Intracluster polymerization reactions within acetylene and

May 5, 1992 - 1968,48,4561. (22) Beltran-Lopez, V.; Castro-Tello, J. J.Magn. Reson. 1982, 47,19. (23) Hurst, G.; Kraft, K.; Schultz, R.; Kreiliek, R. ...
0 downloads 0 Views 762KB Size
J. Phys. Chem. 1992,96,9139-9144 (13) Chen, I.; Ablanvitz, M.; Sharp,J. H. J. Chem. Phys. 1%9,50,2237. (14) Weil, J.; Anderson, J. J. Chem. Phys. 1961, 35, 1410. (15) Iwaaaki, M. J. Mugn.Reson. 1974, 16,417. (16) Schweiger, A. ENDOR of TrunsitionMetal Complexes with eganjc Lfgunds;Springer-Verlag: Berlin, 1982. (17) Greiner, S. P.; Kreilick, R. W. J. Mugn.Reson., in press. (18) Greiner, S. P. Thesis, Department of Chemistry, University of Rochester, 1987. (19) Greiner, S. P.; Baumgarten, M. J. J . Mugn.Reson. 1989, 83, 630. (20) Kneubuhl, F. K. J . Chem. Phys. 1960, 33, 1074. (21) Abkowitz, M.; Chen, I.; Sharp, J. H. J. Chem. Phys. 1968,48,4561. (22) Beltran-Lopez, V.; Caatro-Tello, J. J. Magn. Reson. 1982, 47, 19.

9139

(23) Hurst, G.; Kraft, K.; Schultz, R.;Kreilick, R. W. J . Mugn.Reson. 1982, 49, 159. (24) Brown, T. G.; Hoffman, B. M. Mol. Phys. 1980, 39, 1073. (25) Manoharan, P. T.; Rogers, J. In Electron Spin Resonunce of Metul Complexes; Yen, T. F., Ed.;Plenum Press: New York, 1969. (26) Maki, A. H.; McGarvey, B. R.J. Chem. Phys. 1958, 29, 31. (27) Lau, P. W.; Lin, W. C. J. Inorg. Nucl. Chem. 1975, 37, 2389. (28) Shpolskii, E. V.; Klimova, L. A. Opt. Spectrosc. 1971, 22, 19. (29) Mobius, K.; Frohling, W.; Lendzian, F.; Lubitz, W.; Piato, M.; Winscom, C. J. J. Phys. Chem. 1982,86,4491. (30) Dinse, K. P.; Winscom, C. J. J. Lumin. 1979, 18-19, 500. (31) Dinse, K. P.; Winscom, C. J. J. Chem. Phys. 1978, 68, 1337.

Intracluster Polymerlzatlon Reactions within Acetylene and Methylacetylene Cluster Ions M. Todd Coolbaugb, Stephanie G. Wbitney, Gopalakrishaan Vaidyanathan, and James F. Garvey*l+ Department of Chemistry, Acheson Hall, State University of New York at Buffalo, Buffalo, New York 14214 (Received: May 5, 1992; In Final Form: July 16, 1992)

We report the observation of large (n up to -25) acetylene and methylacetylene cluster ions-the largest yet reported from ionization of neutral (C2H2J,or (CH3CCHJ,clusters. The cluster ion intensity distributions of both systems display prominent magic numbers at n = 3. This finding is indicative of intracluster ion-molecule reactions giving rise to what are most likely benzene and trimethylbenzene ions from (C2H2Jn+ and (CH3CCH),,+ cluster ions, respectively. The acetylene cluster ion intensity distributionsobserved under efficient clustering conditions are further characterized by unexpected features, most notably a sharp break at n = 7 and a strong magic number at n = 14. The mass spectra of methylacetylene display a less dramatic break at n = 7 and a weak magic number around n = 10. These surprising structures may arise as a result of the formation of particularly stable covalently bonded molecular ions formed via intracluster polymerization reactions.

I. ~troducti06 Gas-phase clusters present the experimentalist excellent op portunities to study processes of chemical importance in greatly simplified environments. One such process which has begun to attract rcccnt interest is ionic polymerization reactions. Clusters allow one to study ionic reactions in a solvated environment without complications due to the need to maintain electroneutrality, i.e. there is no need to consider effects due to neutralization reactions or counterions. It is also possible to investigate the chemistry of single ions. Although the use of a molecular beam cluster source results in a distribution of cluster sizes, essentially all of the reaction products observed in the cluster mass spectrum (CMS) arise from reactions of a single ion within the cluster. The study of polymerizationreactionsin clusters pennits one to gain insight into the importance of several factors in determining the course of reaction. Studying the reactivity patterns as a functionof cluster size has shown the crucial role solvent molecules play in the intracluster polymerization reactions by stabilizing the highly excited intermediates formed by the first few addition reactions. We have studied fhe positive ion chemistry of several olefin van der Waals clusters (ethene, 1, l-difluoroethene, and The small clusters: show ample evidence of the effects of solvent-induced changes in the ion chemistry, particularly the stabilization of highly energetic intermediates. We have also presented results indicating that sequential ion-molecule addition reactions (intracluster cationic %olymcrization~)takeplace within the cluster ions.'-' The findings of these studies are entirely consistent with previous bulk phase studies and suggest that clusters can provide insight into the initial stages of ionic polymerization. Several other groups are now also investigating intracluster polymerization reactions, both cationid and anionic! We have recently begun an investigation of the cluster ion chemistry of the alkynes acetylene (ethyne, ACE) and methyl'Alfred P. Sloan Foundation Fellow 1991-1993.

acetylene (propyne, MACE). Ionic polymerization reactions involving these molecules have been implicated in the formation of polyaromatic hydrocarbons (PAHs) in flames' and during pyrolysis of alkynes.* We will be reporting here the largest acetylene "clusterwions observed to date. The CMS of the alkyne clusters shows evidence of ionic intracluster polymerization reactions for both ACE and MACE. The CMS of both acetylene and methylacetylenedisplay magic numbers at n = 3, which may be indicative of the formation of benzene and trimethylbenzene ions, respectively. The CMS of acetylene also display several features at higher sizcs, the most intriguing of which is a magic number at n = 14. This feature may indicate the formation of a particularly stable, covalently bound ion.

II.

Experimental Section The experimental setup has been described in detail elsewhereg and is shown schematically in Figure 1. In brief, it consists of a continuous molecular beam cluster source of the Campargue design coupled to a chamber housing a mass spectrometer. The cluster beams were generated by expansion of gas through a 2%" supersonic nozzle. The nozzle assembly is equipped with a shroud through which fluid from a circulating chiller could be passed. The pertinent expansion conditions-pressure (Po) and temperature (To)-are reported in the figure captions. The acetylene (Scott, 99.696,dissolved in acetone) was expanded neat after passing through an activated charcoal fdter (Matheson 454) to remove as much acetone vapor as possible. This purification step is absolutely necessary in order to observe neat acetylene clusters since the acetone impurity in the beam can lead to the nearly exclusive observation of mixed acetylene/acetone cluster ions.'O Due to safety considerations, expansion pressures Po < 2 atm were utilized for all experiments. Overall clustering efficiencycould be increased by reducing the nozzle tempcraturq however, as Towas decreased, the relative amounts of acety-

0022-365419212096-9139$03.00/00 1992 American Chemical Society

-

9140 The Journal of Physical Chemistry, Vol. 96, No. 23, 1992 20

I

I

cm

I

-

I

1

3

a v >

c In

-

1 1

C

1 .

I

P)

c

-C C

-0

30

80

TABLE I: Acetylene CTW as a Fllllctioll of Ekctrolr hwct h m fragment 15 eV 20 eV 30 eV 70 eV mlz ion 6 11 24 C3+ 0 36 0 2 37 27 195 0 38 7 61 262 39 144 696 1552 2158 0 4 48 14 35 0 49 3 63 191 157 930 50 2469 3528 51 203 1572 4472 6541 1686 52 8326 14953 18308 70 421 53 1376 2492 1 14 55 54 96 1 65 8 54 99 0 73 0 0 25 74 0 0 13 188 75 0 0 31 146 26 76 131 307 412 77 141 700 1681 1878 78 1570 6937 9386 8012 108 79 474 686 634 80 5 19 30 35 2 7 24 35 91 0 1 101 11 20 102 2 15 62 198 25 103 135 337 320 104 110 501 642 590 8 105 41 74 74 2 106 9 20 31 2 4 117 14 28 0 127 2 25 26 4 128 27 109 111 129 1 5 31 35 73 130 323 418 350 131 10 41 53 50 2 10 17 21 132 3 154 18 37 29 155 1 10 11 4 42 210 156 247 207 7 27 36 157 31 1 7 12 158 11 180 3 15 12 11 181 0 0 3 5 18 95 123 182 98 183 3 15 17 15 184 1 4 8 7 2 12 686 206 634 1 4 2 207 4 17 76 208 85 75 3 14 17 17 209 210 2 2 12 7 ~

Quadrupole Ma98 eter

Figure 1. Schematic representation of the molecular beam cluster source and mass spectrometer.

n

Coolbaugh et al.

130 180 230 280 330 380

m/z Figure 2. A typical 70-eV acetylene cluster mass spectrum. The expansion conditions were Po = 1.8 atm and To= 259.5 K. The asterisks mark the positions of stoichiometric acetylene clusters, (C2H2Jnt.

lene/acetone or acetylene/water heteroclusters detected rapidly increased. For all experiments reported here the concentrations of acetone and water in the unclustered beam were measured to be much less than 1%. Mass assignments reveal that the impurity clusters detected when To > 0 OC consisted almost exclusively of acetylene/water clusters yet when To< 0 OC only acetone impurity clusters are observed. The methylacetylene (Applied Gas Technology, 10% in He) was used without further purifcation. Propadiene and isobutene were listed as being present in trace amounts. After being skimmed and collimated, the neutral cluster beam enters the ionizer of a mass spectrometer (Extrel C50; capable of unit mass resolution to 1500 amu) where a small fraction of the neutral species are ionized via electron impact (emission current 0.65 mA), mass filtered, and detected with a particle detector. The electron impact energy (E,) is reported in the figure captions.

III. Results Figure 2 displays a portion of a typical acetylene CMS obtained under expansion conditions which may be expected to lead to efficient clustering. As discussed above, even after purification, cluster ions containing acetone impurities still constitute a significant fraction of the total ion signal, particularly at higher masses. We have previously published a study of these mixed cluster i o d Oand we will therefore concentrate only on the neat acetylene clusters. The stoichiometric acetylene cluster ions, i.e. (C2H2In+,are marked with asterisks in Figure 2. These ions constitute the dominant series of neat acetylene cluster ions and may be observed in Figure 2 up to about n = 25, making these the largest acetylene cluster ions observed to date. It is possible to phenomenologically divide the acetylene CMS into two parts-a low mass part consisting of the clusters smaller than the pentamer and a high mass region consisting of the pentamer and larger clusters. These two groups of clusters may be distinguished by both the general appearance of the CMS and

~~~~

their behavior with respect to the energy of the ionizing electrons. Table I lists the intensities of a number of acetylene cluster ions observed in a moderate expansion (Po = 1.5 atm and To= 300 K) at several electron energies. It can be seen that the fragmentation pattern for the smaller clusters is highly sensitive to electron impact energy (Le., (CJ-I*,)+, x = 0-4and n = 2-5). However, for the larger clusters there is almost no change in the fragmentation patterns of the clusters. This change in the nature of the acetylene clusters for n L 5 has also been observed in other olefinic We note that the (CZnHbl)+ (n 1 5) ions are observed with very low intensities while the (CJ-Ik2)+ (n 1 5) ions are observed with slightly higher intensities. Unfortunately, t h w ions could also be due to (Ck3H&22%OHH20}+, i.e. clusters containing both a single acetone and water impunty. One particularly striking example is the "protonated" acetylene clusters, i.e. the (C,H,+I)+ ions. These ions are present in the E, = 30 and 70 eV CMS, but are essentially absent from the 20 and 15 eV CMS. These effects are now under investigation but can probably be attributed to formation and reaction of C2H+ species. Table I demonstrata that the distribution and/or C2HZ+* of the acetylene clusters with n 1 5 is only weakly dependent on electron impact energy. The magic numbers in the (C2H2In+ion intensity distributions discussed below are observed under all electron impact energies.

The Journal of Physical Chemistry, YO/. 96, No. 23, 1992 9141

Large Acetylene and Methylacetylene Cluster Ions

0

5

15

10

20

25

5

0

10

n

0

5

10

15

20

25

15

20

25

n

15

20

25

5

0

10

n

n

Figure 3. Plots of {C,H,J,+ ion intensitits vs n for several sets of expansion conditions. Data have been corrected for background noise and contributions to ion signals due to the "C isotope. All data were obtained with 70-eV electron energy. Expansion conditions are as foltows: (a) Po = 1.5 atm, To = 275.8 K (b) Po 1.3 atm, To= 268 K; (c) Po = 1.5 atm, To= 268 K (d) Po = 1.8 atm, To = 268 K.

Figure 3 displays plots of the intensities of the acetylene cluster ions vs cluster size, n, for several different sets of expansion conditions. It is clear that expansion conditions favoring the production of larger neutral clusters lead to the emergence of a very prominent magic number in the (ACE],+ ion intensity distributions at n = 3. As we will discuss below, this observation was not a great surprise and was suggested by the photoionization experiments of Shinohara et al.I1 The same cannot be said of the magic number at n = 14 which is apparent in Figure 3. A number of other reproducible features are also found in the high-pressure acetylene cluster ion intensity distributions with the most dramatic of these being the break at n = 7. Many of the CMS show slight maxima or breaks at n = 10 and 12, although these features are not nearly as prominent or reproducible as the features at n = 3 and 14. These results have been duplicated on a number of different occasions. Isobaric interferences due to either water or acetone or its fragments do not occur in the mass range under consideration. We conclude that these ions do not arise as a result of reactions of impurities within the clusters and the (C2H2),+ions are assigned as arising from reactionr in neat acetylene clusters. Our previous studies of acetylene/acetone clusters indicate that in acetylene/acetone heteroclusters the charge becomes trapped on an acetone molecule and leads to production of C,HloO+ cyclic ions.l0 We cannot rule out the possibility that dissociative charge transfer11J2t a b place in acetylene/acetone clusters (Le., such that IE[(ACE),J < IEIacetone]; IE is ionization energy). However, the weak energy dependence of the ion intensity distributions suggests that this effect would not make a significant difference in the observed spectra. We also note that the intensities of the (C,Hd,+ ions display a very strong i m r s e dependence on acetone concentration in the expansion, suggesting that they indeed arise from pure C2H2clusters. Finally, it has been observed that the intensity distribution of the (C2H2J,,+ ion exhibits the same behavior when water is observed as the only impurity. Considering the differing reactivities likely for water and acetone, it seems highly unlikely that the n = 14 magic number can be attributed to impurity reactions. Figure 4 displays a portion of a typical methylacetylene CMS taken under conditions of efficient clustering. Figure 5 displays plots of the intensities of the acetylene cluster ions vs cluster size, n, for several different sets of expansion conditions. In this case

70

1 120

170 220 270 320 370 420 m/z

Figure 4. A typical 50-eV methylacetylene cluster mass spectrum. Expansion conditions were Po = 1.5 atm and To = 259.6 K.

a magic number at n = 3 is observed. We were unable to observe a magic number at n = 14 in the methylacetylene CMS.

IV. Discussion (1) The n = 3 Magic Number Iona The main limitation of any single stage mass spectrometric experiment, such as we have employed in the present experiments, is the lack of direct structural information, i.e. knowledge of an ion's m/z may only uniquely determine the ion's empirical composition. Of course in many instances knowledge concerning the neutral precursor(s) may allow certain possible structures to be postulated while others are ruled out. In the case of a cluster mass spectrum a further complication arises since now there is often more of a question concerning the nature of the bonding forces within the cluster, i.e. covalent vs van der Waals or hydrogen bonding. Acetylene and methylacetylene clusters are thermodynamically unstable with respect to polymerization reactions. However, such reactions are unlikely because of large activation barriers. Such reactions may be expected to be quite facile in (ACEJWIACE+ and {MACEJWIMACE+ clusters since the activation barriers are much smaller for ion-molecule reactions, Le. an acetylene or methylacetylene ion "solvated" within such a cluster will certainly

9142 The Journal of Physical Chemistry, Vol. 96, No. 23, 1992

0

6

12

Coolbaugh et al.

0

18

6

n

12

I

1

I

I

d

X40

0

12

6

18

n

18

0

n

12

6

18

n

5. plots of (C,H,In+ ion intensities vs n for several sets of expansion conditions. Data have been corrected for background noise and contributions to ion signals due to the 13Cisotope. All data were obtained with 50-eV electron energy. Expansion conditions are as follows: (a) Po = 1.5 atm, To = 259.6 K (b) Po = 1.8 atm, To= 259.6 K (c) Po = 2.1 atm, To= 259.6 K (d) Po = 2.7 atm, To= 259.6 K.

react with one of the surrounding molecules. This ion product entirely from dissociative ionization of higher order clusters. The observation of the strong n = 3 magic number in the may also be capable of reacting with the surrounding molecules. One of the more interesting questions which arises in connection acetylene CMS alone does not provide conclusive evidence that with these clusters then concerns the extent of these the C6H6+ magic number ions possess the benzene structure. However, the apparent low reactivity of the observed C&+ ion,18 "polymerization" reactions within the clusters. We have previously studied the cluster mass spectra of several the fact that this structure can be formed without any rearrangements, and the fact that the benzene ion is the lowest energy olefinic molecules which were characterized by the emergence of prominent magic numbers under experimental conditions expected C6H6+ isomer would all seem to favor the benzene structure. to favor efficient ~lustering.'-~In the cases of ethene and 1,lApplying these same criteria to the methylacetylene system difluoroethene, fOr eXample, We O w e d that the G H ~ ~ + / C ~ H B F B +leads one to the conclusion that one should observe a magic (i.e. n = 4) ions were formed with highest probability while a number at n = 3 in this system corresponding to the production dramatic drop in ion intensity beyond the n = 5 ions was observed. of trimethylbenzene ions (eq 2 depicts the analogous gas-phase These results were rationalized in terms of the kinetic of the reaction). This expectation is born out in the experimental intracluster ionic addition ("polymerization") reactions.'-3 Mofindings. nomer molecules will add to the reactive ion in the cluster until C3H4+ 2C3H4 -m c-C~H~Z+ such time as an unreactive ion is formed-that is, until a kinetic AHO = -(7.87 - 8.09) eVI5J9 (2) bottleneck in the reaction sequence is reached. Following this, More direct support for this conclusion may be drawn from the the highly excited product ion formed within the cluster "boils" trimer and tetramer ion "fragment" regions of the methylacetylene off a large number of unreacted monomer molecules leading to CMS. As can be seen in Figure 4, the most abundant nonstoia high probability of observing the bare molecular ions and the chiometric or "fragment" cluster ions correspond to loss of CH3 emergence of the magic numbers in the CMS. radicals from the (MACE)3,4+ clusters. (MACE),+ also displays A similar line of reasoning may be appropriate in connection weaker peaks corresponding to loss of C2H, where x = 2-7, and to the alkyne clusters. In the case of acetylene, formation of losses of 1,3, and 5 H atoms. All of these fragments arc consistent benzene is a well-known "energy trap" in the polymerization with production of trimethylbenzene ions. sequence,14 and it is not surprising to observe a strong magic Table I1 compares MACE CMS data from a low-pressure number emerge at n = 3 in the CMS as larger neutral clusters expansion with the standard 70-eV MS of a number of C9HI2 are produced. A similar interpretation was offered by Shinohara molecules.2o Under these expansion conditions very little ion signal et al." The gas-phase analog of this termolecular reaction is shown is observed above the trimer ion and the trimer ion itself is smaller in reaction 1. than the ions o k e d in the trimer "fragment region" of the CMS, C2H2+ + 2C2H2 C'C6H6' AfP -8.23 eVI5 (1) i.e. 80 < m / z < 120. If we are correct in our understanding of Further support for this hypothesis may be derived from the the intracluster polymerizations, then the C9HIZ' ions are being work of Ng et al., who studied the single photon ionization of produced by dissociative ionization of quite small clusters, e.g. (ACE),.I6 The conclusion drawn from these studies was that the n 3. Under these conditions, the majority of the CgH12+ions apparently still retain enough internal energy to undergo fragacetylene trimer ions isomerized to the same structure(s) as benzene ions produced with the same amount of internal energy. mentation reactions with high probability. The observation of undissociated C6H6+ions in our experiments Because of the fact that the dimer ion and smaller ions may is a direct consequence of the production of large acetylene cluster also be directly produced by dissociative ionization, comparisons ions since the large exothermicity of the ionic addition reactions are only made in the region between the dimer and trimer ions, smuch be dissipated by the evaporative ions of monomers from i.e. m / z = 82-120. The intensities presented in Table I1 have been normalized to that of the most intense ion within the 80 < the cluster. This conclusion is consistent with the recent findings m / z < 120 mass range. Although a comparison such as presented of Booze and Bacr,l7who have shown that the C6H6+ ion signal observed following ionization of neutral acetylene clusters arises here is expected to be rather crude, the overall agreement between

+

-

-

The Journal of Physical Chemistry, Vol. 96, No. 23, 1992 9143

Large Acetylene and Methylacetylene Cluster Ions

TABLE II: Acetykac CMS C o m W to a Variety of c;H,,+Mrss Spectra fragment cluster 1,2,4-trimethyl- 1,3,5-trimethylm/z ion mass spectra benzene benzene 525 119 118 117 116 115 106 105 104

103 102 94 93 92 91 90 89

589 156 11 26 9 32 85 lo00 30 51 12 0 23 18 91 4 12 7.6 x 103

615 151 0 37 4 30 89 1000 30 57 12 0 3 23 87 3 0 1.3 x 104

the intensities of the trimethylbenzene ions and the MACE CMS is quite striking. These data suggest that the C9HI2+ions arising from dissociative ionization of small MACE clusters and trimethylbenzene ions at least pass through a common structure before dissociation. This in turn indicate8 that it should be pawible for C4H12+ions formed in larger MACE cluster ions to rearrange to the stable trimethylbenzene ion structure(s). (2) 'IbeHlgber Ma& N ~ d e r s .(i) Acetyleae. The observation of a relatively strong magic number for n = 14 in the acetylene CMS in intriguing. In view of the highly reactive nature of the cluster components it seems extemely improbable that this ion represents a solvated acetylene ion, i.e. (ACEJI3C2H2+. In view of the obvious propensity of C6H6+ions to form from acetylene clusters, it might be suggested that the (ACE)14+ion represents an (ACE)llC6H6+ion. This assignment is not particularly satisfactory in as much as it is not clear why eleven acetylenes solvating a central benzene ion would possess such extraordinary stability. Although the CMS alone does not provide sufficient information to allow us to come to any definitive conclusions regarding the natures of the (ACE),,3+ ions, it may be useful to consider the present results in light of the known gas-phase chemistry of acetylene. Unlike ethene, which undergoes only limited ionmolecule addition/polymerization reactions, acetylene has been found to undergo very facile addition reactions leading to polymeric products, giving products with n >> 7.21 In the results section it was pointed out that for n 2 5 very little ion signal is observed corresponding to ions of the type (ChHZrl)+. This result is consistent with the known sequential ion-molecule chemistry of acetylene. Brill and Eyler have pointed out that collisional stabilization of C 6 H 2 and C8Hn+ions is favored over H or H2 loss at rather low pressures.22 We therefore speculate that the (C2H2)14+ ion which is observed in our experiments is a covalently bonded species, i.e. C28H28+. At the m e time it would also suggest that the (ACE),+ ions wth 3 In 5 15 are likewise covalently bound ions. This conclusion is also supported by the behavior of ion intensity distributions shown in Figure 3 wherein the n = 3 and 14 magic number ions are s e n to arise simultaneously as expansion conditions increasingly favor the formation larger clusters. This observation would not be expected if the (ACEJI4+ion were a weakly bound complex consisting of acetylene molecules solvating a benzene ion. In this case the extremely large energy release accompanying the formation of the benzene ion would be expected to lead to a wide range of expansion conditions over which only the bare benzene ion would be observed. A magic number due to formation of a fully solvated benzene ion would be expected to arise from much larger clusters. All of the available evidence thus suggests that molecular ion. the (ACE)14+ion is a C2!&+ We can only conjecture as to the structure(s) of the n = 14 magic number ion and the origin of its enhanced stability. One

methyltoluene 316 37 6 23 5 3 89 1000 23 52 9 1 6 33 96 4 14 5.7 X lo4

pethyltoluene 289 30 5 19 5 20 85 1000 25

45 10 0 5 20 91 6 15 7.0 X lo4

isopropylbenzene

n-propylbenzene

1,l-nonadiyne

254

7 74 3 18 4 11 33 382 15 39 4 6 64 435 lo00 53 10 1.6 X lo6

240 1 0 5

8 2 4 2 9 85 lo00 23 60 11 0 1 7 52 2 8 9.8 X 104

1 11 1 39 4 10 3 0 3 111 lo00 8 11

1.8 X 106

possibility is that the n = 14 magic number may arise as a consequence of the geometry of the neutral clusters, i.e. the neutral clusters may take a form which does not allow more than 14 molecules in the cluster to react. This is improbable since the observation of stoichiometric clusters implies that the initially formed product ions are stabilized. Stabilization must Occur by transfer of energy into the cluster modes which leads to rapid heating and evaporative losses of monomers from the cluster ion-i.e. the cluster ions would retain their "memory" of the neutral cluster structure for only a very short time. In all likelihood, the enhancement of the n = 14 cluster ion, as well as the rather unusual shape of the ion intensity distribution for n > 6,must be attributed to the kinetics of the intracluster polymerization reactions. The acetylene cluster ion intensity distributions then indicate that the ionic polymerization reaction sequence in an acetylene cluster ion is very likely to be halted by the formation of a benzene ion (seeabove); Le. benzene formation represents an efficient kinetic bottleneck. Not all of the polymerization reactions are quenched at this point. In some finite number of clusters the ionic reactions are able to form ions of higher molecular weight. The Occurrence of a further magic number at n = 14 would then imply an additional kinetic bottleneck. High-pressure radiolytic polymerization of acetylene produces benzene and a solid polymeric product known as cuprene. The latter product is observed in the form of small, nearly perfect spheres, and several researchers have attempted to elucidate the mechanism for formation of this material.2f25It has been deduced that the formation of cuprene takes place in two steps. In the first step an oligomer is formed which contains approximately 20 carbon units. This oligomer then condenses into liquid drops in the gas phase. These droplets are then induced to undergo complete polymerization by further electron/particle bombardment. The failure of the polymerization to proceed to completion in the first step could not be explained, but a kinetic effect was suggested. In this context the observation of a pawible kinetic bottleneck in the intracluster polymerizations at CBHB+ is certainly intriguing. Our previous experiences with intracluster polymerization reactions suggests that this magic number may be associated with formation of an ion which is particularly stable toward further reaction with the monomer. In the case of the olefin clusters and the n = 3 magic number ions in acetylene and methylacetylene, the stability is apparently associated with formation of cyclic ions. Thus there are some grounds to postulate that the same is true of the C2&8+ ion. These ions, which might then be thought of as hydrogenated carbon clusters CZnHh+,may therefore possess very interesting structures, perhaps similar to the fulleranes. (1)Methylacetylene. Methylacetylene clusters do not display the same magic number at n = 14 as do the acetylene clusters. They do, however, show evidence of a break at n = 7 and a possible magic number at n = 10. Methylacetylene should polymerize as

J . Phys. Chem. 1992,96, 9144-9148

9144

efficiently as acetylene so it is very likely that the (MACE],' ( n > 3) clusters are probably also covalently bonded molecular ions. The reaaon for methylacetylene's differing reactivity pattern may be associated with the prtsence of the methyl group; similar large changes in the CMS were observed upon changing from ethene to propene. It is possible that the bulk of the methyl group decreases the probability of forming larger polymer ions within the clusters by blocking reactive sites or by hindering isomerization reactions leading to the formation of kinetically stable products.

V. Conclmions Ionization of acetylene and methylacetylene cluster ions initiates intracluster polymerization reactions. The observation of magic numbers at n = 3 under conditions of efficient clustering is best explained in terms of formation of benzene and trimethylbenzene ions in acetylene and methylacetylene cluster ions, respectively. The formation of benzene ions, which are stable, relatively inert, cyclic mol& ions, represents a fairly efficient kinetic bottleneck in the intracluster polymerization reactions. The observation of further anomalous features, particularly in the acetylene cluster ion intensity distributions, is indicative of more extensive polymerization reactions.

Acknowledgment. We gratefully acknowledge the financial support of this work provided by the Office of Naval Research and the Alfred P. Sloan Foundation.

References and Notes (1) Coolbaugh, M. T.; Peifer, W. R.;Garvey, J. F. Chem. Phys. Lett. 1990. 168, 337. (2) Garvey, J. F.; Peifer, W. R.;Coolbaugh, M.T. Acc. Chem. Res. 1991, 24, 48.

(3) Coolbaugh, M.T.; Vaidyanathan, G.; Peifer, W. R.;Garvey, J. F. J. Phys. Chem. 1991, 95,8337. (4) Strictly speaking, the reactions observed in cluster ions would be better described as oligomerizationreactions since in most cases they appear to give rise to low molecular weight products containing two to five monomer units. (5) El-Shall, M.S.;Marks, C. J. Phys. Chem. 1991,95,4932. El-Shall, M. S.; Schriver, K. E. J . Chem. Phys. 1991.95, 3001. (6) Tsukuda, T.; Kondow, T. J . Chem. Phys. 1991, 95,6989. ( 7 ) Gerhardt, Ph.; Homann, K. H. J. Phys. Chem. 1990, 94, 5381. (8) Boyle. J.; Pfefferle, L. J . Phys. Chem. 1990, 94, 3336. (9) Peifer, W. R.;Coolbaugh, M.T.; Garvey, J. F.J . Chem. Phys. 1989, 91. 6684. (10) Whitney, S. G.; Coolbaugh, M. T.; Vaidyanathan, G.; Garvey, J. F. J . Phys. Chem. 1991, 95,9625. (11) Shinohara, H.; Sato, H.; Washida, N. J . Phys. Chem. 1990,94.6718. (12) Vaidyanathan, G.; Coolbaugh, M. T.; Garvey, J. F.J. Phys. Chem. 1992. 96. 1589. (13) Garrett, A. W.; Zwier, T. S. J . Chem. Phys. 1992, 96, 7259. (14) See,for example: Allcock, H. R.;Lampe, F. W. Contemporary Polymer Chemistry; Prenticc-Hall: Englewood Cliffs, NJ, 1981. (15) Lias, S. G.; Bartmes, J. E.; Liebman, J. F.; Holmes, J. L.; Levin, R. D.; Mallard, W. G. J. Phys. Chem. Ref. Data 1988, 17 Suppl. 1 . (16) Ono, Y.; Ng, C. Y. J. Am. Chem. Soc. 1%2, 104, 4752. (17) Booze, J. A.; Baer, T. J . Chem. Phys. 1992,96, 5541. (18) Bohme, D. K.; Wlodek, S.; Zimmerman, J. A.; Eyler, J. R. Inr. J . Mass Spectrom. Ion Proc. 1991, 109, 31. (19) The overall exothermicity of (2) is dependent on the exact structure of the trimethylbenzene ion, the heats of formation being 187, 190, and 192 kcalemol-' for the 1,2,4-, 1.33-, and 1,2,3-trimahylbemne ions, respectively. (20) Atlas of Mass Spectral Data; Stenghagen, E., Abrahamson, S., McLafferty, F. W., Eds.; Interdencc: New York, 1969; Vol. 1, p 19. (21) Ion-Molecule Reactions in the Gas Phase: Adv. Chem. Ser. No. 58:' Auslods, P., Ed.; American Chemical Society: Washington, DC, 1968. (22) Brill, F. W.; Eyler, J. R. J . Phys. Chem. 1981, 85, 1091. (23) Briggs, J. P.; Back, R.A. Can. J. Chem. 1971, 49, 3789 and references therecn. (24) Willis, C.; Back, R.A.; Morris, R. H. Can. J. Chem. 1977,55,3288 and references therein. (25) Schmieder, R.W. Radiar. Res. 1984,99,20 and references therein.

Matrix Isolation EPR Study of the Reaction of Ag Atoms with PN, SiS, and GeOt James A. Howard,* Ruth Jones, John S. Tse, Mauro Tomietto, Steacie Institute for Molecular Sciences, National Research Council, Ottawa, Canada KIA OR9

Peter L. Ti"s,*

and Andrew J. Seeley

School of Chemistry, University of Bristol. Bristol, England BS8 ITS (Received: May 6, 1992; In Final Form: August 4, 1992)

Reactions of Io7Agatoms with PN, SiS, and GeO,isoelectronic analogues of CO, in inert hydrocarbon matrices in a rotating cryostat at 77 K have been investigated by electron spin resonance spectroscopy. PN gives one readily recognizable product, the mononuclear monoligand silver(0) complex Ag(PN). The EPR spectrum of this species at 180 K consists of a doublet of doublets of triplets centered at g = 1.9987 with alcn(l)= 1 1 16 MHz, a3l(l) = 183 MHz, and a,,(l) = 14.7 MHz, indicating there is about 61% unpaired 5s spin population on the Ag nucleus. SiS and GeO both give complex spectra with evidence for several mononuclear Ag(0) complexes. There is, however, reasonable evidence for formation of Ag(SiS) and Ag(Ge0). Spin-polarized local density function calculations indicate that Ag adds to the P nucleus of PN, that the most stable geometry is bent with a Ag-P-N angle of 97.4', and that AgGeO and AgSiS have triangular structures. Calculated unpaired spin populations were found to be in good agreement with those obtained experimentally from the EPR spectra.

I.troduction From matrix isolation spectrogcopic (EPR and FTIR) studies it is known that silver atoms complex with CO to give Ag(CO), where x = 1-314, and with S i 0 to give Ag(Si0) and other monosilver complexes of S i 0 oligomers.s Complexation of silver Two atoms by PN has been studied by matrix IR ~pectrosoopy.6~~ species have been identifiad, AgpN and another identified initially as Ag2(p-PN)t but subsequently as Ag(P2N2).7 We now report Issued as NRCC No. 34225.

0022-3654/92/2096-9144$03.00/0

a detailed EPR spectroscopic study of the interaction of Io7Ag atoms with PN molecules in an adamantane matrix at 77 K in a rotating cryostats.9 and with SiS and GeO. This work has enabled the trends in bonding in a series of Ag atom-mono isoelectronic ligand species to be observed. Experimental Section The rotating cryostat technique has been thoroughly described el~ewhere.~*~ PN was prepared by decomposition of P3N5in a molyMcnum pouch at ca.1170 K (the P3N5was prepared by rapid

Published 1992 by the American Chemical Society