrogen bonding through the acetylenic hydrogen atom and a weaker “ side-on ” interaction involving the n electrons of the acetylene have both been i~lentified.~ As well as adsorbed acetylene, Bhasin et u Z . , ~ reported a coloured product on the surface of alumina. They showed evidence of polymerization by the appear- ance of the C=C stretching band in their infrared spectra. The present study was undertaken to elucidate the nature of coloured surface species resulting from surface induced reactions of acetylene.Previous laser Raman work with the alkali cation exchanged series of zeolites A reveal only a weak “ side-on ” interaction. It has been shown that adsorbed acetylene was more strongly held by zeolite KX than NaX or LiX.’ The tendency of acetylene to polymerize on this surface was therefore also investigated. and X EXPERIMENTAL MATERIALS y-alumina (Grade H supplied by Peter Spence Ltd) was used in the powdered form at a mesh size of 100-200. The B.E.T. N2 surface area is reported by the manufacturers to be z 120 m2 g-l. The zeolite used, was prepared using conventional ion exchange methods from the powdered form of the zeolite NaX, supplied as Linde 13X (by Union Carbide). The amount of potassium exchange was determined by flame photometry and found to be 68 %.Acetylene gas (British Oxygen) was further purified through a purification train consisting of a concentrated sulphuric acid trap and a cold trap (ethanol, 200K). The infrared spectrum of the purified gas showed no evidence of impurity. 2542J . HEAVISIDE, P . J . HENDRA, P . TSAI AND R . P . COONEY 2543 INSTRUMENTATION Laser Raman spectra were recorded on a modified Coderg T800 spectrometer using the green (514.5 nm) or the blue (488.0 nm) emission line of a Spectra Physics model 170 argon ion laser as a source. Laser power levels used ranged from 1 to 130mW at the sample. Plasma radiation was removed from the source using an interference filter (Technical Optics). The laser beam was focused onto the sample surface with a cylindrical lens rather than the more conventional spot focusing lens in order to reduce the light flux at the surface.6 A spectral bandpass of 6 cm-1 was used throughout.Photons were detected using a cooled S20 phototube (EMI-9558) and then processed using a Brookdeal-Ortec 5C1 photon counting system. Supplementary spectra were recorded using a Cary 81 spectrometer in the 90" mode and using an unfocused beam of a Coherent Radiation Laboratories 52G argon ion laser as excitation source. The powdered samples were enclosed in a cell of the type described earlier.g PRETREATMENT OF SAMPLE The powdered y-alumina was activated by heating at temperatures ranging from 300- 1200 K in a vacuum of x mmHg. Samples pretreated in the temperature range 700- 870 K, suffer from a high background fluorescence.To minimize this they were heated in oxygen for 24 h,lO* l1 in addition to the normal pretreatment. All samples were cooled to room temperature under high vacuum prior to adsorption. The zeolite was used in the form of discs pressed at pressures of ~404oOo kg cm-2 in a 1 cm diameter stainless steel split die. The discs were cleaned by slowly heating in vacuo until a pressure of M mm was reached. This precaution was taken to avoid structural breakdown of the zeolite by the sudden removal of water at temperatures above 450 K.12* l3 They were then heated to a final temperature of 820 K in 600 Hg of oxygen for 5-24 h and then cooled to room temperature under a pressure of x mmHg. The X-ray diffraction patterns of the activated zeolite showed no evidence for structural changes as a result of this pretreatment.Using a vacuum system, purified acetylene was introduced at atmospheric pressure to the adsorbents which had been pre-cooled with liquid nitrogen. The precooling of the KX samples (not employed in previous studies)' was important for the production of detectable quantities of polymerized product. Spectra of the adsorption complex were recorded at ambient temperature. RESULTS AND DISCUSSION activated at various temperatures are summarized in table 1. y-alumina and zeolite KX (both activated at 820 K) after adsorption of acetylene. The Raman spectra of acetylene adsorbed on y-alumina and zeolite KX both Fig. 1 shows typical spectra of the adsorbent adsorbate complex formed on PHYSISORPTION OF ACETYLENE )"A L U MI N A Physisorbed acetylene which could easily be removed by evacuation was detected on all samples.The v(C=C) stretching vibration was detected, at about 1962 cm-l in all cases. From table 2 it can be seen that this frequency is considerably shifted downwards from the gas phase value (1974 crn-l)14 and this behaviour is character- istic of acetylene interacting weakly with the substrate, through its n-electron~.~* The alternative " end-on " interaction through the acetylenic proton is also known to occur, 4* but this shifts the v(C=C) stretching frequency to a higher value than in the gas phase. This behaviour has been detected by infrared techniques 4 * although no sign of it was detected in our Raman spectra.P U 0 cn sample (activation temp) TABLE RA RAM AN SPECTRA * OF ACETYLENE ADSORBED ON y-A1203 AND ZEOLITE KX (AP/crn-l) polyene unchanged 645 w 1954 vs 3321 w unchanged 635vw 1961 m unchanged 1962 w cream 1962 w brown 1962 vw purple 1962 vw CH v3 V 2 ddormation mode mode 2’3 v2+”3 2v2 1 0 2 5 1 ~ 1130s 1512s out-of-plane planar planar 2245 w, br 2634 w, br 3027 vw, br 1136m 1520rn 2230 w, br 2608 w, br 3025 vw, br 1013 w 1127 s 1517 s 1122 s 1507 s -2200 w, vbr -2625 w, vbr 3000 vw, vbr 1015 w 1117vs 1501 vs 2225 w, br 2612 w, br 2988 vw, br av, very; w, weak; m, medium; s, strong; br, broad. z 0 1: U cd 0 l? cc2545 The intensity of the band due to physisorbed acetylene appeared to drecrease as the activation temperature of the alumina was raised." As the number of surface hydroxyl groups decreases with increasing activation temperature, we suggest that the physisorbed acetylene interacts through the z-system with them and/or water molecules adhering thereto.This would be an example of sorption to specific sites as distinct from mono- and multi-layer sorption. J. HEAVISIDE, P. J . HENDRA, P. TSAI AND R. P. COONEY I I t 1 3500 3000 2500 2000 1500 1000 500 A;tcm-l FIG. 1 .-Raman spectra of acetylene on (a) y-Al,O, (activated at 820 K) and (b) zeolite KX (activated at 820 K). Exciting line, 514 nm Ar+ (130 mw). Slit width, 6 an-'. TABLE 2.-RAMAN SPECTRA OF THE VARIOUS PHASES OF ACETnENE vibration previous work this work gas liquid 0 crystal KX KX (820 K) y-AlzO3 (300 K) v1 v(G--H) 3374 3341 3324 3318 3321 a v2 v(C=C) 1974 1961 1956 1952 1954 1962 ~ 4 , v i G(HrC-C) 612 630 626 a 645 635 (activation temperature) Q Not observed ; b see ref.(14) ; C see ref. (15) ; d see ref. (16) and (17) ; e see ref. (7). ZEOLITE KX Consistent with previous assignments,6 the bands at Av = 645,1954 and 3321 cm-I are attributed to physisorbed acetylene. Table 2 shows the Raman spectroscopic data for the gas,14 liquid,15 crystalline 16* l7 and adsorbed acetylene on zeolite KX (1954 and 3321 cm-I). The frequencies observed are very similar to those already reported for acetylene on zeolite KA6 and for crystalline acetylene. This would suggest that the adsorbed and crystalline phases are closely related as has been previously suggested. On detailed examination of the spectrum, the band at 1954 cm-l was found to be asymmetric on the lower frequency side suggesting more than one band present.After evacuation, consistent with earlier results for A type zeolites,6 the residual maxima had shifted to 1945 cm-l. Similar effects have been obtained previously 6 e * Fig. l(u) shows the sample activated at 820K in which the band at 1962 cm-' can only just be observed.2546 ADSORPTION AND POLYMERIZATION OF ACETYLENE and have been explained in terms of acetylene bonded to two different types of site in the zeolite framework. H-C= - T C I ! i 0 FIG. 2.-Physisorption of acetylene. POLYENE FORMATION If acetylene is left in contact with activated alumina at room temperature the surface is found to discolour. We therefore studied the Raman spectra of these surfaces with a view to elucidating the nature of the coloured surface species.In the present study, strong Raman lines were found in the 1100 and 1500 cm-l regions in some of the spectra (see table 1). The Raman spectra of polyenes 8 * are known to give rise to two characteristic intense bands in these regions and have been assigned to skeletal vibrations l 8 (see fig. 3). Therefore, we suggest that in this case the two bands in our spectra may be assigned to skeletal modes in a polyene. J J J v2 (1500 cm”) FIG. 3.-Skeletal vibration of polyene chains.J. HEAVISIDE, P. J . HENDRA, P . TSAI AND R. P. COONEY 2547 A weak band can be observed in the spectrum of acetylene on zeolite KX (820 K) at 1025 cm-l and at M 1015 on y-alumina (820 and 1200 K). Shirakawa et aZ.,19 have assigned a band in this region of the trawpoly(acety1ene) spectrum to the trans-CH out-of-plane deformation vibration.The Raman spectra that we observe are hence consistent with that of tran~-poly(acetylene).~~ As well as the funda- mentals, the overtones and combination bands of the symmetric polyene-chain vibrations v3 M 1125 and v2 M 1500 cm-l were observed (see table 1). Ivanova et u Z . , ~ O observed an extended overtone structure of the polyene skeletal mode in the resonance Raman spectra of C2H5COO(CH=CH),COOC2H5, n = 6,7 or 8. They asserted that it is only observed when the exciting frequency falls into the interior of the visible vibrational structure of the absorption band. In this study, a very broad absorption band, ranging from 650 nm into the ultra- violet region, was observed in the reflectance spectrum of the purple polyene on y-alumina (1200K).The two exciting laser lines used (514.5 and 488.0nm) were both in the interior of this absorption band and hence it is likely that the spectrum is resonance enhanced. 1170 1070 \ I I 1 I I 1120 1130 1140 1500 1510 1520 v3 v2 AT1crn-l FIG. 4.-Raman band shifts in the polyene bands with varying activation temperatures of pA1203. The first overtones and combination bands of the skeletal modes were observed in both cases. The intensity of the spectrum was found to be slightly dependent on the exciting line used. In fact, the spectrum was enhanced ,by 60 % using the 514.5 nm line compared with the 488.0 nm line. Gravimetric results show that very little polyene was formed on the surfaces I <0.5 mg g-1 A1203 (1200 K)I, but despite this minute amount, the Raman spectrum of the polyene is very intense.This evidence suggests strongly that the spectrum of the polyene is resonance enhanced. Recently, we have recorded spectra from thermally degraded poly(viny1 chloride).2548 ADSORPTION AND POLYMERIZATION OF ACETYLENE? This material is known to suffer dehydrochlorination to produce unsaturated sectors of the polymer chains containing conjugated double bands. This material produces a Raman spectrum similar to that from sorbed acetylene. In P.V.C. the spectrum is known to be resonance enhanced.21 The v2 and v3 frequencies of the polyene are sensitive to the activation temperature of the y-alumina. These bands shift to lower frequencies with increasing activation temperatures.The plot of frequencies (v, and v,) against the activation temperatures is given in fig. 4. Behringer has attributed the variations in the frequency of the vc=c vibrations in the polyene spectra to differing conjugated chain lengths; an increase in chain length resulting in a decrease of frequency. It appears that the colour changes observed on y-alumina at different activation temperatures (see table 1) are also related to the changes in chain length. Longer chain length increases the conjugation and hence the absorption maxima moves further into the visible region.22 We therefore propose that longer polyene chains are formed on y-alumina activated at high temperatures. It is tempting to speculate on the average length of these polyene sequences produced but we are unable to guarantee the cis: trans isomeric configuration of our product or to be sure that sorption does not perturb the skeletal vibrational frequencies.We feel estimates of chain length could be most misleading. As activation is essential for polymerization, it was thought that the formation of polyenes was linked in some simple way with the acidic surface sites. It has been shown that at higher activation temperatures of y-alumina, an increasing number of active surface sites (of which a large proportion are Lewis acid in nature) are formed.23 This would therefore suggest that the weight of polyene formed should increase markedly with increased activation temperature. As explained above, this cannot be confirmed experimentally, as the amount of polymer formed is minute in all cases.We therefore suspect that the polymerization cannot be taking place at Lewis acid sites activing individually. It is possible that pairs of Lewis sites are involved in the polymerization stage of the reaction, as in olefin isomerisation 24* 2 5 but again we suspect that the amount of polyene produced is too small to satisfy this explanation. It is likely that the polymerization step of the reaction occurs by addition to a terminal C=CH2 group which itself could be absorbed to a Lewis acid pair. Further, the polyene molecules themselves may be adsorbed along their length, thus masking many otherwise active sites, and reducing the amount of polymerization, but we accept that this conclusion is far from satisfactory.Our proposal is supported by the low values observed for the frequencies of the polyene skeletal modes. The authors are grateful to the Office of Naval Research (U.S.N.), ARPA, the S.R.C. and the Australian Research Grants Committee for assistance with funds. J. H. would like to thank the S.R.C. for providing a Postgraduate Award. P. T. thanks the Australian Government for providing a Commonwealth Postgraduate Research Award. N. Sheppard and D. J. C. Yates, Proc. Roy. SOC. A, 1956,238,69. L. H. Little, Infrared Spectra of Adsorbed Species (Academic Press, London, 1966), pp. 143-153. W. A. Pliskin and R. P. Eischens, J. Chem. Phys., 1956,24, 482. D. J. C. Yates and P. J. Lucchesi, J. Chem. Phys., 1961, 35,243. M. M. Bhasin, C.Curran and G. S. John, J. Phys. Chem., 1970,74, 3973. Nguyen The Tam, R. P. Cooney and G. Curthoys, J.C.S. Furaduy I, 1976, 72,2577. Nguyen The Tam, R. P. Cooney and G. Curthoys, J.C.S. Faraday I, 1976,72,2592. * T. A. Egerton and F. S. Stone, J.C.S. Furaduy I, 1973, 69, 22.J . HEAVISIDE, P. J . HENDRA, P. TSAI AND R. P. COONEY lo T. A. Egerton, A. H. Hardin, Y. Kozirovski and N. Sheppard, Chem. Comm., 1971, 841. l1 T. A. Egerton, A. H. Hardin, Y. Kozirovski and N. Sheppard, J. Catalysis, 1974,32,343. l2 R. M. Barrer and W. I. Stuart, Proc. Royal Soc. A, 1959,249,464. l3 J. L. Carter, P. J. Lucchesi and D. J. C. Yates, J. Phys. Chem., 1964, 68, 1385. l4 G. Hemberg, Infrared and Rarnan Spectra of Polyatornic Molecules (Van Nostrand, New York, l5 G. Glockler and M. M. Redrew, J. Chem. Phys., 1938,6,340. l6 G. L. Bottger and D. F. Eggers, J. Chem. Phys., 1964,40,2010 ; 1966,44,4366. l7 M. Ito, T. Yokoyama and M. Suzuki, Spectrochim. Acta, 1970,26A, 695. 2549 P. J. Hendra and E. J. Loader, J.C.S. Trans. Faraday SOC., 1971, 67, 828. 1945), pp. 288-90. J. Behringer, Observed Resonance Raman Spectroscopy, ed. H. Szymanski (Plenum Press, New York, 1967), pp. 168-223. l9 H. Shirakawa, T. Ito and S . Ikeda, Polymer J., 1973,4,460. 2o T. M. Ivanova, L. A. Yanovskaya and P. P. Shorygin, Optika i Spektroskopiya, 1965,18,206. 21 D. L. Garrard and W. Maddams, Macromolecules, 1975, 8, 54. 22 H. H. JafTe and M. Orchin, Theory and Applications of Ultra-violet Spectroscopy (Wiley, 23 P. J. Hendra, I. D. M. Turner, E. J. Loader and M. Stacey, J. Phys. Chem., 1974, 78, 300. 24 W. K. Hall and H. R. Gerberich, J. catalysis, 1966,5, 99. 25 I. D. M. Turner, S. 0. Paul, E. Reid and P. J. Hendra, J.C.S. Faradzy I, 1976,72,2829. New York, 1962), pp. 220-41. (PAPER 711848)