J. Phys. Chem. 1992, 96, 6131-6133 University of Jerusalem. The helpful discussions with Dr. J. R. Norris are highly acknowledged.
References and Notes (1) Wasielewski, M. R.;ONeil, M. P.; Lykke, K. R.; PeUin, M. J.; Gruen, D. M. J. Am. Chem. SOC.1991,113, 2114. Shinar, J.; Engel, J. P.; Barton, (2) Lane, P. A.; Swanson, L. S.; Ni, 0.-X.; T. J.; Jones, L. Phys. Rev. Lett. 1992,68, 887. (3) For a full description of EPR direct-detection, spectra accumulation, Levanon, line shape, and spin dynamics determination, see, e.g.: Gonen, 0.; H. J. Chem. Phys. 1986,85,4132. (4) Parker. D. H.: Wurz.P.: Chatteriee. K.: Lvkke. K. R.: Hunt. J. E.: Peliin, M. J.; Hemminger, J. C.; Gruen, D.M.; Stock, L. M. J. Am. Chem. SOC.1991, 113, 7499. ( 5 ) (a) Hare, J. R.;Kroto, H. W.; Taylor, R. Chem. Phys. Lett. 1991, 177, 394. (b) Ebbesen. T. W.; Taniaaki, K.; Kuroshima, S. Chem. Phys. Lett. 1991,181, 501. (c) Leach, S.; V&loet, M.; DesprL, A,; Briheret, E.; Hare, P. J.; Dennis, T. J.; Kroto, H. W.; Taylor, R.; Walton, D. R. M. Chem. Phys. 1992, 160, 451. (6) Dimitrijevic, N. M.; Kamat, P. V. J. Phys. Chem. 1992, 96, 4811. (7) Optical absorption spectra were not taken with LC matrices due to low transmittance of E-7in the nematic phase. (8) For triplet EPR detection in liquid crystals see, e.&: (a) Levanon, H. Rev. Chem. Intermed. 1987,8,287. (b) Regev, A,; Levanon, H.; Murai, T.; Sessler, J. L. J. Chem. Phys. 1990, 92, 4718. (c) Regev, A.; Galili, T.;
6131
Levanon, H. J. Chem. Phys. 1991, 95, 7907. (9) Line shape simulationsof triplet EPR spectra of Cm in LC and toluene indicate anisotropic rotational motion: (a) Gamliel, D.; Levanon, H. J. Chem. Phys., submitted for publication. (b) Regev, A.; Meiklyar, V.; Michaeli, S.; Levanon, H. To be published. (10) (a) Yannoni, C. S.; Johnson, R.D.; Meijer, G.;Bethune, D. S.; Salem, J. R. J. Phys. Chem. 1991,95,9. (b) Tycko, R.; Haddon, R.C.; Dabbagh, G.;Glarum, S. H.; Douglas, D. C.; Mujsce, A. M. J. Phys. Chem. 1991,95, 518. (c) Tycko, R.; Dabbagh, G.;Fleming, R. M.; Haddon, R. C.; Makhija, A. V.; Zahurak, S. M. Phys. Rev. Lett. 1991, 67, 1886. (11) Torrey, H. C. Phys. Rev. 1949, 76, 1059. (12) (a) Kim, S. S.; Tsay, F.-D.; Gupta, A. J. Phys. Chem. 1987,91,4851. (b) Fessmann, J.; Rhch, N.; Ohmes, E.; Kothe, G.Chem. Phys. Leu. 1988, 152, 491. (13) Alba, C.; Busse, L. E.; List, D. J.; Angell, C. A. J. Chem. Phys. 1990, 92, 617. (14) (a) Allemand, P.-M.; Srdanov, G.;Koch, A.; Khemani, K.; Wudl, F.; Rubin, Y.; Diederich, F.; Alvarez, M. M.; Anz, S. J.; Whetten, R. L. J. Am. Chem. SOC.1991, 113, 2780. (b) Krusic, P. J.; Wasserman, E.; Parkinson, B. A.; Malone, B.; Holler, E. R., Jr. J. Am. Chem. Soc. 1991, 113, 6274. (15) Argenhofer, A.; Toporowicz, M.; Bowman, M. K.; Norris, J. R.; Levanon, H. J. Phys. Chem. 1988,92, 7164. (16) (a) Hore, P. J.; McLauchlan, K. A. J. M a p . Res. 1979,36, 129. (b) Hore, P. J.; McLauchlan, K. A. Mol. Phys. 1981,42, 533. (c) McLauchlan, K. A. In Advanced EPR Application in Biology and Biochemistry; Hoff, A. J., Ed.; Elsevier: Amsterdam, 1989; Chapter 10.
Two New Electronic States of CH, Karl K. Irikura,? Russell D. Johnson 111, and Jeffrey W. Hudgens* Chemical Kinetics and Thermodynamics Division, Chemical Science and Technology Laboratory, National Institute of Standards and Technology,t Gaithersburg, Maryland 20899 (Received: April 24, 1992; In Final Form: June 8, 1992) The new states fi(3p) and I(4p) of methylene radical have been detected by mass-resolved, one-color 2 + 1 resonance-enhanced multiphoton ionization. Since it is a strong peak,the 3p state offers a convenient and sensitive means for detecting ground-state methylene radicals. In two photons, the 3p state lies at 31 1.80 nm (64 126 cm-I) in CH2, 31 1.84 nm in CHD, and 312.01 nm in CD2, and the 4p state lies at 269.27 nm (74 254 cm-') in CH2 and 269.36 nm in CD2.
Methylene radicals, CHI (g3BI),are often found in energetic chemical environments. For lack of a convenient detection technique, however, their roles in important processes such as combustion and diamond chemical vapor deposition remain undetermined. In a previous communication, we reported resonance-enhanced multiphoton ionization (REMPI) detection of triplet CH2 through previously known states.' Although this one-color 3 1 REMPI scheme is convenient, a 2 1 REMPI mechanism would be expected to provide superior sensitivity. We have found this to be the case and report the results herein. The apparatus and procedures used to record the one-color, mass-resolved REMPI spectra have been described in detail elsewheree2 The apparatus consists of a flow reactor, a timeof-flight mass spectrometer, and a computer data acquisition system. CH2 radicals are produced in the flow reactor by sequential reactions of fluorine atoms with methane under the conditions used previously to produce ground-state triplet CH2 (g3BI).1REMPI spectra of C, CF, CH, CH3, CHF, and CH2F are also otiser~ed,~ indicating that many chemical reactions occur in the F CH, system. CD2 and CHD radicals are generated from CD, and/or CD3H. Radicals are photoionized by the focused (focal length = 150 or 250 nm), linearly polarized, frequencydoubled output (energy = 1-5 mJ/pulse) of a tunable dye laser. Photoions are mass-analyzed by time-of-flight. The laser dyes (Exciton Chemical Co.)' used in this work are Sulforhodamine 640 and DCM (308-315 and 303-334 nm, pumped with Nd:
+
+
+
*To whom correspondence should be addressed. NIST/NRC postdoctoral associate. t Formerly called the National Bureau of Standards.
YAG) and Coumarin 540A (262-289 nm, pumped with XeCl). Spectra are not corrected for the variation in laser power that occurs over the range of each dye. Frequencies are corrected to vacuum and are calibrated to observed lines of atomic carbon.* The positions of the most prominent peaks are collected in Table I. Figure 1 shows the spectra camed by CH2+( m / z 14), CHD+ ( m / z 15), and CD2+( m / z 16) between 306 and 317 nma9 All evidence indicates that these spectra originate from REMPI of the corresponding neutral methylenes. The ion signal disappears in the absence of methane, or if the microwave discharge that generates fluorine atoms is extinguished. The isotopically labeled methanes yield similar spectra carried by ions of the appropriate masses. Finally, comparison of the spectra carried by the methylene ions with those carried by methyl and methine ions indicates that the ion signals in Figure 1 result neither from photodissociation nor from detector saturation. Since the strong band at 3 11-80nm and the much weaker band at 269.27 nm are only slightly shifted by deuterium substitution, they are a_ssigntd to the vibrational origins of the previously unknown H and I states. The adiabatic ionization energy of CH2 is IP, = 83 851 cm-'(10.396 eV).I0 The new peaks are therefore identified as two-photon transitions to 3p and 4p Rydberg states (6 = 0.64 and 0.62). For comparison, the 3p and 4p quantum defects are 6 = 0.62 and 0.61 for CH3I1and 6 = 0.66 and 0.63 for CH2F.5 The spectra of Figure 1 display no vibrational progressions. The theoretical literature on CH2+indi_catesthat no vibrational progressions should be expected. The-X 3B, ground state of CH2has a bond angle of 133.90,12and the X 2Al ground state of the CH2+ ion has a bond angle of 140.8O.I3 Since the bond angles are similar,
This article not subject to U S . Copyright. Published 1992 by the American Chemical Society
6132 The Journal of Physical Chemistry, Vol. 96, No. 15, 1992
Letters
TABLE I: Prominent Peaks in the fi(3p) Region of the Spectra of CH,, CHD, and CD, and in the t(4p) Region of the Spectra of CH, and CD,, As Observed by Mass-Resolved 2 1 REMPI"
+
CH2
CHD
(air), nm
state energy (vacuum), cm-'
314.79 311.80 311.43 3 10.45 309.45
63 517 64 126* 64 201 64 403 64612
269.79 269.27 265.83
74111 74 254' 75 214
ham
haw
(air), nm R(3p) 311.84 . 311.56 310.03
t(4p)
CD2
state energy (vacuum), cm-I 2 )By Bands 64117* 64 175 64 492
-
2 jB,
(air), nm
state energy (vacuum), cm-I
3 14.67 3 14.49 314.22 314.00 312.01 311.76 310.95 310.66
63 540 63 577 63 632 63 675 64 082* 64 133 64 301 64 360
272.50 269.86 269.36
73 374 74091 74 228.
haw
Bands
OA value marked with an asterisk is an assigned origin.
Two-Photon Energy (cm-') 65000
64000
6 3 0
REMPI signal obtained from CH3 at Ala, = 333.4 nm." In 1983, the discovery of a 2 + 1 REMPI detection scheme for CH3enabled other workers to measure CH3concentrations." The most demanding measurements have been accomplished in environments such as flames1' and diamond chemical vapor deposition reactors,'* where a mass spectrometer cannot be used to eliminate interferences from other species. Such CH3 measurements are successful because few molecules can contribute a REMPI signal at Alaser = 332.4 nm. We believe that the 2 1 REMPI scheme for CH2 (X 3BI) can support similar profile measurements. Most small molecules do not produce a REMPI signal at 31 1.80 nm. During our experiments with the F + CH4 reaction system, we confirmed that C, CF, CH, CH3, CHF, and CH2F radicals were present in the flow tube by observing their REMPI spectra. Of these, only the CH radical contributes even a weak signal at 3 11.80 nm. The REMPI spectra of other flame species such as O2(X32J19 and OHmindicate that these species will not interfere with CH2detection at 31 1.80 nm. Thus, we predict that this new detection scheme for CH2 (X 3B1)will find use in a wide variety of studies, including combustion and diamond chemical vapor deposition.
+
J
I I 1 l I o
306
308
1
1
1
"
310
'
"
'
1
312
4
'
1
i
'
'
314
316
Laser Wavelength (nm) Figure 1. REMPI spectra of 2 'B1 CH2, CHD, and CD2 ( m / z 14, 15, and 16) observed between 306 and 317 nm. The peak at 313.45 nm is due to atomic c a r b ~ n . ~
excitation of the bl electron in CH2 is not expected to lead to a References and Notes vibrational progression in u2. (1) Irikura, K. K.; Hudgens, J. W. J . Phys. Chem. 1992, 96, 518-519. The REMPI spectra in Figure 1 do not show well-resolved (2) Johnson, R. D., 111; Tsai, B. P.; Hudgens, J. W. J. Chem. Phys. 1988, rotational structure. The laser line width (Au = 0.8 cm-l) does 89, 4558-4563. (3) Johnson, R. D., 111; Hudgens, J. W. J. Phys. Chem. 1987, 91, not account for this result. The resolution did not improve as we 6 189-6 19 1. reduced the laser intensity to the threshold of practical signal (4) Chen, P.; Pallix, J. B.; Chupka, W. A,; Colson, S. D. J . Chem. Phys. detection. Rapid destruction of resonant Rydberg radicals through 1987, 86, 516-520. prediisociation or prompt photoionization could broaden rotational (5) Hudgens, J. W.; Dulcey, C. S.; Long, G. R.; Bogan, D. J. J . Chem. Phys. 1987.87.4546-4558. spectra.I4 (6) DiGiuseppe, T. G.; Hudgens, J. W.; Lin, M. C. J . Phys. Chem. 1982, The apparently ill-resolved REMPI spectra may simply reflect 86, 36-41. irregular rotational structure that arises from strong mixing among (7) Certain commercial materials and equipment are identified in this rotational, bending, and electronic motions. Methylene species paper in order to specify the experimental procedure. In no case does such identification imply recommendation or endorsement by the National Institute are nonrigid rotors, and the CH2+ (X 2Al) ion _core of each of Standards and Technology, nor does it imply that the material or equipment Rydberg state has a low-lying excited state (A 2Bl, T, = identified is necessarily the best available for the purpose. 1000-1200 cm-l) with which it is strongly mixed for K, # O.I3 (8) Moore, C. E.Atomic Energy Levels; National Standards Reference Simple simulations based upon energy levels in the l i t e r a t ~ r e I ~ J ~ . ~Data ~ Series, National Bureau of Standards 35; US.Government: Washington, D.C., 1971; Vol. 1. fail to reproduce even the general features of Figure 1. Further (9) The peak at 313.45 nm in Figure 1 is due to the intense (2p3p) IS progress in the analysis of this system awaits high-resolution (2p2) ID transition in atomic carbon.* This resonance is substantially powexperiments and more comprehensive ab initi? predictions. er-broadened under the conditions of Figure 1 . The spectral width (fwhm) We expect 2 1 REMPI through the new H(3p) state to be of this peak is -18 cm-I, as compared with 2-4 cm-l for other C peaks and 2-3 cm-' for rotational lines of the CF molecule. At low laser power, the an excellent detection method for triplet methylene. It is a simple, 313.45-nm C peak is neither saturated nor noticeably broadened. At the laser one-color diagnostic requiring only commonly available lasers. intensities necessary for good detection of CHI, the intense C+ signal severely Under our experimental conditions the total concentration of C1 saturates the photomultiplier of the Daly detector. Because the photomulspecies in th_e laser focus is lo9 6m-j. If all of thete CI species tiplier does not completely recover from saturation within the 760-11s interval between the C+ and CD2+arrival times, the C+ signal produces a vstigial peak were CH2(X 3B1),the detection sensitivity for CHI (X3Bl)_would in the m / r 16 REMPI spectrum of CD2. be lo8 radi~als.cm-~.(laser pulse)-'. Of course, CH2 (X 3B1) (10) Herzberg, G. Can. J . Phys. 1961, 39, 1511-1513. represents only a fraction of the F CHI reaction products. For ( 1 1 ) Hudgens, J. W.; DiGiuseppe, T. G.; Lin, M. C. J. Chem. Phys. 1983, example, the 2 + 1 REMPI signal that we obtained from CH2 79, 571-582. (12) Jensen, P.; Bunker, P. R. J . Chem. Phys. 1988.89, 1327-1332. at Xlaser = 311.80 nm was 5-10 times weaker than the 2 + 1
-
+
-
-
+
J. Phys. Chem. 1992,96,6133-6135 (13) Reuter, W.; Peyerimhoff, S. D. Chem. Phys. 1992, 160, 11-24. (14) Herzberg attributed the nonresolved rotational structure in the B (3d) Rydberg state of I2CH2to predissociation. See: Herzberg, G. Proc. R. SOC. (London) 1961, A262,291-317. (15) Carter, S.;Handy, N. C. Mol. Phys. 1984,52, 1367-1391. (16) Bunker, P. R.; Jensen, P.; Kraemer, W. P.; Beardsworth, R. J. Chem. Phys. 1986,85, 3724-3731.
6133
(17) Smyth, K. C.; Taylor, P. H. Chem. Phys. Lett. 1985, 122, 518-522. (18) Celii, F. G.; Butler, J. E. Annu. Reu. Phys. Chem. 1991,42,643-684. (19) (a) Johnson, R. D., 111; Long, G. R.; Hudgens, J. W . J. Chem. Phys. 1987,87, 1977-1981. (b) Johnson, R. D., 111; Long, G. R.; Hudgens, J. W. J. Chem. Phys. 1988,89, 3930. (20) Forster, R.; Hippler, H.; Hoyermann, K.;Rohde, G.; Harding, L. B. Chem. Phys. Lett. 1991, 183,465-470.
On the Way to Fullerenes: Molecular Dynamics Study of the Curling and Closure of Graphitic Ribbons D.H.Robertson,* D.W. Brenner, and C.T.White Theoretical Chemistry Section, Code 6179, Naval Research Laboratory, Washington,DC 20375-5000 (Received: May 4, 1992)
The short-time behavior of isolated graphitic ribbons is simulated at high temperature using a model hydrocarbon potential. These ribbons show large instantaneous deviations from planarity that often result in the formation of open-ended hollow carbon structures representing good fullerene precursors. While confirming the importance of pentagon formation in the production of these precursors, these results also point to the central role of relatively high temperatures in these processes.
Sparked by the synthetic breakthrough of Kratschmer, Huffman, and co-workers yielding macroscopic amounts of research on the fullerenes (discovered 5 years earlier3) has continued to accelerate. Despite the pace of this research, the mechanisms leading to the facile formation of highly symmetric fullerene cages in the chaotic conditions of the carbon arc remain to be ~larified."~In this Letter, following ideas advanced by Smalley, Curl, Kroto, and co-workers,eE we use molecular dynamics (MD) simulations to study the behavior of graphitic fragments thought to condense in the cooling carbon vapor prior to fullerene formation. These fragments are treated as ribbons as an idealization of their probable low-symmetry, far-from-circular shapes. At the conditions present in the carbon arc, we find that such ribbons can exhibit large thermal fluctuations from planarity without fragmentation. These fluctuations, when s u p plemented by the curvature induced by pentagons, often cause these isolated strips spontaneously to form open-ended hollow carbon structures. Once formed, these hollow structures represent good fullerene precursors. Graphitic ribbons are modeled using a bond-order type empirical hydrocarbon potentiallo which accurately predicts the bonding and energetics of solid diamond lattices and graphite sheets, as well as hydrocarbon molecules, while still allowing reactions to occur. The reactive aspect of the potential is important both to model correctly the reactive edges of these ribbons and to allow for any possible fragmentation at high temperature. Although this potential has already been shown to provide a good model for a wide variety of properties of fullerenes and related struct u r e ~ , to ~~ assess - ~ ~further its reliability for the present study we examined the energetics of the inversion of the corranulene molecule ( C a l 0 ) . This molecule is similar in shape to one-third of a buckminsterfullerenecluster. Recent studies have shown that corranulene can rapidly fluctuate between two curved minima at room temperature14-a behavior that is related to the ribbon motions discussed below. First principles calculations yield an energy difference between the minimum-energy curved structures and a planar intermediate of 8.8-1 1.0 kcal/mol.l5J6 Our empirical potential predicts a difference of 10.3 kcal/mol, in excellent agreement with these calculations. The MD simulations are performed by integrating Newton's equation of motion with an accurate high-order Nordsieck predictor-corrector method.l 7 The starting conditions are generated by assuming the velocities of the individual atoms in the relaxed
carbon ribbon are initially distributed with arbitrary directions in 3-D space according to a Boltzmann distribution. The dynamics of these isolated ribbons are then followed at constant energy to simulate their motion in the carbon vapor between collisions. Assuming the conditions in the carbon arc reported by Haufler et a1.8 (a temperature of 1000-2000 OC and a pressure between 100 and 200 Torr), the average time baween collisions of He, C, or Cm will be on the order of nanosmnds.I* The 250-ps length of our simulations is an order of magnitude less than this collisional time. The initially planar ribbons used in this study are depicted in Figure 1. These ribbons, with sizes ranging from 32 to 108 atoms, have varying lengths and widths. In addition, 1-fold coordinated atoms were included at the edges of most of these ribbons to allow for the dynamic formation of edge pentagon^.^ The effective temperature of an isolated ribbon (herein referred to as the temperature) is defined as the average kinetic energy per degree of freedom in units of temperature K, and u is the standard deviation in the average distance of the atoms to the fragment center of mass. u is large for planar ribbons, varies with changes in curvature, and approaches zero as the atoms become equidistant from the center of mass as in very symmetric fullerenes. At low temperatures all these ribbons remain close to planar, showing only small out-of-plane fluctuations. This is illustrated in Figure 2 for the 108-atom ribbon where the solid line is u for this ribbon plotted as a function of time for a trajectory started at 300 K. The observable fluctuations in the solid line correspond to only minor atomic displacements from planarity-less than 0.5 nm. No pentagonal rings form over the course of this trajectory because the temperature is too low to cause sufficient movement of any of the 1-fold coordinated edge atoms to allow the formation of another bond at the ribbon's edges. However, at higher temperatures pentagonal rings do form as illustrated in Figures 3 and 4. In Figure 3a, u is plotted for a 108-atom ribbon trajectory started at an initial temperature of 1150 K. During the first 15-20 ps, the internal temperature, plotted in Figure 3b, increases to near 2500 K because of the exothermic process of bond formation along the edges of the ribbon producing five-membered rings. During this time, the amplitude of the fluctuations from planarity (Figure 3a) increases dramatically from both this internal heating and the curvature induced by the edge pentagons. These large fluctuations are illustrated in Figure 4 where several snapshots of this trajectory are shown
0022-365419212096-6133$03.00/0 0 1992 American Chemical Society