Effect of Dimer Formation of the Electronic ... - ACS Publications

Aug 24, 1973 - R. W. Chambers, T. Kajiwara, and D. R. Kearns. Effect of Dimer ... Richard W. Chambers,' Takashi Kajiwara, and David R. Kearns*. Depart...
0 downloads 0 Views 895KB Size
R . W. Chambers, T. Kajiwara, and D. R. Kearns

380

Effect of Dimer Formation of the Electronic Absorption and Emission Spectra of Ionic Dyes. Rhodamines and Other Common Dyes Richard W. Chambers,’ Takashi Kajiwara, and David R. Kearns* Department of Chemistry, University of California, Riverside, California 92502 (Received August 24, 1973) Publication costs assisted by the

U.S. Public Health Service

The visible absorption, excitation, fluorescence, and phosphorescence spectra have been measured for the monomeric and dimeric forms of rhodamine B and sulforhodamine B (Figure 1).Dimer fluorescence and phosphorescence spectra were assigned through the use of excitation spectroscopy, and polarized excitation spectra were used to determine the polarization of the dimer absorption bands relative to the dimer fluorescence and phosphorescence emission. The most unexpected finding of these studies is that all SI transition of the dimer have the same polarization. Simmajor bands associated with the lowest So ple exciton theory in which vibronic interactions are neglected cannot account for the polarization of the dimer absorption spectra, but with the inclusion of vibronic interaction it is possible to account for the polarization of all bands except the origin band. To account for the “anomalous” polarization properties of the origin band it is suggested that the two molecules in the dimer are not equivalent, perhaps due to different orientation of phenyl substituents. The theoretical interpretation of the absorption automatically accounts for the anomalous intensity distribution in the dimer emission and the -10-fold increase in the fluorescence lifetime produced by dimerization. Less complete results obtained with some eosin, and acridine dyes serve to confirm the generality of the effects observed with the rhodamine dyes.

-

Introduction

Materials and Methods

The widespread use of dyes in dye laser^^-^ and in various photosensitized reactionss-13 has generated a renewed interest in the excited-state properties of dyes. Since many of the commonly used dyes tend to aggregate and form dimers in solutions, a number of the studies carried out during the past 10 years have been concerned with understanding the spectral effects produced by dimerization. 14-33 In the experimental studies which have been carried out so far, the emphasis has been on measurement of spectral shifts of bands, hypochromism, and changes in vibronic structure produced by dimerization. Insofar as we are aware, there has been no measurement of the polarization of the electronic transitions in the dimers, despite the fact that most theoretical treatments of dimer spectra give strong predictions regarding the polarization of transitions in the dimer. Thus, an important experimental parameter for testing theories of dimer spectra has not previously been measured. Perhaps an even more serious problem was recently raised by the studies of Ferguson and Mau,33 who demonstrated that many spectral changes previously attributed to dimerization were actually the result of an acid-base equilibrium. In the present study we have reexamined the conditions which we previously used to obtain dye monomer and dimer spectra26327 and have specifically tested for the role of pH on the spectra. Polarized emission and excitation spectra have been used to determine the relative polarizations of bands in both the monomer and dimer absorption spectra. As we shall show, the simple exciton theory in which vibronic interactions are neglected cannot account for the polarized dimer spectra. With the inclusion of interactions between vibronic levels of the lowest excited singlet state, it is possible to account for the polarization of the major dimer absorption bands. In order to account for the polarization of the origin band of the dimer fluorescence, however, it is necessary to invoke some sort of local asymmetry.

Chemicals. In the present study we have used the following dyes: sulforhodamine and rhodamine B (ChromaGesellschaft Schmid Co.), eosin B (MCB), eosin Y (Eastman), acriflavine (K and K Laboratories), proflavine (nutritional Biochemicals), and acridine Orange (Baker Chemical). The structures of these different dyes are shown in Figure 1. Generally the dyes were used as received from the manufacturer after a chromatographic examination indicated that impurities present had no deleterious effect on the spectral measurements. Spectroscopic Measurements. Most of the spectroscopic studies were carried out using a previously described fluorophosphorimeter26,27 which permitted both fluorescence and phosphorescence excitation spectra as well as emission spectra to be measured. Polarized excitation and emission spectra were obtained by incorporating a GlanThompson polarizer on the output side of the exciting monochromator and a polarizing sheet in front of the detector. Although a 90” configuration with respect to the exciting and emitted light was used for most of the experiments, a 180” configuration was used in the polarization experiments. Most spectroscopic measurements were carried out at 77°K (although some room temperature data are reported) and several different solvent systems were used. In order to obtain monomer spectra, the dyes were dissolved in either a 9:l mixture of ethanol and methanol or in EPA (diethyl ether, isopentane, and ethanol in a ratio of 5:5:2). To obtain dimer spectra, aqueous solutions containing 10 M LiBr or 10 M LiCl were used since the dyes have a greater tendency to aggregate in this solvent at low temperature. All excitation spectra reported in this paper have been corrected for the nonuniform spectral distribution of the exciting light from the monochromator, but only the rhodamine B and sulforhodamine B fluorescence spectra have been corrected for the spectral response of the detector system (0.25 M Jarrel Ash monochromator and RCA 7265 photomultiplier).

The Journal of Physical Chemistry. Vol. 78. No. 4, 1974

Emission Spectra of Ionic Dyes

381 H

I

ACRlOlNE ORANGE

CH. ACRIFLAVINE

H b

W

k

H

2

PROFLAVINE

Br

'0

COBIN Br

B

EOSIN

Y

0:

r

Figure 1. Structuresof the dyes studied.

In order to have high sensitivity in the polarized fluorescence and phosphorescence excitation spectra measurements and to eliminate polarization effects in the detection system, the emission (phosphorescence or fluorescence) was monitored through a band pass filter rather than a monochromator. Fluorescence lifetime measurements were obtained with a nanosecond lifetime apparatus constructed in this laboratory according to the design of Ware.34 Interference filters were used to separate fluorescence emission from the exciting light. Authentication of Monomer and Dimer Spectra. Depending upon the dye concentration, solvent, and temperature used in the experiments, it is possible to have dyes present as monomers and/or dimers. It was necessary, therefore, to determine which species was responsible for the emission and to establish conditions where monomer or dimer emission could be observed separately. Furthermore, the results of Ferguson and Ma@ regarding pH effects on dye spectra raised serious questions regarding some spectra previously attributed to dimers. For this reason the effect of pH on some low-temperature spectra was also studied. Monomer absorption spectra were easily obtained at room temperature in dilute (10-5 M) alcohol solutions and upon cooling these samples to 77°K there were no changes in the spectra except for expected improvement in the resolutions. From earlier work we know that dimer formation is promoted in aqueous 10 M LiCl or LiBr solutions at 77"K.26,27From comparison of our low-temperature spectra obtained with previously published and assigned dimer spectra, we conclude that the dyes were almost exclusively present in the dimeric state in the concentrated salt solutions at 77°K. The effect of pH on the spectral properties of rhodamine B, sulforhodamine B, and proflavine and the results of these experiments are shown in Figure 2. From a comparison of these spectra with the dimer spectra discussed below, it is clear that the spectra which we attribute to the dimer are not due to a change in the state of protonation of the monomeric form of the dye. In the aqueous solutions which we used to form the dimers a t low temperatures, there is little change in the spectra of the rhodamines over a range of pH extending from about 4 to 12. In the case of the proflavine spectrum the range is smaller,

'7 A

PROFLAVINE

,/-",a

Wave Number ( k cm-' )

e

Wave Number

[

kcm-1

)

Figure 2. A. Effect of p H and temperature on the proflavineabsorption spectrum: a, 1.8 X M in H20-10 M LiCI, pH 3-5 at room temperature, monomer; b, 1.8 X M in H20-10 M LiCI, pH -3 at 77"K, dimer; c, M in H20-13 N NaOH at 77°K; d , M in H 2 0 - 1 3 N NaOH at room temperature. B. Effect of pH on the absorption spectrum of rhodamine B (3.6 X M ) in H20 at room temperature; a, 5 N HCI; b, pH - 0.2; c, pH 1.0; d , pH 3.7; e, pH 13.0; f, 5 N NaOH. Only slight changes in the spectra are observed in the pH range 3.7-13.0.

but it is clear that the spectra of the basic form is different from the dimer spectrum which are obtained in dilute solutions at low temperature. We therefore conclude that the problems which Ferguson and Mau encountered using anhydrous alcohol solutions appear to be absent in the aqueous solution we used. Once authentic monomer and dimer absorption spectra were obtained, the emitting species were easily identified by measuring the fluorescence or phosphorescence excitation spectra of optically dilute samples. In the phosphorescence excitation measurements, the intensity of the The Journal of Physical Chemistry, Voi. 78, No. 4, 1974

R. W. Chambers, T. Kajiwara, and D. R. Kearns

382

,

450

500

,

550

600

650 nm

E N ERGY (kcrn-l)

450

500

I

I

550

600

650 nm

E N ERGY (kcm-1)

M ) dimers in 10 M LiCI-water (solid M. Individual bands (numcurve) and monomers in 9:l ethano1:methanol solution (broken curve), Solutions were approximately bered 1-4) are discussed in the text: A , rhodamine B; B, sulforhodamine 8.

Figure 3. The 77'K absorption spectra of rhodamine B

M ) and sulforhodamine B

phosphorescence was monitored a t the peak of the phosphorescence while the wavelength of the exciting light was continuously varied throughout the visible region of the spectra. Fluorescence excitation spectra were obtained in a similar manner.

1.0 -

yx)

h 0.8 -

r c

-$ 0 6 .-g o d I

d

.

0.2 -

20

19

18

17

16 k c w l

WAVE L E N G T H

Figure 4. The polarized fluorescence excitation and emission spectra of rhodamine B and sulforhodamine B dimers in 10 M LiCI-water solutions at 77'K: A , rhodamine B; B, sulforhodamine B. The monomer emission spectra (shown by the dotted curves) were obtained using 50% ethanol-50% 9 M LiCl water solutions at 77'K. The polarization ratios of the dimer excitation and emission spectra are shown at the top of the figure. In the

fluorescence excitation measurements the second fluorescence band was monitored (650-710 nm). A part of the spectra measured at higher sensitivity is also shown by broken lines to show the detailed structure at lower energy region. The small band which is observed at the high energy side of the dimer fluorescence, and which gives a high polarization ratio, is due to residual monomer fluorescence. T h e small shoulder on the lower energy side which gives a low polarization ratio is due to dimer phosphorescence. The Journal of Physical Chemistry, Vol. 78, No. 4 , 1974

Results and Discussion Using the procedures described above, monomer and dimer absorption and emission spectra were obtained and these results are presented in Figures 3-6. In each case we find that the long wavelength maximum of the absorption spectrum is blue shifted by dimerization and one or two relatively weak shoulders appear on the long wavelength side of the strong dimer band. Once the monomer and dimer absorption spectra were obtained, fluorescence and phosphorescence excitation spectroscopy were used to establish the authentic monomer and dimer emission spectra (compare Figures 3 and 4). It is well known that dimerization or aggregation of dyes usually leads to a great reduction in the fluorescence quantum yield.14~35However, in the case of rhodamine B and sulforhodamine B some residual fluorescence was still observable even though the molecules were almost completely dimerized.36 Through the use of excitation spectroscopy this residual fluorescence, which is shown in Figure 4, was established as authentic dimer fluorescence.36 (Compare Figures 3 and 4.) Rhodamine B and sulforhodamine B appear to be unusual in this regard since none of the other dyes we studied exhibited measurable dimer fluorescence. The fact that dimer fluorescence was observed from these two dyes provided us with a unique opportunity to study the polarization properties of the dimer absorption spectra, and a preliminary account of this work has already been presented.36 Relative polarizations of the absorption and emission bands are a valuable source of information about the nature of the excited states of the dye monomers and dimers. Photoselection measurements were therefore carried out using phosphorescence and fluorescence emission on both monomers and dimers of the various dyes and the results of these experiments are presented in Figures 4 and 6. Since the measurements were conducted at 77"K, depolarization effects due to molecular rotation are absent and dye concentrations (10-5 M ) were sufficiently low

303

Emission Spectra of Ionic Dyes

ACRIDINE

EOSIN Y

ORANGE

1.0-

.6

-

.4

-

2l

1.0

,

.

1

_

B

RHODAMINE

ACRIFLAVINE

-

.

1.0-

-

.4

-~

I

PROFLAVINE

f i t . , /),>\,

2

, ,

,

600

,

,

,

,

,

8 00

700

600

700

,

,

0

500

B

SULFORHODAMINE

Wave le ng t hI nm I Figure 5. A comparison of the monomer and dimer 77'K phosphorescence emission spectra of six representative dyes. The monomer spectra were obtained using M solutions in 1:l ethanol-water containing 10 M LiBr. The dimer spectra were obtained using M in water containing 10 M LiBr. EOSIN Y 10

2

MONOMER

4

m a

MONOMER

p

P

10

n Q

0 . .c..

O" : 025.

200

-

0 m

cwn

P

a-

P

7

n

:;.

300

400

Wavelength

500

600

2 00

300

400

500

IO

Wavelength

Figure 6. A comparison of the polarized phosphorescence excitation for the monomer and dimer of eosin B and eosin Y at 77'K. The monomer spectra were obtained using an aqueous alcohol solution and the dimer spectra were obtained using a 10 M LiCl aqueous solution: A, direct absorption spectrum; P, polarization ratio of the phosphorescence excitation spectrum. The Journal of Physical Chemistry, Vol. 78, No. 4 , 7974

R. W. Chambers, T. Kajiwara, and D. R. Kearns

384

M

D

Figure 7. A schematic drawing illustrating the effect of dimer formation on the spectroscopic properties of a dye according to simple exciton theory. In the dimeric state (D) the transition to the upper state is strong and polarized perpendicular to t h e weaker, lower energy transition.

that depolarization of the emission due to reabsorption effects was eliminated. A comparison of fluorescence polarization data obtained in an EPA glass at 77°K and in glycerol at room temperature demonstrated that uncracked EPA glass does not significantly depolarize the emission. With slow cooling, good quality glasses of aqueous 10 M LiCl were also obtained and used to obtain polarized dimer emission spectra. Interpretation of the Rhodamine B and Sulforhodamine B Dimer Spectra. Since our most complete experimental results were obtained with rhodamine B and sulforhodamine B we discuss these two molecules first. The absorption spectra of the monomeric and dimeric forms of rhodamine B and sulforhodamine B shown in Figure 3 illustrate the spectral changes which are typically noted with many dyes. Dimerization leads to a blue shift of the main absorption band, a slight loss of intensity in the low energy portion of the spectrum (hypochromism), and a reduction in the fluorescence quantum yield (usually complete quenching). Previously dimer spectra such as these have usually been interpreted in terms of a simple exciton mode1,14,35,37-39 the results of which are schematically depicted in Figure 7 . According to this theory the strong band in the dimer spectrum (band 3 in Figures 3 or 4) would be assigned to the allowed transition to the \k+ state and the weaker, lower energy band in the dimer spectrum (band 2 in Figures 3 or 4) should have the opposite polarization and would be assigned as the “forbidden” transition to the 9- state, Band 1, which is easily observed only in the low-temperature dimer spectrum of both sulforhodamine B and rhodamine B had not previously been reported since most of the earlier measurements were carried out a t room temperature. The polarized fluorescence excitation and emission spectra for the dimer forms of these two molecules are shown in Figure 4. In the absence of polarization data it appeared that the dimer absorption spectra could be qualitatively accounted for in terms of a simple exciton picture where the dimer symmetry is sufficiently low that there is incomplete cancellation of the transition dipole moment for the transition to the \k - state. Insofar as the theoretical interpretation of the dimer is concerned, however, the most significant observation is that the two strong absorption bands (bands 2 and 3) in the dimer spectra are positively polarized with respect to the 2’ band in fluorescence. Furthermore, the origin band in absorption (band 1) is obliquely polarized with respect to bands 2 and 3 in absorption, and band 2‘ in emission. The Journal of Physical Chemistry, Vol. 78, No. 4, 1974

Figure 8. A schematic diagram showing the relation between the energy levels of a sandwich dimer (D) and two monomers (M). This diagram has been drawn based on the zeroth-order resonance coupling theory which includes the vibronic effects but not configuration interaction between zeroth-order wave functions. The actual forms of zeroth-order wave functions are given in Table I . The positions of levels were determined based on absorption and fluorescence spectra of rhodamine B, the strongest absorption band and t h e weakest absorption band being fitted to level 5 and level 4, respectively. Relative intensities for various transitions are indicated by t h e thickness of the arrows and the polarizations of the transitions (relative to the strongest band) are indicated by f.

These polarization properties are simply not consistent with the simple exciton model. Since the true origin band has now been observed, band 1 might be reassigned as the transition to the \k- state, but this assignment raises a problem with regard to the interpretation of the two stronger absorption bands (bands 2 and 3) since only one can be assigned as the transition to the 9+ state. The dimer emission properties also pose serious problems for the simple exciton theory in two other important respects. First of all, the relative intensities of the 0 0 and 0 1 bands in the dimer emission spectrum (bands 1’ and 2’) are quite different from those found in the monomer emission spectrum. Secondly, the polarization properties of bands 1’ and 2‘ differ from one another. In order to account for these “anomalies” it is necessary to modify the simple exciton theory by taking into consideration the role of molecular vibrations and the vibronic coupling between various states of the dimer. Several different treatments of the effects of vibronic interaction on the electronic spectra of dimers have been published40-46 so that only the essential results of these treatments need to be summarized here. Depending upon the strength of the electronic coupling between the two molecules there are two different theoretical limits that should be considered. In the weak coupling limit the electronic coupling is assumed to be smaller than the spacing between vibronic levels of the monomer, whereas in the strong coupling limit, the electronic coupling is larger than the spacing between vibronic levels. In view of the large changes which occur on dimerization, it would appear that the strong coupling limit is more appropriate to the case at hand. A schematic diagram showing how the monomer states correlate with those of the dimer in the strong coupling limit are shown in Figure 8. In this diagram the rel-

-

-

Emission Spectra of ionic Dyes

385

TABLE I: Wave Functions Used to Describe the Dimer States in the Strong Coupling Limit (see Figure 8 ) a

Excited Electronic States

*I+ =

~1f(O,O;f)

=

1 7 (p*lpbO+ 2

ppsppbl)xs("o)xb(',o)

q6@1-(1,0;-) i ( ( p s l p b o - (pso'pbl) [Xa('3')Xb('to) f X~.(',~)xb(~,')] q 7 += *l+(l,o; -) = ;(ppB'qbo - qpsoqb') [xs("l)xb(l,o) - X a ( l , O ) X b ( l , l ) ] = $1+(1,0;+) PQ- = @1-(1,0;+) =

$(qnlQbo $('pp.'(0bo

f qaoPbl)

[X~("")Xb('")

f Xn(130)Xb(1,1)]

[Xa('")Xb("o)

-

+ ppsoqpb')

Xa("o)Xb(131)]

is the nth excited-state function. xa(n,v) is the vibrational wave function for a r ~is' the electronic wave function for the ground state of molecule a, and molecule a in its nth electronic state, and vth vibrational level. The corresponding wave functions for molecule b are indicated in a similar fashion. The complete wave functions for the dimer are described by the functions Qn*(v,0,*) where the superscript indicates the overall (electronic vibrational) symmetry.

+

ative intensities of the various transitions and their polarization with respect to the strongest band in the absorption spectra have also been indicated to facilitate comparison with the experimental data. (Also see Table I.) The strongest band in the dimer absorption spectrum is assigned as a transition to state \k5 and the polarizations of all other transitions are specified relative to this transition. (In the simple exciton theory this would correspond to a transition to the \k+ state.) Since \ k 5 is the second electronic state there is no band in the emission spectrum corresponding to this absorption band. The moderately intense longer wavelength in the dimer absorption spectra (band 2) is correspondingly assigned as the transition to the vibrationally excited state of the lowest electronic state of the dimer. In the strong coupling model this is a transition to \k7 and it receives its intensity by vibronic coupling with the higher energy 9 5 state. It therefore is predicted to have the same polarization as band 3, in agreement with experiment. The corresponding band in emission corresponds to the transition from 9 4 to the vibrationally excited ground state \k3 and it too should have a positive polarization. Band 2' in the emission spectrum can clearly be identified as this transition. This assignment also permits us to understand the anomalous intensity of this band as is discussed below. The weak origin band in the dimer absorption and emission spectrum can be assigned as a transition between states \k1 and 9 4 . If the dimer has a center of symmetry or a plane of symmetry, theory predicts that this transition should be forbidden or partially allowed and polarized opposite to the intense band 3. The fact that this transition is polarized obliquely to the main dimer band is therefore not accounted for, and as long as the dimer is assumed to have some symmetry there is no way that positive polarization can be introduced into this transition so that it would have roughly .equal contributions from positive and negative polarizations. Conceivably the weak origin band could be a vibrationally induced band made allowed by a small-frequency mode belonging to a nontotally symmetric representation of dimer symmetry. However, the maximum of this weak band and the maximum of the higher energy band of dimer fluorescence are separated by only about 300-400 cm-1 and this is nearly the same as the Stokes shift for corresponding

+

monomer bands. Also, the absorption and fluorescence bands are not as sharp as expected for vibronically induced forbidden band. Thus, this mechanism seems to be unlikely. It is unlikely that this weak origin band and the corresponding higher energy fluorescence band are respectively superposition of two bands which have mutually perpendicular transition dipole moments, because of the facts discussed above and because the intensity ratio of the higher energy band to the lower energy band of dimer fluorescence is smaller than that of monomer fluorescence. We are forced to conclude that the two molecules in the dimer complex are slightly different from each other. The angular orientation of the phenyl substituents offers one obvious source of assymetry. This may result in a slightly different transition probability for the same transition in the two different molecules, and together with slight deviation from parallel arrangement, will result in an oblique orientation of the transition dipole moment for the transition to the 9 4 state of the dimer. Finally, we have to explain the anomalous intensity of the 0-1 band of the dimer fluorescence (at 678 nm for rhodamine B and at 690 nm for sulforhodamine B). As we noted above, the polarization of this band is easily accounted for if we assume that it corresponds to a vibronically induced transition from \k4 to \k3. This assignment also accounts for the large increase in intensity of the 0-1 band relative to the 0-0 band, which occurs on dimerization (0.6 for the dimer compared to 0.19 for the monomer of rhodamine B and 0.9 compared with 0.25 for sulforhodamine B) as the following analysis indicates. The intensity of band 2 in the absorption spectrum is derived from vibronically induced mixing between the 9 7 and 9 5 states which, experimentally, are separated from each other by only about 1000 cm-1. The corresponding vibronically induced band in emission (band 2') corresponds to a transition between states \ k 4 and \k3 and it derives intensity uia vibronic coupling between state \k4 and state \kg which is located over 3000 cm-1 away. Because of this larger separation (factor of over 3) vibronic coupling effects on the 0-1 band in emission are reduced by a factor of about 10 compared with the corresponding absorption band. In the dimer absorption spectrum the 0-1 band is about 10 times stronger than the 0-0 band. However, because of the factor of 10 difference arising The Journal of Physical Chemistry. Vol. 78. No. 4 . 1974

386

from the AE factor, we predict that the two bands would be about equal in intensity in the emission spectrum. ExperimentaIly, the ratio of the intensity of the 0-1 band to the intensity of the 0-0 band is 0.6 for rhodamine B and 0.8 for sulforhodamine B, in reasonably good agreement with our approximate estimate. Our interpretation of the dimer absorption spectrum is thus found to give a reasonable account of the enhancement of the 0-1 band relative to the 0-0 band in the dimer emission spectrum as compared with the monomer emission spectrum. Our interpretation is that the emission spectrum is comprised of a weak electronic transition which corresponds to the 0-0 band and a vibronically induced transition which corresponds to the 0-1 band. In view of this assignment and the relationship between absorption coefficients and radiative lifetimes, we further predict that the radiative lifetime of the dimer should be a factor of 8-10 times longer than the monomer, based on a comparison of the integrated intensity in the monomer spectrum. The observed lifetime of the monomer fluorescence of rhodamine B is reported as 6.2 nsec in ethanol and the quantum yield 0.97.47 Experimentally, we find a value of 3.4 nsec at room temperature in ethanol-glycerol and 5.3 nsec at 77°K. By comparison, the lifetime of the dimer in an aqueous solution is 38 nsec at 77”K, corresponding to at least a sevenfold increase in the radiative lifetime as a result of dimer formation. This agrees quite well with the estimated value, and confirms that the strong band 3 in the dimer absorption spectrum corresponds to a transition to an upper electronic state of the dimer which is not involved in the emission spectrum.35 Monomer and Dimer Phosphorescence from Acridine and Xanthene Dyes. Some time ago we reported the observation of dimer phosphorescence from acridine orange, rhodamine B, and eosin Y.26 These observations have now been extended to several other dyes and these results are presented in Figure 5 along with corresponding results for the monomer. The comparison of the monomer and dimer phosphorescence spectra presented in Figure 5 reveals that dimer formation generally leads to a -30-nm red shift in the peak of the phosphorescence. Because of this spectral shift it was of interest to determine whether or not dimerization affected the phosphorescence decay times. Phosphorescence lifetimes were measured for both monomers and dimers of acridine orange, eosin Y, and sulforhodamine B. Within experimental error dimerization has no effect on the phosphorescence lifetime of these three dyes indicating that either dimerization has no effect on either the radiative or nonradiative transition rates from the triplet state or (less likely) that the rates are fortuitously changed in such a manner that they exactly cancel one another. The phosphoi-escence lifetime for acridine orange and sulforhodamine B were also measured at 1.8”K and found to be virtually identical with those measured at 77°K. The motivation for this latter experiment lay in the possibility that the triplet states of the dimer might be slightly split due to an exciton interaction, and therefore at low temperatures emission might be observed from the lower component with an altered lifetime. The observation that there is no change in lifetime indicates that either the splitting of the triplets is less than 2 cm-1 or that the radiative lifetimes of both triplet components are comparable. Spin polarization effects also appear to be absent. As with the rhodamines, direct absorption spectra were The Journalof Physical Chemistry. Voi. 78. No. 4, 1974

R. W . Chambers, T. Kajiwara, and D. R. Kearns

compared with the phosphorescence excitation spectra to establish the authenticity of the emitting species. In general, the phosphorescence from these compounds is polarized out-of-plane so that it is not possible to learn anything about the relative polarization of the in-plane polarized singlet-singlet transitions of the dimers. The halogenated xanthenes are exceptions to this generalization since they exhibit an in-plane polarized phosphorescence. With eosin Y and eosin B the monomer phosphorescence is polarized positively ( P = +0.1) with respect to the first strong singlet-singlet band of the monomer (Figure 6). In the dimer the polarization of the phosphorescence is slightly reduced, but it is nevertheless still a positive polarization with respect to the two bands in the dimer absorption spectrum which appear to be the counterparts of bands 2 and 3 in the dimer absorption spectra of the rhodamines (see Figure 6). The true origin band (counterpart of band 1) is apparently too weak to be seen in the eosin Y dimer spectrum and this is consistent with the fact that the no dimer fluorescence is observed from this compound. Consequently, we believe that our interpretation of the dimer spectra is not restricted to the rhodamines, which are special in exhibiting dimer fluorescence as well as dimer phosphorescence, but rather that we are dealing with a rather general phenomenon which applies to many other similar dye systems. Summary In this paper we have presented the results of measurements on the polarized absorption and emission of several dyes in the dimeric state. The rhodamine dyes are unusual in that the dimers are fluorescent, and this permitted us to determine for the first time the relative polarizations of various dimer absorption and emission bands. We find that neither the dimer absprption nor the emission spectrum can be accounted for in terms of the simple exciton model commonly used to interpret the electronic spectra of dye dimers. The spectra can, however, be satisfactorily interpreted in terms of a strong coupling model in which vibronic interactions are included. The fact that the 0-1 band in fluorescence is intensified compared to the 0-0 band as a result of dimerization raises some question about the frequently discussed stabilization energies of excimers of some aromatic hydrocarbons such as perylene and pyrene.48 These stabilization energies have usually been evaluated based on the energy difference between the maximum of the monomer fluorescence and the maximum of the excimer fluorescence. The present study suggests that the maximum of the excimer fluorescence might correspond to a “0-1” band rather than the “0-0” band, and this leads to -1300 cm-1 smaller values of stabilization energy than those usually accepted.

Acknowledgment. The support of the U. S. Public Health Service (Grant GM 10449) is gratefully acknowledged. References a n d Notes (1) Submitted in partial fulfillment of the requirements for the Ph.D. degree in Chemistry at the University of California, Riverside. (2) P. P. Sorokin,etal., l 6 M J., 11, 139, 148 (1967). (3) (a) D. A. Leonard, Appl. Phys. Lett., 7 , 4 (1965); (b) D. A. Leonard, R. A . Neal, and E. T. Gerry, /bid., 7, 175 (1965). (4) M. Bass, T. F. Deutsch, and M. J. Weber, Appl. Phys. Lett., 13, 120 (1968). (5) G. I. Farmer, B. G. Huth, L. M. Taylor, and M. R. Kagan, Appl. Opt., 8 (2),363 (1969). (6) H. Sameison, Electronics, 142 (1968). (7) R. A. Keiler, l E E E J . Quantum Electron., 8 (7), 411 (1970).

Theory of Circularly Polarized Emission (8) D. R. Kearns, Chem. Rev., 71,395 (1971). (9) J. D. Spikes and R. Livingston, Advan. Radiat. Biol., 3, 29 (1969). (10) D. C. Chatterjee and E. A. Noltmann, Eur. J. Biochem., 2, 9 (1967). (11) E. Scoffone, G. Galiazzo, and G. Jori, Biochem. Biophys., 38, 16 (1970). (12) H. E. A. Kramer and A. Maute, Photochem. Photobiol., 15, 25 (1972). (13) F. Leterrier and P. Douzou, Photochem. Photobiol. 8, 369 (1968). (14) T. Fdrster and E. Konig, Z. Elektrochem., 61, 344 (1957). (15) L. V. Levshin and V. K. Gorshkov, Opt. Spektrosk., 10,401 (1961) (16) L. A. Ignat'eva, L. V. Levshin, T. D. Osipova and Y. M . Polukhin. Opt. Spektrosk., 13, 219 (1962). (17) K. L. Arvan and N. E. Zaitseva, Opt. Spektrosk., 11, 38 (1961). (18) G. P. Gurinovich and T. I. Stelkova, Biofizika, 8, 229 (1963). (19) L. V. Levshin and I . S. Lonskaya, Opt. Spektrosk., 11, 148 (1961). (20) L. V . Levshin and E. G. Baranova. Opt. Spektrosk., 6, 31 (1959). (21) L. V. Levshin and V. G. Bocharov, Opt. Spektrosk., 10,330 (1961). (22) V . G . Bocharov and L. V. Levshin, In. Akad. Nauk SSSR, Ser. Fiz., 27, 591 (1963). (23) L. V. Levshin, lzv. Akad. Nauk SSSR, Ser Fiz., 29, 1299 (1965). (24) L. V . Levshin and D. M. Akbarova, Zh. Prikl. Spektrosk., 2, 43 (1965). (25) L. V. Levshin and D. M. Akbarova, Zh. Prikl. Spektrosk., 3, 326 (1965). (26) R. W. Chambers and D. R. Kearns, J. Phys. Chem., 72, 4718 (1968). (27) R. W. Chambers and D. R. Kearns, Photochem. Photobioi., 10, 215 (1969). (28) D. J. Blears and S. S. Danyluk. J. Amer. Chem. SOC., 89, 21 (1967).

307 (29) S. J. Davidson and W . P. Jencks, J. Amer. Chem. SOC., 91, 225 (1969). (30) R. E. Ballard and C. H. Park, J. Chem. SOC.,A, 1340 (1970). (31) K. K. Rohatgi and A. K. Mukhopadhyay, Photochem. Phofobiol., 14, 551 (1971). (32) E. Braswell, J. Phys. Chem., 72, 2477 (1968). (33) J. Ferguson and A. W. H. Mau, Chem. Phys. Lett., 17, 543 (1972). (34) W. R. Ware, "Creation and Detection of the Excited State." A. Lamola, Ed., Marcel Dekker, New York, N. Y. 1970. (35) E. G. McRaeand M. Kasha, J. Chem. Phys., 28,721 (1958). (36) T. Kajiwara, R. W. Chambers, and D. R. Kearns. Chem. Phys. Lett., 22, 37 (1973). (37) E. G. McRae and M. Kasha. "Physical Processes in Radiation Biology," Academic Press, New York, N. Y., 1964, p 23. (38) M. Kasha, H. R. Rawls, and M. Ashraf El-Bayoumi, Pure Appl. Chem., 11,371 (1965). (39) A. S. Davydov, "Theory of Molecular Excitons," translated by M. Kasha and M. Oppenheimer, Jr., McGraw-Hill, New York, N. Y., 1962. (40) W. T. Simpson and D. L. Peterson, J. Chem. Phys., 26, 588 (1957) (41) A. Witkowski and W. Moffitt, J. Chem. Phys., 33,872 (1960). (42) R. L. Fulton and M. Gouterman. J. Chem. Phys., 35, 1059 (1961). (43) R. L. Fulton and M. Gouterman. J. Chem. Phys., 41, 2280 (1964). (44) M . G. Sucre, F. Geny, and R. Lefebvre. J. Chem. Phys., 49, 458 (1968). (45) J. H. Young, J. Chem Phys., 49,2566 (1968). (46) E. G. McRae,Aust J. Chem., 14, 329 (1961). (47) S. J. Strickler and R. A. Berg, J. Chem. Phys., 37 814 (1962). (48) J. B. Birks, "Photophysics of Aromatic Molecules," Wiley-lnterscience, London, 1970.

Theory of Circularly Polarized Emission from Molecules Displaying Rotary Brownian Motion' Joseph Snir and John A. Schellman* Chemistry Department, University of Oregon, h g e n e , Oregon 97403 (Received September 78, 7973) Publication costs assisted by the University of Oregon

Formulas are obtained for the extent of circular polarization of light produced by fluorescence emission. Photoselection and its relaxation by rotary Brownian movement are taken into account. For simplicity, the model is restricted to spherical molecules and a simple dipole coupling mechanism. On the basis of this model, it is shown that the ratio of circularly polarized emission to total emission is independent of the relaxation of photoselection. The dependence of the absolute intensity of circularly polarized emission is evaluated as a function of the optical parameters and the relaxation time for rotary Brownian movement. The relaxation correction is small.

Introduction The use of optical activity as a reflection of molecular conformation and changes in molecular conformation has been established for many years. The measurement and interpretation of optical activity came into particular prominence when it became convenient to measure optical rotation and later circular dichroism within absorption bands. In measuring circular dichroism, one determines the difference in absorbance by the molecule of left and right circularly polarized light. Recently, interest has developed in the possibilities of the reverse process, L e . , the measurement of the circularly polarized component of emitted light from asymmetric molecules. Experimental observations have been obtained with crystals of sodium uranyl acetate2 and with solutions of both organic compounds (hydrandindone and thiohydrandindone) and inor-

ganic complexes [Cr(en)3](C104)3 by Emeis and OosterhofS3These pioneering investigations have mostly been made with systems which guaranteed a fairly strong signal of circularly polarized light in the emission p r o ~ e s s . ~ Recently, an instrument has been developed by Steinberg and Gafni which provides the capability of measuring circularly polarized light when it comprises only one part in l o 4 of the total intensity.51 The first application of this method (which uses stress modulation techniques now common in CD measurements) has been to the difficult problem of molecules bound to enzymes.bb This technical development, which requires only relatively simple modifications of a spectrofluorimeter, widely extends the applicability of the method. The principal question of interest at the present time is the nature of the new information which will come from the new measureThe Journal of Physical Chemistry. Vol. 78. No. 4. 1974