J. Phys. Chem. 1987, 91, 2637-2645 with a. expressed in Gauss (gJ3, erg G-l), and
where He is the resonance field for a free electron. Equations B2 and B3 may each be rearranged to give
2637
Adding and subtracting eq B2 and B 3 converts Hi into h,. Rearranging gives 21HohA0 = (B9) (21 + 1)Ho - h+ 2H0 = h+
* [h+2+ 21Aoh-]’/2
B y substituting for A. from eq B9 in eq B10 and simplifying, it may be shown that (21
which is analogous to eq 5 of ref 24, and 2Ho[Ho - H*1 A0 = f (21 1)Ho - H+
+
(B7)
similar to eq 4 of ref 47. We now extend these previously derived results to obtain an analytic solution for Ho in terms of the two transition fields H,. Define h+ = 1 / 2 ( H + + H-) 1 h- = -(H- - H+) 21
Note that, in the limit of small A o / H o ,h+
-
(B8)
Ho and h-
-
Ao.
+ l ) H o = (1 + l ) h + * 1[h+2+ (21 + l ) h - 2 ] ’ / 2
(B11)
which is the desired result. Accordingly, from the measured H,, it is easy to calculate Ho from ( B 1 1 ) and then find A. from, for example, (B7). The conventional parameters a, and go may then be obtained from eq B4 and B5. For sufficiently large Ao,the low-field FSR transition field (H+) becomes negative: see eq B6 and ref 24. For this situation, the dominant low-field feature may be an “NMR transition” between m,, m = -112, -1 and m,, m = -112, -1 1 .24 With this particular N M R transition field denoted by H,,, then it is straightforward to show that eq B2 through B11 still apply provided H+ is replaced in both rare gas24 by -H,. This latter situation pertains to and nitrogen matrices. Registry No. Cu,, 66771-03-7; N1. 7727-37-9; 63Cu, 14191-84-5.
+
Potential Probes of Metal Cluster Oxlde Quantum Levels. Optical Signatures for the Oxidation of Small Metal Clusters M, (M = Cu, Ag, B, Mn) R. Woodward, P. N. Le, M. Temmen, and J. L. Gole* High Temperature Laboratory, Center f o r Atomic and Molecular Science, and School of Physics, Georgia Institute of Technology, Atlanta, Georgia 30332 (Received: June 17, 1986)
A source configuration which lies intermediate to a low-pressure effusing molecular beam and a high-pressure flow device
is used to generate large concentrations of small metal clusters in a highly oxidizing environment. Clusters are formed from a high metal flux source (KKnudscn C l), which creates the seed for the initial phases of a cluster-forming environment, and are further agglomerated in a controlled argon or helium flow at room to liquid nitrogen temperature. The high-flux noneffusive source has been combined with techniques which have proven valuable in studying chemiluminescent processes across a wide pressure range to probe the chemiluminescent emission from several processes including the oxidation of small copper (Cu, 0 3 )silver , (Ag, + 0 3 )boron , (B, NO2,N20),group IVA (group 14), and early transition-metal clusters. From these studies we have obtained the first quantal information on the energy levels and optical signatures of several metal cluster oxides, MnOY. The present study outlines the potential for chemiluminescent probes of metal cluster oxide quantum levels, not only within themselves but as a means of suggesting future laser fluorescent probes of the metal cluster oxides.
+
+
Introduction As the current volume demonstrates, there is now widespread interest and a growing effort to understand several aspects of the structure and properties of atomic, molecular, and ionic clusters. Within this grouping, metal clusters represent unique intermediate states of matter, the analysis of whose properties should reveal much about the growth of atoms into small metal particles,’ the development of features in the bulk metallic phase,2 and, within (1) Bauer, S. H., private communication. See also, Kung, R. T. V.; Bauer, S. H. 8th International Shock Tube Symposium, Stoller, J. L., Ed.; Chapman and Hall: London, 1971. Keifer, J.; Lutz, B. J. Chem. Phys. 1966, 44, 658. Stever, H. G. “Condensation Phenomena in High Speed Flows” In Princeton Series on High Speed Aerodynamics and Jet Propulsion, Vol. 111, Princeton University: Princeton, N J 1970; Section F. Frurip, D. J. “Light Scattering and Absorption by Nucleating Metal Vapors”. Ibid. Freund, H. J.; Bauer, S. H. “Homogeneous Nucleation in Metal Vapors. 11. Dependence of the Heat of Condensation on Cluster Size”. Ibid. Bauer, H.; Frurip, D. J. “Homogeneous Nucleation in Metal Vapors. 111. A Self Consistent Kinetic Model”. Ibid. (2) Kittle, C. Introduction to Solid State Physics; Wiley: New York, 1971, 4th ed. Solymar, L.; Walsh, D. Lectures on the Electrical Properties of Materials; Oxford University Press: London, 1975. Harrison, W. Solid State Theory; McGraw-Hill: New York, 1970.
themselves, a very intriguing dynamic behavior. Further, sufficient evidence now exists to demonstrate that metal clusters are of importance to the fundamental mechanisms of catalysis and numerous chemical conversion^.^ Thus, the basic properties (geometry, bond strength, reactivity) of small metallic clusters, M , ~
(3) (a) Fischer, T. E. Physics Today, May 1974,23. (b) Robinson, A. L.; Science 1974, 185, 772. (c) Slater, J. C.; Johnson, K. H. Phys. Today, Oct 1974, 34. (d) Muetterties, E. L. Bull. Soc. Chim. Belg. 1975,84, 959. 1976, 85, 451. (e) Thomas, M. G.; Beier, B. F.; Muetterties, E. L. J. A m . Chem. Soc. 1976, 98, 1296. (f) Demitras, G. C.; Muetterties, E. L. J. Am. Chem. Soc. 1977,99, 2796. (9) Robinson, A. L. Science 1976,194, 1150. (h) Sinfelt, J . H. Arc. Chem. Res. 1977, 10, 15. (i) Sinfelt, J . H. Science 1977, 195, 641. (j) Schaefer, H. F. Acc. Chem. Res. 1977, 10, 287. (k) “Advances in Metal Cluster Chemistry Detailed”, Chem. Eng. News, Apr 3, 1978, p 20. (1) “Catalyzing Reactions of Small Abundant Molecules”, Chem. Eng. News, March 1, 1976, p 17. (m) See, for example, Bond, G. C. Sut$ Sci. 1969, 18, 11. (n) Robinson, A. L. Science 1974, 185, 772. Schaefer, H. F. Arc. Chem. Res. 1977,10, 287. Nature (London) 1978, 274, 17, and references in these reviews. (0)Robinson, A. L. Science 1976, 194, 1150, 1261. Muetterties, E. L. Science 1977, 196, 839. (p) See, for example, Johnston, R. D. Adu. Inorg. Radio-Chem. 13,471. Also Cotton, F. A. Acc Chem. Res. 1978, 1 1 , 225, and references therein. (9) See, for example, Cotton, F. A.; Chisholm, M H . Chenz. Eng. News, June 28, 1982, p 40
0022-36541871209 1 -2637$0 1.5010 0 1987 American Chemical Society
2638 The Journal of Physical Chemistry, Vol. 91, No. 10, 1987 (2 I n < 6 ) , have become the subject of intense theoretical and experimental s t ~ d y . ~ , ~ While it is clear that the electronic and reactive properties of small metal clusters lie intermediate to those of the atom and those of the bulk metallic phase, the detailed aspects of this picture especially with respect to internal mode structure have not yet been colored. Quantum level probes which have already indicated the need for models including vibronic coupling6 as well as demonstrating the unique dynamic behavior of small metal clusters as a function of temperature are slowly emerging7 It is now (4) (a) Metal Bonding and Interactions in High Temperature Systems, Gole, J. L., Stwalley, W. C., Eds.;American Chemical Society: Washington, DC, 1982; ACS Symp. Ser. No 179. (b) "Diatomic Metals and Metallic Clusters", Symp. Faraday SOC.1980, 14. (c) Gole, J. L. 'The Gas Phase Characterization of the Molecule Electronic Structure of SmalI Metal Clusters and Cluster Oxidation" In Metal Clusters, Moskovits, M., Ed.; Wiley: New York, 1986. (d) Garland, D. A.; Lindsay, D. M. J . Chem. Phys. 1983, 78, 2813. (e) Gerber, W. H. Ph.D. Thesis, Universitat Bern, Switzerland, 1980. Gerber, W. H.; Schumacher, E. J . Chem. Phys. 1978, 69, 1692. See also: Bull. Am. Phys. SOC.1982, 27, 304. (f) Schulze, W.; Becker, H. V.; Minkwitz, R.; Manzel, K. Chem. Phys. Lett. 1978, 55, 59. (g) DeLella, D. P.; Taylor, K. V.; Moskovitz, M. J . Phys. Chem. 1983,87, 524. (h) Howard, J. A.; Preston, K. F.; Sutcliffe, R.; Mile, B. J . Phys. Chem. 1983, 87, 536. (i) Hilpert, K.; Gingerich, K. A. Ber. Bunsen-Ges. Phys. Chem. 1980, 84, 739. (j) Howard, J. A,; Sutcliffe, R.; Mile, B. J . Am. Chem. Soc. 1983, 105, 1394. (k) Knight, W. D. Surf. Sci. 1981, 106, 172. (I) Knight, W. D. Hela. Phys. Acta 1983, 56, 521. (m) Ozin, G. A.; Huber, H.; Mitchell, S . Inorg. Chem. 1979, 18, 2932. (n) George, A. R. Bull. Am. Phys.Soc. 1983,28,285. (0) Clemenger, K.; deHeer, W. A. Bull. A m . Phys. Sot. 1983, 28, 285. 1983, 28, 1321. (p) Saunders, W.; deHeer, W. A. Bull. A m . Phys. SOC.1983, 28, 1344. (4) Powers, D. E.; Hauser, S. G.; Geusic, M. E.; Micholopoulos, D. L.; Smalley, R. E. J . Chem. Phys. 1983, 78, 2866. (r) Genzel, L.; Martin, T. P.; Krebig, U. Z . Phy. B 1975, 21, 399. (s) Hofmann, M.; Leutwyler, S.; Schulze, W. Chem. Phys. 1979, 40, 145. (t) Moskovits, M.; Ozin, G . A. Cryochemistry; Wiley: New York, 1976. (u) Ozin, G. A. Catal. Rec. Sci. Eng. 1977, 16, 191. (v) Robinson, A. L. Science 1974, 185, 772. (w) Demitras, G. C.; Muetterties, E. L. J . A m . Chem. Soc. 1977, 99, 2796. (x) Sinfelt, J. H. Arc. Chem. Res. 1977, 10, 15. (y) Muetterties, E. L. Science 1977, 196, 839. (z) Muetterties, E. L. Bull. SOC.Chim. Belg. 1975, 84, 959. (aa) Muetterties, E. L. Bull. SOC.Chim. Belg. 1976, 85, 451. (bb) Meutterties, E. L.; Rhodin, R. N.; Band, E.; Brucker, C. F.; Pretzer, W. R. Chem. Reo. 1979, 79, 91. (cc) Band, E.; Muetterties, E. L. Chem. Rec. 1978, 78, 639. (5) Martins, J. P.; Buttet, J.; Car, R. Phys. Reu. B 1985, 31, 1804. (b) Richtsmeier, S. C.; Dixon, D. A.; Gole, J. L. J . Chem. Phys. 1982,86, 3942. (c) Richtsmeier, S. C.; Hendewerk, M. L.; Dixon, D. A,; Gole, J. L. J . Phys. Chem. 1982, 86, 3937. (d) Martins, J. L.; Car, R.; Buttet, J. J . Chem. Phys. 1983, 78, 5646. (e) Flad, J.; Stoll, H.; Pruess, H. J . Chem. Phys. 1979, 71, 3042. See also, Chem. Phys. 1983, 75, 331. (f) Dietz, E. R. Ph.D. Thesis, University of California, Berkeley, 1980. Dietz, E. R. Phys. Rev. A 1981, 23, 751. (g) deHeer, W. A. Bull. A m . Phys. Soc. 1983, 28, 285. (h) Rice, M. J.; Schneider, W. R.; Strassler, S . Phys. Rev. B 1981, 24, 554. (i) Richtsmeier, S.; Gole, J. L.; Dixon, D. A. Proc. Natl. Acad. Sci. U.S.A. 1980, 77, 561 I . (j) Schaeffer, 111, H. F. Arc. Chem. Res. 1977, 10, 287. (k) Bauschlicher, Jr., C.; Bagus, P.; Schaeffer, 111, H. F. I B M J . Res. Deu. 1978, 22, 213. (I) Companion, A. L.; Steible, D. J.; Starshak, A. J. J . Chem. Phys. 1968, 49, 3637. (m) Companion, A. L.; Steible, D. J.; Starshak, A. J. J . Chem. Phys. 1968,49,3637. (m) Companion, A. L. Chem. Phys. Lett. 1978,56,500. (n) Pickup, B. T. Proc. R . Sot. London, Ser. A 1973, 333, 69. (0)Gelb, A.; Jordan, K. D.; Silbey, R. Chem. Phys. 1975, 9, 175. (p) Davies, D. W.; Del Conde, G. Mol. Phys. 1977, 33, 1813. (4) Lindsay, D. M.; Herschbach, D. R.; Kwiram, A. L. Mol. Phys. 1980, 39, 529. (r) Bagus, P. S.; Del Conde, G.; Davies, D. W. Faraday Discuss., Chem. SOC.1977,62, 321. (s) Kendrick, J.; Hiller, I. H. Mod. Phys. 1977, 33, 635. (t) Gole, J. L.; Childs, R.; Dixon, D. A,; Eades, R. A. J . Chem. Phys. 1980.72,6368. (u) Fripiat, J. G.; Chow, K. T.; Boudart, M.; Diamond, J. R.; Johnson, K. H. J . Mol. Catal. 1975, 1, 59. (v) Anderson, A. B. J . Chem. Phys. 1977, 66, 5108. (w) Baetzold, R. C. J . Chem. Phys. 1971,55,4355. (x) Baetzold, R. C. Adu. Catal. 1976,25, 1. (y) Goddard, W. A.; Walch, S. P.; Rappe, A. K.; Upton, T. H.; Melius, C. F. J . Vac. Sot. Technol. 1977, 14, 416. (2) Bachman, C.; Demuynk, J.; Veillard, A. Gazz. Chim. Ital. 1978, 108, 398. (aa) Messner, R. P.; Knudsen, S. K.; Johnson, K. H.; Diamond, J. B.; Yank, C. Y. Phys. Reo. B 1976, 13, 1396. (bb) Messmer, R. P.; Caves, T. C.; Kao, C. M. Chem. Phys. Lett. 1982, 90, 296. (6) (a) Gerber, W. H. Ph.D. Thesis, Bern University, Switzerland, 1980. (b) Gerber, W. H.; Schumacher, E. J . Chem. Phys. 1978, 69, 1692. (c) Schumacher, E.; Gerber, W. H.; Harri, H. P.; Hofmann, M.; Scholl, E. Preparation, Electronic Spectra, and Ionization ofMeta1 Clusters, Gole, J . L., Stwalley, W. C., Eds.; American Chemical Society: Washington, DC, 1982; ACS Symposium Series No. 179, p 83, metal bonding and interactions with emphasis on the alkali metals. (d) Hermann, A,; Hofmann, M.; Leutwyler, S.; Schumacher, E.; Woste, L. Chem. Phys. Lett. 1979, 62, 216. (e) Delacretaz, G.; Woste, L. Surf. Sci. 1985, 156, 710. (f) Delacretaz, G.; Grant, E. R.; Whetton, R. L.; Woste, L.; Zwanziger, J. W. Phys. Rev. Lett. 1986, 56, 2598. (g) Thompson, T. C.; Trular, D. G.; Mead, C. A. J . Chem. Phys. 1985, 82, 2392.
Woodward et al. apparent that it will be difficult to probe resultant quantum levels even with extremely sensitive laser spectroscopic techniques. If intense cluster beam sources are developed, however, it may be possible to overcome some of these difficuItie~.~*~ These cluster sources must be designed to overcome the substantial loss mechanisms including predissociation and rapid intramolecular conversion and relaxation which deplete excited-state populations at a rate competing effectively with those spectroscopic probes that can be applied to the analysis of quantum level structure. We have been concerned with the development of intense sources producing cold metal clusters and with the characterization of cluster molecular electronic s t r u ~ t u r e , ~dynamic - ~ * ' ~ behavior,' and Studies of gas-phase metal cluster oxidation afford the opportunity to characterize the intermediate region bordered on one side by the gas-phase oxidation of metallic atoms and dimers and on the other by the surface oxidation of the bulk metallic phase. It has been suggested that these studies may provide information useful for the assessment of short- and long-range factors affecting surface ~ x i d a t i o n . ' The ~ products of metal cluster oxidation may be studied by using a combination of chemiluminescent (product formation in excited electronic states for highly exothermic oxidation) and laser fluorescent techniques, although observations of the internal mode structure associated especially with the polyatomic products of oxidation may be plagued by the rapid depletion of excited-state populations, either before the emission of a monitoring photon (chemiluminescence) can occur or before an appropriate, usually multiphoton, laser spectroscopic probe can be made operative (for example, quantum level probes by TPI spectro~copy'~).Here again, the development of intense metal cluster sources offers a viable means for overcoming these loss mechanism^.^.^^ Although the results may be complex, the potential value of such studies, as a means of evaluating the local environments which characterize surface oxidation, provides an impetus for the development of sources appropriate for the study of metal cluster oxidation products and their internal state distributions. Quantum level probes of the products of metal cluster oxidation have been carried out in two distinct configurations. In one configuration a stream of metal clusters formed through the "supersonic expansion" of the metallic element of interest is made to intersect a tenuous atmosphere of a given oxidant (beam gas (7) Crumley, W. H.; Hayden, J. S.; Gole, J. L. J . Chem. Phys. 1986, 84, 5250. (8) The impetus must be to overcome loss mechanisms by generating concentrations such that -lo5 to lo' photons can be emitted from the fluorescence zone. (9) Gole, J. L.; Green, G. J.; Pace, S. A,; Preuss, D. R. J . Chem. Phys. 1982, 76, 2247. Gole, J. L. "Laser Induced Photodissociation of Trimeric Sodium" In Proceedings ofthe International Conference on Lasers '81 (New Orleans, Louisiana), STS Press: McLean, VA; p 244. Hayden, J. S.; Woodward, R.; Gole, J. L. J . Phys. Chem. 1986, 90, 1799. (IO) Preuss, D. R.; Pace, S. A.; Gole, J. L. J . Chem. Phys. 1979, 71, 3553. (1 I ) (a) Crumley, W. H.; Gole, J. L.; Dixon, D. A. J . Chem. Phys. 1982, 76, 6439. (b) Hayden, J. S.; Woodward, R.; Crumley, W. H.; Gole, J. L., to be submitted. (12) Gole, J. L.; Woodward, R.; Hayden, J. S.; Dixon, D. A. J . Phvs. Chem. 1985, 89, 4905. Note that this piesent approach bears some resemblance to the liquid-nitrogen-entrainment-induced agglomeration of metal clusters used by Stein and co-workers and Solliard. These authors formed much larger aggregates which they studied using electron diffraction techniques. See for example, de Boer, B. G.; Stein, G. D. Surf. Sci. 1981, 106, 84. Solliard, C. Ph.D. Thesis, Ecole Polytechnique Federal de Lausanne, Switzerland, 1983. See also Sur5 Sci. 1981, 106, 58. J . Phys. 1977, CZ, 167. We wish to form smaller clusters in a more controlled manner. Further, the approach resembles that recently employed by Bondebey and co-workers to study small diatomic metal molecules. See for example: Gole, J. L.; English, J. H.; Bondybey, V. E. J . Phys. Chem. 1982, 86, 2560. Bondybey, V. E.; English, J. H. J . Chem. Phys. 1981, 74, 6978. Bondybey, V. E.; English, J. H. J . Chem. Phys. 1982, 76, 2165. Bondybey, V. E.; Schwartz, G. P.; English, J. H. J . Chem. Phys. 1983, 78, 11. Bondybey, V. E. J . Chem. Phys. 1982, 77, 3771. Bondybey, V. E.; English, J. H. Chem. Phys. Lett. 1983, 94, 443. ( 1 3) Gole, J. L. "Formation and Oxidation of Intense Metal Cluster Beams and Flows" In Proceedings ofrhe International Workshop on Ionized Cluster Beam Techniaues (ICBT) '86: D 8 5 . (14) Taylo;, T. N.; Campbe1l:C. T.; Rogers, Jr., J. W.; Ellis, W. P.; White, J. M.Surf. Sci. 1983, 134, 529. ( 1 5) See ref 6c, 6d, 6f, and Morse, M. D.; Hopkins, J. B.; Langridge Smith, P. R.; Smalley, R. E. J . Chem. Phys. 1983, 79, 5316.
Quantal Probes of Metal Cluster Oxidation
The Journal of Physical Chemistry, Vol. 91, No. 10, 1987 2639
/
\
lo
..... ....
Zrtube
Ta sheet
Srcloth
a
P\
LN2 tank
Side inlet 4
W basket heater
r Reserv ior
4
3
I -
-
Metal flux
3r Metal flux
b Figure 1. (a) Schematic of metal entrainment-agglomeration oxidation device showing tungsten basket heater, insulation, entrainment region, and oxidation region. (b) Outline of entraining gas cooling system.
configuration), the products of reaction being studied by a combination of chemiluminescent and laser fluorescent techniques. As a prototype the Nan + X (Cl, Br, I) reactions has been under study in our laboratory.” The Na3-X (Cl, Br, I) metatheses are found to produce Na2* (electronically excited products) with sharp bimodal excited-state vibrational distributions, the emission from these states resembling that characterizing optically pumped alkali dimer lasers.16 In this discussion, we will focus on a second and possibly more versatile source configuration which lies intermediate to a low-pressure molecular beam and high-pressure flow device.12 Clusters are formed from a high metal flux source and further agglomerated by an entraining argon or helium flow at room to liquid nitrogen temperature. Using this source, we have successfully obtained the first quantal information on the energy levels and optical signatures of several metal cluster oxides and select halides (M,O, M,X), the qualitative nature of which we now consider. Our focus is to be distinguished from recent very exciting studies in which small to intermediate size clusters have been generated in flow reacting with reagents in another continuous or pulsed flow stream under high pressure (- 300-500 Torr) conditions in a modified merged flow environment. The products in the flow have been measured mass spectrometrically; (16) Cobb, S.; Woodward, R.; Crumley, W. H.; Gole, J. L., work in progress. (1 7) (a) Geusic, M. E.; Morse, M. D.; Smalley, R. E. J. Chem. Phys. 1985, 82, 5901. (b) Richtsmeier, S. C.; Parks, E. K.; Liu, K.; Pobo, L. G.; Riley, S. J. J. Chem. Phys. 1985,82, 3659. (c) Parks, E. K.; Liu, K.; Richtsmeier, S. C.; Pobo, L. G.; Riley, S. J. J. Chem. Phys., in press. (d) Whetton, R. L.; Cox, D. M.; Trevor, D. J.; Kaldor, A. J. Chem. Phys. 1985, 89, 5666. (e) Trevor, D. J.; Whetton, R. L.; Cox, D. M.; Kaldor, A. J. Am. Chem. Soc. 1985, 207,519. (f) Rolfing, E. A.; Cox, D. M.; Kaldor, A. J. Chem. Phys. 1984,82, 3322. (g) Whetton, R. L.; Cox, D. M.; Trevor, D. J.; Kaldor, A. Phys. Rev. Lett. 1985, 54, 1494.
however, this technique has yet to provide a direct measure of structural or dynamic properties although important kinetic information has been extracted. Here, we focus on both spectroscopic information and the intriguing dynamic behavior which appears to characterize the oxidation reactions of small metal clusters. The purpose of the following discussion is to summarize (1) the experimental techniques for generating large concentrations of small metal clusters in a highly exothermic oxidizing environment and (2) information which we have garnered on the quantum levels of metal cluster oxide compounds, MxO,,. We suggest that the present studies outline the potential for chemiluminescent probes of metal cluster oxide quantum levels, not only within themselves but also as a means of suggesting future laser fluorescent probes of the metal cluster oxides. Experimental Considerations To approach the study of the internal mode structure of metal cluster oxides and halides, we extrapolate experience gained in the study of chemiluminescent metal atom oxidation reactions,’* focusing on the development of “high flux” continuous metal flow sources. We use this high flux continuous metal flow to create large concentrations of small metal clusters by forming an environment which is intermediate to that of a low “source” pressure effusive device, producing primarily atoms, a small percent of (1 8) See for example: Preuss, D. R.; Gole, J. L. J. Chem. Phys. 1977,66, 2994. Gole, J. L.; Preuss, D. R. J. Chem. Phys. 1977,66,3000. Dubots, L. H.; Gole, J. L. J. Chem. Phys. 1977, 66, 779. Lindsay, D. M.; Gole, J. L. J. Chem. Phys. 1977,66,3886. Gole, J. L.; Pace, S. A. J. Chem. Phys. 1980, 73, 836., Chalek, C. L.; Gole, J. L. J. Chem. Phys. 1976, 65, 2845. Gole, J. L.; Pace, S. A. J. Phys. Cheh. 1981,65,2651. Gole, J. L. Annu. Rev. Phys. Chem. 1976, 27, 525.
2640
Woodward et al.
The Journal of Physical Chemistry, Vol. 91 No. 10, 1987 I
diatomics, and, in some cases, small percentages of p~lyatomics,’~ and those conditions which prevail subsequent to the agglomeration of the metallic plasma formed in laser vaporization as it is entrained in a continuous or pulsed rare gas flow at high pressures. The latter technique, especially when operated in the pulsed supersonic expansion mode, produces a wide diversity of much larger clusters (vs. effusive source) although at small ( 107/cm3) concentration.20 In operating a metal source at temperatures or through containment designs such that the Knudsen number associated with the source is much less than one, we create the seed for the initial phases of a cluster-forming environment. The high metal flux, which within itself can lead to agglomeration to form clusters, is further agglomerated to small clusters by an entraining argon or helium flow at room to liquid nitrogen temperature. The agglomeration-entrainment device is depicted schematically in Figure la. Here, a tungsten basket heater (R. D. Mathis) is used to heat a metal in a particularly designed crucible to a temperature producing a vapor pressure between one and three orders of magnitude greater than that employed for effusive operation. The basket heater is wrapped in zirconia cloth (Zircar, Florida, NY) and surrounded concentrically by (1) a tantalum heat shield and (2) a cylindrical heavy-walled zirconium tube (Zircar). Both the top and bottom of the basket heater zone are heavily insulated with zitconia cloth. This extra insulation allows the ready operation of the tungsten basket heater at temperatures consistently at the upper limits of its performance specifications. The metal flux issuing from the lower crucible chamber is entrained in a rare gas (He or Ar, Airco 99.998%) flow ranging in pressure from 100 to 3000 mTorr, dependent upon the metal under study and the optimization of the chemiluminescent oxidation processes which are of interest, the agglomeration of the metal to form small clusters occurring both as a result of the high metal flux and as the metal flow is cooled by the variable-temperature entrainment gas. In order to vary the temperature of the entrainment gas which, as it enters the device depicted in Figure la, is brought to the desired operating condition, the entire upper assembly depicted in the figure, with which the gas is in intimate contact, is maintained at the desired temperature. This is accomplished with the systems depicted in Figure 1b. For cooling to temperatures approaching 196 K, methanol is continuously pumped through all cooling lines after passing through a dry ice slush bath. For cooling to lower temperatures, liquid nitrogen is allowed to flow at varying rates through these cooling lines. The choice of a cooling system is dictated by several parameters which will be considered in following discussion. At a suitable point above the flow, an oxidant intersects the entrained metal clusters, entering either from a concentric ring injector inlet as depicted in Figure l a or from a nozzle perpendicular to the flow and elevated above the cooled upper region of the oven assembly (Figure lb). Typical oxidant pressures ranged from 10 to 100 mTorr. Here, for metatheses which are sufficiently exothermic, a chemiluminescent flame may be formed. In the studies which we will outline here, concentrations of metal, carrier gas, and oxidant are adjusted in a controlled manner to maximize or minimize the intensities of various potential emN
(19) See for example references in Gole, J. L. ‘The Gas Phase Characterization of the Molecular Electronic Structure of Small Metal Clusters and Cluster Oxidation” In Metal Clusters, Moskovits, M., Ed.; Wiley: New York, 1986. Gole, J. L.; Stwalley, W. C. “Characterization of Alkali Metal Aggregation from Dimer to Bulk” In Advances in Atomic and Molecular Physics, in press. See also, for example (a) Hilpert, K.; Gingerich, K. A. Ber. Bunsen-Ges. Phys. Chem. 1980, 84, 739. (b) Gingerich, K. A.; Cocke, D. L.; Finkbeiner, H. C.; Chang, C. A. Chem. Phys. Lett. 1973, 18, 102. (c) Gingerich, K. A.; Cocke, D. L.; Choundary, U. V. Inorg. Chem. Acta 1975, 14, 147. (d) Neubert, A,; Ihle, H. R.; Gingerich, K. A. Presented at 6th International Conference on Thermodynamics, Merseburg, D.D.R., Aug 26-29, 1980. ( e ) Kingcade, J. E.; Dufner, D. C.; Gupta, S . K.; Gingerich, K. A. High Temp. Sci. 1978, 10, 213. (0 Gingerich, K. A,; Cocke, D. L.; Miller, F. J . Chem. Phys. 1976, 64,4027. (g) Kingcade, Jr., J. E.; Choudary, U. V.; Gingerich, K. A. Inorg. Chem. 1979, 18, 3094. (h) Drowart, J.; DeMaria, G. J. Chem. Phys. 1959,30, 308. (i) Kant, A,; Strauss, B. J Chem. Phys. 1966, 45, 822. (20) Liu, K.; Riley, S., private communication.
TABLE I: Parameters Associated with the Generation of Metal Cluster Flows metal (purity, source)
Ag (99.99, Cerac) Cu (99.9, Fisher) B (99.7, Alfa) Mn (99.95, Fisher)
metal source operating temp, K (vapor press., Torr) 1400-1700 -1650-1960 2520-2800 1350-1600
(10-’-5) (10-I-2) (IO-’-1) (IO-’-5)
effusive operation temp, K (press., Torr)
51400 51650 52520 51300
(IO-]) (IO-]) (lo-]) (98%) have been used in these studies. Ozone was generated and used as previously Spectra were taken with a 1-m Czerny-Turner scanning spectrometer operated in first order with a Bausch and Lomb 1200 groove/mm grating blazed at 5000 A. A dry ice cooled EM1 9880 photomultiplier tube was used to obtain the majority of the spectra depicted in the following discussion. In addition, a dry ice cooled RCA 4840 phototube was used. The photomultiplier signals were detected with a Keithley 417 fast picoammeter whose output signal (partially damped) drove a Leeds and Northrup stripchart recorder. All spectra were wavelength calibrated. ( 2 1 ) Lindsay, D. M.; Gole, J. L. J . Chem. Phys. 1977. 66, 3 8 8 6
The Journal of Physical Chemistry, Vol. 91, No. 10, 1987 2641
Quantal Probes of Metal Cluster Oxidation
,
7200
8
7100
-
7000
7
I
6200
6600
,
6100 16000
5900
5800
AA
I
6000
7500
I
I
7000
6500
6000
5500
5000
Figure 2. (a, bottom) Chemiluminescent spectrum resulting from the oxidation of a moderate copper flux (KKnudacn < 1) entrained in room temperature argon. The C u O A2Z+-X211, A’?Zt-X211, and b2Zt-X211 emission features are identified in the figure. Systems I-V do not correlate with CuO and are believed associated with metal cluster oxide emission systems. Spectral resolution is 8 A. (b, top) Expanded views as indicated. See text for discussion.
Results and Discussion Influencing Parameters. There are a number of factors in addition to metal flux, entrainment gas cooling rate, pumping speed, and oxidant configuration which determine the nature of the fluorescence spectra that can be generated by the techniques described herein. A process must be of sufficient exothermicity to populate the excited electronic states of the products of a given metathesis, be they polyatomic and/or diatomic emitters. A significant limiting factor on the success of the present experiments is the magnitude of the quantum yield for fluorescence once the excited electronic states of the polyatomic molecules on which we focus are populated. How large can a metal cluster be in order that we are able to observe its chemiluminescent oxidation to form a metal cluster oxide or halide? At some point, even with the requisite reaction exoergicity,22quantum yields will drop to zero as nonradiative processes strongly dominate the system. This will be influenced by transition moments, the strength of coupling to nonradiative channels, and the density of states associated with a given system, as it increases with atomic mass. We monitor the competition between radiative and nonradiative channels and through this effort also assess those systems which will, in the future, be amenable to laser fluorescent probes. In a given system, there will be a variety of conditions under which cluster distributions optimal for oxidation to form strongly emitting metal cluster oxides will be produced. As experimental conditions are varied, shifting cluster distributions, we proceed through optimal regions for the excitation of a given metal cluster oxide fluorescence spectrum. We thus engage in a search for those experimental conditions optimizing a given metal cluster oxide spectrum. Given sufficient exothermicity we must gradually shift the metal source cluster distributions through either a buildup of metal flux or increased (entrainment) cooling rate. In the following examples, we contrast results obtained upon the oxidation of effusive metal atom flows and those changes which occur as agglomeration ensues. (22) As product metal cluster oxide size increases, excited electronic states will be red shifted, decreasing exothermicity requirements. This may, however, be balanced by a decrease in the exothermicity of the oxidation process.
Cu, + 0,.Depicted in Figure 2 is the emission spectrum obtained when the output from a high-flux copper source (operating at KKnudscn < 1) is entrained in room temperature argon, this mixture being subsequently oxidized with ozone. The spectrum displays features which can be associated with several excited states of diatomic copper oxide (CuO) but is, in fact, dominated by a number of other band systems in the range 5200-5500 (I), 5700-6200 (11), 6900-7100 (111), and 6500-7800 A (IV). In addition, we find a moderately sharp emission feature (V) extending from 6300 to 6450 A. This is in contrast to the situation at lower copper fluxes where emission from the CuO A2Z+-X211, A”Zf-X211, and 62Z-X211 band systems23 is much more proemission features dominating the nounced, the A2Z+ and spectrum.24 Systems I, 11, and I11 (Figure 2) are characterized by what appear to be short progressions in low-frequency 120-1 50-cm-l modes or sequence structure. We favor the former interpretation for sequence frequency separations of this magnitude imply that extended progressions in the modes with which they are associated will be observed. These are apparently absent and systems I, 11, and I11 do not appear to extend over wide spectral regions. Based on the behavior which we will now exemplify, we assign systems I-V to metal cluster oxide emitters. The Cu + O3 CuO O2 metathesis is 1.76 f 0.05 eV ex other mi^.'^ This corresponds to 14 198 cm-I which does not represent a sufficient energy to populate the CuO A22+ state = 16492 ~ m - ’ voo(A-X211112) ;~~ = 16 213 (VW(A’Z+-X~II~,~) ~ m - ’ ) ’unless ~ the -2000 cm-I of additional energy necessary can be obtained from internal energy in O3and atomic copper (Emt(03) + E,,(Cu)) or the relative initial translational energy of the copper-ozone encounter.26 Similarly, the reaction of room tem-
-
-
+
(23) See, for example, Huber, K. p.; Herzberg, G. Molecular Spectra and Molecular Structure-Constants of Diatomic Molecules; Van NostrandReinhold: New York, 1979. (24) Temmen, M.; Woodward, R.; Devore, T. C.; Gole, J. L., work in progress. ( 2 5 ) Based on a CuO bond energy of 2.79 eV (Smoes, S.; Mandy, F.; Vander Auwera-Maheiu, A.; Drowart, J. Bull. Soc. Chim. Belg. 1972,81,45) and an O3bond energy of 1.03 eV (Rossini, F. D. Selected Values of Chemical Thermodynamic Properties; National Bureau of Standards: Washington, DC.).
2642
The Journal of Physical Chemistry, Vol. 91, No. 10, 1987
J
/-
/
jhl
I
\,
/
.-~
8000
~~-
70p0__--___._
-.
60G2
~
.-.. \^"2/
-
'
- - -#
.~~~~ 5000
Figure 3. (a-c) Chemiluminescent spectra resulting from the oxidation of a moderate copper flux (ICKnudssn < 1) entrained in liquid nitrogen cooled argon at various stages of cooling increasing from (a) to (c)' where liquid nitrogen cooling is at a maximum. In (a) the CuO A2Z+-X211, A'2Z+-X211, and 6*Z+-XZlI band systems are apparent although the A-X and A'-X band systems are muted relative to Figure 2. Systems I-V are assigned to Cu,O (x 2 2). Spectra b and c indicate that the CuO A-X and A'-X band systems are lost and the 6*Z+-X211 band system is muted relative to systems I-V. Spectral resolution is 14 A. See text for discussion.
perature copper dimer (Cu,) with ozone in a multicenter reaction producing CuO and CuO, is not expected to increase the reaction exoergicity substantially since the Cu, dissociation energy (2.03 eV),' and the CuO, bond strength are expected to be similar.28 The formation of the CuO A2Z+ and Af2Z+states must therefore result from a more complex process which, under the conditions of Figure 2, involves the reaction of vibrationally hot copper dimer or larger copper clusters. The Cu3 bond energy (D(Cu-Cu2) is substantially smaller than that of the dimer (D(Cu-Cu2) = 1.02 eV),'" and a multicentered process involving the trimer could yield the excited A2Z+ and A',Z+ states of CuO. In contrast to the A2Z+ and A%+ states, the CuO 6%+ state can be formed as a result of the Cu-O3 metathesis. The chemiluminescence from the Cu, + 0, metathesis has now been characterized across a wide pressure range extending from single collision conditions (beam-gas configuration) to the intermediate pressure range exemplified in Figures 2 and 3. The single collision results, which will be considered in more detail at a latter time, bear summary here. Under single collision conditions, the chemiluminescent spectrum resulting from the reaction of a copper beam with ozone displays a weak b u t clearly structured feature in t h e green,29a moderately structured and dominant feature in the region 5700-6400 A, and a weak emission feature at -7700 A. A ( 2 6 ) 2000 cm-' of excitation easily exceeds the sum E,,,(Cu)
+ E,d03)
+ ET'which one can readily associate with a copper beam at 1600 K inter-
acting with room temperature ozone in a nonthermalized beam-gas environment. (27) Hilbert, K. Ber. Bunsen-Ges. Phys. Chem. 1979,83, 161. (28) We estimate the Cu-02 bond strength to be 1.8-2 eV based on the bond strength for NaO, Na O2of 1.6-2.0 eV and the ionic character associated with CuO, (Bauschlicher, C. W., private communication). (29) Under certain circumstances, the copper-ozone system displays structured features in both the blue and green spectral regions extending from 4600 to 5600 A. The extension of spectral emission to wavelengths shorter than 5200 A is believed due to an electric discharge i n the region of the oven source.
-
+
Woodward et al. portion of the structured features in the green region of the spectrum extending from 5200 to 5600 8, appear to correspond to system T in Figure 2, whereas the 5700-6200-A feature appears to correlate with system 11. The 7700-A feature most certainly correlates with the emission system of CuO. The single collision chemiluminescent features all display a first-order ozone reactant dependence. However, while the temperature dependence of the 7700-h; feature indicates that it results from the reaction of a ground-state copper atom, the features observed at wavelengths shorter than 6200 8, appear to increase in intensity at a rate too fast for the chemiluminescence to depend solely on ground-state copper atoms. It appears that the emitter may be formed in the bimolecular reaction of metastable copper atoms which form a very small percentage of the beam or, more likely, vibrationally hot copper molecules. The single collision results correlate with and complement the multiple collision agglomeration studies. Under the conditions by which the spectra in Figure 2 were obtained, the copper beam is thermalized to a temperature between 500 and 700 K. Thus, the reaction of copper atoms at sufficiently high thermal energies to populate observed band systems at wavelengths shorter than 7000 8, seems unlikely. Metastable ,D copper atoms might be present under the conditions associated with Figure 2; however, the absence of band systems below 5200 A and the observation of some emission in the 4600-5200-A region29under certain single collision conditions casts doubt on this possibility. On the other hand, vibrationally hot copper dimers and higher copper n-mers are ideally suited to form copper oxide in a multicentered process with ozone to produce CuO and CuO, or higher order Cu,02 species. (Although the origin of a CuO, emission system can be found in the region of system 11,30it corresponds to a transition from the B state to a low-lying A state and is characterized by the dominance of a long progression in a Cu-0 stretch at -600 cm-l. Further, if the 5700-6200-A feature corresponds to Cu02, we find no indication of the CuOz B-X system at To 20 700 cm-' which should be observed.) If Cu,, vibrationally hot, reacts to form CuO A2Z+and A',Z+, a decrease in vibrational excitation should result in the quenching of the CuO A2Z+and A'2Z+-X211 features as subsequent studies with liquid-nitrogen-cooled argon entrainment demonstrate. Given that clusters are formed in high metal flux entrainment, cluster agglomeration should be enhanced as the temperature of the argon entrainment gas is decreased. Figure 3 demonstrates the effect of such a cooling process for a moderate and effectively constant copper flux. In Figure 3a-c, the liquid nitrogen flow and cooling rate are increased. In Figure 3a, the CuO A2Z+-X211 and A"Z+-X211 emission bands observed in Figure 2 are again monitored; however, the A2Z+ features are muted somewhat relative to A"Z+; the 6*2+-X211 system is readily observed. As the liquid nitrogen flow rate and cooling are increased (Figure 3b), a condition which should (1) enhance the cooling of vibrationally hot species in the copper beam and (2) increase agglomeration, the CuO A2Z+ and At2Z+ emission features virtually disappear while the S2Z+-X211emission system is still apparent. In contrast, systems I-V are enhanced. As the liquid nitrogen cooling is further increased, the 62Z+-X211and system I emissions are virtually quenched and the spectrum is dominated by systems 11-IV which have not been observed previously. The most significant change resulting from the transformation from single to multiple collision conditions observed in the Cu, + O3system is the continual growth of the 6500-7800-A emission region relative to other spectral features. In fact, these systems (111, IV) are rather muted under low copper flux multiple collision conditions and virtually absent under low copper flux single collision conditions. That the observed behavior is coincidental with increased agglomeration suggests that band systems 11-V and probably system 13' should be associated with the formation of excited states
-
(30) Bondybey, V. E.; English, J . H. J . Phys. Chem. 1984, 88, 2247. (31) It is noteworthy that the emission from system I at first increases in intensity with increasing liquid nitrogen flow (Figure 3) and then decreases. This would seem to indicate its formation via reaction with a small copper cluster that is subsequently lost as the distribution shifts to larger cluster sizes.
The Journal of Physical Chemistry, Vol. 91, No. 10, 1987 2643
Quantal Probes of Metal Cluster Oxidation
-----
,
4800
4400
4600
4200
4000
Figure 4. Chemiluminescent spectra resulting from the ozone oxidation of a moderate silver flux (KKnudacn < 1) entrained in room temperature argon showing emission corresponding to the A g o A211-X211 band system (4000-4300 A) and what is believed to be an A g 2 0 emission system (4400-4800 A). Spectral resolution is 5 8,. See text for discussion.
of the metal cluster oxide via metathesis with small copper agglomerates. Ag, 0,.We have previously reported the initial phases of a study of the silver-ozone system in the multiple collision agglomeration mode.', Here, we briefly summarize these results and outline the extension of initially reported efforts. In contrast to the chemiluminescence observed when a copper beam intersects a tenuous atmosphere of ozone in a beam-gas configuration, the corresponding silver-ozone interactions do not populate excited electronic states of the products of metathesis. This is not surprising for an effusive silver beam, containing predominantly silver atoms and a small concentration of silver dimer, can react with ozone in a process whose exoergicity is -1.25 eV for the reaction of ground state atoms (D00(AgO)32 - D(O-0,))25excluding the internal internal energies of ozone and silver and the relative initial translational energy, E;, of the reactants.33 The sum of all available energy is insufficient to produce the observed excited states of the metal monoxide. In contrast to atomic copper which has a low-lying metastable ,D state (2D5/2. 1 1 203 cm-I; 2D3/2, 13 245 cm-I), the lowest-lying metastable ,D state of the silver atom (2D5/2,30 242 cm-I; 2D3/2, 34714 cm-') will not be sufficiently populated at the temperatures associated with effusive beam conditions (Table I) and therefore need not be considered. Further, even the multicentered reaction of silver dimer with ozone, viz. Ag, + O3 A g o + A g o 2 (1) is also not sufficiently exothermic to form the excited states of A g o since the A g o z (Ag-02)34 bond strength is approximately equal to or slightly less than the Ag, (1.63 eV)35dissociation energy. Thus, based strictly on thermodynamic considerations, the interaction of an effusive silver beam with ozone produces no known electronically excited products. Figure 4 demonstrates results obtained as the silver-ozone system is taken in a controlled manner from single to multiple collision conditions and we operate in the metal agglomeration mode. The spectra in Figure 4 were obtained when a moderate silver flux (KKnudsen < 1) entrained in room temperature argon in the pressure range 15C-700 mTorr is oxidized with ozone. The observed emission corresponds to A g o in the region 4000-4200 A (some continuation (weak features) in 4200-4400-A range) and is believed to correspond to Ag20 in the region -4300-4800 A. The A g o bands correspond to (0,O) transitions originating from predissociating excited A211 spin-orbit components and terminating in ground-state X211 spin-orbit level^.^^^^^ The band at 4096 8, is assigned to the transition and that at 4125 8, is assigned to the 2111j2-2111j2 transition.
+
-
~~
~
(32) Based on an Ago bond strength of 2.29 eV. See ref 25. (33) Elnt(03)+ E,,,(Ag) + ET' will not exceed 0.15 eV even at the highest temperatures (1800 K) associated with an effusive oven system operating in a beam-gas configuration. (34) We estimate the bond strength of Ago2 Ag + O2as 1.5 eV based on the bond strength for N a 0 2 Na + O2 of 1.6-2.0 eV and the ionic character observed for CuO, (Bauschlicher, C. W. private communication). (35) Gingerich, K. A. J . Cryst. Growth 1971, 9, 31. (36) Uhler, U. Ark. Fys. 1953, 7, 125. (37) The identity of the ground state has been confirmed by ab initio calculations. See Bauschlicher, Jr., C. W.; Nelin, C. J.; Bagus, P. S . J . Chem. Phys. 1985, 82, 3265.
-
-
ym
15-
3-
-
~ _ _ ~ .
-~
7
,700
a00
9xDA
Figure 5. Chemiluminescent spectra resulting from the ozone oxidation of (a) a moderate silver flux (Khudxn < 1) entrained in room temperature argon and (b) a high silver flux entrained in room temperature argon. Both spectra consist of the A g o B211-X211 band system and certain silver atom emission features. At high silver flux the A g o emission system is considerably quenched relative to the atomic emission features and the A g 2 0 emission features in Figure 4. Spectral resolution is 5 A. See text for discussion. 4.0
(b)
mo
--
**-- --
-
____
&e
Figure 6. Portion of the chemiluminescent emission spectra from the
oxidation of high silver fluxes (KKnuLcn< 1) entrained in room temperature argon showing a closeup of the observed emission in the region 5500-7300 8, believed to correlate with the metal cluster oxides Ag,O, (x 2 2). The upper spectrum (b) is obtained at considerably higher silver flux vs. spectrum (a). Spectral resolution is 7 8,. See text for discussion.
Although the A g o excited electronic state cannot be formed from the reaction of silver atoms or dimers with ozone, the multicentered reaction of silver trimer Ag, + O3 A g 2 0 + A g o 2 (2)
-
can be sufficiently exothermic to populate excited states of A g o if the formation of Ag,O, from Ag2 and 0, releases 3.0 eV of energy. A large energy release is possible because of the low trim'er, Ag-Ag2 bond strengthIga (0.99 eV) and because of the possibility of some vibrational excitation in the trimer.38 The major consideration is that the formation of A g o A 2 n requires a metathesis involving silver clusters, Ag,, x I3.38 The spacings between the resolved vibrational peaks in the region 4400-4800 8, do not follow a regular progression increasing to a maximum of -260 cm-I. 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 Ag,O to what is thought to be a linear ground state.39
-
(38) The formation of A g 2 0 can occur directly from the reaction of the dimer with O,, Ag, 0, Ag20 + 02,however, even if we assume that a symmetrical species is formed and that the two Ag-0 bond strengths in symmetrical Ag,O are the same as that in Ag-O it is unlikely that the reaction will lead to the formation of Ag,-O in the excited electronic state from which emission is observed unless an unlikely substantial vibrational excitation in Ag2 (-1.0 eV) is transferred into electronic excitation. The reaction in the absence of vibrational excitation releases 1.92 eV of energy which is clearly not sufficient to form Ag20 with enough energy to luminesce in the observed range. The trimer can also react
+
-
Ag,
+ 0,
-
Ag2O
f
Ago2
(5)
to form Ag20 with enough internal energy to account for the observed chemiluminescence. (39) Based on comparison with Li20 or Na,O we would predict a linear ground state of D,,, 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.
2644
Woodward et al.
The Journal of Physical Chemistry, Vol. 91, No. 10, 1987 Bo I
&*NO2-
67W
1100
rmo
f,W
2700
Figure 7. (a), (b) Chemiluminescent spectra resulting from the multiple
collision oxidation of boron with NO2 showing the BO A2n-X2Z+ and BO2 A-X and B-X emission systems. (c) Chemiluminescent spectrum < 1) entrained resulting from the oxidation of a high boron flux (KKnudssn in dry ice cooled argon. The spectrum c is characterizedby the presence of a new emission feature at X > 5900 A correlated with the reaction of boron agglomerates B, ( x I 2). Spectral resolution is 8 A for (a), 5 A for (b), and 6 A for (c). See text for discussion.
Figure 8. Closeup of X > 5900 A emission region associated with the reaction of boron agglomerateswith NO2. The spectrum appears to be a combination of both sequential structure and a long progression or sequence structure corresponding to a higher frequency mode or a lowfrequency (- 142 cm-') vibrational mode. Spectral resolution is 5 A. See text for discussion.
As the silver flux is increased, the features emanating from the predissociating A211 state of A g o are quenched40 and features tentatively associated with the silver cluster oxides become more p r o n o ~ n c e d . ~Simultaneously, ~ a similar behavior characterizes the monitored chemiluminescence corresponding to the A g o 3500-8, band system (B211-X211) depicted in Figure 5 . In an ozone-rich en~ironment,~' the spectral features depicted in the 3200-3700-A region are dominated by the A g o emission feature. However, in a silver-rich environment, this emission feature is strongly quenched relative to both the Ag* emissions and those features which dominate the spectrum at longer wavelength. A further increase of the silver flux (the generation of higher-order clusters) leads to some quenching of the features observed in the 4400-4800-A region40 and an enhancement of spectral features in the range 5000-7500 A. Figure 6 depicts a portion of the spectra over the range 5500-7400 i% now obtained in the high silver flux configuration. While the spectrum in Figure 6a is complex in the region 5500-6340 A it clearly simplifies to a series o f bands separated by between 150 and 175 cm-'. The overall frequency separations in the range 5500-7000 i% are indicative of silver-silver stretch, bending, or torsional modes; however, more definitive statements will require considerable further evaluation under a wider variety of conditions leading to the generation of the observed spectral features. Finally, we note that as the silver flux is again increased (Figure 6b) further spectral complication arises with the definite emergence of a new redshifted feature at X > 6800 A. The results outlined here in conjunction with previous studies of the 5000-6000-A region suggest that the observed features result, at least in part, from AgxO, x 1 3, although considerable further experiments and analysis will be necessary to establish definitive correlations. As one might anticipate, the observed emission features appear to red shift as the size of the metal cluster oxide increases. B, + NO,, N 2 0 . The results obtained for copper and silver represent specific examples of a general trend. In Figure 7, we provide a further comparison of the significant changes which can be observed using the agglomeration technique in conjunction with the much higher temperature boron system. We have studied the B-N02 and B-N20 systems across a wide pressure range from single to multiple collision condition^.^^ Under both single and multiple collision conditions (B + NO2 + Ar), the boron-NO, and BO* reaction leads to the formation of BO2* (A%, BZZ+)
(A2n) excited electronic states as we exemplify in Figure 7, a and b. When an intense boron beam is agglomerated in dry ice cooled argon and subsequently reacts with NO2, the spectrum is found to consist largely of a modified "BO2"emission system43which overlaps BO emission features (BO A211-X22) emanating from v' = 0, 1, 2, A%, and what appears to be a new system at X > 5900 8, which begins to dominate the total chemiluminescent spectrum as the boron flux is increased substantially. The spectrum is apparent in Figure 7c taken under conditions such that first-order light at wavelengths below 3700 8, has been filtered (second-order features absent to 7400 8,). A closer view of the new emission system is presented in Figure 8. At least two AU = 40 cm-l sequence groupings separated by -440 cm-l and a second long progression (or sequence grouping) with Au 142 cm-l are observed and tentatively correlated with emission from the asymmetric BBO molecule,44 Mn, + 0,.In Figure 9 we summarize one further comparison which again emphasizes the significant changes induced using the agglomeration technique. We have studied the manganese-ozone reaction (Mn + O3 MnO 0,) across a wide pressure range from single to multiple collision condition^.^^ Under both single and multiple collision conditions (Mn + O3+ Ar, He), the Mn-0, reaction leads to the formation of the lowest-lying MnO* A7Z+ excited electronic state46and the observation of a strong chemiluminescent signal corresponding to the MnO A7Z-X7Z band
(40) See ref 12a also Woodward, R.; Le, P. N.; Gole, J. L. to be published. (41) The observed Ago emission features are found to be strongly dependent on the relative silver flux and ozone concentration (through a series of experiments monitoring the effect of changing ozone concentrations). Some contribution to the observed spectral features in the range 4200-4400 A may result from short progressions in the previously identified Ag2 A-X band system; however, a definitive statement will await further spectral analysis. (42) Ohllson, Bengt; Greene, Edward; Gole, J. L. Chem. Phys., to be submitted. See also Green, G. J.; Gole, J. L. Chem. Phys. Lett. 1980, 69, 45. Hanner, A. W.; Gole, J. L. J . Chem. Phys. 1980, 73, 5025.
(43) It is not clear that all of the features associated with the region of the BOz spectrum in Figure 1Oc can readily be associated with BO,;however, at significantly higher NO1 concentrations, there appears to be a convergence to a spectrum strongly dominated by the same features associated with BO2 emission in Figure 10, a and b. (44) Woodward, R.; Devore, T. C . ; Gole, J . L., work in progress. (45) Devore, T. C.; Burkholder, T.; Gole, J. L., work in progress. (46) Both experiment and theory indicate that the A'Z: excited state is the lowest-lying excited state of manganese oxide. Devore, T. C., private communication.
i
9225
_-
I L
ea1 5
Lu2 5
,
i
'.
1% I
/-, w w - -
-A I
---
a-- xa-----Figure 9. Chemiluminescent spectra resulting from the multiple collision ozone oxidation of manganese atoms to form MnO* and manganese molecules to form Mn,O* where x is most likely 2. Spectral resolution is 3 A. See text for discussion
-
-
+
J . Phys. Chem. 1987, 91, 2645-2649 system. When an intense manganese beam is agglomerated in liquid-nitrogen-cooled helium and subsequently interacts with 03, the intense MnO A-X emission system which extends from -500 to 700 nm is accompanied by a new system extending from -670 to at least 1000 nm which grows relative to the MnO A-X system with increased entrainment gas cooling. This new system, whose first-order spectrum is depicted in Figure 9, is believed to result from the interaction of manganese clusters and is tentatively associated with moderate progressions in two modes of the ground electronic state of MnzO ( M I I ~ + O - ) . ~ ~
Summary and Conclusions We have emphasized the development of a new source for the study of metal cluster oxidation and the analysis of metal cluster oxide internal mode structure. The results which we have outlined are very new and much work remains to be done to clarify the band systems and dynamics responsible for the observed metal cluster oxide emissions. In contrast to a preliminary mass spectrometric analysis of the silver system which is also characterized by a much more readily defined thermodynamics, only limited mass spectrometry has yet been performed on the copper, boron, and manganese systems to clarify precisely the cluster distributions characterizing these systems. In addition, the more complex group IVA (group 14)47cluster oxide systems are also under study. These studies are in their infancy; however, we (47) In this paper the periodic group notation (in parentheses) is in accord with recent actions by IUPAC and ACS nomenclature committees. A and B notation is eliminated because of wide confusion. Groups IA and IIA become groups 1 and 2. The d-transition elements comprise groups 3 through 12, and the p-block elements comprise groups 13 through 18. (Note that the former Roman number designation is preserved in the last digit of the new numbering: e.g., 111 3 and 13.)
-
2645
believe that we have implemented a general approach for the aggregation of atoms to small metallic agglomerates which can subsequently be oxidized to form the metal cluster oxides. These examples provide a flavor for the broad approach to the formation of the metal cluster oxides of a variety of metals ( M = Cu, Ag, Si, Ge, B, AI, early transition metals, ...) and the characterization of their quantum level structure. Further, we have now obtained C12, Br2 preliminary chemiluminescent data on the Ag,, Si, systems. Here, the Ag, Si CI,, Br2 reactions are dark (no chemiluminescence) whereas the high metal flux metal clusterhalogen molecule interactions are the source of emission. We intend to study this emission and, in correlation with an available technology for forming halogen atoms," will definitely extend our efforts to investigate the Ag,, Si, + C1, Br, I reaction system. It is hoped that these studies will provide information which will be useful for the assessment of short- and long-range factors affecting surface oxidation as well as aiding the characterization of those metal cluster oxides which play an important role in catalytic systems. Finally, we should note that the internal structure which we are mapping using chemiluminescence should subsequently be studied by laser vaporization techniques to generate and interrogate the metal cluster oxides.
+
+
Acknowledgment. It is a pleasure to acknowledge the National Science Foundation, the donors of the Petroleum Research Fund, administered by the American Chemical Society, and the Silver Institute for their partial support of the research effort described herein. Registry No. 03,10028-15-6; NO2, 10102-44-0; N,O, 10024-97-2; CUO, 1317-38-0; Ago, 1301-96-8; Ag20, 20667-12-3; BO,12505-77-0; BO2, 13840-88-5; MnO, 1344-43-0; BBO, 12045-60-2; Mn,O, 10647197-0; Ag,O, 80487-82-7; Ag, 7440-22-4; Mn, 7439-96-5; B, 7440-42-8;
CU,7440-50-8.
Size Dependence of the Structural Properties of Transitlon-Metal Aggregates from an Empirical Interatomic Potential Scheme Louis Marvillet and Wanda Andreoni* Institut de Physique ExpBrimentale, Ecole Polytechnique FBdPrale de Lausanne, PHB- Ecublens, 1015 Lausanne, Switzerland (Received: June 17, 1986)
We apply empirical interatomic potentials (Finnis-Sinclair potentials) to the calculation of the cohesive energy of transition-metal clusters of different sizes, in several structures. The evolution of the cohesive energy and of the structural parameters as a function of the cluster size is analyzed at T = 0. The validity of the Finnis-Sioclair potential for the structural determination of bulk solids and surfaces is also examined.
1. Introduction Structural information on metal clusters is still rather limited. Only a few characteristic size effects have been identified in noble and transition-metal aggregates, e.g. the contraction on average of the interatomic distances with decreasing number of atoms N and a tendency to prefer noncrystalline structures below a certain size. In particular, electron diffraction patterns reveal an "average" contraction of the lattice parameter with decreasing size for supported clusters of Pt and Au with face-centered-cubic structure' and extended-X-ray-absorption-fine-structure (EXAFS) measurements indicate shrinkage of the nearest-neighbor distances with decreasing size for a number of metal aggregates (Cu: Ni,2 Cr,3 Fe,3 and Ag3). Only some indication exists for the stability Present address: Physics Department, MIT-CMSE, Cambridge, MA 02139. * Present address: IBM-Ziirich Research Laboratories, CH-8803 Riischlikon, Switzerland.
0022-3654/87/2091-2645$01.50/0
of structures from that of the solid4in particles of Mo and W and for icosahedral-type of arrangements in gold cluster^.^ The theoretical study of structural properties of noble and transition-metal clusters has so far developed along two lines: (1) first-principle calculations for aggregates of molecular size6 ( N (1) Solliard, C.; Fliieli, M. Surf. Sci. 1985, 156, 487. (2) Apai, G.; Hamilton, J. F.; Stohr, J.; Thompson, A. Phys. Reu. Letr. 1979, 43, 165. (3) Montano, P. A.; Schulze, W.; Tesche, B.; Shenoy, G. K.; Morrison, T. I. Phys. Reo. B 1984,30,672. Montano, P. A,; Purdum, H.; Shenoy, G. K.; Morrison, T. U.; Schulze, W. Surf. Sci. 1985, 256, 228. (4) Iwama, S.; Hayakawa, K. Surf. Sci. 1985, 156, 85. (5) Solliard, C.; Buffat, Ph. J . Phys. 1977, C2, 167. Solliard, C. Ph.D. Thesis, Ecole Polytechnique Fedtrale de Lausanne, 1983 (unpublished). (6) Bernholc, J.; Holzwarth, N. A. W. Phys. Reu. Lett. 1983, 50, 1451. Delley, B.; Ellis, D. E.; Freeman, A. J.; Baerends, E. J.; Post, D. Phys. Reu. B. 1983, 27, 2132. Lee, K.; Callaway, J.; Kwong, K.; Tang, R.; Ziegler, A. Phys. Rev. B 1985.31, 1796. Andreoni, W.; Martins, J. L. Surf. Sci. 1985, 156, 635.
0 1987 American Chemical Society
