l0 January 1997

CHEMICAL PHYSICS LETTERS

I ELSEVIER

Chemical Physics Letters 264 (1997) 265-272

Laser-induced optoacoustic studies of the non-radiative deactivation of ICT probes DMABN and DMABA Nitin Chattopadhyay l, Johan Rommens, Mark Van der Auweraer, Frans C. De Schryver Laboratory of Molecular Dynamics and Spectroscopy, Department of Chemistry, Katholieke Universiteit Leaven, 200 F Celestijnenlaan, B-3001 Heverlee, Belgium Received 22 July 1996; in final form 4 November 1996

Abstract Laser induced optoacoustic spectroscopy has been used to determine the relative contribution of the intersystem crossing and internal conversion to the non-radiative energy dissipation from the two excited ICT probes p-N,N-dimethylaminobenzonitrile (DMABN) and p-N,N-dimethylaminobenzaldehyde (DMABA) in solvents of different polarity. It is observed that while intersystem crossing is the main non-radiative deactivation process for D M A B N in all the solvents studied this is not the case in polar solvents for D M A B A .

1. Introduction Since its discovery by Lippert [1], the dual luminescence of DMABN and some related compounds in polar solvents has stimulated photophysicists and chemists to study this class of compounds. One mechanism correlates the appearance of the lower energy emission in DMABN phenomenologically to a torsional motion about the donor-acceptor single bond following charge transfer in the excited state [2]. Interest in the phenomenon has been enhanced by the assumption of dual luminescence in the properties of fluorescence probes and laser dyes, the isomerisation of polyenes and rhodopsin, molecular switching devices, charge separation in photochemi-

I On leave from the Department of Chemistry, Jadavpur University, Calcutta - 700 032, India.

cal energy conversion etc. [3]. Based on the observation of the occurrence of intramolecular charge transfer in a series of 4-(dialkylamino)benzonitriles, even in apolar solvents, a different model has been proposed by Zachariasse [4,5]. In this model the dual luminescence is interpreted as a solvent-induced pseudo Jahn-Teller effect [6]. A recent CASSCF (complete active space self-consistent field) calculation, however, seems not to support fully such an alternative interpretation as the only contributing mechanism to the charge transfer state [7]. The energy of the locally excited (LE) and relaxed (ICT) singlet, as well as the triplet states of DMABN and DMABA are more or less known both from theory as well as from photochemical studies in a selection of solvents and in rigid environments [8-23]. The literature indicates that for the ICT fluorophores upon increasing the polarity of the environment a low energy emission (ICT band) grows at the cost of the high energy (LE band) emission. The

0009-2614/97/$17.00 Copyright © 1997 Elsevier Science B.V.B.V. All fights reserved. PH S 0 0 0 9 - 2 6 1 4 ( 9 6 ) 0 1 3 5 4 - 1

266

N. Chattopadhyay et al. / Chemical Physics Letters 264 (1997) 265-272

overall fluorescence quantum yield, however, decreases. A similar effect has also been observed in microheterogeneous organised environments [10,2426]. This decrease in the radiative yield has been explained by an enhancement in the non-radiative decay of the excited fluorophore to the ground state through the stabilised ICT state [8-10]. Participation of the triplet states has also been suggested. However, there is no systematic study relating the channels by which the excited species return to the ground state to the solvent polarity. In this Letter, we have used the laser-induced optoacoustic spectroscopic (LIOAS) technique to determine the relative proportion of the intersystem crossing (ISC) and intemal conversion (IC) in the molecular dissipation of energy from the excited species. Two fluorophores DMABN and DMABA respectively, which are considered to form ICT, have been studied in solvents of different polarity. 2. E x p e r i m e n t a l

Each of the compounds DMABN, DMABA and 2-hydroxybenzophenone (all from Aldrich) were recrystallised twice from methanol. The purity of the compounds was checked by spectroscopic methods as well as TLC. The solvents, acetonitrile (ACN, Lab-scan), tetrahydrofuran (THF, Fluka), di-n-butyl ether (DBE, Merck), 1,4-dioxane (DOX, Merck) and n-heptane (HEP, Aldrich) were all spectroscopic grade and were used as received. 2-Hydroxybenzophenone (HBP) was used as the reference compound [ a (fraction of total excitation energy liberated as heat) = 1.0] for the photoacoustic measurements [27,28]. In each solvent, three solutions of the probes, with optical densities 0.08, 0.10 and 0.12 were prepared. The absorbances of the reference solutions were the same as that of the sample solutions within the limit of the experimental error. The solutions were degassed by bubbling with argon for 15 min. The o ' s were determined with respect to the reference for the three solutions in a particular solvent and the mean value was taken. The fluorescence quantum yields of the fluorophores were determined using quinine sulfate in 0.1 N H2SO 4 as reference (th = 0.54) [6]. The absorption and fluorescence spectra were recorded on a Perkin Elmer Lambda 6 U V / V I S

spectrophotometer and a Spex Fluorolog fluorimeter respectively. The LIOAS experiments, where the sample was excited at 320 nm, were performed using the apparatus described [27-29] using a piezo-electric PZT detector and a 1 mm pinhole.

3. Results

The emission spectra of DMABN in different solvents are presented in Fig. 1. It is obvious from the figure that with an increase in the solvent polarity, the ICT yield increases at the cost of the LE emission with a gradual decrease in the overall fluorescence quantum yield for the fluorophore. In a photoacoustic experiment the integrated amount of heat emitted within the time resolution of the apparatus is detected. For our apparatus this time window amounts to 1 ixs. The fluorophore (DMABN or DMABA) excited to its first excited singlet state can deactivate to the ground state through fluorescence (high energy LE emission as well as lower energy ICT emission) or internal conversion or intersystem crossing to the lowest triplet state. As the triplet species is much longer lived (in the range of micro or milliseconds) the amount of energy stored in the triplet will not contribute to the amount of prompt heat detected by the LIOAS. Fig. 2 depicts a typical photoacoustic signal obtained for a solution of DMABA in THF (O.D. = 0.1 and laser energy = 15 ixJ). 2.NIE+06

- HEP . . . . . . DBE --DOX --THF ~ACN

2 00E'H~ - '.,' i = a

~

150E*06 •

1.00E*~.

!\

/

\

5.00E*05

000E"~-O0 320

340 360 380 400 420

440 460 480

Wavelength

500

520

540 560

580 600

(nm)

Fig. 1. Room temperatureemission spectra of DMABN in different solvents (Aexc = 320 nm). The solvents are indicated in the inset.

N. Chattopadhyay et al. / Chemical Physics Letters 264 (1997) 265-272 200

267

300 -

150

~

250-

.. " e ' ' " m

100 200 -

~"

50

.~

o 2

~.

. - " .e"

~

• ..*'"

]

`1~o.

i

`100.

D""

•

.

3

-.ca -tO0

50

-150

• ..~."e

-

.4'"

.a"

. .e"

0

-200

0

4

8

Time(/is)

12 Laser energy

16

20

24

(ITS)

Fig. 2. Photoacoustic signal obtained from a solution of D M A B A in THF (OD = 0.1 and laser energy = 15 p.J).

Fig. 3. Dependence of the photoacoustic signal on the laser energy. ( ) for D M A B N and ( . . . . . . ) for HBP (the reference) in acetonitrile. The optical densities at Aexc (320 nm) were 0.1 for both solutions.

The fraction of absorbed energy (Eabs) that is detected by LIOAS (Eth) is given by the value of the experimental quantity a which equals Eth/Eab s. Since, with the exception of the decay of the excited triplet state, all other decay processes take place within the time resolution of the LIOAS, the overall quantum yield of intersystem crossing, ~b~sc, from the excited singlet state (LE and ICT) can be obtained from

replaced by ( E e) and calculated from

ot'Eex c = E e x c -

~)fEf -

(~tsc ET ,

(1)

where Eex c, Ef and E r are the energy of the excitation, fluorescence and triplet respectively and ~be and thlSC are the overall quantum yields of fluorescence and intersystem crossing respectively. To take into account the large width of the emission spectra Ef is

(Ef)

=

flr( v)vdv/flf(

v)dv,

(2)

where If(v) is the spectral distribution of fluorescence and v is the energy in wavenumbers. The value of < E l > will be close to the emission maxima for molecules yielding sharp and symmetric fluorescence spectra which is not the case with our probes in all the solvents studied. Fig. 3 shows a plot of the amplitude of the first maximum of the optoacoustic signal ( H ) as a function of laser energy (E0). The linearity indicates the absence of any biphotonic process and ground state depletion [29]. According to the Eq. (3) below, a can be calculated as the ratio of the slopes (of the

Table 1 Photoacoustic and fluorometric data for D M A B N and D M A B A in different solvents Solvent

e

ot

DMABN

ACN THF DBE DOX HEP

38 8.2 = 3 2.2 1.9

0.526 0.435 0.326 0.454 0.320

DMABA

ACN THF DBE DOX HEP

38 8.2 = 3 2.2 1.9

0.834 0.666 0.432 0.273 0.288

a In ethanol glass, Ref. [12].

~bf

Ef (era - I )

Eex¢ ( c m - i)

E T ( c m - i)

~blSc

~bnc

q~nc/~blsc

0.022 0.046 0.12 0.076 0.13

20630 23620 27800 24800 28930

31219 31219 31219 31219 31219

24000 24000 24000 24000 24000

a a a a a

0.60 0.69 0.74 0.63 0.73

0.38 0.26 0.14 0.29 0.14

0.64 0.38 0.19 0.46 0.19

0.0027 0.0009 --- 0 0.0026 -= 0

18610 23230 24060 -

31219 31219 31219 31219 31219

21000 21000 21000 21000 21000

b b b b b

0.24 0.50 0.84 1.07 1.05

0.75 0.50 0.16 0 0

3.08 1.02 0.18 0 0

b In P E M A film, Ref. [16].

268

N. Chattopadhyay et aL / Chemical Physics Letters 264 (1997) 265-272

plot of H versus E o) obtained for the sample and the reference, characterised by ct = 1.0. n = KEoot(1 - 1 0 - a ) ,

(3)

where H, A, E 0, K correspond to the experimentally obtained amplitude of the photoacoustic signal, the absorbance of the solution of the sample or the reference, the incident laser energy (IJ,J) and a constant which depends on the geometry of the experiments and the thermoelastic properties of the medium. Using this method, the ot's have been determined for DMABN and DMABA in different solvents (Table 1). The triplet energies of the probes are taken directly from the literature. Although these values correspond to the energies in the frozen glass (for DMABN) [12] or in film (for DMABA) [16], a more enhanced relaxation of the triplet state in less viscous media would only change the triplet energy to a minor extent and hence induce a decrease in the values of ~blc and an increase in ~blsc. However, the trends observed for the solvent dependence of those properties would remain. The quantum yields of intersystem crossing (~blsc) and internal conversion (~b]c) have been calculated and are presented in Table 1.

energy is much more interesting (and also complex) due to the involvement of the nrr * as well as "rr~r * triplet states. In their recent pressure and light induced luminescence studies, Drickamer and coworkers have shown that the lowest triplet has n-rr * character in a poly(ethyl methacrylate) (PEMA) film at atmospheric pressure while under pressure the "rrrr * character increases changing its emitting properties [16]. From the phosphorescence studies in condensed phases it is known that the nature of the emitting triplet state may be changed from n-rr * to ~r-rr * with the variation in the solvent polarity [30]. However, the existing literature does not suggest any ICT character for the lowest triplet state of DMABA. Thus for both DMABN and DMABA, we assume that the lowest triplet state is of LE nature. Considering the charge distribution of the excited states, we can conceive that during the relaxation processes the electrostatic constriction in the volume (AV) is small for S~c --> S O (IC), but large for SI -> SI~cT as well as SIjcT --> T transitions (see Scheme 1).

FC

51

0 SI

•

LE , kls C kcE~cr , ~CT,,E ' ~' ICT

51

ISC~,.

4. Discussion LE kf

Excitadon

From Table 1, it is clear that intersystem crossing (ISC) is the most important deactivation channel for the excited singlet state of the fluorophores. Since ISC involves triplet states it is pertinent to start our discussion with the properties of the triplet state. For DMABN both locally excited and ICT triplet states are postulated [12-14]. Although time-resolved infrared spectroscopic results indicate that the amount of ICT triplet formed is proportional to the amount of the ICT singlet [ 14], nanosecond transient studies indicate that the lowest triplet does not have strong CT character [12]. The dipole moment of the triplet state is considerably less than the dipole moment of the excited singlet state, indicating that the triplet state is slightly more polar than the ground state of the molecule [21-23]. For DMABA, the participation of the triplet states in the non-radiative dissipation of the excited state

ICT klc

ICT kf kT

I !

i v •

i 0

So

T

V~

The A V ' s for the latter two steps will, however, be opposite in sign and thus, to the extent that the net AV between the triplet state and the ground state can be neglected, cancel each other. Although the slightly larger dipole (2 to 3 debye) of the triplet state compared to the ground state [21-23] will allow for a small AV due to the electrostriction, this effect will be small compared to the volume change induced by thermal expansion. Hence one can attribute the experimentally observed optoacoustic signal to a good

N. Chattopadhyay et a l . / Chemical Physics Letters 264 (1997) 265-272

approximation to the temperature changes induced by the non-radiative energy dissipation by the excited probe. Table 1 shows that as we move from lower to higher polarity solvents, the fluorescence quantum yield of DMABN decreases regularly with a deviation for 1,4-dioxane; which is known to provide, in general, a more polar microenvironment than the bulk (indicated by the e value only). The estimated Sf values agree with the available data in ACN, DOX and HEP [21,31,32]. This decrease in the fluorescence yield in more polar solvents is related to an increase in the non-radiative decay of the excited fluorophore through the stabilised ICT state [10]. Table 1 also reflects that as we move towards more polar solvents the quantum yield of internal conversion (~btc) increases, substantiating the earlier experimentally unverified proposition [10]. For DMABA the fluorescence quantum yield is extremely low in all the solvents. It is, however, clear that while the emission yield for DMABN is greater in the less polar solvents, DMABA is almost nonfluorescing in these solvents (DBE, HEP). This is most likely due to the interaction of the n'rr * triplet state with the singlet state. As the ICT singlet does not contribute in these apolar environments [4,5,10], the proximity of the nTr* triplet state with the S~ state facilitates the deactivation through this nrr * triplet state. This is supported by the large enhancement in the value of ~blsc in the these solvents and it is revealed that practically all the excitation energy is dissipated through the non-radiative ISC channel (thlsc -=- 1.0) making the molecule almost non-fluorescent. It is pertinent to point out that the photoacoustic data in apolar solvents are free from the approximation coming out of the electrostriction (AV) effect since the effect is absent in these solvents. From the energy values of the states it is seen that the ISC process is not very exothermic as the energy difference is small, maximum --4000 cm -~ for both probes in all the environments studied. Thus a change in the solvent polarity is not expected to modify the process (and kts c) much. However, since with a change in the solvent polarity the nature of the emitting singlet state is changed (LE or ICT) it is not so straightforward to comment on the variation of this property upon variation of the solvent polarity

269

Table 2 Fluorescence lifetimes (in nanosecond) and radiative and nonradiative rate constants for DMABN in different solvents Solvent

kf

ACN THF DOX HEP a See text.

ktc (107 S- 1)

kls c (10 s S- l)

~(ns)

Ref.

( 1 0 7 S- 1)

0.50 1.05 1.96 3.82

8.64 5.91 7.47 4.12

1.36 1.57 1.62 2.15

4.4 4.4 3.88 3.4

[32] a [4] [12] b

b r value in n-hexane has been used as such.

in a detailed way. Considering the equation kls c

=

t.ICT + (1 --

ot KISC

I", ~LE

ol )KIS C ,

we see that kls c --~ISCt'~CT in ACN solution while kts c = kLEc in apolar solvents. Since the electronic character of the state is changed upon changing the solvent polarity, this should be taken into account when the variation of kls c is discussed. Similarly, the variation of klc with the variation in solvent polarity also becomes less predictable quantitatively. From the literature values of the fluorescence nanosecond (ns) lifetimes 2 of the DMABN in different solvents (ACN, DOX and HEP) the rate constants of the fluorescence, ISC and IC processes are calculated and presented in Table 2. In n-heptane the ICT species is formed to a limited extent and, hence, the ns lifetime can be associated to the LE species. In a similar way, in the ACN solution, the ns lifetime corresponds to the ICT state, as there is hardly any LE emission. Thus, the rate c o n s t a n t s ( k f , klc and kls c) can be considered to correspond to the LE state and the ICT state in HEP and ACN solvents respectively since in the respective solvents the other species is virtually non-existent. In the other solvents, however, we have two different emission bands. In these solvents, the life-

2 A biexponential decay is observed for either of the LE or the ICT emissions, the fast component being a few picoseconds only in ordinary solvents and at room temperature [4,33,34]. The nanosecond lifetime obtained upon analysing the decay as a single exponential equals, within the experimental error limit, the long ~(ns) obtained upon analysing the fluorescence decays as biexponential. This r can, however, be considered as an average effect coming from the two species as there exists a rapid equilibrium between the two excited states [33,34].

270

N. Chattopadhyay et aL / Chemical Physics Letters 264 (1997) 265-272

time values can be considered as an average over the two excited singlet states as there is a rapid equilibrium between the two excited species [35]. Thus, the different rate constants in these solvents should be considered as a weighted average over the two emitting singlet states (vide Appendix). The calculation of the rate constants reveals that with an increase in the solvent polarity the rate constant for the nonradiative internal conversion (IC) deactivation process (k~c) increases while that of the ISC process (kls c) decreases appreciably (see Table 2). However, both effects cancel each other and hence the net lifetime of the fluorophore is not strongly solvent dependent. The lowering in the fluorescence quantum yield in the polar solvents is ascribed to the significant decrease in the kf value in these solvents.

5. Conclusions A laser induced optoacoustic spectroscopic study reveals that the non-radiative channels have a prominent role in the photophysics occurring from the ICT state. The intersystem crossing (ISC) is the main deactivation channel for DMABN irrespective of the solvent. However, for DMABA in polar solvents, the ICT state prevails and is deactivated mainly by internal conversion. The participation of the n ~ * triplet state of DMABA in apolar solvents has been proposed based on the high value of the ISC yield in these solvents.

Acknowledgements NC would like to thank the Research Council, KU Leuven for a fellowship. JR thanks IWT for financial support. MVDA is an 'Onderzoeksdirecteur' of the Belgian 'Fonds voor Kollektief Fundamenteel Onderzoek'. The continuing support of FKFO and DWTC through IUAP III-40 and IUAP II-16 is gratefully acknowledged.

that can be reduced to a bicompartmental system. In this system 'average' rate constants for the excited state decay processes can be determined when the relaxation of the excited state equilibrium (kEt and kl2) between 1" and 2* is fast compared to the excited state decay processes k01 and k02 (thermodynamic limit). ~

k21 k12

ll, o,

l

k02

When it is assumed that compartment 2 is empty in the ground state the quantum yield of a photophysical or photochemical process (fluorescence, intersystem crossing, internal conversion) occurring from l* and 2* with a rate constant kj and k 2 respectively, is given by

klY

61:-62

ktY

X Y - k21kt2

ko2k21-t- kolkl2 + kolk02 '

k2 k21

k2 k21

XY

k21kl2

--

k02 k21 -I- kol kt2 -I- kol k02 '

with X = k01 + k2t and Y = ko2 + k12. The overall yield 4) -- 6j + 62 for this process is now:

6=

ktY + k 2k 2 t

ko2 k21 -I- k01kl2 4- k01k02

Under conditions where k02 << kl2 this equation can be simplified to

6--

klk~ 2 + k2k21 koEkEl + kolk12

Dividing the denominator and numerator by k2~ + k12 and putting P = (klE)/(k12 -I- k21) yields k IP + k2(1 - P )

Appendix A. 'Average' rate constants in bicompartmental systems Often photophysical processes occur in systems

6=

k02(1

-

p ) + kolP

In the case where both species 1 and 2 are present in the ground state, a fraction /3 of the excitation

N. Chattopadhyay et al. / Chemical Physics Letters 264 (1997) 265-272

will be absorbed by 2 and a fraction (1 - / 3 ) will be absorbed by 1, the above expressions become (1 -/3)k~Y 4'1 =

ko2k21 + kotkl2 + kink02

+

/3klk12 ko2 k2t + kolk~2 + kojko2

in the ground state does not influence the quantum yields. For a two compartmental system the fluorescence decay of 1" and 2* is always described by an algebraic sum of two exponential decaying terms with decay rates "/1.2- Note that, contrary to the amplitudes, YL2 do not depend upon /3

y,.2=½(X+Y+[(X-y)2-4kI2k2,]'/2}.

and

4'2 =

(1 - /3 )k2k21 ko2 k21 + kolkl2 4" kolko2 /3k2 X

4"

When ko2 ,~: kt2 and kot << k21, ')'1 and "Y2 can be approximated by [35]

Yl ~

ko2k2j + kolkl2 + kolk02 '

(1 - /3 )klY + /3klkl2

k02 k2j + kolkl2 + kink02 '

or

k, [(1 -/3)k02 + k,2]

k2[k21 +/3k0! ] k02 k2~ + kolkl2 + kolk02

k l [ ( 1 - - / 3 ) k o 2 + k 1 2 ] 4- k2[k21 4-/3kol ] ko2k21 + kolkl2 + kink02

Under conditions where (1 -/3)k02 << kl2 and flkol << k2~ this equation can be simplified to

4'=

klkl2 + k2k21 k02k21 + k0lkl2 "

This can again be written as:

4)=

Yl

4'Yl = k i P + k 2 ( 1 - P)" This means that when ko2 ~:: k12 and ko~ "~¢ k21, the product, 4'TI, can be considered as a weighted average value for the rate constant k.

References

This yields for 4)

4'=

kiP+k2(1 -P)

ko2k2j + kmkl2 + koiko2

and 4'2 =

= Pkol + (1 - P ) ko2 ,

In this case the condition (1 -/3)ko2 ~ kl2 and

4'= (1 - / 3 ) k 2 k 2 1 +/3k2X

4'j=

k21 4- k12

/3ko! "~ k21 is fulfilled afortiori. Hence:

k02 k21 + kolkl2 + kink02

and 4'2 =

kojkl2 + ko2k21

]12 = k12 + k21 •

or

4'1 =

271

kiP+k2(1 -P) ko2(l - P ) + k o l P "

Hence under the condition (1 -/3)ko2 << kl2 and /3ko~ << k2~ a possible population of compartment 2

[I] E. Lippert, W. Liider and H. Boos, in: Advances in molecular spectroscopy, ed. A. Mangini (Pergamon Press, Oxford, 1962) p. 443. [2] K. Rotkiewicz, K.H. Grellmann and Z.R. Grabowski, Chem. Phys. Lett. 19 (1973) 315. [3] E. Lippert, W. Rettig, V. Bonacic-Koutecky, F. Heisel and J.A. Mieh6, Adv. Chem. Phys. 68 (1987) 1. [4] W. Schuddeboom, S.A. Jonker, J.M. Warman, U. Leinhos, W. K~hnle and K. Zachariasse, J. Phys. Chem. 96 (1992) 10809. [5] Th. vonder Haar, A. Hebecker, Y. ll'ichev, W. K~hnle and K. Zachariasse, Proceedings of the 54th international meeting of physical chemistry, ed. A. Tramer (American Institute of Physics Press, New York, 1996) p. 295. [6] A. Onkelinx, F.C. De Schryver, L. Viaene, M. Van der Auweraer, K. lwai, M. Yamamoto, M. Ichikawa, H. Masuhara, M. Maus and W. Rettig, J. Am. Chem. Soc. 118 (1996) 2892. [7] L. Serrano-Andr~s, M. Merch~in,B.O. Roos and R. Lindh, J. Am. Chem. Soc. 117(1995) 3189. [8] S. D~ihne, W. Freyer, K. Teuchner, J. Dobkowski and Z.R. Grabowski, J. Lumin. 22 (1980) 37.

272

N. Chattopadhyay et al. / Chemical Physics Letters 264 (1997) 265-272

[9] J. Dobkowski, E. Kirkor-Kaminska, J. Koput and A. Siemiarczuk, J. Lumin. 27 (1982) 339. [10] K. Bhattacharyya and M. Chowdhury, Chem. Rev. 93 (1993) 507. [11] A.-D. Gorse and M. Pesquer, J. Phys. Chem. 99 (1995) 4039. [12] G. K6hler, G. Grabner and K. Rotkiewicz, Chem. Phys. 173 (1993) 275 [13] D. Majumdar, R. Sen, K. Bhattacharyya and S.P. Bhattacharyya, J. Photochem. Photobiol. A 73 (1993) 177. [14] M. Hashimoto and H. Hamaguchi, J. Phys. Chem. 99 (1995) 7875. [15] Z.A. Dreger, J.M. Lang and H.G. Drickamer, Chem. Phys. Lett. 232 (1995) 351. [16] Z.A. Dreger, J.M. Lang and H.G. Drickamer, J. Phys. Chem. 100 (1996) 4646. [17] D.R. Keams and W.A. Case, J. Am. Chem. Soc. 88 (1966) 5087. [18] T. Takemura and H. Baba, Bull. Chem. Soc. Jpn. 42 (1969) 2756. [19] S. Nagaoka and N. Hirota, Bull. Chem. Soc. Jpn. 56 (1983) 3381. [20] W. Rettig, D. Braun, P. Suppan, E. Vauthey, K. Rotkiewicz, R. Luboradzki and K. Suwinska, J. Phys. Chem. 97 (1993) 13500. [21] R.J. Visser, C.A.G.O. Varma, J. Konijnenberg and P.C.M. Weisenbom, J. Mol. Struct. 114 (1984) 105. [22] P.C.M. Weisenborn, C.A.G.O. Varma, M.P. De Hass and L.M. Warman, Chem. Phys. Lett. 129 (1986) 562.

[23] J. Lipinski, H. Chojnacki, Z.R. Grabowski and K. Rotkiewicz, Chem. Phys. Lett. 70 (1980) 449. [24] A. Nag, R. Datta, N. Chattopadhyay and K. Bhattacharyya, Chem. Phys. Lett. 157 (1989) 83. [25] S. Kundu and N. Chattopadhyay, J. Mol. Struct. 344 (1995) 151. [26] S. Kundu and N. Chattopadhyay, J. Photochem. Photobiol. A 88 (1995) 105. [27] P. Van Haver, L. Viaene, M. Van der Auweraer and F.C. De Schryver, J. Photochem. Photobiol. A 63 (1992) 265. [28] M. Van der Auweraer, L. Viaene, P. Van Hayer and F.C. De Schryver, J. Phys. Chem. 97 (1993) 7178. [29] P. Van Haver, N. Helsen, S. Depaemelaere, M. Van der Auweraer and F.C. De Schryver, J. Am. Chem. Soc. 113 (1991) 6849. [30] H. GSmer and H.J. Kuhn, J. Phys. Chem. 90 (1986) 5946. [31] K. Rotkiewicz, Spectrochim. Acta 42A (1986) 575. [32] K. Rotkiewicz, Z.R. Grabowski and J. Jasny, Chem. Phys. Lett. 34 (1975) 55. [33] J.M. Hicks, M.T. Vandersall, Z. Babarogic and K.B. Eisenthai, Chem. Phys. Lett. 116 (1985) 18. [34] Y. Wang, M. McAuliffe, F. Novak and K.B. Eisenthal, J. Phys. Chem. 85 (1981) 3736. [35] W.R. Ware, D. Watt and J.D. Holmes, J. Am. Chem. Soc. 96 (1974) 7853.