J . Chem. Soc.. Faradoy Trans. 1. 1982. 78, 241 1-2421 Effect of Substituent on the Behaviour of the Excited Singlet and Triplet States in Carbonyl Derivatives of Anthracene of the Type 9-X*CO.A BY SATOSHI HIRAYAMA Faculty of Textile Science, Kyoto Technical University, Matsugasaki, Sakyo-ku, Kyoto, 606 Japan Received 24th August, 198 1 Kinetic and spectroscopic studies relevant to the excited singlet and triplet states have been carried out on eighteen carbonyl compounds of anthracene, principally of the type 9-X CO . A (where A is anthracene), aiming to reveal distinctly how a substituent X controls the excited-state behaviour. The compounds examined are classified into six groups based on the nature of X. Group I compounds, where X is an alkyl group, are non-fluorescent in any solvent at room temperature, but they start to fluoresce strongly near 77 K with fluorescence lifetimes of ca.10 ns. When X is capable of conjugation with the carbonyl group, the derived compounds (group 11) totally lack fluorescence even at 77 K. The carbonyl groups of anthracene-9-carbaldehyde and 2-methylaceanthren-1-one (group 111) differ from the others in that they form hydrogen bonds in protic solvents. Only group V compounds, with an electron-donating substituent such as an amino, methoxy or hydroxy group, are modestly fluorescent at room temperature. Intersystem crossing from S, to Trill+, whose location relative to S, is affected by X, appears to play an important role in the appearance of fluorescence. Group VI compounds show that the introduction of another substituent at the 10-position also affects the fluorescence properties. In contrast to the wide variation seen in the fluorescence properties, the lowest triplet states are found to be similar to each other and only slightly different from those of anthracene.The shape of the phosphorescence and triplet-triplet absorption spectra and the lifetimes and energies of the lowest triplet states are also similar to those of anthracene. This similarity suggests that the lowest triplet states of the examined compounds are not of nn* but of m* character, originating from the anthracene group. The very weak or practically absent fluorescence in those organic compounds which consist of both pi and lone-pair electron systems has often been explained in terms of the so-called El-Sayed rule.' This states that intersystem crossing of the type nn* rw, ~ n * or vice versa is the most important process in depleting the lowest excited singlet state S,, although in some exceptional cases an internal conversion to the ground state would become dominant.2 In particular, when the radiative transition from S, is allowed, as is frequently so for SIIIn* (but not for Slnn*), the lack of fluorescence means that very efficient intersystem crossing occurs on the picosecond time-scale .In anthracene it is well recognized that chemical substitution affects the radiative properties through a substantial change in the intersystem crossing rate from S1,3 while the radiative transition probability itself remains practically constant. When a substituent contains a carbonyl group the effect of substitution may well be amplified, since it adds nn* states, which are expected to play an important role as intermediate states in the radiationless process depleting S,.This paper is a study of such a substituent effect on the photophysical processes of the rneso-substituted carbonyl derivatives of anthracene of the type 9-X * CO *A, where A and X stand for anthracene and the substituent, respectively. By changing X systematically, it will be shown how the substituent controls the radiative properties of the derived compounds. 241 12412 EXCITED-STATE BEHAVIOUR OF SUBSTITUTED ANTHRACENE EXPERIMENTAL The absorption spectra were obtained on a Shimadzu 210A spectrophotometer and the fluorescence and excitation spectra were recorded on a Shimadzu RF 502 spectrofluorophoto- meter provided with a Hamamatsu R 928 photomultiplier.The excitation spectrum of anthracene paralleled its absorption spectrum, showing the high precision of the excitation spectra obtained. The fluorescence lifetimes were measured with a Hitachi MPF 4 spectro- fluorometer equipped with a 10 ns flash-lamp and processing devices, as described in a previous paper. The triplet-triplet (T-T’) absorption spectra at 77 K were taken either spectrophotographically or photometrically by means of conventional flash excitation. A rectangular sample cell (10 x 20 x 50 mm with a 20 mm path length) fitted with a glass pipe for evacuation was made of Pyrex and its upper half was kept in good contact with a metal jacket filled with liquid nitrogen.The whole part of the cell (the glass cell and metal jacket) was immersed in a transparent Dewar vessel which was also filled with liquid nitrogen. Its level, however, was kept slightly below the bottom of the metal jacket so that the lower half of the cell was in touch with liquid nitrogen but bubbling did not interfere the light path of an interrogating beam. When the triplet lifetimes were > 1 ms, a commercially available standard camera flash was found to be very useful to observe a transient absorption or its decay. When the triplet lifetimes were Q 1 ms, a Q-switched ruby laser (30 ns pulse width) was used as an excitation s ~ u r c e . ~ The decay curves were monitored with a Techtronics 475 oscilloscope. The triplet lifetimes were calculated from the decays of T-T’ absorption monitored at the maximum absorption wavelengths. The phosphorescence spectra were recorded on a home-built phosphorometer which consisted of a Hamamatsu R 928 photomultiplier, a Ritsu ML 20 monochromator and a rotating can with two windows.A 500 W high-pressure mercury arc was used as an excitation light source. The spectra obtained were not corrected for the spectral sensitivity of the apparatus. The phosphorescence lifetimes were also measured with this phosphorometer, but signals were fed in a Union RA-450 data processor and thus smoothed curves were obtained for display on an X-Y recorder. In several cases the triplet-state lifetimes obtained from T-T’ absorption decays were compared with those from the phosphorescence decays in order to establish the identity of the origin of the phosphorescence and T-T absorption. All of the anthracene derivatives except 9-anthryl methoxy ketone were synthesized according to the literature and purified either by sublimation in uucuo or by thin-layer chromatography.Relevant references to the synthesis of the compounds are given in table 1. The solvents used were either of a spectroscopic or a guaranteed grade. The sample solutions were not degassed since no significant oxygen effect was observed at 77 K. The concentrations of the solutions were ca. mol dm-3 for T-T’ absorption and phosphorescence spectrum measurements and ca. 2 x mol dm-3 for fluorescence and lifetime measurements. Thus the excitation spectra obtained are practically free from reabsorption.RESULTS A N D DISCUSSION CLASSIFICATION OF THE COMPOUNDS The eighteen carbonyl compounds examined are classified into six groups based on the nature of the substituents, as given in table 1 . The first group of compounds (called group I etc. hereafter) are derived from 9-X - CO - A, where X is an alkyl group, i.e. 9-anthryl methyl ketone (l), 9-anthryl ethyl ketone (2), and 9-anthryl propyl ketone (3). The compounds in which X is capable of conjugation with the carbonyl group are gathered in group 11. They are 9-anthryl vinyl ketone (4), 9-anthryl phenyl ketone (5), 9-anthryl styryl ketone (6) and 9-anthryl 1 -naphthyl ketone (7). Anthracene- 9-carbaldehyde [X = H, (S)] and 2-methylaceanthren-1-one (9) are placed in group 111, since in these molecules the planes expanded by the carbonyl groups are in a broad sense co-planar rather than orthogonal to the anthracene ring.Thus, from aS. HIRAYAMA 2413 TABLE 1 .-SPECTROSCOPIC PROPERTIES AT ROOM TEMPERATURE OF THE CARBONYL DERIVATIVES OF ANTHRACENE EXAMINED absorptiona fluorescenceb? compound 0-0 band 0-0 band v(C=O) (9-X * CO * A) /nm /nm 77 Kd /cm-l ref. I n JII IV V VI 9-Me - CO - A (1) 9-Pr * CO - A (3) 9-Et * CO * A(2) 9-CH2:CH*CO*A (4) 9-Ph. CO * A (5) 9-C6H5 * CH : CH * CO .A (6) 9-Naph - CO * A (7) (Naph = Naphthyl) (9) (see text) 9-H*CO*A (8) 9-CH2Br - CO * A (10) 9-CH3 - CHBr - CO A (1 1) 9-NH,.CO.A (12) 9-COOH-A (13) 9-CH30. CO A (14) 9-Ph. CO- 10-C 1 * A (15) 9-Ph.CO-1O-NO;A (17) 9-Ph. CO- 1 0-CN * A (18) 9-Ph * CO- 1 0-Br * A (16) 382 382 38 1 382 383 382 384 400 418 383 384 380 380 38 1 394 394 388 406 none none none none none none none 500e 470"g 2o none none 390 388 4 76 none none none 450 387 1703 6 387 1708 6 387 1703 7 none 1663 8 none 1669 9 none 1633 10 none 1660 11 463" 1682 12 430 1703 8 430 1729, 1700 7 none 1703 7 387 387 410 13 14 15 none 1670 16 none 1670 16 none 1670 16 415 1673 17 a In cyclohexane; 'none' indicates that the fluorescence quantum yield is < lop4; in cyclohexane except for compounds 8 and 9 for which ethanol was used; in ethanol; due to the hydrogen-bonded form.The italicised values are for the absorption or emission maxima. molecular-structure point of view, these two compounds are in sharp contrast to the other compounds, where the carbonyl groups are approximately perpendicular to the anthracene ring.In the fourth group, two a-bromoalkyl anthryl ketones [(lo) and (ll)] are listed. These compounds can be derived by the bromination of the corresponding compounds in group I. In contrast to the parent compounds, the compounds thus derived undergo several interesting intramolecular photochemical reactions which are thought to reflect the presence of efficient energy-dissipation channels in the carbonyl derivatives of anthracene.sv l8 In group V we list the compounds which have electron-donating substituents such as amino, hydroxy and methoxy groups, i.e. anthracene-9- carboxamide (12), anthracene-9-carboxylic acid (13), and its methyl ester (14). Compounds of the sixth group are derived from (5) by introducing another substituent at the 10-position.ABSORPTION SPECTRA The wavelengths of the 0-0 absorption bands measured in cyclohexane at room temperature are summarized in table 1. When the absorption spectrum is too broad to determine the position of the 0-0 band unambiguously, the maximum absorption wavelengths are given instead. For any 9-substituted compound other than those in24 14 EX C I TED-S T ATE BE H AVI OUR 0 F SUBSTITUTED AN THR A CE NE group 111, the position of the 0-0 band shifts to the red (by ca. 8 nm) only slightly compared with that of anthracene. This minute substituent effect on the absorption spectrum arises from the large steric hindrance caused by peri-hydrogen atoms which put the carbonyl group in a position perpendicular to the anthracene ring in the ground state, making the conjugation of the C=O double bond with the anthracene group insignificant.Consequently, the substituent effect on the absorption spectra is nothing more than that due to a methyl group.19 On the other hand, in group I11 compounds this conjugation is very effective because their molecular structures favour co-planarity, making the 0-0 band (or the maximum) shift to the red larger (by ca. 30 nm). To illustrate this difference, the absorption spectra of 5, 7 and 8 are shown in fig. 1. In addition to this large red shift, group I11 compounds are characterised by a large solvent effect on the electronic spectra, since the carbonyl groups in these compounds can form hydrogen bonds in protic solvents. In fact the fluorescence in protic solvents is dominated by the hydrogen-bonded forms in these compounds, as will be shown later.A detailed spectroscopic study of 9 has been given elsewhere. 2o wavelength/nm FIG. 1.-Absorption spectra of 5 (-.--), 7 (-) and 8 (----) in cyclohexane at room temperature. The absorption spectrum for 8 is exceptionally broad and red-shifted compared with the other compounds. In contrast, compounds other than group I11 do not exhibit a specific solvent effect on the electronic spectra which might arise from the presence of a carbonyl group within the molecule. By introducing another substituent at the 10-position, the red shift mentioned above is expected to increase in an additive way, as is exemplified by the 0-0 absorption wavelengths of the group VI compounds shown in table 1.The effect is greatest for the carbonitrile group.21 These additional shifts appear to play an important role in the radiationless processes seen in these molecules, as will be discussed later. FLUORESCENCE SPECTRA There has long been discussion as to which kinds of substituents increase or decrease fluorescence efficiency.Z2* 23 The carbonyl group is widely recognized as a quencher of fluorescence rather than an enhancing group. At first glance, this seems to be the case for the compounds listed in table 1. All compounds other than those in groups I11 and V are practically non-fluorescent, at least at room temperature, irrespective ofS. HIRAYAMA 2415 300 350 400 450 500 550 wavelength/nm FIG. 2.-Fluorescence and excitation spectra of 3 observed in EPA at 77 K.For comparison, the fluorescence spectrum of 12 in ethanol at room temperature is shown by the broken line. TABLE 2.-FLUORESCENCE LIFETIME IN EPA AT 77 K compound +sa 1 2 3 9 12 13 14 18 ~~ ~ 11.2 11.4 11.6 9.9 13.1 11.2 ( 1 2. 5)b 11.1 a The standard deviations of zf are 0.1-0.3 ns. The fluorescence from 10 was too weak to determine a reliable lifetime. Except 8, all of the other compounds not listed in this table are non-fluorescent even at 77 K. In cyclohexane at 22 OC. See T. C . Werner, T. Matthews and B. Soller, J. Phys. Chern., 1976, 80, 533. solvent polarity. However, when lowering the temperature, we see that such a simple empirical argument does not necessarily hold. For instance, group I compounds become strongly fluorescent near 77 K.24 As a typical example, the fluorescence spectrum of 3 is shown in fig.2 together with that of 12. The fluorescence lifetimes measured at 77 K in EPA are given in table 2 for most of the compounds fluorescent at 77 K. They are of the order of magnitude expected for anthracene derivatives whose electronic transitions to S , are allowed (with radiative rate constant k , z lo8 s-l), indicating that the fluorescence quantum yields at 77 K are of an appreciable magnitude. In other words, radiationless processes, which are so important at room temperature, are virtually frozen at 77 K. With a conventional fluorophotometer, the lowest detection limit of fluorescence would be a quantum yield2416 EXCITED-STATE BEHAVIOUR OF SUBSTITUTED ANTHRACENE of ca.and hence the lack of fluorescence implies that the fluorescence quantum yield is < lod4. Thus an increase in the fluorescence quantum yield up to, say, 0.5 indicates the enhancement of fluorescence by a factor of 5 x lo3 in going from room temperature to 77 K. With this enormous increase, group I compounds would provide a very interesting case among those compounds whose fluorescence is temperature-sensi tive. 25 9 26 Group I1 compounds exhibit a great contrast to group I compounds with respect to the temperature dependence of fluorescence. The conjugating groups X examined vary widely from the vinyl to naphthyl groups but the derived compounds are always totally non-fluorescent even at 77 K.? How well this holds can be seen by examining the effect on fluorescence caused by breaking a conjugation between the carbonyl group and X, e.g.as in 9-anthryl phenethyl ketone, which is obtained by saturating the a-double bond of 6.1° It behaves like the group I compounds, i.e. non-fluorescent in solution at room temperature but highly fluorescent at 77 K. Thus we see that conjugation of the group X with the carbonyl group is essential to make the molecule totally non-fluorescent even at 77 K. 0. a Q) 2 -fl 3 0.4 9 wavelength/nm FIG. 3.-Absorption, fluorescence and excitation spectra of 8 observed in ethanol at 77 K. The absorption spectrum is shown by the dotted line. The red-shifted and structured excitation spectrum (thin solid line) monitored at 490 nm is caused by the hydrogen-bonded form, whose absorption is not prominent in the absorption spectrum.It was carefully confirmed by examining the excitation spectrum at several concentrations of 8 that the disagreement between the absorption and excitation spectra is not due to experimental artifacts such as inner-filter effects. Group I11 compounds are fluorescent in solution at ambient temperature, but their yields are low20 and the fluorescence arises mainly from hydrogen-bonded forms. To show this, the fluorescence spectrum of 8 and its excitation spectrum are presented in fig. 3 together with the absorption spectrum at 77 K. Apparently the absorption and excitation spectra do not coincide and the red-shifted and more structured excitation spectrum is due to the hydrogen-bonded form. t Note that this holds only for 9-substituted anthracenes.In the case of 1- or 2-carbonyl substituted anthracenes, e.g. 1 - and 2-Me. CO * A or 1 - and 2-Ph * CO * A, appreciable fluorescence has been observed.2’. 28 One of the reasons for this difference might be that the above-mentioned compounds are able to form hydrogen bonds in protic solvents, in contrast to those of groups I and 11.S. HIRAYAMA 2417 Contrary to the expectation that group IV compounds behave like those of group I, the consequence of bromination is remarkable. As has been reported previously,s* 21 both compounds undergo facile intramolecular photochemical reactions ; 10 yielding 9-bromoanthracene and aceanthren- 1 -one as the major photoproducts and 11 yield- ing 4 and 9, whose relative yields depend critically on temperature.Since both reactions of 10 require large structural change, it is not surprising to find that these reactions are completely prohibited in rigid matrices at 77 K. On the other hand, photo- dehydrobromination of 11 to yield 4 occurs readily even in frozen matrices at 77 K. Presumably because of this difference (10) starts to fluoresce weakly near 77 K, but i l l ) is practically non-fluorescent even at 77 K. Certain vibrational modes, which in the extreme lead to the dehydrobromination reaction, may well act as efficient promoting modes of radiationless transition from S, of 11 to the ground state. Group V compounds present another interesting example which illustrates how remarkably the nature of the substituent X controls the fluorescence. All compounds in this group are modestly fluorescent in solution at room temperat~re,'~ in contrast to compounds in the other groups.Taking account of the electron-donating nature of the substituent~,~~ it is quite likely that these groups shift the Tnn* state to a higher energy than that of S,, making the process of Slnn* mTnll* unfavourable energetically. T-T' ABSORPTION SPECTRA A N D TRIPLET-STATE LIFETIMES A T 77 K The T-T' maximum absorption wavelengths and the lowest triplet-state lifetimes determined by decay of the T-T' absorption are listed in table 3. The T-T' absorption spectra of 5 and 9 are compared in fig. 4 to illustrate the spectra in the two extremes with respect to the extent of conjugation to which the carbonyl groups contribute. When a carbonyl group is not in conjugation with the anthracene, the T-T' absorption spectra are similar to that of anthracene and have well-resolved peaks at ca.407 and TABLE 3.-TRIPLET-STATE PROPERTIES OBSERVED IN EPA AT 77 K FOR THE CARBONYL DERIVATIVES OF ANTHRACENE triplet-state T-T' absorption phosphorescence compound lifetime/ms Amax/nm 0-0 band/cm-l 1 2 3 4 5 6 7 8 9 10 12 13 15 16 17 29.5 32.1 34.2 31.3 33.1 34.3 30.7 (29.8)" 1.7 5.9 (5.6)u (20.9)" 39.6 34.5 3.8 0.17 16.05 429 430 430 430 430 430 (broad) 427 (broad) 455 45 5 433 43 1 432 430 43Y - 14 720 14 730 14 740 14 720 14 730 14 740 14 690 13 790 13 830 14 350 14 820 14 760 14 160 14 140* 14 390 a From the phosphorescence decays. * The phosphorescence spectrum was measured by removing the rotating can from the phosphorophotometer.2418 EXCITED-STATE BEHAVIOUR OF SUBSTITUTED ANTHRACENE 1.aJ c .fl 0. s1 I, m '0.. "'%..o. Q.,, - 400 410 420 430 440 450 460 470 480 wavelength/nm FIG. 4.-T-T absorption spectra observed for 5 (-) and 9 (--O--) in EPA at 77 K. The two peaks at ca. 407 and 430 nm are characteristic of the compounds examined here except for 8 and 9, although some may show broader absorptions. 430 nmO3O The lifetimes centre around 30 ms, which is again close to that of anthra~ene.~~? 30 On the other hand, when a carbonyl group comes into conjugation with the anthracene group, the T-T' absorption spectra become broad with largely red-shifted absorption maxima, as is shown by the broken line in fig. 4. Furthermore, the lifetimes become much shorter than those for compounds in groups I, 11, V and VI.Thus again the compounds in group I11 are in great contrast to the other carbonyl derivatives with respect to the shape of the T-T' absorption spectrum and triplet-state lifetime. Those compounds containing a heavy atom such as chlorine or bromine naturally show shorter triplet-state lifetimes owing to the heavy-atom-induced T,-S, transition. PHOSPHORESCENCE SPECTRA AT 77 K The 0-0 band of the phosphorescence in anthracene lies in a long-wavelength region owing to the low energy of the lowest triplet state, T,. The substituent effect on the energy of T, has been thought to be minute and, in fact, the phosphorescence spectra observed for the carbonyl derivatives of anthracene are very similar to that of anthracene in their spectral shape and energy (14927 cm-l for anthra~ene,~~ see table 3) except that the phosphorescence in group I11 compounds are very weak and largely red-shifted.The phosphorescence spectra of 7 and 9 in EPA at 77 K are shown in fig. 5. The observation of phosphorescence is of great significance from the following two points. First, we can locate the T, state and identify its electronic nature. Taking into account the low energies of T, and the similarity between the shapes of the phosphorescence spectra and that of anthracene, it can safely be said that the lowest triplet states of the compounds studied here are of nn* rather than nn* character.t The shortening of lifetimes found in the group I11 compounds may be due to a slight mixing of Tnl* with the lowest triplet state T1,,, , since the coplanar configuration of the carbonyl group against the anthracene ring is favourable for this mixing, as has previously been The lowering in T,-state energies of these compounds (ca.1000 cm-l) would not be large enough to cause a large increase in the non-radiative rate constant k,, of Tlnl*-SO to give such shortened triplet-state lifetimesP2 Secondly, t The energy gap between S, and T, (ca. 1 1000 cm-l) is too large for T, to be assigned as originating from an nn* state.S. HIRAYAMA 2419 A wavelength/nm FIG. 5.-Phosphorescence spectra observed for 7 (-) and 9 ( * - - .) in EPA at 77 K. In either case the separation between the first and second band groups is in the range 1400 cm-’, indicating the electronic states of the lowest triplet states are of KK* character.Compared with 7, the phosphorescence in 9 is weaker, and hence an expanded spectrum (ca. 8 times) is given in this figure. it should be emphasized that most of the rneso-substituted anthracenes studied here give triplet states in appreciable yields even at 77 K. Other rneso-substituted anthracenes with no carbonyl group are reported to yield no triplet state at 77 K;33 i.e. #f is practically unity at this temperature, since the pathway of S,-T, (n = 2 or 3) intersystem crossing is completely shut off at low temperatures because of its endothermicity. Typical examples are 9-methylanthracene and 9,lO-dichloro- anthra~ene.~~ In fact neither phosphorescence nor T-T’ absorption was observed in these compounds with the apparatus employed in the present paper.Thus, the rneso-substituted carbonyl compounds which start to fluoresce at low temperatures are unique among the anthracenes in the sense that their df values at 77 K are still less than unity. CONCLUSIONS We have so far seen how chemical substitution determines excited-state behaviour in compounds of the type 9-X - CO .A. Although numerous cases are known in which chemical substitution affects the kinetics, reactivities, spectroscopic properties, etc. in a way predicted by theory, only a limited number of examples34 are known where a substitution effect on the excited-state behaviour in single-family compounds has been so widely examined as in the present study. The fluorescent properties found in the compounds of groups I-VI are so widely varying that, at a first glance, it may appear to be difficult to provide a unified picture of the radiationless processes pertinent to these compounds.When we look at the results more closely, however, we see that the location of the triplet nn* state could well play a critical role in determining the lack of fluorescence in many of these compounds, although its location has not yet been disclosed spectroscopically for2420 EXCITED-STATE BEHAVIOUR OF SUBSTITUTED ANTHRACENE them. Consequently, in order to explain the non-fluorescent properties seen for several carbonyl derivatives, it is reasonable to invoke a very rapid intersystem-crossing process of the type Slnn*nnryTnn, ,which is faster than the radiative one by at least three orders of magnitude.' A previous picosecond-laser study of intersystem crossing in some carbonyl derivatives of anthracene has revealed that triplet-state population occurs on the picosecond time-~cale.~~ 35 Electron-donating substitutents such as NH, shift the nn* level to higher energy,29 making the process Slnn* 'vvYT,,, less efficient.As a result fluorescence can be observed, as is seen in group V compounds. When X is a conjugating group, it does so with the carbonyl group, although the anthracene part itself cannot come into conjugation because of steric hindrance. The triplet nn* level will then be lowered enough to be located below Slnn*. Thus it would be expected that Slnn*-+Tnn* would become fast enough to make the compounds non-fluorescent even at 77 K. However, when Slnn* is already low enough, as is found in 18, the fluorescence process could start to compete with radiationless processes.21 Another important factor in determining the fluorescence efficiency may be related to rotational motion of the carbonyl group upon photoexcitation, as is implied from the photochemical reactions found in compounds of group IV.The restriction of the rotational motion caused by lowering the temperature would serve to shut off the efficient radiationless transition. For group I compounds this appears to play a dominant role in the appearance of fluorescence at very low temperatures. This problem, however, has already been fully discussed elsewhere. Finally, note that most of the compounds presented here deserve further detailed quantitative study, e.g. by picosecond laser photolysis, because of their unique rapid radiationless processes.It would also be interesting to test the present classification of carbonyl derivatives of anthracene for compounds other than those mentioned here. I am grateful to Prof. K. Hamanoue for the use of a Q-switched ruby laser. I also thank Mr Y. Kajiwara for his help in measuring the triplet-state lifetimes using the ruby laser. S. K. Lower and M. A. El-Sayed, Chem. Rev., 1966, 66, 199. H. Giisten, M. Mintas and L. Klasinic, J. Am. Chem. SOC., 1980, 102, 7936. A. Kearvell and F. Wilkinson, Transitions Non-Radat. Mol., 20th Reunion SOC. Chim. Phys., 1969; J. Chim. Phys., special no., 1970, 125. S. Hirayama, J. Am. Chem. SOC., 1981, 103, 2934. K. Hamanoue, S. Hirayama, T. Nakayama and T.Teranishi, J. Phys. Chem., 1980,84, 2074. E. L. May and E. Mosettig, J. Am. Chem. SOC., 1948, 70, 686. T. Matsumoto, M. Sat0 and S. Hirayama, Bull. Chem. SOC. Jpn, 1975, 48, 1659. 'I R. Calas and R. Lalande, C. R. Hebd. Seances Acad. Sci., 1958, 246, 277. !3 P. H. Gore and J. A. Hoskins, J. Chem. SOC., 1964, 5666. lo H. J. Williams, J, Chem. SOC., Perkin Trans. I , 1973, 1852. l2 L. F. Fieser and J. L. Hartwell, J. Am. Chem. SOC., 1938, 60, 2555. l3 E. L. May and E. Mosettig, J. Am. Chem. SOC., 1948, 70, 1077. l4 H. G. Latham Jr, E. L. May and E. Mosettig, J. Chem. SOC., 1948, 1079. l5 T. C. Werner and D. M. Hercules, J. Phys. Chem., 1969, 73, 2005. P. Rona and U. Feldman, J. Chem. SOC., 1958, 1737. J. W. Cook, J. Chem. SOC., 1926, 1282. P. J. Gore, J. A. Hoskins, R. L. W. Lefevre, L. Radom and G. L. D. Ritchie, J. Chem. SOC. B. 1967. 227. T. Matsumoto, M. Sat0 and S. Hirayama, Chem. Phys. Lett., 1974, 27, 237. M. Martynoff, M. Chauvin, M. Grumez and N. Lefevre. BUN. SOC. Chim. Fr.. 1958, 164. T. Matsumoto, M. Sat0 and S. Hirayama, Chem. Phys. Lett., 1973, 18, 563. 'O S. Hirayama, Bull. Chem. SOC. Jpn, 1975, 48, 1127; 2653.S . HIRAYAMA 242 1 22 Th. Forster, Fluoreszenz der Organischer Verbindungen (Vanderhoeck und Ruprecht, Gottingen, 23 D. M. Hercules, Fluorescence and Phosphorescence Analysis (Wiley, Univ. of Tokyo Press, 1966). 24 T. Matsumoto, M. Sat0 and S. Hirayama, Chem. Phys. Lett., 1972, 13, 13. 25 T. Wismonski-Knitlel, G. Fisher and E. Fisher, J. Chern. Soc., Perkin Trans. 2, 1974, 1930. 28 Pill-Soon Song, Quae Chae, M. Fujita and H. Baba, J. Am. Chern. Soc., 1976,9?3, 819. 27 S. Hirayama, Rev. Phys. Chem. Jpn, 1972, 42,49. 2B T. Tamaki, Bull. Chem. Soc. Jpn, 1978, 51, 2817. 28 R. Ditchfield, J. E. Del Bene and J. A. Pople, J. Am. Chem. Soc., 1972, 94, 703. 30 M. V. Alfimov, N. YaBuben, V. L. Glagolev, E. S. Kuyumdzhi, Yu. V. Pomazan and V. N. Shams- 31 M. R. Padhye, S. P. McGlynn and M. Kasha, J. Chem. Phys., 1956, 24, 588. 32 J. B. Birks, Photophysics of Aromatic Molecules (Wiley-Interscience, London, 1970), p. 147. 33 E. C. Lim, L. D. Laposa and J. M. H. Yu, J. Mol. Spectrosc., 1966, 19, 412. 34 S. Murata, C. Iwanaga, T. Toda and H. Kokubun, Ber. Bunsenges. Phys. Chern., 1972,76, 1176. 35 S. Hirayama and T. Kobayashi, Chern. Phys. Lett., 1977, 52, 55. 1951). hev, Opt Spectrosc., 1977, 42, 267. (PAPER 1/1361)