MICHAEL KASHA DISCUSSION OF THE LOWEST SINGLET TRANSITION I N NAPHTHALENE AS A AND REMARKS ON THE HIGHER SINGLET LEVELS * FORBIDDEN TRANSITION A,, - Al, BY H. SPONER AND (THE LATE) GERTRUD P. NORDHEIM Received 18th July, 1950 The absorption system of naphthalene a t 3200-2goo A is interpreted as IAl, -lAlg transition. The infra-red active vibration of 568 cm.-l is tentatively suggested as vibration that makes the forbidden transition allowed in second order. In making this analysis, experimental data on absorption and fluor- escence in the different states of aggregation of naphthalene and new results obtained for heavy naphthalene were used. The 0-0 band of the next higher absorption system at ~ ~ O O - Z ~ O O A is located at 35910 cm.-l and of the third absorption system at about 45100 cm.-l, both for the vapour.A brief com- parison with theoretical results is given. The near ultra-violet absorption of naphthalene has been studied in the vapour phase,b 2 in liquid 2, 3 and in rigid glass solutions,* and in the solid state 5 (crystal at zoo K). The spectrum consists of two absorption regions lying fairly close together. The long wave part ranging from about ~ ~ O O - Z ~ O O shows in the vapour sharp narrow bands, whereas the bands from 2900-2500 A are broad and diffuse. Moreover, the pattern in which the bands appear is different in the two regions. The measurements in liquid solution give an average extinction coefficient E of 200-300 for the first region and of 5000-6000 for the second region. Furthermore, at 2200 a very strong and broad band with little indication of structure was observed 2 .6 with E N IOO,OOO. The absorption spectrum of the solid * This research was supported by the Office of Naval Research. 1 Henri and de Laszlo, Proc. Roy. SOG. A , 1924, 105, 662. de Laszlo, 2. physik. Chem., 1925. 118, 369. For example, Morton and de Gouveia, J . Chem. Soc., 1934, ~ I I ; Mayneord Kasha and Nauman, J . Chem. Physics, 1949, 17, 516. Klevens and Platt, J . Chem. Physics, 1949, 17, 479. and Roe, Proc. Roy. Soc. A , 1935, 152, zgg. 6 Prjkhotjko, J . Physics, U.S.S.R., 1944. 8, 257.20 SINGLET TRANSITION IN NAPHTHALENE contains discrete relatively sharp bands, but looks different from the vapour spectrum whether obtained with natural or with polarized light. Fluorescence of naphthalene has been studied mostly in liquid and solid solution^.^^ It consists of a number of bands in the region 3000-3650 A.Fluorescence of the vapour was excited 8 by irradiation with wavelengths below 3000 A showing narrow bands between 3000 and 3340 superim- posed on a continuous background. Fluorescence and absorption bands coincide in the overlapping region. The fluorescence of crystalline naph- thalene 0 (at zoo K) consists of a complicated spectrum of many, mostly narrow, bands. From studies of the fluorescence lifetime of naphthalene in rigid glass solution Kasha and Nauman 4s concluded that the weak long wavelength absorption is a forbidden transition. Calculations of the electronic terms of naphthalene have been carried out with the bond orbital method taking into account valence structures with no or one I ‘ long ” bond.1° With this approximation one obtains as the first two excited singlet states a B,, and a B,, level. Transitions from the symmetrical ground state to the B,, state are allowed with the transition moment in the long molecular axis (x axis), and to the B,, state with the moment in the short (y axis).Earlier more complete cal- culations by Blumenfeld, l1 taking all 42 canonical structures into account, give for the first excited level a symmetric one, lA1#. The transition from the ground state to this level is forbidden. Blumenfeld obtains as next excited levels a lBSu and a lBg, term. Craig la also using the valence bond method, has likewise predicted an A,, as lowest excited singlet level, followed by a lBsu, lB1,, 1B,, term, in that order.Platt’s 13 one-dimensional free electron picture leads to the same order of the allowed levels as the bond orbital method, whereas calculations with the standard molecular orbital method (German, l4 Coulson,ls Davydov 16) and a slightly modified molecular orbital method (Simpson 17) reverse this order. In this approach the symmetric level should lie at higher excitation energy. In the most recent treatment of the energy states of naphthalene by Jacobs 1* an antisymmetriced molecular orbital calculation was carried out, taking into account the interaction between the different configurations formed by the molecular orbitals. In this approach a B,, and a B,, are obtained as lowest singlet levels with a very small separation between them.The first excited lAl, state appears 1’4eV above the B,, state. Oscillator strengths were calculated for the first three allowed transitions and compared with experimental values omitting the first weak absorption at 3zoo-2gooA. The agreement is rather good. Analysis of the Lowest Electronic Singlet Transition in Naphthalene at 3200-2900 A.-(a) OCCURRENCE OF 0-1 and 1-0 BANDs.<onsidering the weakness and structure of the first absorption in naphthalene, and the lifetime of the fluorescence, this system can be explained best as a transition forbidden by symmetry. Possibilities of interpreting the low intensity by cancellation of positive and negativz terms of the same order 7 Compare P. Pringsheim, Fluorescence and Phosphorescence (Interscience Publishers, New York, 1949)~ 408.Pringsheim, Ann. Acad. Warsaw, 1938, 5 , 29. Obreimov and Shabaldas, J . Physics U.S.S.R., 1943, 7, 168. Remarks of M. Kasha a t O.N.R. Spectroscopy Conference, Jan., 1948. lo Sponer and Nordheim, ONR Contract N60ri-107, T.O. I (Annual Report, l1 Blumenfeld, J . Physic. Chem. U.S.S.R., 1947, 21, 529. l2 Craig, Proc. XIth Int. Congr. Chem. (Lond.) (July, 1947). l3 Platt, J . Chem. Physics, 1949, 17, 484. 1* German, J . Physics U.S.S.R., 1944, 8, 276. l6 Coulson, Proc. Physic. Soc., 1948, 60, 257. l6 Davydov, J . Expt. Theor. Physics U.S.S.R., 1947, 17, 1106. Simpson, J . Chem. Physics, 1948, 16, 1124. l8 Jacobs, Proc. Physic. SOC. A , 1949, 62, 710. June, 1948)-H. SPONER AND GERTRUD P. NORDHEIM 21 in the transition moment integral have also been mentioned.lol 1 7 If the long wavelength system arises from a transition forbidden by symmetry, it is obvious that absorption and fluorescence bands should show the same charactcristic difference as the near ultra-violet transition in benzene, i.e. a strong 0-1 band and a w-eak 1-0 band in absorption and a weak 0-1 and a strong 1-0 band in fluorescence.Experimental data on ab- sorption and on fluorescence in the different states of aggregation w2re consulted and the interpretation chosen is presented in Table I. There is no doubt that the narrow band group at 3080.1-3081.3 A in the vapour represents the 0-1 transition. It is very strong in absorption and weak in fluorescence. The band at 3080.4 A or 32454 cm.-l is taken as main band.Likewise, the band at 3172-4-3173'4 A occurs strongly in fluorescence and weakly in absorption and is here considered as 1-0 transition. Already Pringsheim * noticed these intensity ratios which would be in agreement with a forbidden transition. 31513 cm.-l is taken as main band of the 1-0 transition. The intensity ratio of the two bands as taken from absorption plates obtained by Cooper la is 1/20. From spectrogram No. 169 in the catalogue of ultra-violet spectrograms a ratio of about I to 30 may be estimated for the solution.20 These ratios will be used later in the evaluation of the 0-0 band. As seen in Table I the band positions in the solution are shifted to the red by about 310 cm.-1 as compared to the vapour. For rigid glass solutions Kasha and Nauman's values have been used.TABLE I.-0-1, 1-0 AND 0-0 BANDS OF LONG WAVELENGTH SYSTEM OF NAPHTHALENE Bands 0-1 1-0 Separation 0-0 Vapour abs., ref. I fluor., ref. 8 32454 cm.-l 31513 crn.-l 941 cm.-l (32080 cm. -l) Liqu. Solution (hexane) abs., ref. I, 20 fluor., ref. 7 32145 crn.-1 31173 cm.-l 972 crn.-l (31757 cm.-l ?) Rigid Glass abs., ref. 4 fluor., ref. 4 32 104 cm. -l 31144 cm.-]. 960 crn.-l 31680 crn.-l Crystal abs.. ref. 5 fluor., ref. g --- - 31062 cm.-1 ? (31062 cm.-1 ?) - I 1 The corresponding assignments in the crystal spectra are much more difficult to achieve. These spectra are extremely complex and not too well understood. Four different electronic transitions were assumed in the absorption analysis and three were used for the explanation of the fluores- cence spectrum.Q The absorption spectra were taken with natural light and with polarized light where the polarization vector was either parallel to the u or b axis of the crystal.In the long wavelength part there are bands which appear weakly or not at all ir, one case and strongly in the other, but above 33400 cm.-l the spectrum is strong only when the polariza- to the a or b axis 21 of the crystal. In the long wavelength part there are bands which appear weakly or not at all in one case and strongly in the other, but above 33400 the spectrum is strong only when the polariza- tion vector is parallel to the a axis and it is practically missing for the other position of the Nicol. In spite of the excitation by polarized light as means of differentiation, we found it not possible to obtain a satisfactory analysis.It was expected that, if the analysis presented here is valid for the vapour and solutions, it would also, at least in its main features, apply to the solid. However, the characteristic features may become obscured because inter- 19 Cooper, Master Thesis (Duke Univ., 1948); 2O American Petroleum Institute Research Project 44 at the National Bureau of Catalogue of UEtra-violet S$ectrograms, Serial No. I 69, Naphthalene Standards. liqu. sol.).22 SINGLET TRANSITION IN NAPHTHALENE action with crystal forces (the molecules are arranged in skew packing "') may cause the breakdown of the selection rule. It will also lead to splitting of the energy levels 22 and the spectrum will look more compli- cated than that in the vapour phase.For naphthalene, the fluorescence spectrum is less complex than the absorption spectrum. Obreimov and Shabaldas stated that the fluorescence spectrum has a sharp beginning with a line at 31062 cm.-l and they succeeded in interpreting all stronger lines with the use of known Raman frequencies belonging to totally symmetric vibrations. Small frequencies u p to about IOO cm.-1 were identified with lattice vibrations found in the Raman effect of naphthalene crystals. 23 From the previous discussion it would seem logical to assign the narrow band at 31062 cm.-l to the 1-0 transition in the crystal spectrum. How- ever, the situation becomes confused when considering the absorption spectrum. The groups in the spectrum do not stand out as clearly as in the fluorescence spectrum and the overlapping of the second absorption region is adding to the complexity of the spectrum.A new analysis by Prikotjke 24 was not available to the authors except a very brief abstract in Chem. Abstr. in which no mention is made of any figures. Interpreting the band 31062 as o--o band would account for the appearance of this band in fluorescence and absorption but it does not lead to a satisfactory analysis of the spectrum. It would also give a very large red shift as compared to the vapour and solution spectra. Work on the spectrum of the crystal is being continued. (b) FIXATION of 0-0 BAND.-The separation between the 0-1 and 1-0 bands in the different states of aggregation are also included in Table I. This separation represents the sum of the vibrational frequencies in the lower and upper electronic state that makes the forbidden transition allowed.Assuming that it is a transition Alg-Alg, it is clear that excita- tion of a ps, vibration will produce a transition moment in the long molec- ular axis and a pZu vibration a moment in the short axis. A moment perpendicular to the molecular plane (along the z axis) could be produced by excitation of a plu vibration. Fixation of the 0-0 band is not easy. The three types of vibration which can produce the necessary moment are Raman inactive. Infra-red measurements are known in the literature Z 5 only above 600 cm.-l. It seems more likely that the separation 941 cm.-l would involve a vibration below that value. Dr. E. K. Plyler and Miss Mary A.Lamb have very kindly taken the infra-red spectrum of naphthalene in solution from 4000 to 360 wave numbers at the National Bureau of Standards. Their absorption curves show four frequencies in the critical region : 618 (m), 568 (w), 476 (vs), 361 (w).* Of these, as far as numerical values are concerned, 568 and 476 seem most plausible. If a 50 yo drop of the frequency in the upper state can be assumed, then the 618 is also eligible. Using these values with the separation 941 cm.-l in the gas (although they have been obtained from solution data), this would give an upper state frequency of 323, or 373 or 465 cm.-l respectively, and correspondingly, would place the 0-0 band at 32131, or 32081 or 31989 crn.-' Although faint and narrow bands have been measured by de Laszlo in all three places, this has not much significance.Cooper, who studied the absorption in dependence of pressure and not of temperature, did not 21 Robertson, Proc. Roy. SOC. A , 1933, 142, 674 : Abrahams, Robertson and White, Acta Cryst., 1949, 2, 238. 22 Davydov, J. Expt. Theor. Physics, U.S.S.R., 1948, 18, ZIO. Gross and Vuks, J . Physique Rad., 1936, 4, 113 ; Nature, 1929, 124, 692. 24Prikhotjko, Nut. Acad. Sci. U.S.S.R., Ser. Phys., 1948, 12, 499, reviewed in Chem. Abstr., 1950, 4, 433. 25 For example, Lambert and Lecomte, Ann. Physique, 1933, 18,329 ; Barnes, Gore, Liddell and Williams, In@-red Spectroscopy (Reinhold, New York, 1944), P. 5:. m = medium, vs = very strong, w = weak,H. SPONER AND GERTRUD P. NORDHEIM 23 observe a band in any of these positions.The solution spectrum obtained by de Laszlo shows a peak at 31757 crn.-l which divides the distance of 970 into 584 + 388 cm.-l, but a similar peak has not been obtained by other authors (for example, ref. 20). The interpretation of the 31757 cm.-l peak as 0-0 band must be considered as uncertain, particularly in view of its intensity. Examining the data obtained in rigid glass solution, it is seen at once that the 0-0 band does occur weakly in absorption and in fluorescence. Interaction with the glass medium causes a breakdown of the selection rule. Unfortunately, the measurements are accurate to only f 30 cm.-l. Actually, the two 0-0 bands deviate by about IOO cm.-l from each other. In the absorption curve the distance from the 0-0 band to the next very strong band is 367 cm.-l, and the corresponding distance in the fluorescence curve is 482 cm.-1.In Table I the values for the 0-1 and 1-0 bands (rigid glass) are the wave numbers of the strongest bands in the spectra and are more reliable than the weak o---o peaks. With this in mind and assuming that the ratio of the distance 367 and 482 has some real significance, the position of the o---o peak is estimated at about 31680 cm.-l which is the averaged value reported in Kasha and Nauman's paper. No additional information can be obtained from the spectra of crystal- line naphthalene as was indicated before. (c) SUPPORT OF THE ANALYSIS AND DETERMINATION OF THE INTER- ACTING VIBRATION FROM THE SPECTRUM OF HEAVY NAPHTHALENE.- Although the preceding analysis (intensities, numerical relations, Boltnnann factor from intensity ratio of 0-1 and 1-0 bands) points to a vibration in the neighbourhood of 600 cm.-l rather than to the strong 476 cm.-l as interacting vibration, the latter cannot be definitely excluded because of the only roughly known positions of absorption and fluorescence peaks in liquid and solid solutions. To facilitate the assignment, a sample of completely deuterated naphthalene was prepared by L.Corrsin * in this laboratory and the ultra-violet absorption spectrum was taken by C. D. Cooper. The first noticeable feature was the greater diffuseness of the spectrum as compared to that of light naphthalene. There was probably a small percentage of not completely deuterated compound in the sample which would produce close lying bands and somewhat obliterate details.However, the edges on the ultra-violet side of the bands would correspond to the completely deuterated substance. It was found that the 0-1 band in heavy naphthalene is shifted by only about IOO cm.-l to the violet as compared to ordinary naphthalene while the corresponding shift in benzene is 180cm.-l. (The shift in mono-deuterobenzene is 31 cm.-1). The shift to the violet results from zero point vibrations. The separation between the 0-1 and 1-0 bands is about 910 cm.-l in heavy naphthalene as compared to 941 in the light compound. The infra-red frequencies were taken from curves obtained again by Dr. E. K. Plyler and Miss Mary A. Lamb at the National Bureau of Standards using our samples in solution.In the region pertinent to our dicussion these vibra- tional frequencies were observed : 594 (m), 566 (vw), 541. (w), 422 (vw), 403 (s), 328 (w). It seems reasonable to assume the following correspond- ences between light and heavy naphthalene vibrations : unlikely that the vibration will have a much larger frequency in the excited Light Heavy 618 (m) 594 (4 568 (w> 541 (4 476 (4 403 (4 361 (w) 328 (w) The separation of 910 cm.-l rules out the participation of the 403 vibration in the mechanism of making the A,,-A,, transition allowed because it is * A paper by L. Corrsin is in the course of preparation.24 SINGLET TRANSITION IN NAPHTHALENE electronic state. For 594 and 541, the ratio of the vibrational frequencies in the two states would be about the same for both naphthalenes.Measure- ments in rigid glass solution agree better with using the 568 than the 618 cm.-1 vibration, but the wide limits of accuracy in these spectra do not entirely exclude the higher value. A consideration of the symmetrics of these vibrations may permit a further selection. The question to which symmetry class the interacting vibration belongs is identical with the question in which direction the “ forbidden ” transition is polarized. Firstly, it is expected and assumed that this vibration is a carbon vibration. It is furthermore assumed that the vibrational transition moment will be in the molecular plane and more specifically, in the long axis ( x ) since the next allowed transition (Alg-B3,J should be polarized in the x direction. This makes the interacting vibra- tion one of type pSu.Either vibration (618 or 568) could belong to this type, From intensity arguments the 618 seems a better choice.* The large drop of almost 50 yo in wave number in the upper electronic state would have to be explained. Although a drop of this magnitude was assumed to occur in benzene 2E for the p2, carbon vibration, it is surprising for the p3% mode. Therefore, until further support in favour of the 618 vibration has been found, the 568 vibration is used here tentatively as the interacting vibration. While this refers to the strong part of the long wavelength system, it is quite possible that weak bands which do not fit in this scheme result from a vibrational moment in the short axis ( y ) and involve a vibration. In Table I1 there are collected the main vibrational frequencies occur- ring in the spectrum produced through interaction of a ,€Isu vibration.All are interpreted as carbon vibrations. Remarks on the Higher Electronic Singlet Levels-As mentioned in the introduction, a second absorption region occurs in naphthalene at zgoo-2500 A. The beginning of this absorption (0-0 band) was given in Kasha and Nauman’s paper as 33736 cm.-1. This value was taken from Prikhotjko’s absorption measurements in the crystal. Klevens and Platt place the onset of the second absorption in solution at about 35,000 cm.-l. From a careful study of our own plates we have chosen the strong broad vapour band at 35910 cm.-1 as 0-0 bandln of a second ab- sorption system. It seems impossible to correlate this value with the 33736 in the solid because of the too large shift. In looking for a possible lower transition we had noticed on our plates a rather strong group of bands located at 34060-33940 cm.-l with the bands spaced about 40 cm.-l apart.They look different from the other naphthalene bands, H group of Henri and de Laszlo appearing strongly in the vapour and in solution. A corresponding group in heavy naphthalene is missing. Dr. Corrsin suggested the group might be due to an impurity which was removed with the sulphonation in the preparation of heavy naphthalene. After re- fluxing the ordinary naphthalene with sodium, the peculiar band group had disappeared.? The spectra lcck now more uniform, and it has been possible to extend the analysis of the lAl,-lAlo system a little further toward shorter waves, and of the second stronger system to longer waves from the 0-0 band 35910 cm.-1.The second system is considered to be the predicted 1A lg-lB 3u transition. Although measurements in the spectrum of heavy naphthalene have not been completed, it may be mentioned here that a preliminary calculation indicates a shift of about *Tentative assignments of the infra-red active vibrations of the two naphthalenes will be published by L. Corrsin soon. 2E Garforth, Ingold and Poole, J. Chem. SOC., 1948, 491. -f Naphthalene which had been sublimed and recrystallized several times still gave a positive test for sulphur before refluxing with sodium. Dr. Corrsin’s suspicion that the impurity might be benzothiophene was recently verified spectroscopically by C.D. Cooper. 3 These results will be published with C. D. Cooper elsewhere.H. SPONER AND GERTRUD P. NORDHEIM 25 130 cm.-l of the 0-0 band in the second system of heavy naphthalene as compared to light naphthalene. The principal vibrational frequencies found in the transition A ,,-B,, are included in Table 11. Except for 930, they are considered as belonging to carbon vibrations. Coming back to the absorption spectrum of the crystal, photograph 4a, Plate I in Prikhotjko’s paper suggests from the intensity distribution of bands that the region 33445-36000 cm.-l may consist of two different transitions. It is tentatively suggested here that the state B,, splits in the crystal into two states, one covering a spectral transition from about 33440-34500, and the other from about 34500 to above 36000 cm.-l.The third absorption region of naphthalene occurs at 2220 A-2000 A (~~ooo-~oooocm.-~). It is more than 10 times stronger than the second. Structure in this system was obtained in the vapour on u.-v. sensitized plates. in this region but shows in addition a few fainter bands on the long wave side which are lost in the sharp rise of the absorption curve in the liquid. It agrees with the general contours of the spectrum of the liquid TABLE II.-VIBRATIONAL FREQUENCIES (CM. -l) OCCURRING IN THE FIRST Two SINGLET SYSTEMS OF NAPHTHALENE. LONG WAVELENGTH SYSTEM Alg-A1, Ground State 1 Excited State 1 Raman I i I- (373) 701 474 I 997 I SHORTER WAVELENGTH SYSTEM AI0-B3, Symmetry of Vibration 512 762 940 1380 I022 Although work on this system is still in progress, a vaIue of 451oocm.-l mav be suggested as possible location of the o--o transition.Comparison with Theory .-The occurrence of a forbidden transition lAlg-lAII as lowest electronic transition in the naphthalene molecule is in agreement with calculations based on the valence bond method.ll9 l2 The tentative position of the o--o band (Table I) is at 3116 (32080 cm.-l), while Blumenfeld calculated 3193 and Craig’s value is 2830 A. The good agreement with Blumenfeld’s value is probably fortuitous. The analysis presented here is also consistent with results obtained from Platt’s free electron model. The location of the 0-0 band fits well on his curve for the lL, transition in polyacenes.6 Both the valence bond and molecular orbital method agree on a B,, level as upper state of the first allowed transition.From these predictions and not from the analysis, the 35910 band was identified with the 0-0 transition of the system 1A1,-1B3u. A distinction between a Bzu and B,, level from the spectrum is difficult as essentially the same vibrational structure may be expected in the two transitions. The close agreement of the position of our lBQu level (2784 A) with Blumenfeld’s calculated position is again more or less fortuitous. Craig’s calculations place it somewhat lower at 2665 A. There seems, however, to be disagreement26 PYRIDINE HOMOLOGUES with the latest results obtained with the molecular orbital method. Cal- culations of Jacobs would place the B,, level at about 3420 A, that is, in the neighbourhood of the forbidden absorption system. These calcula- tions further predict a B,, level very close to the B,, state. Transitions to these levels from the ground state would give overlapping spectra. Careful search was made for a second system between the long wave forbidden spectrum and the allowed system starting at 35910 cm.l This seemed important in view of the appearance of the crystal spectrum in that region, as was discussed before. Now it is true that there are a number of unexplained bands in the critical region of both transitions, but almost all the stronger bands (or rather groups of narrow bands) were found to fit into one of the two systems. There are also a very few fainter bands which look different from the other naphthalene bands in either system and which, remembering the fate of the H group (notation of Henri and de Laszlo), might still belong to an impurity. Taking all this into account, the existence of another system in the transition region between the two identified systems cannot be excluded until all the bands have been accounted for, but this existence is not considered very probable. The third, very intense transition is polarized in the long axis as was shown in previous experimental The authors would like to express their indebtedness to Dr. E. K. Plyler and Miss Mary A. Lamb of the National Bureau of Standards for taking the infra-red spectra of light and heavy naphthalene in solution. and theoretical research. Department of Physics, Duke University, Durham, North Carolina. 27 Jones, Chem. Rev., 1947, 41, 353.