Roger E. Gerkin and Arthur M. Winer
692
Electron Paramagnetic Resonance Absorption in Oriented Phosphorescent 2,3-Benzocarbazole and 1,2,3,4-Tetrahydroanthracene at Magnetic Fields below 65 G 1 Roger E. Gerkin” and Arthur M. Winer2 Department of Chemistry, The Ohio State University, Columbus, Ohio 43270
(Received November 12, 7973)
Electron paramagnetic resonance absorptions have been observed from phosphorescent 2,3-benzocarbazole, an impurity which is ubiquitous in commercial chrysenes, and from phosphorescent 1,2,3,4-tetrahydroanthracene (THA), an impurity which was identified in symmetric octahydroanthracene (OHA). Triplet-state zero-field splittings for 2,3-benzocarbazole oriented in single-crystal p-terphenyl and in single-crystal chrysene a t -85 K, and for THA oriented in single-crystal OHA a t -85 and -273 K, have been determined using data obtained a t magnetic fields below -65 G. Multiple resonance absorptions (five for each of the two transitions studied) were observed for 2,3-benzocarbazole in p-terphenyl whereas only single signals were observed for each of the corresponding transitions for 2,3-benzocarbazole in the chrysene host structure. Only single signals were observed from THA in single-crystal OHA, and its zerofield splittings and spin Hamiltonian parameters were found to be quite similar to those of naphthalene, the largest aromatic substructure of THA.
tained for 2,3-benzocarbazole in N,N-dimethylformamides7 In the course of an extensive low-field electron paraChrysene-dlz of nominal isotopic purity 98% was obmagnetic resonance study of triplet chrysene oriented in tained from Merck Sharp and Dohme, Ltd. (MSD. Lot p-terphenyl or symmetric octahydroanthracene (OHA),3-5 No. AD-024), and was subjected to column chromatogratwo tenacious impurities were encountered in sample maphy on Woelm neutral alumina, the only purification proterials used in the preparation of chrysene mixed crystals. cedure shown to be effective in removing essentially all of The first of these two impurities, 2,3-benzocarbazole, octhe 2,3-benzocarbazole impurity. The removal was demcurs in the guest material, chrysene, in concentrations as onstrated by the absence, visually determined, of the high as 10% even in nominally very pure commercial sam- characteristic orange phosphorescence of 2,3-benzocarbaples.6-7 The second impurity, identified as 1,2,3,4-tetrahy- zole in single crystals of chrysene photoexcited a t -85 K. droanthracene (THA), was found to occur persistently in (Four additional impurity bands were resolved on the colOHA, even after this host material had been zone-refined, umn and were also separated from the chrysene-dn.1 chromatographed, or recrystallized. In order to remove Some chrysene-dlz was used unpurified as a “source” of any doubt concerning the possible origin from impurity 2,3-benzocarbazole in an unknown state of deuteration in triplet states of multiple resonance absorptions observed mixed crystals of chrysene-dlz inp-terphenyl. for chrysene in p-terphenyl or OHA,3-5 it was deemed deScintillation grade p-terphenyl was obtained from Arasirable to investigate the low-field resonance spectra of pahoe Chemicals Inc. and was used without further purifi2,3-benzocarbazole and THA. We here report the results cation, since no (impurity) phosphorescence was detected of these brief investigations of phosphorescent 2,3-benzo- visually from single crystals photoexcited at -85 K, nor carbazole oriented in single crystals of p-terphenyl and was magnetic resonance observed from such crystals. chrysene, and of phosphorescent 1,2,3,4-tetrahydroanthra1,2,3,4-Tetrahydroanthracene (THA) was synthesized cene oriented in single crystals of symmetric octahydroanfrom 9,lO-dihydroanthracene according to the procedure thracene. of Orchin.10 The THA obtained was chromatographed on Woelm neutral alumina using benzene eluant and was reExperimental Section crystallized from ethanol. Symmetric octahydroanthracene (OHA) obtained from Apparatus and Methods. The basic low-field spectromeChemical Procurement Laboratories, Inc., was zone-reter and the experimental methods employed have been fined 50 passes, chromatographed, or recrystallized twice described in detail e l ~ e w h e r e . ~ ~ 8Optical -9 spectra obfrom ethanol. Further discussion concerning the use of tained for purposes of identification of species were taken OHA is given below. on a Turner spectrofluorophosphorimeter described preMixed Crystals. Single crystals were grown, cleaved, v i o u ~ l y .Precision ~ determinations of zero-field splittings and mounted in resonant cavities as described previouswere made for crystal samples in brass resonant cavities ly.8 The temperature of crystals illuminated by the unfilimmersed either in liquid nitrogen boiling a t ambient tered radiation of the A-H6 source and contained in cavipressure, or in distilled water-ice slush baths. ties immersed in liquid nitrogen was taken to be 85 f 3 K Sample Preparation. 2,3-Benzocarbazole employed as a on the basis of previous measurements made in this laboguest material in mixed crystals with p-terphenyl or chryratory. Such temperatures are designated canonically as sene was obtained from Chemical Procurement Laborato-85 K in the text. ries Inc. and was purified by vacuum sublimation. The The compositions and nominal guest concentrations of wavelength of the fluorescence maximum observed for this mixed crystals employed in these studies are shown in purified material in EPA glass was 4030 A, a value in good Table I. agreement with a previously reported value of 4070 A ob-
Introduction
The Journal of Physical Chemistry, Voi. 78. No. 7, 1974
693
Epr of Triplet Benzocarbazole and Tetrahydroanthracene
TABLE I: Composition and Nominal G u e s t Concentrations of Mixed Crystals Employed in This S t u d y Boule no.
Host material
199
p-Terphenyl p-Terphenyl
169
347
p-Terphen yl p-Terphenyl Chrysene Chrysene OHA OHA OHA
348
OHA
266 252 104 254
226 241
Host purification
Mode of occurrence of guest
Guest material@)
Chrysene-d 12b Chrysene-dd 1,2-Benzocarbazole
None. Nonea
Added Added Added incidentally as an impurity in the chrysene-d12 Added Added Host impurity Host impurity Host impurity Host impurity Added
2,3-Benzocarbazole 2,3-Benzocarbazole 2,3-Benzocarbazole 2,3-Benzocarbazole THA THA THA
Nonea Nonea None None None None Recrystalled twice from ethanol Recrystalled twice from ethanol
Nominal guest concn, mol
0.1
0.9 N O .I d
1.1
0.08 10d 10d
e e
2 .O
Nonef
No phosphorescent impurities were detectable in this material (see text). Purified by column Chromatography to remove essentially all 2,3-b~nzocarbazole Unpurified. Estimate based on nominal 10-12 mol t/o of 2,3-benzocarbazole found in commercial chrysenes. See re€ 6 and 7. e Not determined. Impurities reduced but not completely removed (see text). a
(see text). f
Treatment of Data The spin Hamiltonian chosen to interpret the magnetic resonance data is X = H*gI/3I*Sf DS:
f
E(SA2 - S,")
(1)
consistent with our previous usage.8 The field dependence of the associated energy levels and the method of least-squares fitting of the resonance data have been presented previously.8 Magnetic Resonance Observations for 2,3-Benzocarbazole p-Terphenyl Host. Assignments to chrysene-dlz triplet states (oriented in p-terphenyl single crystals) of multiplet patterns of resonance absorptions associated with sets of slightly different zero-field splittings at -0.116, -0.062, and -0.053 cm-1 were made in detailed analyses given in ref 4 and 5 . The component signals in the multiplet patterns were interpreted as arising from sets of differently oriented chrysene-a'lz triplets which occupy several kinds of sites of substitution at which there are significantly different molecular crystal fields giving rise to the observed differences in the zero-field splittings of the component signals. In addition to these three well-characterized multiplets containing a total of 20 resonance absorptions, two other sets of multiple resonance absorptions were observed from mixed crystals of chrysene-dlz in pterphenyl. For one set of signals the corresponding zerofield splittings were shown to range between 0.0847 and 0.0858 cm -I, while zero-field splittings corresponding to absorptions in the other multiplet ranged in value from 0.0636 to 0.0644 cm-l. At least five resonance absorptions were observed in each of these patterns, and each absorption within a multiplet was shown to be associated with a slightly different zero-field splitting by changing the microwave frequency by small increments so that the centers of the individual resonance signals were moved successively to zero magnetic field. When an authentic sample of 2,3-benzocarbazole (Figure l a ) was used as the sole guest in p-terphenyl mixed crystals, multiplet structure patterns were observed at
H
Figure 1. Structure of (a) 2,3-benzocarbazole and (b) carba-
zole.
-0.085 and at -0.064 cm-I from such crystals, and component signals in these patterns were shown to correspond one-to-one with components of the multiplet patterns observed at -0.085 and -0.064 cm-1 from impure chrysene in p-terphenyl mixed crystals. No absorption patterns were observed, however, at the three energies corresponding to the chrysene-dlz zero-field splittings. Conversely, when p-terphenyl mixed crystals prepared with chromatographed chrysene-dlz were studied, the three absorption patterns attributed to chrysene-dlz were observed but the two patterns attributed to 2,3-b'enzocarbazole were not. These results confirm that 2,3-benzocarbazole was indeed not the source of any of the resonance absorptions attributed to chrysene in our previous analyses495 but was the ~ observed origin of the -0.085- and ~ 0 . 0 6 4 - c m -multiplets in those studies. The correspondence between multiplet signals observed from authentic 2,3-benzocarbazole in p-terphenyl, and those observed from 2,3-benzocarbazole incorporated in p-terphenyl uia its occurrence as a contaminant in chrysene-& suggests the following self-consistent alternatives: (a) that the 2,3-benzocarbazole present as an impurity underwent little or no deuteration during the perdeuteration of its chrysene carrier with the consequence that the present experiments give no information about a possible deuterium isotope effect in 2,3-benzocarbazole; (b) that the 2,3-benzocarbazole was perdeuterated during the predeuteration of its chrysene carrier with the consequence that the present experiments establish that phosphorescent 2,3-benzocarbazole, in contrast to most other triplet The Journal of Physical Chemistry, Voi. 78, No. 7, 1974
Roger E. Gerkin and Arthur M .Winer
694
tings at -0.085 cm-I may be associated with ID - E ) / h c and those at -0.064 cm-I may be associated with ID El lhc. Consistent with this assignment, the values of the spin Hamiltonian parameters for 2,3-benzocarbazole in chrysene a t -85 K are )Dl/hc = 0.07483 cm-I and IEJ/hc = 0.01111 cm-I, each with standard deviation 0.00003 cm-1. We are not aware of any previous experimental determinations of the spin Hamiltonian parameters of 2,3-benzocarbazole, either for glass or mixed single-crystal systems. However, IDJ/hc and IEl/hc for carbazole (Figure Ib) in glassy ether have been measured as 0.102 and 0.007 cm-l, respectively.13
b
+
Figure 2. p-Terphenyl crystal structure
states studied in this laboratory,llJ2 manifests essentially no deuterium isotope effect in its zero-field splittings at -85 K. Chrysene Host. Singke crystals of unpurified chryseneh12 grown by the Bridgman method were of very poor quality. Nonetheless, when photoexcited a t -85 K such crystals manifested intense orange phosphorescence, as did single crystals of 2,3-benzocarbazole in p-terphenyl. Two weak resonance absorptions were observed from the (2,3-benzocarbazole in) chrysene crystals. The zero-field splittings associated with the absorptions were determined and found to be 0.08594 cm-1 with standard deviation 0.00003 cm-I and 0.06372 cm-I with standard deviation 0.00002 cm-l, respectively. Thus, the approximate values found for the (multiple) zero-field splittings of 2,3benzocarbazole in p-terphenyl host are very similar to these more precise values determined for the two individual signals observed from the chrysene host structure. A quite small host-structure shift in zero-field splittings3 is indicated by a comparison of the average values of the multiple 2,3-benzocarbazole splittings observed for p-terphenyl host with the single values obtained for chrysene host. The observation of multiple resonance absorptions from phosphorescent 2,3-benzocarbazole in p-terphenyl, but only single absorptions at corresponding energies for the chrysene host structure, parallels results we have obtained for other triplet states studied in at least two host structures of which one was p-terphenyl. For example, chrysene-& in p-terphenyl manifests at least seven resonance absorptions corresponding to the ID - E ( / h c zero-field transition but only two absorptions for the same transition when oriented in the OHA host s t r ~ c t u r e A . ~ comparable reduction in number of signals was obtained for chrysenecllz on going from p-terphenyl to biphenyl host.4 We interpret these observations in terms of the propensity of a guest triplet state to adopt a unique orientation at a substitutional site in a given host structure, e.g., 2,3-benzocarbazole in chrysene, or conversely to adopt multiple stable orientations, as is the case for every guest which we have studied in p-terphenyl. If 2,3-benzocarbazole occupies substitutional sites in the p-terphenyl host structure (Figure 2) so that its fine structure axes are approximately parallel to the molecular axes of p-terphenyl, then provisionally (on the basis of satisfaction of the zero-field selection rules)4 the zero-field splits4
The Journal of Physical Chemistry, Voi. 78, No. 7, 1974
Magnetic Resonance Observations for 1,2,3,4-Tetrahydroanthracene(THA) in Symmetric Octahydroanthracene (OHA) As discussed in a previous paper,4 extensive purification of OHA, including zone-refining and chromatography, failed to remove an impurity which gave rise to a yellowgreen phosphorescence emission from photoexcited single crystals of OHA, both at -85 and at -273 K. Only for OHA recrystallized several times from ethanol was significant reduction in the impurity, as evidenced by reduced phosphorescence intensity, obtained. Two magnetic resonance absorptions were detected from photoexcited single crystals of unpurified OHA and the zero-field splittings associated with the absorptions were determined at -85 K to be 0.111420 and 0.083237 cm-l, respectively, with estimated uncertainties for replicate determinations -1 x 10-5 cm-:. Magnetic resonance lifetimes measured in experiments a t -0.111 and -0.083 cm-I were found to be approximately 1 sec at -85 K. The absorption line width for the signal near -0,111 cm-l was approximately 15 G. In an experiment in which the resonant cavity was immersed in a distilled water-ice slush and a bandpass filtel.4 for A-H6 excitation was employed, the resonance absorption associated with the zero-field splitting value 0.11142 cm-1 at -85 K was determined semiquantitatively to have a corresponding zero-field splitting at 273 K of 0.1097 cm-l. Thus, an approximate zero-field splitting temperature coefficient of -9 x cm-I K - I was observed forthe transition at -0,111 cm-1. The similarity of the magnitudes of the triplet energy (as evidenced by the observed phosphorescence), zerofield splittings, magnetic resonance lifetime, and temperature coefficient of the ( D - E ) zero-field splitting of the impurity to the values of these properties for naphthalene suggested that the impurity might be THA, shown in Figure 3a. An additional basis for this suggestion was that in the catalytic hydrogenation of anthracene both THA and OHA (Figure 3b) are produced. Further verification of the identity of the impurity was made difficult, however, by the behavior of OHA boules grown by the Bridgman method14 from OHA recrystallized twice from ethanol and then doped with authentic samples of THA. Such boules consistently shattered violently, after exerting sufficient force to break the crystalgrowing tubes some hours after they emerged from the heated region of the air-core furnace. In contrast, undoped boules of recrystallized OHA remained ice-clear and retained definite cleavage planes months after being grown. Two groups of investigators have described phase transformations in OHA similar to those observed in the present research, but neither group discussed the role of impuri-
Epr of Triplet Benzocarbazole and Tetrahydroanthracene
Figure 3. Structure of (a) 1,2,3,44etrahydroanthracene and ( b ) symmetric octahydroanthracene.
ties in the processes.15J6 A thorough investigation of this problem lay outside the scope of this research. Inasmuch as whole boules of authentic THA in recrystallized OHA could not be produced, the best single crystal pieces which could be salvaged from shattered boules were employed in double-crystal experiments. A special double-window resonant cavity was constructed and then divided by an opaque partition.* A single-crystal of THA in recrystallized OHA was placed behind one window, and a crystal from a boule grown from the same recrystallized OHA (but not doped with THA) was placed behind the other window. Rotation of the resonant cavity by T permitted alternately photoexciting one, but not the other, crystal piece. In such two-crystal experiments no signals were observed from single crystals of undoped recrystallized OHA whereas for the same host material which had been doped with THA, resonance absorptions were observed at -0.111 and -0.083 cm-l. Precision zero-field splitting determinations for the two observed signals a t -85 K yielded values of 0.111424 and 0.083228 cm-l which may be compared with values of 0.111420 and 0.083236 cm-1, respectively, determined (as discussed above) for the resonance absorptions observed at -85 K from the impurity
695
in unpurified OHA. These precision determinations of zero-field splittings establish the identity of the OHA impurity to be THA. Due to the comparatively poor quality of OHA crystals obtained in these studies and the lack of a detailed crystal structure determination for OHA it was not possible to assign the observed zero-field energies for THA to corresponding zero-field transitions on the basis of a rigorous investigation of satisfaction of the zero-field selection rules as a function of orientation. By analogy to naphthalene, however, it is plausible to assign the 0.11142- and 0 . 0 8 3 2 3 - ~ m -splittings ~ to ID - El/hc and ID + El/hc, respectively, This leads to values of IDl/hc and IE(/hc for THA in OHA of 0.09733 and 0.01410 crn-', respectively, at -85 K. (For naphthalene in biphenyl the corresponding values are approximately 0.0993 and 0.0155 cm-l, respectively.) To our knowledge the precision zero-field splittings and spin Hamiltonian parameters presented here for THA are the first such data reported for a partially hydrogenated aromatic hydrocarbon. References and Notes This work was supported in part by the National Science Foundation. Statewide Air Pollution Research Center, University of California, Riverside, Calif. 92502. R. E. Gerkin and A. M . Winer, J. Chem. Phys., 47,2504 (1967). R. E. Gerkin and A. M. Winer, J. Chem. Phys., 56,1359 (1972). R. E. Gerkin and A. M. Winer, J. Chem. Phys., 58, 1360 (1973). P. Mabiiie and N. P. Buu-Hoi, J. Org. Chem., 25, 1937 (1960). D. F. Bender, E. Sawicki, and R. M . Wilson, Anal. Chem., 36, 1011
(1964). R. E. Gerkin and P. Szerenyi, J. Chem. Phys., 50,4095 (1969). A. M. Winer, Ph.D. Dissertation, The Ohio State University, 1969. M. Orchin, J. Chem. SOC.,66,535 (1944). R. E. Gerkin and A. M. Winer, J. Chem. Phys., 50,3114 (1969). A. S. Culiick, R. E. Gerkin, D. L. Thorseii, and A. M. Winer, to be submitted for publication. S. Siegei and H. S. Judeikis, J. Phys. Chem., 70,2201 (1966). J. N. Sherwood and S. J. Thompson, J. Sci. lnstrum., 34, 42
(1965). V. N. Vatuiev and A. F. Prikhot'ko, Fiz. Tverd. Tela, 7,42 (1965). Yu. V. Mnyukh and M. A. Tseneva, Dokl. Akad. Nauk SSSR, 162,
326 (1965).
The Journal ot Physical Chemistry, Vol. 78,No. 7, 1974