Physical Chemistry - American Chemical Society

---

The Journal of

Physical Chemistry

0 Copyright 1995 by the American Chemical Society

VOLUME 99, NUMBER 38, SEPTEMBER 21,1995

LETTERS Sot-Promoted Chemisorption and Oxidation of Propane over Pt(ll1) Karen Wilson, Christopher Hardacre: and Richard M. Lambert" Chemistry Department, Cambridge University, Cambridge, CB2 IEW, England Received: May 9, 1995; In Final Form: July 21, 1995@

SO2 chemisorption on oxygenated Pt( 11 1) enormously enhances the dissociative chemisorption and subsequent combustion of propane. This activation of the metal surface is induced by an adsorbed sulfoxy species which is formed at '220 K. In the absence of adsorbed SO2 the sticking probability of propane is immeasurably small. However in the presence of S02, the precursor-mediated initial sticking probability rises from -0.02 at 300 K to -0.15 at 160 K: all of the hydrocarbon is irreversibly adsorbed. n-Butane and n-heptane show similar behavior: methane and ethane do not.

Introduction Alkane chemisorption on clean transition metals is in general a slow process due to the inefficiency of the initial dissociative event which involves H atom abstraction by the surface:' this is usually the rate-limiting step in the catalytic conversion of alkanes. The effects of SO2 in hydrocarbon oxidation are important in determining the performance of catalytic combustion systems and catalytic convertors. Reactor studies of W A1203 and Pt-Rh/A1203/Ce02 catalysts show that inclusion of SO2 in the gas feed poisons the oxidation of CO and propene, while actually promoting the oxidation of p r ~ p a n e . ~Infrared .~ analysis of these catalysts following exposure to 0 2 and S02, indicated the presence of S042- on the support but not on the Pt component, leading to the suggestion that the promotion effect must be support mediated.2 However, there is little or no understanding of the chemical mechanisms involved. Under ultrahigh-vacuum (UHV) conditions propene and other alkenes readily chemisorb and decompose on Pt( 11l),4.5enabling their oxidation to be studied by temperature-programmed reaction (TPR) spectro~copy.~.~ This is not possible with propane because Pt( 111) is totally inactive toward dissociative

' Present address: Department of Chemistry, Queen's University, Belfast BT9 5AG, Northem Ireland. * Corresponding author. Abstract published in Aduunce ACS Abstracts, September 1, 1995. @

chemisorption of the alkane under such condition^.^.^ Here we show that chemisorption of SO2 on oxygen-precovered Pt( 111) generates a surface which is highly active for the dissociative chemisorption of propane under UHV conditions, even at 300 K, thus permitting an investigation of the subsequent oxidation catalysis under well-defined conditions. The results unambiguously demonstrate that SO2 directly promotes the catalytic activity of the metal without any mediation by a support. They also suggest that a sulfate/sulfite species on the Pt surface is responsible for the crucial initial step of alkane chemisorption.

Experimental Section The LEED/Auger/TPRS UHV apparatus and method of sample preparation have been described in detail elsewhere.'O Sample cleaning was achieved by Ar+ ion bombardment and treatment with oxygen at 800 K. LEED and Auger spectroscopy were used to verify that the crystal surface was well ordered and clean before commencing chemisorption and reaction measurements, a heating rate of 17 K s-I being employed for the latter. 0 2 (99.995%), C3Hs (99.95%,) and SO2 (99.98%) were supplied by M. G. Distillers. No sulfur Auger data are reported, due to the susceptibility of SO2 and SO, species to electron stimulated desorptioddissociation. I I

0022-3654/95/2099-13755$09.00/0 0 1995 American Chemical Society

Letters

13756 J. Phys. Chem., Vol. 99, No. 38, 1995

300

500

'

900

700

I 3

Pre-annealing Temperature (K) Figure 2. Oxidation product yields following 160 K C3H8 adsorption over P t ( l l 1 ) precovered by 100 langmuirs 0 2 (at 300 K) and 24 langmuirs SO2 (at 160 K) as a function of annealing temperature of the 0 2 / S 0 2 adsorbate overlayer.

v)

3

2

200

400

600

800

Figure 1. (a) TPR following 300 K adsorption of 6 langmuirs C3H8 on Pt( 11 1) precovered by 100 langmuirs of 0 2 (saturation dose) and 24 langmuirs of S 0 2 . (b) TPR following 160 K adsorption of 6 langmuirs of C3Hs on Pt( 11 1) precovered by 100 langmuirs of 0 2 and 24 langmuirs SO2 (dosed at 300 K).

Results The interaction of propane with both clean and oxygen precovered Pt(ll1) at 160 and 300 K was studied by TPR spectroscopy. No detectable adsorption occurred, in agreement with earlier studies of alkane adsorption on clean Pt( 11l).839 Likewise, at both temperatures, no C3H8 adsorption occurred on SO2-precovered Pt( 111). However, stepwise exposure of Pt(ll1) to 0 2 and then SO2 at 300 K produced a surface that was very active for dissociative adsorption of C3H8: the initial sticking probability for chemisorption of propane at 300 K increased from an immeasurably small yalue on the clean Pt surface to 0.02 & 0.01 on the activated surface. (The latter value was estimated by quantifying the desorption yield of CO and C02 produced after a known propane exposure.) The reverse dosing sequence (SO2 then 02) was entirely ineffective. Figure l a shows typical TPR data obtained after presaturation with oxygen at 300 K, followed by 24 langmuirs of SO2 and 6 langmuirs of propane, also at 300 K. CO2 is observed at 440 K along with SO2 at 400 K and 540 K. Post-reaction Auger analysis showed no detectable S , C, or 0 after such TPR experiments, indicating complete combustion of the adsorbed hydrocarbon and restoration of a clean surface. Survey experiments revealed that n-butane chemisorption exhibits similar O,/ SO2 promotional behavior, whereas n-heptane underwent chemisorption and oxidation in the presence of preadsorbed oxygen alone: in this latter instance SO2 behaved as an additional source of O(a),increasing the extent of hydrocarbon oxidation. Methane and ethane showed no 02/S02 promotional effect at 160 or 300 K; presumably reflecting the higher C-H bond strength in these two molecules. The former result is also in agreement with studies of methane oxidation over Pt-Rh/Ce02/Al203 catalyst^.^ Reactively formed H20 desorbs from Pt( 111) I300 K;6,13 therefore to characterize the reaction more thoroughly, the 300 K 02/S02 pretreated crystal was cooled to 160 K prior to C3H8

adsorption. The resulting TPR spectrum is shown in Figure lb, from which it can be seen that H20, C02, and CO desorbed at 320, 440, and 500 K, respectively. A low-temperature shoulder at 320 K is also apparent in the 44 amu data-this is characteristic of CO oxidation over Pt( 111).21 At 160 K the propane initial sticking probability increased to 0.15 f 0.05. Exposure of the Pt( 11l)/O(a)system to SO2 at 160 K did not induce subsequent adsorption of C& indicating that there is a threshold temperature for formation of the surface sulfoxy species responsible for C3H8 adsorption. To determine this threshold temperature, SO2 was adsorbed on the Pt( 11l)/O(a) surface at 160 K, this was then preannealed to 220, 250, 300, 400, and 600 K prior to C3H8 exposure at 160 K. In the subsequent TPR sweeps the extent of C3H8 uptake and oxidation increased sharply after annealing to 250 K, continued to increase with preannealing temperature, and passed through a maximum at 300 K (Figure 2). This variation in reactivity with increasing annealing temperature may be ascribed to increased surface concentration of the active SO, species; the trend in CO and C02 production indicates that this SO, species is stable up to -500 K and destroyed by -600 K. Substantial changes in the 64 amu (S02+)desorption spectra occurred when C3H8 was adsorbed on the 02/S02 precovered surface. Curves A and B in Figure 3 show the 64 amu desorption following adsorption of 0 2 and SO2 alone and after subsequent adsorption of C3H8 at 160 K, respectively. In the absence of C3H8 two SO2+ desorption states are observed at 370 and 560 K. The high temperature state coincides with a peak in the 80 amu spectrum (S03') and the low-temperature peak may be assigned to SO2 desorption, in agreement with earlier work on S02/02 coadsorption on Pt(l1 l).I4 Following adsorption of C3H8, two new peaks at 330 K and 440 K appeared in the 64 amu spectrum, and the 560 K peak was attenuated. The 560 K peak continued to decrease with increasing C3H8 exposure and was eventually extinguished at high C3H8 coverages (oxygen-lean conditions), at which stage CO appeared as a product. The 440 K 64 amu (S02+) and 44 amu (C02+) peaks were always coincident (e.g., Figure 3), and there is a strong correlation between the intensities of these two features as shown in Figure 4, which also shows the corresponding results for CO and H20. The integrated intensities of the 440 K 64 and 44 amu peaks were obtained from experiments conducted under partial oxidation conditions (i.e., 560 K 64 amu peak not present; CO formation observed). This procedure was followed in order

Letters

J. Phys. Chem., Vol. 99, No. 38, 1995 13757 peratures increase the lifetime of the molecular precursor state, increasing the probability of it encountering an active site. There is some uncertainty in the literature regarding the nature of SO2 adsorption on clean Pt( 111) and no detailed information is available about the coadsorption of 0 2 and SOZ. Wassmuth concluded that SO2 adsorbs dissociatively at 160 K forming SO, and Oa, which undergo recombinative desorption at 300 In the light of their HREELS and XPS data, White et a1.I8 argue in favor of molecular adsorption at 130 K, followed at 300 K either by desorption as Son, or dissociation to form adsorbed S, SO, and S04. Our results indicate that preadsorbed oxygen is a necessary condition for Sopinduced activation of the surface toward propane chemisorption. In the presence of oxygen we find a threshold temperature of >220 K for the formation of the activating species. The identity of this species is not certain, although our observation that SO2 dosed at 300 or 160 K does not activate the surface indicates that it cannot be S, SO, or S02.l8 Pt is an efficient catalyst for the oxidation of SO2 to SO, which may be desorbed as a stable reaction p r ~ d u c t . ' In ~ accord with this, the 560 K SO2+ peak (Figure 3B) and the associated coincident 80 amu signal (s03+, Figure 3A) are ascribed to SO3 desorption. In these experiments, SO2 was dosed at 300 K, which is above the dissociation temperature for molecular SO2.l8 Hence as the 370 K 64 amu feature is not associated with any coincident S03+ peak, and we assign it to SO2 desorption resulting from 0, SO, recombination, as suggested by Wassmuth et a1.l6.l 7 In the presence of adsorbed propane (Figure 3B) this feature is masked by a 330 K 64 amu peak: since C3H8 chemisorption almost certainly involves H abstraction, we tentatively suggest that the 330 K SO2+ peak is due to adsorbed HSO, which undergoes decomposition to SO2. (In this connection it should be noted that Leung et al. have suggested that SO;! and H2S can react to form H2SO3 over Cu( Furthermore, this is consistent with the pronounced decrease in SO3 desorption induced by the presence of coadsorbed propane, which strongly suggests the occurrence of a reaction between the hydrocarbon and the precursor to SO3 desorption (Figure 3). The correlation between the integrated intensity of the 440 K S02+ feature and the C02 desorption yield (Figure 4) points to the formation of a complex between the SO, species and either the hydrocarbon fragments or reactively formed CO. Note that the CO2 peak exhibits a shoulder at -320 K which corresponds to oxidation of CO, by 0,,2'indicating that some of the reactively-formed CO is adsorbed directly onto the metal surface and not incorporated into a SO,-containing complex. CO desorption (Figure lb) is observed only under oxygen-lean conditions (Le., in the absence of the 560 K SO2+ peak due to so3 desorption). This CO peak occurs at the same temperature (500 K) as that observed during TPR oxidation of ethene and propene on R: therefore it probably results from the oxidation of adsorbed carbon atomS.6.7.20 Our observations may be rationalized in terms of the following scheme. 1. Initial dissociative adsorption of propane involving H abstraction by SO,: K.I69l7

200

400

'

600

800

'

Figure 3. Comparison of 64 amu desorption obtained following coadsorption of io0 langmuirs 02 and 24 langmuirs of SO2 (dosed at 300 K and then cooled to 160 K), with that obtained following 160 K adsorption of 3 langmuirs of C3H8 on the 02/SOz overlayer. The 44 amu desorption is also shown to illustrate the correlation between the 440 K SO, and C02 peaks.

+

440 K 64 amu peak area 0

1

2

3

4

5

Figure 4. Oxidation product yields as a function of 440 K 64 amu intensity. See text.

to eliminate interference from the 560 K 64 amu peak which occurs only under total oxidation conditions. The initial exposure of 0 2 was varied (1- 100 langmuirs) prior to exposure of 24 langmuirs of SO2 at 300 K, and 6 langmuirs of C3H8 at 160 K . This combined with the results shown in Figure 3 suggest formation of a surface complex between the SO, species and the hydrocarbon, either upon chemisorption or upon subsequent reaction of the latter.

Discussion Promotion of propane oxidation has previously been studied only over supported Pt c a t a l y ~ t s , ~ . ~ .the l ~ . 'results ~ being interpreted in terms of support-mediated effects due to the presence of sulfate on the oxide phase; this view is at least consistent with infrared observations which show that sulfated alumina adsorbs propane.2 However, the present results clearly demonstrate that in the presence of chemisorbed oxygen, SO2 dramatically enhances the chemisorption and oxidation of propane on platinum in the absence of any effects due to a support phase. Thus the initial sticking probability of propane on the oxygenated, S02-activated surface is -0.15 at 160 K and 0.02 at 300 K, whereas on the or oxygenated metal surface it is immeasurably small at both temperatures. There are two possible routes to the dissociative chemisorption of alkanes - direct or precursor mediated di~sociation.~ A decrease in sticking probability with increasing surface temperature is consistent with the latter mechanism: lower adsorption tem-

SO, +'C3H,

-

+

C3Hy HSO,

2. At higher temperatures C-C bond cleavage occurs to form CH, fragments which are oxidized to yield H20, CO, and COz. 3. Coincident desorption of SO2 and C02 at 440 K suggests decomposition of a surface complex. No coincident H20 desorption occurs, which indicates that the complex does not contain any H, Le., is of the form CO-SO,. This could

13758 J. Phys. Chem., Vol. 99, No. 38, 1995 References and Notes

decompose according to CO-SOx-C02

+ SO2 +

Letters

(X

- 2)0,

(-440 K)

At high propane coverages (oxygen-lean system) incomplete oxidation occurs. No excess oxygen is available for SO3 formation, and the associated 560 K amu desorption feature is no longer observed. CO is then produced by the oxidation of surface carbon at the characteristic temperature associated with this process, as observed during catalytic hydrocarbon oxidation.6.7.20 The identity of the SO, species and further aspects of the mechanism of propane adsorption and reaction on Pt( 111) are currently under investigation by HREELS and XPS.I9

Conclusions A Pt( 111) surface precovered by 0 2 and SO2 at 300 K is highly active for the dissociative chemisorption and subsequent combustion of propane. The precursor-mediated initial sticking probability of propane on the SO,-activated surface varies between -0.02 (300 K) and -0.15 (160 K). Support-mediatedeffects are not necessary for SO2 promotion of Pt-catalyzed propane combustion. The threshold temperature required for formation of the activating SO, species is '220 K. This species appears to form a surface complex with carbon-containing surface intermediates. Similar promotion effects occur with n-butane and n-heptane but not with methane and ethane.

Acknowledgment. C.H. and K.W. acknowledge the award of a UK EPSRC Research Fellowship and EPSRC Research Studentship, respectively. This work was supported under EPSRC Grant No. GWJ00632.

(1) Garin, F.; Hilaire, L.; Maire, G. In Catalytic Hydrogenation-Studies in Surface Science & Catalysis; Elsevier: Amsterdam; Vol. 27, p 145. (2) Yao, H. C.; Gandhi, H. S.; Stephien, H. K. J. Catal. 1981, 67, 231. (3) Ansell, G. P.; Golunski, S. E.;.Hatcher, H. A,; Rajaram, R. R. Catal. Lett. 1991, 11, 183. (4) Avery, N. R.; Sheppard, N. S. Proc. R. SOC. London 1986, A405, 1

I.

( 5 ) Salmeron, M.; Somorjai, G. A. J. Phys. Chem. 1982.86, 341. (6) Steiniger, H.; Ibach, H.; Lehwald, S. SUI$ Sci. 1982, I 1 7, 685. (7) Wilson, K.; Hardacre, C.; Lambert, R. M. Manuscript in preparation (8) Firment, L. E.; Somorjai, G. A. J. Chem. Phys. 1977, 66, 2901. (9) McMaster, M. C.; Arumainayagam, C. R.; Madix, R. J. Chem. Phys. 1993, 177, 461. (10) Hardacre, C.; Roe, G. M.; Lambert, R. M. Surf: Sci. 1995,326, 1. (11) Wassmuth, H.-W.; Ahner, J.; Hofer, M.; Stolz, H. Prog. Surf: Sci. 1993, 42, 257. (12) Gandhi, H. S.; Shelef, M. Appl. Catal. 1991, 77, 175. (13) Ogle, K. M.; White, J. M. Surf: Sci. 1984, 139, 43. (14) Astegger, St.; Bechtold, E. Surf: Sci. 1982, 122, 491. (15) Hubbard, C. P.; Otto, K.; Gandhi, H. S.; Ng, K. Y. S. J. Catal. 1993, 144, 484. (16) Hofer, M.; Hillig, S.; Wassmuth, H. W. Vacuum 1990, 41, 102. (17) Kohler, U.; Wassmuth, H. W. Surf: Sci. 1982, 117, 668; Su$ Sci. 1983, 126, 448. (18) Sun, Y.-M.; Sloan, D.; Alberas, D. J.; Kovar, M.; Sun,Z.-J.i White, J. M.; Surf: Sci. 1994, 319, 34. (19) Wilson, K.; Hardacre, C.; Lambert, R. M., manuscript in preparation. (20) Schafer, L.; Wassmuth, H. W. Surf: Sci. 1989, 208, 55. (21) Gland, J. L.; Kollin, E. B. J. Chem. Phys. 1983, 78(2), 963. (22) Leung, K. T.; Zhang, X. S.; Shirley, D. A. J. Phys. Chem. 1989, 93, 6164.

JP95 12932