J . Phys. Chem. 1985, 89, 4905-4908
+ M = N2 + 0 + M N20 + 0 = N2 + 0 2
NzO
+ 0 = NO + N O N20 + H = N, + OH
N20
from our study of N 2 0 decomposition and the N20-H2 reaction? the calculated t , values are in good agreement with the observed ones, as shown in Figure 4. The revised rate constant expressions used are k3 = 7.0 X l O I 4 exp(-28000/RT) cm3 mol-’ s-l, k, = 5.6 X l O I 4 exp(-28000/RT) cm3 mol-’ S-I, and kS = 1.5 X l O I 4 exp(-15000/RT) cm3 mo1-l s-l. When these values are used, the rate constant calculated for k6 becomes 1.6 X l o i 4 exp (-50300/RT) cm3 mol-’ s-l. Both the OH* concentration and ignition delay time t , calculated with these values fit the experimental data well as shown in Figures 2 and 4. Systematic errors in the instrumentation and kinetic model (including the rate constant of the quenching reactions) influence the k6 value. The rate constant expression for k6, therefore, may be assigned generous error bounds as (1.6 f 1) X 1014exp(-50300/RT) cm3 mol-’ s-I. A comparison of observed with computed OH* emission profiles using the mechanism including the revised rate constants for k3, k4, kS,and k6 is shown in Figure 5. The calculated and observed curves are in fairly good agreement. Since our k6 expression is derived from fitting the intensity maxima only (Figure 2), the agreement with the time dependence (Figures 4 and 5) provides an independent confirmation of k6 and also of the predictive ability of the mechanism and rate constant expressions used.
(2) (3) (4)
(5) with literature values for their rate constantss, the maximum OH* concentrations calculated at high temperatures are much smaller than observed, as shown in Figure 2. Therefore, a large portion of OH* is formed by some other reactions. The simplest assumption is N 2 0 H = OH* N 2 A”,= 17 kcal/mol (6)
+
4905
+
Though reaction 6 is endothermic, it might be responsible for production of OH*. In computing OH* concentrations, it is necessary to assume the quenching rate constant of OH* by each species. The values reported previously6were used, and we assumed that NO and N2 have the same quenching rate as O2 and that N 2 0 has same s quenching rate as H 2 0 . We adopted the value of 0.7 X reported previously6 as the radiative lifetime of OH*. When the value of k6 = 1.1 X 1014exp(-50300/RT) cm3 mol-I s-l is used as the rate constant of reaction 6, the OH* concentrations calculated fit the experimental data best, as shown in Figure 2. Ignition delay times t , were defined as times between reflected shock arrival and maximum OH* emission intensity. A plot of t , vs. 1 / T for each mixture is shown in Figure 3. The t , values calculated with the literature rate constants are longer than observed. This is considered to come from a real shortcoming of the rate constant expressions in the N20-H2 reaction. When we use revised rate constant expressions for k,, k4, and k5 obtained
Conclusion Chemiluminescent OH* formation in H2 oxidation by N 2 0 over the temperature range 1400-2000 K was found to be mainly due to H 0 + M = OH* M and N 2 0 + H = N 2 + OH*. The expression k6 = (1.6 f 1) X lOI4 exp(-50300/RT) cm3 mol-’ was determined for reaction 6. With this rate constant, the OH* emission intensities and ignition delay times were satisfactorily interpreted.
+
(8) D. L. Baulch, D. D. Drysdale, D. G. Horne, and A. C. Lloyd, “Evaluated Kinetic Data for High Temperature Reactions”, Vol. 2, Butterworths, London, 1973.
+
(9) Y. Hidaka, H. Takuma, and M. Suga, Bull. Chem. SOC.Jpn., 58,2911 (1985).
Gas-Phase Oxidation of Silver: The Reaction of Silver Clusters with Ozone James L. Gole,* R. Woodward, J. S. Hayden, High Temperature Laboratory, Center for Atomic and Molecular Science, and School of Physics, Georgia Institute of Technology, Atlanta, Georgia 30332
and David A. Dixon Central Research and Development Department, E . I. DuPont de Nemours and Company, Wilmington, Delaware 19898 (Received: June 26, 1985) In a device which produces a metallic flow intermediate to that of a standard effusive source (atoms and small percent diatomics) and laser vaporization-plasma formation followed by rare gas entrainment at pressures exceeding several torr (wide cluster distribution), the reaction of small silver clusters, M,, n 2 3, with ozone has been observed in the gas phase at pressures in the millitorr range. Chemiluminescent emission from the products Ago and what appear to be Ag20 and Ag,O, where x 2 3, has been monitored. We associate a red shift of the spectral features with the increasing size of the metal cluster oxide. The spectral data combined with supplementary thermodynamic information demonstrate that Ago* with enough energy to account for the observed chemiluminescence cannot be produced through the reaction of either Ag or Ag, with 0 3 .The smallest cluster whose reaction can yield excited states of Ago is the trimer. The formation of Ag20* can also be achieved through reaction of the trimer; however, it may also be accounted for via reaction of higher clusters. There has been widespread and growing effort to understand the structures and properties of small free metallic clusters. An increasing number of experimental characterizations1 are now beginning to balance the impressive array of theories which have been applied to these systems.2 Small clusters have been generated in flow systems, reacting with reagents in another continuous or pulsed flow stream under high pressure (-30-500 torr) conditions in a modified merged flow e n ~ i r o n m e n t . ~The products in the flow have been measured mass spectrometrically; however, this technique provides no direct measurement of structural or dynamic properties. In contrast, we have recently detected chemiluminescence from the oxidation of sodium clusters by halogens under single-collision condition^.^ This latter experiment is ad0022-3654/85/2089-4905$01.50/0
vantageous in that it provides both dynamic and spectroscopic information about the products of the metathesis. The chemistry of silver is of substantial technological importance in both photographic5 and catalytic processes.6 This is exemplified through the use of supported silver in the epoxidation of ethylene and bulk silver in the dehydrogenation of methanol to give formaldehyde. The bulk of experimental studies, thus far, on “naked” silver clusters have been done in low-temperature matrices.’,* We report here initial results from a study of the oxidation of gas-phase silver clusters with ozone. The present approach provides a novel route to the preparation of small metal cluster oxides of silver, Ag,O. The apparatus used in this study is a modified version of that 0 1985 American Chemical Society
4906
The Journal of Physical Chemistry, Vol. 89, No. 23, 1985
used in previous studies of metal atom oxidation.’ We wish to create an environment which is intermediate to that of a lowpressure effusive source and those conditions which prevail subsequent to the creation of the plasma formed in laser vaporization ( I ) (a) ‘Metal Bonding and Interactions in High Temperature Systems”, J. L. Gole and W. C. Stalley, Eds., American Chemical Society, Washington, DC, 1982, ACS Symp. Ser. No. 179. (b) “Diatomic Metals and Metallic Clusters”, Symp. Faraday Soc., 14 (1980). (c) J . L. Gole, “The Gas Phase Characterization of the Molecule Electronic Structure of Small Metal Clusters and Cluster Oxidation” in ‘Metal Clusters”, M. Moskovits, Ed., Wiley, New York, in press. (d) D. A. Garland and D. M. Lindsay, J . Chem. Phys., 78, 2813 (1983). (e) W. H . Gerber, Ph.D. Thesis, Universitat Bern, Switzerland, 1980; W. H. Gerber and E. Schumacher, J . Chem. Phys., 69, 1692 (1978). See also Bull. A m . Phys. Soc., 27, 304 (1982). (f) W. Schulze, H. V. Becker, R. Minkwitz, and K. Manzel, Chem. Phys. Lett., 55, 59 (1978). (g) D. P. DiLella, K. V. Taylor, and M. Moskovits, J . Phys. Chem., 87, 524 (1983). (h) J. A. Howard, K. F. Preston, R. Sutcliffe, and B. Mile, J . Phys. Chem., 87, 536 (1983). (i) K. Hilpert and K. A. Gingerich, Eer. Eunsenges. Phys. Chem., 84, 739 (1980). Q) J. A. Howard, R. Sutcliffe, and B. Mile, J . A m . Chem. SOC.,105, 1394 (1983). (k) W. D. Knight, Surf: Sci., 106, 172 (1981). (1) W. D. Knight, Helv. Phys. Acta, 56, 521 (1983). (m) G. A. Ozin, H . Huber, and S. Mitchell, Inorg. Chem., 18,2932 (1979). (n) A. R. George, Bull. A m . Phys. SOC.,28,285 (1983). (0)K. Clemenger and W. A. deHeer, Bull. Am. Phys. SOC.,28,285 (1983); 28, 1321 (1983). (p) W. Saunders and W. A. deHeer, Bull. Am. Phys. Soc., 28, 1344 (1983). (9) D. E. Powers, S. G. Hauser, M. E. Geusic, D. L. Micholopoulos, and R. E. Smalley, J . Chem. Phys., 78, 2866 (1983). (r) L. Genzel, T. P. Martin, and U. Kreibig, Z . Phy. E , 21, 399 (1975). (s) M. Hofmann, S. Leutwyler, and W. Schulze, Chem. Phys., 40, 145 (1979). (t) M. Moskovits and G. A. Ozin, “Cryochemisitry”, Wiley, New York, 1976. (u) G. A. Ozin, Card. Rea. Sci. Eng., 16, 191 (1977). (v) A. L. Robinson, Science, 185, 772 (1974). (w) G. C. Demitras and E. L. Muetterties, J . A m . Chem. Soc., 99, 2796 (1977). (x) J. H. Sinfelt, Acc. Chem. Res., 10, 15 (1977). (y) E. L. Muetterties, Science, 196, 839 (1977). (z) E. L. Muetterties, Bull. SOC.Chim. Eelg., 84, 959 (1975). (aa) E. L. Muetterties, Bull. SOC.Chim. Eelg., 85, 451 (1976). (bb) E. L. Muetterties, R. N. Rhodin, E. bdnd, C. F. Brucker, and W. R. Pretzer, Chem. Rea., 79, 91 (1979). (cc) E. Band and E. L. Muetterties, Chem. Rea., 78, 639 (1978). (2) (a) J. P. Martins, J. Buttet, and R. Car, Phys. Reu. E, 31, 1804 (1985). (b) S. C. Richtsmeier, D. A. Dixon, and J. L. Gole, J . Chem. Phys., 86, 3942 (1982). (c) S. C. Richtsmeier, M. L. Hendewerk, D. A. Dixon, and J. L. Gole, J. Phys. Chem., 86, 3937 (1982). (d) J. L. Martins, R. Car, and J. Buttet, J . Chem. Phys., 78, 5646 (1983). (e) J. Flad, H . Stoll, and H. Pruess, J. Chem. Phys., 71, 3042 (1979). See also Chem. Phys., 75, 331 (1983). (f) E. R. Dietz, Ph.D. Thesis, University of California, Berkeley, 1980; E. R. Dietz, Phys. Reu. A , 23, 751 (1981). (g) W. A. deHeer, Bull. Am. Phys. SOC., 28, 285 (1983). (h) M. J. Rice, W. R. Schneider, and S. Strassler, Phys. Rec. E, 24, 554 (1981). (i) S. Richtsmeier, J. L. Gole, and D. A. Dixon, Proc. Narl. Acad. Sci. U.S.A., 77, 561 1 (1980). Q ) H. F. Schaeffer, 111, Acc. Chem. Res., 10, 287 (1977). (k) C. Bauschlicher, Jr., P. Bagus, and H. F. Schaeffer, 111, IEM J . Res. Dec., 22, 213 (1978). (I) A. L. Companion, D. J. Steible, and A. J. Starshak, J . Chem. Phys., 49, 3637 (1968). (m) A. L. Companion. Chem. Phys. Lerr., 56, 500 (1978). (n) B. T.Pickup, Proc. R . SOC.London, Ser. A , 333, 69 (1973). (0)A. Gelb, K. D. Jordan, and R . Silbey, Chem. Phys., 9, 175 (1975). (p) D. W. Davies and G . Del Conde, Mol. Phys., 33, 1813 (1977). (4) D. M. Lindsay, D. R. Herschbach, and A. L. Kwiram, Mol. Phys., 39, 529 (1980). (r) P. S. Bagus, G. Del Conde, and D. W. Davies, Faraday Discuss., Chem. SOC.,62, 321 (1977). (s) J. Kendrick and I. H. Hiller, Mod. Phys., 33, 635 (1977). (t) J. L. Gole, R. Childs, D. A. Dixon, and R.A. Eades, J . Chem. Phys., 72,6368 (1980). (u) J. G. Fripiat, K. T. Chow, M. Boudart, J. R. Diamond, and K. H. Johnson, J . Mol. Catal., 1, 59 (1975). (v) A. B. Anderson, J . Chem. Phys., 66, 5108 (1977). (w) R. C. Baetzold, J . Chem. Phys., 55, 4355 (1971). (x) R. C. Baetzold, Adu. Catal., 25, 1 (1976). (y) W. A. Goddard, S. P. Walch, A. K. Rappe, T. H . Upton, and C. F. Melius, J . Vac. SOC.Technol., 14, 416 (1977). (2) C. Bachman, J. Demuynk, and A. Veillard, Gazz. Chim. Ifal., 108, 398 (1978). (aa) R. P. Messmer, S. K. Knudsen, K. H . Johnson, J. B. Diamond, and C. Y. Yank, Phys. Reo. E, 13, 1396 (1976). (bb) R. P. Messmer. T. C. Caves, and C. M. Kao, Chem. Phys. Lett., 90,296 (1982). ( 3 ) (a) M. E. Geusic, M. D. Morse, and R. E. Smalley, J . Chem. Phps., 82, 5901 (1985). (b) S. C. Richtsmeier, E. K. Parks, K . Liu, L. G. P o b , and S. J. Riley, J . Chem. Phys., 82, 3659 (1985). (c) E. K. Parks, K. Liu, S. C. Richtsmeier, L. G. Pobo, and S. J. Riley, J . Chem. Phys., in press. (d) R. L. Whetton, D. M. Cox, D. J. Trevor, and A. Kaldor, J . Chem. Phys., 89.5666 (1985). (e) D. J. Trevor, R. L. Whetton, D. M. Cox, and A. Kaldor, J . A m . Chem. SOC.,107, 518 (1985). (f) E. A. Rolfing, D. M. Cox, and A. Kaldor, J . Chem. Phys., 81, 3322 (1984). (g) R. L. Whetton, D. M. Cox, D. J. Trevor, and A. Kaldor, Phys. Rec. Lerr., 54, 1494 (1985). (4) W. H. Crumley, J. L. Gole, and D. A. Dixon, J . Chem. Phys., 76,6439 (1982). (5) “The Physics of Latent Image Formation in the Silver Halides”, A. Baldereschi, W. Czaja, E. Tosatti, and M. Tosi, Eds., World Scientific, Singapore, 1984. “The Theory of the Photographic Process”; T. H. James, Ed.. MacMillen, New York, 1977. (6) “Ethylene and Industrial Derivatives”, S. A. Miller, Ed., Ernest Benn Ltd., London, 1969. (7) See G. A. Ozin, Faraday Symp., Chem. SOC.,14, 7 (1980). (8) W. Schulze and H. Abe, Faraday Symp., Chem. SOC.,14, 87 (1980).
Letters as it is entrained in a continuous or pulsed rare gas flow at high pres~ure.’‘’~ The former technique produces atoms and a small concentration of dimers whereas the latter technique produces a wide diversity of much larger clusters although at small ( 107/cm3)concentration. By operating our metal source at temperatures or through containment designs such that the Knudson number associated with the source is much less than one, we create the seed for the initial phases of a cluster-forming environment. Thus, in the present study, we combine a high flux noneffusive metal source and those techniques which have proven invaluable in studying chemiluminescent processes across a wide pressure rangeg to produce controlled substantial metal cluster concentrations and probe the chemiluminescent emission from their reaction with ozone. In the present experiment, silver is heated in a specially designed crucible to temperatures between 1400 and 1700 K corresponding approximately to a gas-phase silver flux between 5 X 10l6 and 5 X IO1* particles/cm2.s) at the oven orifice, the concentrations at the upper limit being well in excess of that generated from an effusive source. The silver flux is entrained in a rare gas (He or Ar) flow (Pt0,,,= 100-500 millitorr) and agglomeration to form small silver clusters via collisions occurs both as a result of the high metal flux and as the high silver concentration is cooled by the room temperature entrainment gas. On the basis of previous s t ~ d i e sthe , ~ final internal rotational temperature of the clusters will be considerably lower than the source from which they exit (-TRot I800 K). At a suitable point above the flow, ozone is mixed with the entrained silver clusters and a chemiluminescent flame is formed. The chemiluminescence which is usually characterized by rotational temperatures between 500 and 700 K9 is detected with an appropriate spectrometer photomultiplier arrangement in a standard configuration similar to that used in previous experiments9 The three spectra shown in Figure 1 were generated as a function of increasing silver flux. The spectra in Figure 1a were C 1). The observed obtained with a moderate silver flux (KKnudsen emission corresponds to A g o in the region 4000-4200 8, (some continuation in 4200-4400-A range) and is believed to correspond to Ag20 in the region 4200-4800 A. The A g o bands are assigned following previous worklo*] to predominantly (0,O) transitions, originating from predissociating excited A* II spin-orbit components to ground-state X211 spin-orbit levels. The band at 4096 8, is assigned to the 2r13/2-2113,2transition and that at 4125 A is assigned to the 211,,i2-2111i2transition. The resolved vibrational structure (see also Figure 1 b) between 4200 and 4800 8, has not previously been observed. The spacings between the peaks do not follow a regular progression; however, the number of observed features indicates a substantial change in at least one molecular parameter (bond angle or bond length) upon transition. We tentatively correlate the emission with a transition from an excited state of A g 2 0 to what is thought to be a linear ground state.12 Although the observed frequency spacing is not regular, it is possible that a portion of the features correspond to a bending mode progression. Further analysis is underway. As the silver flux is increased (Figure lb), the features emanating from the predissociating A*II state of A g o are quenched and the features due to Ag20 become more pronounced. A further
+
(9) See, for example, D. R. Preuss and J. L. Gole, J. Chem. Phys., 66,2994 (1977); J. L. Gole and D. R. Preuss, J . Chem. Phys., 66, 3000 (1977); L. H. Dubois and J. L. Gole, J . Chem. Phvs., 66, 779 (1977); D. M. Lindsay and J . L. Gole, J . Chem. Phys., 66, 3886 (1977); M. J. Sayers and J. L. &le, J . Chem. Phys., 67, 5442 (1977); J . L. Gole and S. A. Pace, J . Chem. Phys., 73,836 (1980); C. L. Chalek and J. L. Gole, J. Chem. Phys., 65,2845 (1976); J. L. Gole and S. A. Pace, J . Phys. Chem., 65, 2651 (1981); J. L. Gole, Annu. Reti. Phys. Chem., 27, 525 (1976). ( I O ) U. Uhler, Ark. Fys. 7, 125 (1953). (1 1) The identity of the ground state has been confirmed by a b initio calculations. See C. W. Bauschlicher, Jr., C. J. Nelin, and P. S. Bagus, J . Chem. Phys., 82, 3265 (1985). (12) Based on comparison with L i 2 0 or N a 2 0 we would predict a linear ground state of D-h symmetry. One might conceive of a C,.configuration, AgAgO; however, the formation of this species is difficult to rationalize on energetic grounds. Detailed quantum chemical calculations are now in progress at Los Alamos Scientific Laboratory (R. L. Martin and P. J. Hay) to evaluate these states and assess whether the ground state is in fact linear.
Letters
The Journal of Physical Chemistry, Vol. 89, No. 23, 1985 4907 react to form the excited electronic states of the metal monoxide. Further consideration of established thermodynamic^'^ demonstrates that no cluster of silver can react with ozone via the abstraction of a single silver atom Ag,
+ O3
-
Ago
+ Ag,.., + 0,
(1)
to yield A g o in an excited electronic state. Since the difference in the bond strengths of A g o and O3 (04,) is only 1.2 eVI5 (favoring the metal oxide), a dark product of the form Ag,Oz must be formed in order that sufficient energy be released to form A g o with at least 3-3.5 eV of internal energy. Because the strong luminescence from A g o is observed at lower silver fluxes, the A g o is likely generated from small clusters. The reaction of silver dimer
+ 03
Ag2
-+
AgO
+ Ago2
(2)
is also not sufficiently exothermic to form the excited states of A g o since the A g o z (Ag-O2)I7 bond strength is approximately equal to or slightly less than the Ag, (1.63 eV)I4 dissociation energy. The reaction of silver trimer with ozone Ag,
Figure 1. Chemiluminescentemission from the oxidation of entrained (argon, helium) silver metal fluxes with ozone. (a) Moderate metal flux with KKnudam < 1 showing emission from Ago (4000-4300 A) and Ag,O (4200-4800 A). (b) With increased silver flux the Ago features
(4000-4300 A) are quenched and the spectrum is dominated by Ag,O fluorescence. (c) At even higher silver flux the Ag,O features decrease in intensity and new features (-5000-6000 A) associated with Ag,O ( x Z 3) onset in the spectrum.
increase of the silver flux (the generation of higher-order clusters) leads to the quenching of the A g 2 0 features and the formation of a new, previously unobserved, emission system (or emission systems) extending from 5000 to 6000 A. There are two clear groups of spectral features superimposed on what appears to be an almost continuous grouping of overlapped yet reproducible spectral features. It is possible that these features will be more apparent after further apparatus improvements lead to an increase in intensity. At present, we observe a doublet near 5800 A and a group of three to four weak bands corresponding to a short progression. A preliminary analysis of the 5000-6000-A region suggest that the observed features result from Ag,O, n 2 3, although considerable further experiments and analysis will be needed. As one might anticipate, the observed emission appears to red shift as the size of the metal cluster oxide increases. Those A g o emission features observed in the present study, not only at 4000 A but also at 3500 A (B-X), are found to be strongly dependent on the relative silver flux and ozone concentration (through a series of experiments monitoring the effect of changing reactant concentrations). A g o emission is favored by the combination of the highest achievable ozone concentrations and the lowest silver concentrations, whereas both the 4000- and 3500-8, systems are quenched when the ozone flow (low even at the lowest silver fluxes employed in the present study) is introduced into a silver-rich system. The bond strength of is 2.29 eV (52.8 kcal/mol) as measured by mass spectrometric techniques. l 4 Thus the reaction of a single silver atom with ozone cannot release sufficient energy to form electronically excited silver oxide. Silver clusters must (13) See, for example, K. P. Huber and G. Herzberg, "Molecular Spectra and Molecular Structure-Constants of Diatomic Molecules", Van Nost-
rand-Reinhold, New York, 1979. (14) S. Smoes, F. Mandy, A. Vander Auwera-Maheiu, and J. Drowart, Bull. SOC.Chim. Belg., 81, 45 (1972).
+ O3
-
Ago
+ Ag202
(3)
could be sufficiently exothermic to populate excited states of A g o if the formation of Ag,Oz from Ag, and 0, releases 3.0 eV of energy. A large energy release is possible because of the low trimer, Ag-Ag,, bond strength,I8 and because of the possibility of some vibrational excitation in the trimer. The formation of Ag20 can occur directly from reaction of the dimer with O3 Ag,
+ O3
-
Ag,O
+ 0,
(4)
However, if we assume that the two Ag-0 bond strengths in Ag20 are the same as that in Ag-0, it is unlikely that reaction 4 will lead to the formation of Ag20 in the excited electronic state from which emission is observed unless an unlikely substantial vibrational excitation in Ag, (= 1.0 eV) is transferred into electronic excitation. Reaction 4 in the absence of vibrational excitation releases 1.92 eV of energy which is clearly not sufficient to form Ag,O with enough energy to luminesce in the observed range. The trimer can also react via
-
Ag,
+ O3
-
AgzO + A g o z
(5)
to form A g 2 0 with enough internal energy to account for the observed chemiluminescence. Process 5 is exothermic by 4.4 eV. Note that reactions 3 and 5 are complimentary suggesting that Ag3 might react to form either excited states of A g o or, at other times, excited states of Ag,O. An alternate possibility, consistent with the observed emission behavior as a function of silver and ozone concentrations and the predicted structures of higher silver clusters,zc involves the formation of Ag,O via the reaction of an Ag, c l ~ s t e r ' ~ The formation of the excited states of the higher metal cluster oxides would appear to result from metatheses involving larger metal clusters.20 It is more likely that clusters of odd atom combination will react to yield the observed fluorescence since, (15) F. D. Rossini, "Selected Values of Chemical Thermodynamic Properties", National Bureau of Standards, Washington, DC. (16) The bond strength for Ag, is 1.63 eV and that for O4 to give 0, 0 is 1.03 eV. (17) We estimate the bond strength of Ago2 Ag 0,as 1.5 eV based on the bond strength for NaO, Na O2 of 1.6-2.0 eV and the ionic character observed for CuO, (C. W. Bauschlicher, private communication). (1 8) The Ag2-Ag bond strength is 0.62 eV. See ref 2c. (19) Indeed, quantum chemical calculations which predict the structure of the Ag, ground state and very low-lying isomeric forms (D2d,DZh,and Dqh configurations) lend support to this possibility. See ref 2c. (20) Also consistent with the present observations, the reaction of a silver atom with vibrationally excited ground-state A g o formed via reaction 3 might also produce electronically excited Ag,O. Further studies are underway to assess this possbility.
-
+
-
+
+
4908
-
J . Phys. Chem. 1985, 89, 4908-4914
for small clusters, the process Ag, Ag,-, + Ag requires less energy if n is odd, the potentially closed-shell configurations showing more stability.*O The variation of spectral features with pressure (Figure 1) and the supplementary thermodynamic analysis would seem to indicate that we- have observed the ixidation of silver clusters. Subsequent experiments are underway to better characterize the silver flux,
to search for the identity of dark products, and to examine the effects of increased agglomeration. Acknowledgment. J.L.G. acknowledges partial support of this research by the National Science FmndatiOn. Registry No. Ag, 7440-22-4; Oj,10028-15-6; Ago, 1301-96-8; Ag,O, 20667-12-3.
ARTICLES Kinetics of the NO 4- CO Reaction on Clean Pt: Steady-State Rates R. L. Klein,? S. Schwartz, and L. D. Schmidt* Department of Chemical Engineering and Materials Science, University of Minnesota, Minneapolis, Minnesota 55455 (Received: February 19, 1985; In Final Form: June 10, 1985)
The kinetics of reaction between NO and CO on clean polycrystalline Pt are studied for temperatures between 300 and 1200 K, pressures between 1 and lo-* torr, compositions from Xco = 0.0005 to 0.99, and rate variations over a factor of lo8. Surfaces are shown to be clean before reaction and are found to contain only monolayers of NO or CO following cooling and pumpdown after reaction at any pressure. No nonstoichiometric residues of reactant or contaminant species are observed on the surfaces under any conditions. In excess NO (Xco < 0.01) the steady-state kinetics can be fit quantitatively assuming a Langmuir-Hinshelwood bimolecular rate expression with a heat of adsorption of NO of 17 kcal/mol. For Xco > 0.05, the reaction can be fit by a bimolecular rate expression, rR PNo/Pco, at low temperatures, but at high temperatures the rate is proportional to PNO and independent of Pco. Near stoichiometric ratios the reaction appears to be limited by the adsorption rate of NO at high temperatures. For pressures between 10” and lo-’ torr the rate can be fit quantitatively by the same LH mechanism valid at high pressures with a heat of adsorption of CO of 30 kcal/mol. These results strongly suggest that this reaction on Pt is a true bimolecular reaction between adsorbed NO and CO rather than unimolecular NO decomposition with CO scavenging of adsorbed oxygen because the kinetics fit only the former mechanism and because rates are up to lo4 higher than those of NO decomposition. N
Introduction The reaction NO
+ CO
-
1/2N2+ C 0 2
(1) over supported Pt, Pd, and Rh is the major reaction of NO removal in the automotive catalytic converter.’ Reaction in the converter occurs for reactant partial pressures up to 10 torr for widely varying compositions of NO and CO at temperatures up to -800 K. Rhodium is a superior catalyst to Pt, presumably due to its greater ability to dissociate NO. This reaction has been studied extensively2-’ on polycrystalline and single-crystal Pt, although no consensus on detailed rate expressions or mechanisms (unimolecular vs. bimolecular ratelimiting steps) has emerged. The (1 l l ) plane of Pt is totally unreactive2s3 for both unimolecular NO decomposition and the NO C O reaction under ultrahigh vacuum (UHV) conditions, while the (100) and some high index planes appear to cause complete reaction of the limiting reactant at low coverage^.^,^ One mechanism for the NO C O reaction, proposed by Bell et al.8,9on Pt and other noble metal catalyst surfaces, is a sequence initiated by the dissociation of adsorbed NO, followed by CO scavenging of coadsorbed oxygen which would otherwise poison the surface for further adsorption and reaction. This process would occur through the steps: NO,,) NO,,) (2- 1)
+
+
C0,d
-- cow
(2-2)
‘Present address: UOP, Des Plaines, IL.
0022-3654/85/2089-4908$01.50/0
Neither C 0 2 nor N2 is believed to chemisorb on Pt although N atoms are observed to be weakly chemisorbed.lO-” Therefore, if the rate-limiting step were eq 2-3, the Langmuir-Hinshelwood (LH) rate expression should be
where kR is the surface reaction rate coefficient
(1) Taylor, K. C. GM Research Publication GMR 4190 PCP 192. (2) Gorte, R. J.; Schmidt, L. D. Surf. Sei. 1981, 111, 260. (3) Gorte, R. J.; Schmidt, L. D.; Gland, J. L. Surf. Sei. 1981, 109, 367.
(4) Lesley, M. W.; Schmidt, L. D., to be published. ( 5 ) Lambert, R. M.; Comrie, C. M. Surf. Sci. 1974, 46, 61. (6) Adlhoch, W.; Lintz, H. G. Surf. Sei. 1978, 78, 58. (7) Singh-Boparai, S. P.; King, D. A. ‘Proceedings of the 4th International
Conference on Solid Surfaces and Third European Conference on Surface Science, Cannes, France, 1980”; Vide, Couches Minces, 1980, No. 201 Suppl. (8) Lorimer, P.; Bell, A. T. J. Cural. 1979, 59, 223. (9) Hecker, W.L.; Bell, A. T. J. Catul. 1983, 84, 200. (IO) Shigeishi, R. A,; King, D. A. Surf. Sci. 1977, 62, 379. ( 1 1) Schwaha, K.; Bechtold, E. Surf. Sci. 1977, 66, 383.
0 1985 American Chemical Society
