for u = 1 of HCN.
-
0) making it possible to obtain the IRCL spectrum
Discussion and Conclusions The H + ICN reaction primarily leads to HCN(O,m,n) + I, but there is a minor channel producing H N C + I. The H N C is
directly formed from the reaction 1b, rather than by isomerization. cm3 The rate constant for H C N formation is (2.0 f 0.4) X C'a t room temperature. The HCN(uI,u2,u3)distribution offers direct evidence into the dynamics of the H + ICN reaction. The highly inverted distribution is in strong contrast with the H X distributions obtained in unimolecular elimination or displacement which implies that the reaction does not proceed through a bound HICN intermediate. At first glance this may seem surprising because the triply bonded C N offers a position where the H could add. In the H + ClCN case the calculations found HClCN to be bound by only 6 kcal mol-'. We believe that HICN would be bound by even less than 6 kcal mol-', or probably it is not bound a t all. Comparing the H C N distribution from H + ICN, BrCN, and ClCN would be interesting. If there is a systematic change in the stability of the intermediates, the dynamics of these reactions would be affected and lead to different HCN distributions. Our preliminary results from H + BrCN seem to suggest a lower , indicating a greater role for a bound intermediate. The partitioning between u3 and v2 could offer additional insight into the dynamics. Unfortunately, our interpretations are strongly affected by our assumption that v2 relaxation should be faster than u3. Nevertheless, substantial bending excitation is required to fit the spectra. It is likely that H either directly attacks the C or it adds to the I end and then migrates to C to form H C N + I. In both cases H will more or less add to the C perpendicularly and transform to the linear H C N molecule, as H C N and I separate. This can clearly produce bending mode excitation as has ~~
4193
J. Phys. Chem. 1991, 95,4193-4195
~
~~
(16) Arunan, E.; Wategaonkar, S.J.; Setser, D. W. J. fhys. Chem. 1991, 95, 1539.
been observed. Since the C N probably moves as one unit relative to the light H atom, significant excitation of the C N stretch is very unlikely. The vibrational distribution and observed in migratory reactions and direct displacement reactions from H + BrF and IF are worth comparing to the present data. The for all these reactions are about the same, ranging from 0.5 to 0.7, but the observed distribution and rotational excitation differ significantly for the two microscopic pathways. The migratory channel results in broader rotational distribution, whereas the displacement channel has less rotational excitation. With HCN, under our experimental conditions it is not possible to infer anything about rotational distribution, but migration may affect the bending mode excitation. As we learn more about HCN(uI,u2,u3)state to state relaxation processes,l' these interpretations should be possible. The IRCL of H C N offers considerable opportunity to study the dynamics of reactions yielding HCN as a product. With higher resolution experiments, it will be possible to obtain better population assignments. The v3 emission not only provides information similar to that from H X emission for inferring details about the dynamics but also reveals new information from excitation in the bending mode. For example, the reaction of H atom with tertbutyroisonitrile18 gives only (O,O,1) (O,O,O) emission without any bending excitation. In this case the H may approach the C end and displace butyl to form H C N via a bound intermediate. Another possibility would be to study the H C N from abstraction of H from a series of molecules by CN. The IRCL of H C N should develop into an important area of chemical reaction dynamics to complement the currently active SEP studies."*
-
Acknowledgment. This work was supported by the National Science Foundation, Grant C H E 8822060. We thank Dr.Glenn Lo for providing the XYCALC program which was used to convolute the simulated spectra. (17) Smith, I. W. M.; Warr, J. F. J . Chem. SOC.,Faraday. Trans. 1991,
87,205. (18) Wategaonkar, S.J.; Setser, D. W. Unpublished results.
Observatlon of Magic Numbers within Ar, H20+ Heteroclusters Copal Vaidyanathan, M. Todd Coolbaugh, William R. Peifer, and James F. Garvey* Department of Chemistry, Acheson Hall, State Vniuersity of New York at Buffalo, Buffalo, New York 14214 (Receiued: March 12, 1991)
We present in this communication the observation of magic numbers for the heterocluster ion Ar,,H20+ for n = 3, 9, and 15 which we attribute to the stability of the cluster ion. On comparison with the calculations of Bohmer and Peyerimhoff for ArJZe', we feel this data suggests that, in the case of ArnH20+,the water cation sits at the center of the argon cluster, representing the first example of an oxygen-containing molecule exhibiting such behavior. However, while Bohmer and Peyerimhoff predict a magic number for Ar12Xe+,we do not observe one for Ar12H20+.We postulate that this difference may be due to the water cation distorting the structure of the solvating argons. The study of van der Waals clusters has aroused a great deal of interest in recent years'" due to the unique position that clusters hold between the gas and condensed phases of matters6*' Theoretical and experimental studies of these isolated clusters have provided considerable insights8s9into changes in their physical and ( 1 ) Levy, D. H. Adv. Chem. fhys. 1981, 47, 323. (2) Ng, C. Y. Adv. Chem. fhys. 1983, 52, 152. (3) Miirk, T. D.; Castleman Jr., A. W. Ado. At. Mol. Phys. 1985,20,65. (4) Jortner, J. Ber. Bunsen-Ges. fhys. Chem. 1984,88, 188. ( 5 ) Atomic and Molecular Beam Methods; Scoles, G., Ed.; Oxford University Press: New York, 1988. ( 6 ) Physics and Chemistry of Small Clusters;NATO AS1 Series B Jena, P., Rao, B. K., Khanna, S. N., Eds.; Plenum: New York. 1987; p 158. (7) Cartlrman, Jr.. A. W.; Miirk, T. D. Gaseous fon Chemistry and Mars Spctrometry; Futrell. J. H., Ed.; Wiley Interscience: New York, 1981.
0022-3654/91/2095-4193$02.50/0
chemical properties as a function of cluster size. Mass spectrometric studies of gas-phase clusters, in particular, have proved extremely useful through the observation of magic numbers (i.e., anomalous enhancements of cluster ion signal for a specific cluster size). The observation of a magic number typically indicates a particularly stable cluster ion stoichiometry, from which the cluster ion geometry can often be inferred. In the case of molecular systems this stability is typically attributed to a stable, closed solvent shell, hydrogen bonded directly to a central cation.'*'4 In recent years the study of magic numbers (8) Hagena, 0. F.; Obert, W. J . Chem. fhys. 1972,56, 1793. (9) S.D. Stein Memorial Issue. J. Phys. Chem. 1987, 91. (10) Stace, A. J.; Moore, C. J. J . Phys. Chem. 1982, 86, 3681.
0 1991 American Chemical Society
4194 The Journal of Physical Chemistry, Vol. 95, No. 11, 1991
Letters
80 eV
n
. . . . . . . . . . . . . . . . . . . .
OJ
I
9
5
7
9
I1
13
I6
17
19
21
n
0
,
I
, . , 3
6
,
, . , . , . , . ,.
*
. , .
7
I7
19
9
11
19
15
n
Figure 1. Plot of normalized ion intensities as a function of n, for the cluster ion ArnH20+,as a function of electron energy (Po = 3.8 atm). Magic numbers are noted by numbering of individual data points. Note at lower electron energies where monomer evaporation is expected to occur to a smaller extent, the magic number structure becomes more pronounced.
has also been used to help uncover new ion-molecule chemistry within van der Waals c l ~ s t e r s . ’ ~ J ~ The theoretical and experimental study of magic numbers has also been of considerable interest in rare-gas cluster^.^^-^^ The first experimental evidence for the existence of such magic numbers was reported by Recknagel and co-workers.17 The stability and the structure of such atomic and molecular clusters has generated considerable debate centering on whether the intensity fluctuations found in the mass spectra reflect the size distributions in the neutral or the ionic clusters. It has now been established that these magic numbers are due to the cluster ion stability and it has been (1 1) Echt, 0.;Morgan, S.;Dao, P. D.; Stanley, R. J.; Castleman, Jr., A. W.Ber. Bunsen-Ges. Phys. Chem. 1984, 88, 217. (12) Shinohara, H.; Nishi, N.; Washida, N. J . Chem. Phys. 1985, 83,
1939. (13) Lifshitz, Ch.; Louage, F. Chem. Phys. Lerr. 1989, 93, 5633. (14) Peifer, W. R.; Coolbaugh, M. T.; Garvey, J. F. J. Chem. Phys. 1989, -91. ., 6684. - - - ..
(15) Coolbaugh, M. T.; Pcifer, W. R.; Garvey, J. F . Chem. Phys. Leu. 1990. 168. 337. (16) Garvey, J. F.;Peifer, W.R.;Coolbaugh, M. T. Acc. Chem. Res. 1991, 24, 48. (17) Echt, 0.; Sattler, K.; Recknagel, E. Phys. Reo.Leu. 1981, 47, 1121. (18) Stephens, P. W.;King, J. G. Phys. Rev. Lerr. 1984, 53,2390. (19) Ding, A.; Hesslich. J. Chem. Phys. Leu. 1983, 94, 54. (20) Harris, 1. A.; Kidwell, R. S.;Northby, J. A. Phys. Reu. Leu. 1984, 53, 2390. 121) Searcv. J. 0 Fern. J. B. J . Chem. Phvs. 1974. 61. 5282. (22) Castlckan jr;, A. W.; Kay, B. D.; HeAan, V.;Holland, P. M.; Mark, T. D. Sur/. Sei. 1981, 106, 179. (23) Stace, A. J.; Moore, C. Chem. Phys. Loit. 1983, 96, 80. (24) Nagashima, U.; Shinohara, H.; Nishi, N.; Tanaka, H. J . Chem. Phys. 1986,84, 209. ( 2 5 ) Kuntz. P. J.: Valldorf. J. 2.Phvs. D 1988. 8. 195. (26) Haberland, H. Sur/. Sci. 1985,-156, 305. (27) Echt, 0.;Kreislc, D.; Knapp, M.; Recknagel, E. Chem. Phys. Len. 1984. 108, 401. (28) H a r e , M. R.Adu. Chem. Phys. 1979, 40, 49.
concluded that the positive charge is localized on either a highly vibrationally excited dimerZ5or a trimer ion29*30 in the center of the cluster. The cluster ion can then undergo a size reduction by the successive evaporation of single atoms as long as sufficient internal energy is available. This sequential decay of metastable argon cluster ions has been observed by Schierer and Miirk” and it has also been shown by Stace and MooreZ3and Echt et alaZ7 that the ion abundance anomalies in the mass spectra of (H20), and Ar, are due to delayed evaporation of monomers following electron impact ionization. The theory that fragmentation processes also play an important role in the structure of the mass spectra is supported by the experimental studies of Buck and Me~er.~~ There have also been a number of experimental and theoretical studies on mixed clusters of the type Ar,M, where M is either an atomic or a molecular ion solvated by n argon atoms. Extensive studies of S t a ~ have e ~ ~shown that the nature of the nucleating species dictates the symmetry of the stable cluster and hence the magic numbers. For example, Ar,I+ displays magic numbers a t n = 12 and 18, whereas Ar,12+ has a local maximum only at n = 17.27 These results were explained in terms of an icosahedral packing with the charge being either on the I atom or on the I2 molecule and the argon atoms being arranged around this charge center. The study of a number of other Ar,M+ systems, where M is an oxygen-containing polyatomic molecule or a fragment ion, also exhibited magic numbers at n = 12 and 18. On the basis of the fragmentation pattern of these heterocluster ions, Stace suggested that the dopant ion “sits on” rather than “within the (29) Bohmer, H. U.;Peyerimhoff, S.D. Z . Phys. D 1986, 4, 195. (30) Hcsslich, J.; Kuntz, P. J. Z . Phys. D 1981, 2, 251. (31) Schicrer, P.; Mark, T. D. Phys. Rev. Lerr. 1987, 59, 1813. (32) Buck, U.; Mcyer, H. J. Chem. Phys. 1986,84,4854. (33) Stace,A. J. J. Phys. Chem. 1983,87,2286; J. Am. Chem. Soc. 1984, 106,4380; Chem. Phys. Lett. 1985,113,355; J . Am. Chem. Soc. 1985,107, 155; J . Phys. Chem. 1987, 91, 1509.
The Journal of Physical Chemistry, Vol. 95, NO. 11, 1991 4195
Letters
1 as a function of electron impact energy. A similar plot is shown in Figure 2 for different expansion pressures. Both of these plots show prominent magic numbers” for n = 9 and weaker ones for n = 3, 5 , 7, and 15. A Monte-Cqrlo simulation of ArJe+ clusters by Bohmer and PeyerimhofP8predicted that clusters with n = 9,12, and 18 should show pronounced stability, with n = I5 showing moderate stability. The results of these calculations are in general a reement with the photoionization results of Ding and -workers, who observed prominent magic numbers at n = 7, 12, and 18. The structures predicted by Bohmer and PeyerimhofP8 for the Ar12Xe+and Ar,&e+ cluster ions are an icosahedron and a double icosahedron, respectively. The structure proposed for ArI5Xe+is intermediate between an icosahedron and a double icosahedron where the three extra atoms are not only bound to the inner shell but among themselves as a tetrahedron. The special stability of the Ar9Xe+ cluster ion is due to a structure where all the argon atoms are at the same optimum distance from the central Xe+ ion as well as from each other (1.8 A). The reason for this special stability is that nine argon atoms is the maximum number that can be placed around a xenon atom without any of the bonds being stretched or shrunk. That is, all the atoms occupy the most favorable position with respect to the Xe+. We feel the comparison between Ar,,Xe+ and Ar,,H20+ is quite interesting since both the Xe atom and H 2 0 molecule have ionization potentials lower than that of Ar (by 3.63 and 3.14 eV, respectively). As a result, in both cases the charge is localized on the dopant molecule. On this basis we feel the stability of Ar9H20+is due to nine argons equivalently solvating a central water cation. We note that the major difference between our results for Ar,H20+ and the theoreticaP and e~perimental)~ results for ArJe+ is our inability to observe a magic number for Ar12H20+. A simple rationale for this difference could be the fact that the presence of the water cation in some way modifies the arrangement of argon atoms around it. Indeed neutron diffraction patterns of pure and H20-doped vapor-deposited rare-gas solids have been reported to be significantly different. These changes have been attributed to a change from a face-center cubic (fcc) structure for pure rare gas crystals to a disrorted hexagonal close-packing (hcp) structure when doped with water.”’ For pure Ar films the fcc and hcp structures differ only in the stacking sequence with the fcc being slightly energetically favored. One of the important factors involved in the transformation observed in water-doped solid Ar is believed to be the nature of the interaction between the water and the rare-gas environment.4I Such interactions could also be of considerable importance for small water-doped argon clusters as well. In conclusion, the similarity in magic numbers between ArJe+ and Ar,H20+ cluster ions is indicative that the water cation sits at the center of the argon cluster. However, the variation from the expected magic number sequence in the case of Ar12H20+ suggests the possibility that the first shell may be distorted from the expected icosahedron by the nature of the ion-neutral interactions. This system we feel is an excellent candidate for theoretical as well as further experimental studies. Acknowledgment. We gratefully acknowledge the financial support of this work provided by the Office of Naval Research.
d
-= I
S
S
7
0
I1
IS
IS
IT
IO
S
S
7
0
11
IS
IS
I7
10
Imo 900
cm
400
I
n Figure 2. Plot of normalized ion intensities as a function of n, for the cluster ion Ar,,H,O+, as a function of expansion pressure (electron energy = 25 eV). Magic number are noted by numbering of individual data points.
cluster”.33 This result has also been observed by Scoles and c o - w o r k e r ~ . ~In~ this letter we present evidence for an oxygencontaining molecule ( H 2 0 ) which appears to sit at the center of an Ar,H20+ heterocluster. Our molecular beam apparatus consists of a Campargue-type beam source.3s The setup for generating mixed continuous expansions using a split-flow pattern has been described earlier.36 Argon (Irish, 99.98%) was passed through a gas purifier before being bubbled through a reservoir containing water (nanopure) subjected to repeated freeze-pump-thaw cycles to remove dissolved gases. The mass spectrometer is an Extrel C-50 instrument capable of unit mass resolution up to m / z = 1400 and the electron emission current for these experiments was kept at 0.65 mA. The plot of the intensity of the various Ar,H20+ ions, normalized with respect to the most intense peak, is shown in Figure (34) Gough, T. E.; Mingel, M.; Rountree, P. A,; Scoles, G . J . Chem. Soc.
1985, 83, 4958.
(35) Campargue, R.; Lebehot, A. Inr. Symp. Rurefied Gus Dynum., 9th 1914, 11, 1. Campargue, R. J . fhys. Chem. 1984,88,4466. (36) Vaidyanathan, G.; Coolbaugh, M. T.; Peifer, W. R.; Garvey, J. F. J . Chcm. fhys. 1991, 94, 1850.
(37) While Ar9H20+ is isobaric with (H20)21+we do not believe any of the ion intensity at m / r = 378 amu is due to a (H20)21+contaminant since we do not observe (HzO),Ht for n > 5 and typically the protonated signal is always much greater then the unprotonated signal (Le., no signal is observed for m / z = 319). (38) Bohmer, H.U.; Peyerimhoff, S. D. 2.fhys. D 1988, 8, 91. (39) Holub-Krappe, E.; Gantefor, G.; Broker, G.; Ding, A. 2.fhys. D 1988, 10, 319. (40) Langel, W.; Schuller, W.; Knozinger, E.; Fleger, H. W.; Fleger, H. J. J . Chem. Phys. 1988,89, 1741. (41) Knhinger. E.; Schuller, W.; Langel, W. Furuduy Discuss. Chem. Soc. 1988.86, 285. Cohen, S.S.;Klein, M. L. In Inerr Guses; Klein, M. L., Ed.; Springer-Verlag: Berlin, 1984; p 87.