Spectrum of the Tropyl Radical BY B. A. THRUSH AND J. J. ZWOLENIK* Dept. of Physical Chemistry, University of Cambridge Received 28 th December, 1962 The spectrum of the tropyl radical (C7H7) has been observed in the flash photolysis of ditropyl and of ditropyl sulphide. This spectrum consists of a Rydberg series in the ordinary ultra-violet converging to an ionization potential of 6237 eV. It is concluded that the ground state of the tropyl radical and the excited states observed have planar symmetrical structures like that of the C7HT ion. A value of AHl(C7H:) = 209h3 kcal/mole is deduced, which is about 13 kcal/mole less than that for the isomeric benzyl ion. Hiickel’s simple molecular orbital theory 1 applied to the n: electron system of monocyclic polyenes of the type (CH), established the (4n+2) rule, which pre- dicts high relative stability for those monocyclic, coplanar systems of trigonally hybridized atoms containing (4n + 2) n: electrons, where n is an integer.The stability of C&6, C5Hy and C7HT have all been explained as the case where n = 1 and the six electrons completely fill the three available bonding orbitals leaving the anti- bonding orbitals vacant. The cyclopentadienyl and the tropyl radicals are not expected to possess the stability of the corresponding carbanion and the carbonium ion, but flash photolysis in the vapour phase should allow production and identifi- cation of these radicals. Theoretical calculations of the spectra of both the cyclopentadienyl and the tropyl radicals have been made by Longuet-Higgins and McEwen.2 The absorption spectrum attributed to the cyclopentadienyl radical, produced by flash photolysis of cyclopentadiene and ferrocene under isothermal conditions, has been reported 3 and found to be in good agreement with the theoretical predictions.Recent mass- spectrometric work 4 has yielded ionization potentials for both radicals. For the tropyl radical, produced in the pyrolysis of ditropyl, a low ionization potential of 6.60+0.1 eV was found. This low value agrees well with the ionization potential of 6.41 eV calculated by Streitwieser.5 Since the unpaired seventh n: electron in the tropyl radical occupies an antibonding orbital, transitions involving the ex- citation and the removal of that electron should occur without a large change in molecular configuration.The loss of this electron should occur readily to yield the more highly stabilized tropylium cation. In this work we have observed the absorption spectrum of the tropyl radical from two parent molecules, ditropyl (bis-7-cycloheptatrienyl) and ditropyl sulphide (bis-7-cycloheptatrienyl sulphide) which were selected because they both show strong absorption in the ultra-violet where the photolytic energy absorbed is more than the 30-40 kcal/mole needed to rupture the appropriate bond for tropyl radical formation.4~ 6 EXPERIMENTAL A conventional flash photolysis apparatus with a quartz reaction vessel 50 cm long and 1.8 cm diam. was employed with a 2000 J photolytic flash of 30 psec duration.7 The * National Science Foundation Postdoctoral Fellow 1960-62.196B . A . THRUSH AND J . J . ZWOLENIK 197 duration of the 100 J, horizontal, end-on spectroscopic flash 8 was 7 psec. Hilger E. 498, E. 517 and E. 484 quartz spectrographs were used depending on the wavelength region studied. The absorption spectra were recorded photographically using Ilford Selochrome plates, sensitized if required with Kodak U.V. sensitizer no. 8269. A standard Cu arc was used for wavelength calibration. Since an adequate vapour pressure of the parent substances could only be obtained at elevated temperatures, the cylindrical reflector surrounding the reaction vessel and photolytic flash lamp was heated electrically. This apparatus and the tube connecting the reaction vessel to the reservoir of the parent substance were maintained at about 10°C warmer than the reservoir.Ditropyl vapour from a bulb at 52°C was admitted to the re- action vessel where it was diluted with excess inert gas. The ditropyl sulphide (which has a lower vapour pressure and is less stable than ditropyl) was placed in a U-tube which was maintained at 70°C and the inert gas was passed slowly over it into the reaction vessel. Optical density measurements using Emax = 6-7 x 103 at 2600 A showed that ditropyl has a vapour pressure of 0.1 mm Hg at 52"C, in good agreement with the previously reported vapour pressure of 0.1 mm Hg at 50T.4 The pressure of ditropyl sulphide used in this work was estimated to be about 0.05 mm Hg, assuming a similar extinction coefficient. In all cases 300-400 mm Hg of " oxygen-free " nitrogen was used as inert gas. The samples of ditropyl and of ditropyl sulphide used were generously provided by Prof.Hyp J. Dauben, Jr., of the University of Washington, Seattle. RESULTS AND DISCUSSION When 0.1 mm Hg of ditropyl in excess nitrogen is flashed photolyzed, three transient absorption bands are observed. These bands appear and decay together and therefore belong to the same species. They are listed in table 1 and a low dispersion spectrum taken on a Hilger E.484 spectrograph shown in fig. 1. Their maximum intensity is reached in about 40 psec and they can still be detected faintly after 200 psec. TABLE 1 n obs. calc. difference approx. width of band (cm-1) (cm-1) (crn-1) (cm-1) 3 150 38,500 38,501 - 1 4 30 43,654 43,626 + 28 5 30 45,991 46,O 1 9 - 28 The two longer wavelength bands (2600 A and 2292 A) have also been identified in the flash-photolysis of ditropyl sulphide in excess nitrogen ; the 2173 A band could not be identified with certainty in this case owing to strong continuous ab- sorption in this region.The only other transient absorption observed in any of these experiments was due to CS, the (0,O) band of which at 2575.6 appeared very weakly in the experiments with ditropyl sulphide. No other bands of CS or of Sz could be detected. On the basis of experiments on the flash photolysis of com- parable amounts of CS;! in excess nitrogen, it was concluded that CS was a very minor product of the photolysis of ditropyl sulphide; it may arise from the decom- position of a C7H7S radical formed in the primary act.The positions of the bands listed in table 1 agree well with a Rydberg series of the form where iz = 3, 4, and 5. This series converges to the very low ionization potential of 6.237+_0-01 eV, and is believed to be the first molecular Rydberg series observed outside the vacuum v (cm-1) = 50329 - R/(n + 0-046)2,198 SPECTRUM OF THE TROPYL RADICAL ultra-violet. Further members of the series could not be observed due to greatly increased absorption by the parent molecule at shorter wavelengths and to the decrease in strength which occurs with the higher members of a Rydberg series. The only species with such a low ionization potential which would be expected to be formed from both these parent molecules is the tropyl (cycloheptatrienyl) radical, and there can be little doubt that this spectrum is that of the tropyl radical.Further, the ionization potential agrees well with that calculated by Streitwieser 5 (6.41 eV) and satisfactorily with the mass spectrometric value of 6.6fO-1 eV deter- mined by Harrison, Honnen, Dauben and Lossing4 when it is remembered that such determinations commonly give values significantly greater than do spectro- scopic 9 and photo-ionization 10 methods. No vibrational structure was detected in any of the electronic transitions. The band at 2600 A was diffuse with a width of about 150 cm-1. The other two bands were much sharper, being ca. 30cm-1 wide. Owing to strong absorption by the parent molecule in these experiments, it would not have been possible to detect vibrational structure of the n = 4 and n = 5 transitions which had an intensity of less than one-tenth or one-fifth of that of the main band respectively. Nevertheless, the absence of strong vibrational structure shows clearly that the tropyl radical and the Rydberg states observed have closely similar structures, confirming the view that the most weakly bound electron in the radical is largely antibonding in character and has little effect on the structure of the radical.The tropyl ion has been shown to have a planar symmetrical D7h strucfure,lI and since the configuration of the Rydberg states must approach that of the positive ion, the lack of vibrational structure and the sharpness of the spectra show that the ground state of the tropyl radical and the Rydberg states observed have structures which are very close to symmetrical and planar.Since the unpaired electron occupies a degenerate orbital in the ground state of the tropyl radical, this state and any degenerate upper states of this radical would be expected to show Jahn-Teller distortion.12.13 The narrowness of the spectra observed indicates that such dis- tortions are small in this case, as might be expected from the non-bonding character of the degenerate orbital. A recent calculation gives 0.86 kcal/mole for the dis- tortion energy of the ground state of this radical.30 It is also interesting to note the small quantum defect (0.046) of the Rydberg states observed. The very low ionization potential observed for the tropyl radical is important in comparing the stabilities of the isomeric tropyl and benzyl ions, since mass- spectrometric studies of tropyl compounds give a value AHf = 217f6 kcal/mole for the tropyl ion which is not significantly lower than that for the benzyl ion (220f3 kcal/mole).4 This is a surprising result since there is clear evidence that benzyl ions formed with small excess energies from toluene 149 15 and from benzyl radicals 16 rapidly isomerize to the tropyl ion in which all the hydrogen atoms are equivalent, and that this ion is formed at the threshold from benzyl halides, where the appear- ance potential of the C7HT ion is 6-10 kcal/mole lower than that calculated from bond energies and the ionization potential of the benzyl radical.17 In comparing the heats of formation of the benzyl and tropyl ions, both mass- spectrometric and spectroscopic ionization potentials have to be used.Unfor- tunately, little comparable data on free radicals have been obtained from the two sources. The only other spectroscopic values of free radical ionization potentials are I(CH3) = 9-843 eV and I(CH2) = 10.396 eV determined by Herzberg.18. 19 The former agrees well with mass-spectrometric values of 9-95,20 9-85,21 and 9-88 22 eV, but the latter does not agree well, the accepted mass-spectrometric value being 11.8 eV.21 This discrepancy may be due to the very reactive nature of the methylene[To facepage 198.B . A . THRUSH A N D J . J . ZWOLENIK 199 radical. While molecular ionization potentials based on Rydberg series are gener- ally lower than those determined mass-spectrometrically, there is no obvious regularity in the difference.There is, therefore, no reason to assume that the mass-spectrometric ionization potential of the benzyl radical is in error by the 0-3 eV discrepancy observed with the tropyl radical. It is perhaps significant that Streitwieser's calculations 5 using a modified molecular orbital method, in which some of the coefficients depend on mass-spectrometrically determined ionization potentials, give excellent agreement with most mass-spectrometric ionization poten- tials, but predict a value of 6-41 eV for the tropyl radical, which is 0-2 eV below the mass-spec tr ometric value. A value of I(C6H5CH2) = 7.76 eV is well authenticated and appears to be the true vertical ionization potential of the benzyl radical.16, 23 The recommended value of A&(C&CHz) = 43 kcal/mole 249 25 is supported by recent work on the pyrolysis of toluene.26 These data give AHf(c6H~cHt) = 222 kcal/mole. The heat of formation of the tropyl radical has been estimated by Dauben, Lossing and co-workers 4 from a calculated value for ditropyl vapour (94.6 kcal/ mole) and mass-spectrometric observation of the temperature at which it decom- poses to yield tropyl radicals. Their data indicate D(C7H7-C7H7) = 35+5 kcal/mole and hence AHf(C7H7) = 65 k 3 kcal/mole, which with Z(C7H7) = 6.237 eV gives AHf(C7Ht) = 209+3 kcal/mole. Thus, the rearrangement of the benzyl ion to the tropyl ion is about 13 kcal/mole exothermic while the corresponding radical rearrangement is about 22 kcal/mole endothermic.TABLE 2.-APPARENT AHj-(C,H+) FROM APPEARANCE POTENTIAL MEASUREMENTS compound toluene ethyl benzene propyl benzene di benzyl cycloheptatriene 7-methyl cycloheptatriene benzyl chloride benzyl bromide benzyl iodide A.P. CC,H', eV 11.8 11.25 11.23 10.53 < 10.1 < 9.5 10.35 9-67 9.23 AHf of compound AHf(C7Hf) kcal/mole kcal/mole 12.0 7.1 1.9 28.0 43.5 37.2 4.0 19 28 232 234 236 228 Q224 <224 214 21 5 21 5 ref. 27 27 15 27 4, 28 4 17, 29 17, 24 17, 24 Table 2 gives heats of formation of the ion of formula C7HT based on its appear- ance potential from various compounds containing benzyl or tropyl groups, and their heats of formation used in the calculation. The < sign indicates evidence that the C7H; ion was being formed with excess energy. For the cycloheptatriene derivatives, this threshold is surprisingly high compared with that for the benzyl halides, where the dissociative ionization which must yield tropyl ions occurs with only ca.6 kcal/mole excess energy. (This excess energy would be zero or negative if the mass-spectrometric value of I(C7H7) were used.) The observation that the apparent heat of formation of C7Ht for benzyl hydrocarbons is greater than that of benzyl ion does not prove that this ion is produced directly in the dissociative ionization. Investigation of the metastable peak corresponding to the process C7H$+C7Hf+H for deuterated toluenes shows that randomization of the hydrogen is almost com- plete before dissociation occurs.14200 SPECTRUM OF THE TROPYL RADICAL Our spectroscopic observations of the structure and ionization potential of the tropyl radical are clearly in good agreement both with theoretical predictions and with mass-spectrometric studies of compounds containing benzyl and tropyl groups.The authors thank the National Science Foundation for the award of a post- doctoral fellowship to J. J. 2. 1 Huckel, 2. Physik, 1931,70, 204; 2. Elektrochem., 1937,43,752. 2 Longuet-Higgins and McEwen, J. Chem. Physics, 1957, 26, 719. 3 Thrush, Nature, 1956,178, 155. 4 Harrison, Honnen, Dauben and Lossing, J. Amer. Chem. SOC., 1960,82,5593. 5 Streitweiser, J. Amer. Chem. SOC., 1960, 82, 4123. 6 Dauben, private communication. 7 Norrish and Thrush, Quart. Rev., 1956, 10, 149. 8 Thrush, J. Phot. Sci., 1960, 8, 232. 9 Field and Franklin, Electron Impact Phenomena (Academic Press, New York, 1957). 10 Watanabe, J. Chem. Physics, 1957, 26, 542. 11 Fateley, Curnutte and Lippincott, J. Chem. Physics, 1957, 26, 1471. 12 Jahn and Teller, Proc. Roy. SOC. A, 1937, 161, 220. 13 Longuet-Higgins, Opik, Pryce and Sack, Proc. Roy. SOC. A, 1958, 244, 1. 14 Rylander, Meyerson and Grubb, J. Amer. Chem. Soc., 1957,79, 842. 15 Meyerson and Rylander, J. Chem. Physics, 1957, 27,901. 16 Pottie and Lossing, J. Amer. Chem. SOC., 1961, 83, 2634. 17 Meyerson, Rylander, Eliel and McCollum, J. Amer. Chem. SOC., 1959, 81, 2606. 18 Herzberg and Shoosmith, Can. J. Physics, 1956, 34, 523. 19 Herzberg, Proc. Roy. SOC. A, 1961,262,291. 20 Lossing, Ingold and Henderson, J. Chem. Physics, 1954, 22, 621. 21 Langer, Hipple and Stevenson, J. Chem. Physics, 1954,22, 1836. 22 Osberghaus and Taubert, 2. physik. Chem., 1955,4,264. 23 Farmer, Henderson, McDowell and Lossing, J. Chem. Physics, 1954, 22, 48. 24 Benson and Buss, J. Physic. Chem., 1957, 61, 104. 25 Sehon and Szwarc, Ann. Rev. Physic. Chem., 1957, 8,439. 26 Price, Can. J. Chem., 1962, 40, 1310. 27 Schissler and Stevenson, J. Chem. Physics, 1954, 22, 151. 28 Finke, Scott, Gross, Messerly and Waddington, J. Amer. Chem. SOC., 1956,78, 5469. 29 Kirkbride, J. Appl. Chem., 1956, 6, 11. 30 Hobey and McLachlan, J. Chem. Physics, 1960,33, 1695.