838
LESLTES.FORSTER 9h'u DANIELDUDLEY
Vol. 66
shown by the large decrease in A S * of both reactions on going from quinoline to 8-methylquinoline. Comparing lines 1 and 4 of Table I1 it will be seen that the AH* values of both reactions decrease on going from aniline to o-toluidine. This is consistent with the increase in nucleophilicity of the amine as a result of the +I effect of the methyl group. At the same time a rather large decrease takes place in the entropy of activation of both reactions, indicative of the strong steric effect of the ortho substituent. In all these solvents the AS" values for the cinnamalmalonic acid reaction are considerably smaller than are those for malonic acid, corresponding with the greater .size of the derivative. In every case, also, the cinnamalmalonic acid reaction has a lower enthalpy of activation, in line with its greater -24.0 -20.0 - 16.0 - 12.0 -8.0 -4.0 0 +4 acidity. AS* (e.u./mole). The abnormally high value of AH* of the cinFig. 1.-The enthalpy-entropy relationship for the decarboxylatiqn of malonic acid and cinnamalmalonic acid in namalmalonic acid reaction in X,N-dimethylvarious amines. Symbols: 0,cinnamalmalonic acid in aniline (line 5 of Table 11) suggests the possibility cinnamalmalonic acid in quin- that it is the mono-anion, and not the un-ionized aniline and its derivatives; 8, oline and 8-methylquiuoline; (D, malonic acid in various di-acid, which is involved here in the rate-deteramines. mining step. Since the anion would exhibit less are those for malonic acid, a reflection of the greater association than the di-acid, the reactive species would be smaller, and hence the AS" of the reacbulk of the malonic acid derivative. The AH* values for both the malonic acid and the tion should have a higher positive value, as is incinnamalmalonic acid reaction are about 2 kca1.l deed the case. mole lower in 8-methylquinoline than in quinoline, Acknowledgments.-(1) The support of this rea fact which is consistent with the increased nucleo- search by the National Science Foundation, Washphilicity of the quinoline derivative (see lines 2 and ington, D. C., is gratefully acknowledged. (2) 6 of Table 11). The strong steric effect of the The synthesis of the cinnamalmalonic acid was methyl group in the 8 position of quinoline is carried out by Donald McCoy.
THE LUMINESCENCE OF FLUORESCEIN DYES' BY LESLIES. FORSTER AND DANIELDUDLEY Department of Chemistry, University of Arizona, Tumon, Arizona Received September 23"1951
The fluorescence and phosphorescence yields and triplet state lifetimes have been determined for a number of halogenated fluorescein dyes in EPA solutions, The lifetimes decrease progressively as the number of halogen atoms is increased in both the bromine and iodine series of derivatives and the lifetime always is less in the iodine than in the corresponding bromine derivative. The results indicate that the diminution of the fluorescence yield that accompanies bromine and iodine substitution is not due primarily to an increase in intersystem crossing but rather to increased internal conversion from the excited singlet state to the ground state.
Introduction Considerable evidence has been accumulated, since the pioneering study of Lewis and Kasha,2 to support the view that a general energy level scheme is applicable to organic molecules.s Excitation to a high vibrational level of the lowest excited singlet state, S', is followed by the rapid radiationless removal of this vibrational energy by the surroundings. The short-lived (10-8-10-9 see.) state, s', then decays radiatively, S' --c S (k2), or nonradiatively s' + S ( h ) to the singlet ground state, S, or alternatively to the lowest tri let state, S' + T(lc4), The relatively long-livef triplet state
(2 sec.) then is converted radiatively, T + S (Tcs), or non-radiatively T + S (Ice) to the ground state. Non-radiative transitions between electronic states generally are termed internal conversions, although if a change in multiplicity is involved, the term intersystem crossing may be used. Considerable interest has been manifested in the effect of environment on these internal conversion processw, but little quantitative information has been acc~mulated.~It has been shown that halogen substitution increases the efficiency of 8' --c T in the order I > Br > C1 > H.6 This presumably is due to the increased spin-orbit coup-
(1) Supported by a grant from the Atomic Energy Commission. (2) G. N. Lewis and M . Kasha, J . Am. Chem. Soc., 66, 2100 (1944). (3) M. Kasha and S. Moalynn, Ann. Rev. Phys, Chen., 7 , 403 (1956).
(4) H. Sponer, Radiation Research, S u p p l . 1, 558 (1969); E. J. Bowen, Discussions Faraday Soc., 27, 40 (1959); G. Jackson, R. Livingston, and A. Pugh, Tvans. Faraday Soc., 66, 1636 (1960). (5) M. Kasha, Radiation Reaearoh, S u p p l . 2, 243 (1960),
May, 1962
THELUMINESCEWE OF FLUORESCEIS DYES
ling that accompanies the substitution of atoms with large 2. However, kinetic evidence has been obtained which indlicates that the reduction in fluorescence efficiency (S' += S) that accompanies the halogenation of fluorescein dyes is not primarily ' ' but rather of inthe result of increased S' += 1 creased non-radiative s' + S.6 It is the purpose of the present work to present further evidence bearing on this problem. Experimental The dyes were purified by paper chromatography.' Geometric isomers were not separated. The bromine derivatives were identified spectrophotometrically.* The absorption maximum shifts to longer wave lengths aa the number of bromine atoms is increased. The iodine derivatives were identified by assuming the same shift with increasing iodination. This shift is correlated with decreasing migration velocity on the paper chromato ram. The spots were cut out and the dyes eluted with 5k ammonia. The samples were prepared by evaporation of the water followed by addition of EPA to the dry salt. The samples were protected from light as m w h as possible during this procedure to minimize photodeicomposition. The samples were contained in a 1-cm. cuvette which was placed in a cryostat that was a modification of a previously described d e ~ i g n . ~The 4358 A. excitation was obtained from a filtered AH-4 lamp and the emission at 45' was passed through a Farrand grating monochromator to a cooled Dumont 6911 photoniultipler tube. The output of this tube was fed into a recorder coupled to the wave length drive of the monochromator. The monochromator-photomultiplier system was calibrated against a standard tungsten lamp. I n the 510-580 mp region the calibration curve so obtained agreed ( & l o % ) with the calibration curve determined by utilizing quinine sulfate in conjunction with the known emission distribution for this substance.10 The dye concentrations were kept in the 10-5-10-6 M range to minimize reabsorption, sglf-quenching, and dimerization. The absorbance a t 4358 A. was measured a t 25" in a 10-cm. cell and the absorbance in 1-cm. path length computed from this measurement. By using the above procedure and assuming a value of C ~ F = 0.92 for the fluorescence yield of aqueous fluorescein dianion, @F = 0.18 was obtained for the aqueous eosin dianion. This can be compared with the directly measured value of 0.19." The phosphorescence lifetimes, T , were measured in EPA glasses at -183" by the flash technique. A Dumont 6911 photomultiplier aho was used in this apparatus to increase sensitivity in the red and near infrared regions.
839
The results are given in Table I. The emission spectra can be grouped into three classes. FlClz and F1 exhibit negligible phosphorescence; FlBr4 and F1Br4C14have poorly resolved phosphorescence bands; while the remaining compounds show well resolved phosphorescence bands at - 183'. The phosphorescence yields, @p, of FlBr4 and FlBrdC14 consequently are less reliable than those of the other dyes. The light absorption was not measured ~ T can be conat -183' and only x = @ p / @ and sidered as primary da:a at -183'. The fluorescence yields, @F, are the primary data at 25'. The @p then can be computed by assuming @F is independent of temperature. This assumption has been verified for F1, F1Br4, and FIIa dianions and presumably is valid for the other halogenated fluoresceins.l4 It has been shown that oxygen has no effect on the fluorescence of fluorescein,12but it cannot be assumed that the phosphorescence also is unaffected by oxygen, even at low temperature and in a rigid medium.6 We compared the phosphorescence of a carefully degassed eosin solution with that of an air-saturated solution and found no quenching by oxygen. However, the room temperature photodecomposition of alcoholic solutions of eosin is reduced considerably in the presence of oxygen15 and we found that more reliable results could be obtained by the use of airsaturated solutions. TABLE I LUMINESCENCE OF DIANIONSOF FLUORESCEIN DYES IN EPA
Dye
Xmax-
phosXmax
fluorescence (mu,
phorescence (ma -183')
25O)
x =
cP~a
(25')
*P/@F
(-183')
7
(msec. -183')
..
X/T
(rel.)
F1 527 .. 0.83 0 *. FlBr 535 634 .60 .13 50 22 FIBrz 540 650 .29 .21 44 48 F1Br4 549 (690)b .40 ,082 9.4 86 FlI 531 642 .15 .67 15.8 420 FlIS 544 667 .054 1 05 10.4 1000 F1IS 549 671 ,061 0.71 5 . 1 1400 FlI, 560 686 ,066 .40 1 . 3 3100 F1Br4C14 572 (660)* .56 .15 5 6 270 FlClz 538 .. .79 0 .. .. a Referred to aqueous fluorescein dianion @F = 0.92. Phosphorescence band insufficiently resolved to determine maximum accurately.
Results A meaningful interpretation of dye luminescence data can be malde only if the emitting species can be identified. I n aqueous soluiion fluorescein can exist as the cation, mono- and dianions, and neutral molecule. Only at a sufficiently high p H is the dianion the predominant species.12 The absorption and emission spectra for each dye are, except Examination of Table I reveals the following for a small solvent shift, very nearly the same in trends: (i) r diminishes progressively with haloEPA as in the corresponding ammoniacal solution. genation in both the bromine and iodine series. This fact coupled with the observation that di- This is to be contrasted with the results for naphmerization of fl'uorescein and eosin dianions be- thalene derivatives, where no simple correlation comes appreciable only when the concentration ex- was fo~nd1~~1'; (ii) the lifetime of a given bromine ceeds M,13 leads to the conclusion that in each derivative is 4-7 times as long as the lifetime of the case the predominant species is the dianion. corresponding iodine derivative; (E) x / r increases (6) A. Adelman and G. Oster, J . Am. Chem. Soc., 7 8 , 3977 (1956). regularly with halogenation in both the bromine and (7) E. Lederer and M. Lederer, "Chromatography," Elsevier, Amsterdam, 1951, p. 236. (13) T. Fdrster and E. Konig, Z. Elektrochem., 61, 344 (1957). ( 8 ) C. Graichen and 6. C. Molitor, J . Assoc. Ofic. Agr. C'hemasts, 4a, (14) V. Zanker and H. Rammensee, Z. physzk. Chem. (Frankfurt). 149 (1959). 16, 168 (1960). (9) T. 0. Jones and J. Willard, Rev. Sei. f n s t r . , 27, 1037 (1956). (10) W. Melhuish, J . Phys. Chem., 64, 762 (1960). (111 G. Weher and F. Teale, Trans. Faraday Sac., 5S, 646 (19571, (12) L. Lindyvist, Atrktu. K E V W16, , 79 (1960),
(15) M. Imamura and M. Koizumi, Bull. Chem. Sac. J a p a n , 29, 899 (1958). (16) D. S. MdCluid, J . Chem. P n y ~ . ,17, 905 (17) V. h m o l a e t ttiid A. Terstsiri, 8. chim. phys., 55, 698 (1948).
(1649).
840
D. M. MARCHAXD T. HENS~IALL
iodine series and is 20-40 times larger for a given iodine derivative than for the bromine analog. Discussion The number of independent data determined in this study are insufficient to determine unambiguously the effect of halogenation on the individual rate constants kz-lc6. However, from the several trends indicated above some meaningful conclusions may be drawn. From the relation T = l/ka -Iks it can be seen *hat kg k6 increases with halogenation and is larger for the iodine derivatives than for the bromine derivatives. The increased spin-orbit coupling due to iodine would increase 4. The effect of such substitution on k6 cannot be determined from the lifetime data alone, but there is evidence that chlorine substitution on naphthalene increases both ks and ks by a factor of five.18 The integrated absorption, a measure of the natural lifetime of S’, T~ = l/kz, varies by less than 30% within the group of halogen derivatives in aqueous The values of T~ computed from @F = 7/70 vary by a factor of two between ff uorescein and e0sin.1~ In the following discussion we will assume that k2 is nearly the same for all of the dyes in EPA solution. The intersystem crossing ratio, x, sometimes is nearly equal to This is not necessarily true in the case of the fluorescein dyes where we cannot he certain that 4 Br > I. The kinetic results have shown that in aqueous solutions the decrease in 4% is due to the increase of k8. The parallel trends suggest that a similar situation prevails in EPA solutions. This conclusion is supported further by examination of the trends within the group of iodine derivatives. @F = k2/(k2 ka l e d ) and since @Fis about 0.1,1/@~ E (ka Ic4)/lcz. k&s increases progressively with the number of iodine atoms, while @F does not decrease in a similar progression. It would seem that if IC4 >> ks such a trend would exist. One final point must be emphasized. The decrease in @F is not accompanied by an increase in @p. The sum @pp @F decreases with substitution in the order H = C1 > Br > I, therefore at least one of the non-radiative processes, S’ -t S or T -+ S, becomes more important in this sequence.
++ +
+
THE KINETICS OF CYCLIZATION OF SOME %,a’-DIPHENIC ACIDS IN SULFURIC ACID BY D. M. MARCH‘ AND T. HENSHALL Department of Chemistrg, Sir John Cass College, London, England Received Septembsr 66,l 9 S i
The kinetics of cyclization of some 2,2’-diphenic acids in concentrated sulfuric acid have been studied using a spectrophotometric method of analysis. Simple first-order kinetics were found in each case, and also a linear dependence of log k on the Hammett acidity function. The energy and entropy parameters were found to vary in a characteristic manner. A mechanism is proposed and is discussed.
The kinetics of cyclization of 2,2‘-diphenic acid to fluorenone-4-carboxylic acid2 have been studied in aqueous sulfuric acid, and in sulfuric-acetic acid mixtures over the ranges 77-100y0 and 60-100% w./w., respectively. The measurements also have been extended to certain 5,5’-disubstituted-2,2‘-diphenicacids con(1) Material taken from a thesis in partial fulfillment of the requirements for the Ph.D., London University, 1961. (2) C, Qraebe and C. Aubin, Ann., 247, 261 (1888).
taining substituents of well defined electronic character; these were studied in the aqueous medium only. In every case, a linear dependence was found of log k on the Hammett acidity function Ho but the slopes differed from unity. The activation energies and entropies have been found to depend markedly on the solvent composition; a probable interpretation is offered (Discussion).
