The time dependence of the low-temperature ... - ACS Publications

---

4626

J. Phys. Chem. 1984,88, 4626-4631

-

suggested that the (n 5s) charge-transfer state should be responsible. However, charge-transfer states involving the s orbital do not spin-orbit couple with the lowest triplet stateZoand should not act as a perturbing singlet state. These authors also proposed the (4d ?r*) reverse chargetransfer state as another candidate. Carsey and McGlynn also included these reverse charge-transfer states in their computation, and the results seem to indicate that the contribution from the reverse charge-transfer states is rather important. We are, however, of the opinion that the reverse

charge-transfer states leading to Ag2+NOz2-are at quite high energy and therefore do not contribute to the radiative transition from the lowest triplet state. Further, inclusion of the reverse charge-transfer states conflicts with the finding12that the colors of nontransition-metal salts is largely correlated with electron affinity of the metal cation. For these reasons, we believe that the (n 5p,) charge-transfer state proposed above is the mQst reasonable perturbing charge-transfer state.

-

-

Acknowledgment. We thank Professor Takeshi Nakajima and Professor S.P. McGlynn for stimulating discussions and criticism. Registry No. AgN02, 7783-99-5; Ag+, 14701-21-4.

-

(20) In a recent article, the role of the (n 5s) charge transfer-state is denied. K.E. Gotberg and D. S. Tinti, Mol. Phys., 47, 97 (1982).

The Time Dependence of the Low-Temperature Fluorescence of Tryptophan7 Eva F. Gudgin-Templeton and William R. Ware* Photochemistry Unit, Department of Chemistry, The University of Western Ontario, London, Ontario, Canada N6A 587 (Received: November 29, 1983; In Final Form: March 20, 1984)

The time dependence of the fluorescence of tryptophan (Trp) and indole in ethylene glycol/water solution has been investigated in the temperature region 77-300 K. It has previously been oberved that the large Stokes shift in the indole emission disappears at low temperature. Timeresolved spectra of indole as a function of temperature show that while at 77 K and room temperature, there is time dependenceof the fluorescence spectrum; in the temperature region where the fluorescence spectrum is changing, there is a red shift in the emission with time after excitation. Similar results are observed for Trp, except that at room temperature the early time spectrum is blue-shifted with respect to the late time spectrum. This is due to the differing spectral distribution of the two components in the Trp fluorescence decay. The fluorescence decay of indole is single exponential at room temperature and below 250 K, while it is nonexponential in the region of spectral change. The decay of Trp is double exponential at room temperature; in the region 150-250 K the decay is nonexponential, and below 150 K Trp decays with a single lifetime identical with that of indole. These results are discussed in terms of some of the models which have been proposed to explain the double-exponentialdecay of some Trp derivatives, and it is suggested that excited-state complexation with the solvent which lowers the 'Laelectronic state below the ILb state is necessary before double-exponentialdecay kinetics will be observed for Trp derivatives in general.

Introduction The low-temperature emission properties of tryptophan (Trp) are of great interest in attempting to understand the reason for its double-exponential decay kinetics in solution at room temperature. The emission spectra, lifetimes, and fluorescence quantum yields of indole and tryptophan in polar solvents are extremely temperature sensitive. However, while spectra and quantum yields as a function of temperature over a wide range have been reported previo~sly,'-~ it is only recently that the time dependence of the low-temperature fluorescence of indole derivatives has been investigated. Three recent reports have appeared which describe investigations of this aspect. Lakowicz and Balter6 have measured the time-resolved fluorescence spectra of N-acetyltryptophanamide (NATA) in propylene glycol from 205 to 3 13 K. Lami' has measured the fluorescence lifetime of indole and 2,3-dimethylindole in glycerol down to 213 K. Meech et a1.* have observed the time-resolved fluorescence spectra and fluorescence decay of 1,3-dimethylindole in 1-butanol from 85 to 280 K. In all of these cases, the fluorescence spectrum was observed to blue shift as the temperature was lowered. In the temperature region where the maximum of the steady-state spectrum was shifting, a red shift of the emission maximum with time after excitation was observed with the use of time-resolved spectroscopy. They concluded that the fluorescence in fluid polar solvent at high temperature arises not from a pure 'La state but from a state with significant intramolecular charge-transfer character. All of the indole derivatives used in the above studies give single-exponential decay at room temperature. The time de+ Publication

pendence of the low-temperature fluorescence of other indole derivatives such as Trp, which decay with a double exponential, has not been reported. The timeresolved spectra and fluorescence decays of tryptophan and indole in ethylene glycol/water glass as a function of temperature have been investigated here in an effort to elaborate on the mechanism responsible for the double-exponential decay kinetics of tryptophan,

Experimental Section Materials. L-Tryptophan (99%, Aldrich) was recrystallized three times from 1-propanol. Indole (99+%, Aldrich) was recrystallized once from ethanol. Ethylene glycol (EG, Fisher reagent) was purified by adding activated charcoal and distilling under vacuum at 5 mmHg. Ethylene glycol/water (EGW) solutions were 1:l (v/v) ethylene glycol to triply distilled water. Solutions were generally lC3M in the fluorophore for front-face measurement of fluorescence and la" M otherwise. All solutions were air saturated and were prepared immediately before use. (1) J. Eisinger and G. Navon, J. Chem. Phys., 50, 2069 (1969). (2) L.C. Pereira, I. C. Ferreira, M. P. F. Thomaz, and M. I. B. M. Jorgc, J. Photochem., 9,425 (1978). (3) C. Conti and L.S . Fmter, Biochem. Eiophys. Res. Commun., 57,1287 (1974). (4) W.C. Galley and R. M.Purkey, Proc. Natl, Acad. Sci. U.S.A., 67, 1116 (1970). (5) S. Suzuki, T. Fujii, A. Imai, and H. Akahori, J. Phys. Chem., 81,1592 (1977). (6) J. R. Lakowicz and A. Balter, Photochem. Photobiol., 36, 125 (1982). (7) H. Lami, 11, Nuovo Cimento SOC.Ita/. Fis. E, 63B, 241 (1981). (8) S. R. Meech, D.Philips, and A. G. Lee, Chem. Phys. Lett., 92,523 (1982).

No. 326.

0022-3654/84/2088-4626%01.50/0 , , I

0 1984 American Chemical Societv -

Low-Temperature Fluorescence of Tryptophan

The Journal of Physical Chemistry, Vol. 88, No. 20, 1984 4627

Spectra and Decay Measurements. Absorption spectra were measured on a Cary 15 spectrophotometer. Steady-state fluorescence spectra were recorded on a Perkin-Elmer 650-40 spectrophotometer. Lifetime and time-resolved spectral measurements were performed using laser excitation in order to achieve improved time resolution; also, it was necessary to have sufficient fluorescence intensity to permit a polarizer to be placed on the emission side, set at the “magic” angle, to eliminate effects due to rotational depolarization. A few lifetime measurements are also reported which were made using flash lamp or laser excitation without any polarizers. Fluorescence lifetime measurements were performed using either of two instruments. One is the PRA-3000 nanosecond spectrofluorometer system (described in detail in ref 9) which uses a hydrogen flash lamp with a full width at half-maximum of 2-4 ns. Excitation and emission monochromators with 25-nm bandpass were generally used; the emission monochromator can be replaced by a cutoff filter to observe the total emission. A lownoise cooled photomultiplier observes the fluorescence; its &pulse single-photon response is quite insensitive to wavelengths between 300 and 400 nm.9 The other system used was a modified PRA-3000 system using laser excitation, which has been previously described in detail.*O The excitation source consisted of a mode-locked argon ion laser (Coherent Radiation CR-8) which synchronously pumped a rhodamine 6G dye laser (Coherent Radiation Model 590). The mode locker was driven by a Coherent Radiation Model 467 driver. The repetition rate of pulses from the dye laser was 92.8 MHz. The dye laser pulses were frequency doubled with an angle-tuned KDP crystal (Cleveland Crystal Co.). The resulting wavelength used for excitation was usually 283 nm. When reduction of the repetition rate was desired, a Coherent Radiation low-voltage Pockels cell (Model 28) was inserted into the beam, which reduced the repetition rate by a factor of 400. For high repetition rates, the “start” and ”stop” pulses were interchanged. The full width a t half-maximum of the laser pulse profile as observed by the photomultiplier tube was less than 500 ps at the wavelengths used. The experimental decay data I(t) are stored in the multichannel analyzer. Z(t) is the convolution I ( t ) = g(t)*F(t)

(1)

where F(t) is the true fluorescence decay. The instrumental response function g ( t ) is the convolution where Zo(Ae,,t) is the excitation flash profile and RPM(Aem,t) is the response function of the photomultiplier tube at the emission wavelength. Because R ~ isMinsensitive to wavelength in this case, g ( t ) can be obtained from scattering light at the excitation wavelength, without wavelength corrections being necessary. Suspensions of alumina or Ludox were used with the MINC-I 1 computer; the best-fit decay law F(t) was obtained by iterative convolution with either a single- or double-exponential decay function of the form F ( t ) = CA1e-‘/rt

(3)

I

Acceptability of fit was judged from the xzvalue, the randomness of the weighted residuals for the fit, and the autocorrelation function. In all cases where multiexponential fits are reported, they were necessary in order to lower the x2 value below 1.2 and to randomize the residuals. The relative quantum yield of fluorescence of each component was calculated from the equation (4)

A typical decay measured with laser excitation is illustrated in ~~

(9) E. Gudgin, R. Lopez-Delgado, and W. R. Ware, Can. J . Chem., 59, 1037 (1981). (10) W. R. Ware, M. Pratinidhi, and R. K. Bauer, Rev. Sci. Instrum., 54, 1148 (1983).

100

10

30 Time/ ns

50

Figure 1. Sample decay with laser excitation. The sample is 10” M Trp methyl ester in pH 5.0 0.1 M acetate buffer at 20 “C. The decay data are shown as dots, the fitted decay line is drawn through the data, and the dashed line is the laser pulse profile. The retrieved parameters for the fit were T~ = 0.25 ns, r2 = 0.89 ns, I, = 0.24, and x2 = 1 . 1 7 .

Figure 1; the decay is that of L-tryptophan methyl ester at pH 5, and the fitted curve (solid line) is superimposed on the data. The dashed curve is the laser pulse profile distorted by the detection system. Low-Temperature Measurements. Low-temperature measurements were performed using either an Oxford Instruments Model CF-204 cryostat or a home-built cryostat, both of which used cold nitrogen gas as coolant. In the Oxford Instruments cryostat, the sample was in a helium atmosphere and was cooled by heat transfer to the N2coolant; a heater element next to the sample provided the thermostating. In the homemade cryostat the sample was cooled directly by the N 2 gas, which was mixed with warm N2 to achieve the desired temperature. The temperature was measured with a thermocouple inserted directly in the sample, and the sample was allowed to equilibrate for 20 min at each temperature. The sample was contained in a 2-mm quartz Suprasil cuvette supported by a brass holder. Front-face excitation was used due to the cracking of the EGW glass formed at low temperature. No difference was observed between measurements performed by cooling the sample directly to 77 K and then heating or by using gradual cooling. A solvent blank was measured at room temperature and 77 K to ensure that it did not contribute to the fluorescence. Spectra are uncorrected for the spectral response of the system. Time-Resolved Spectra. Time-resolved spectra were measured by using the modified PRA-3000 with laser excitation. The multichannel analyzer was used in the multichannel scaling mode, and the emission monochromator was scanned with a stepping motor. The time window was set with discriminators on the analyzer. Time zero is defined as the beginning of the laser pulse profile, which has a full width at half-maximum of less than 500 ps, as observed by the photomultiplier tube. Multiple scans were performed to improve statistics. The emission monochromator was set at a IO-nm band-pass, and due to the statistical noise in the data, fine structure in the spectra was not resolved. The pockels cell was used to reduce the repetition rate of the laser excitation in all time-resolved spectral measurements, so that no fluorescence generated by previous excitation pulses was observed at early times. Results Time-Resolved Spectra. It has been previously reported’ that the emission maximum of Trp shifts to the blue as the temperature is lowered and the spectrum becomes less broad; by about 100 K no further effect of temperature is observed. Phosphorescence is observed for A,, > 400 nm. As the shape of the emission changes at lower temperature, the emission intensity and quantum yield increase, although not to the same degree.’ These observations are also true in general for 1-MeTrp and 1,2-dimethylindole.’ The spectral changes for Trp are most pronounced in the temperature region 170-230 K, while the quantum yield of

4628 The Journal of Physical Chemistry, Vol. 88, No. 20, 1984

Gudgin-Templeton and Ware

hl n m

Figure 2. Time-resolved fluorescence spectrum of indole in EGW as a function of temperature. Spectra were taken in the time window 0-1.8 ns after excitation.

290

330

370 hinm

h l nm

Figure 6. Time-resolved fluorescence spectrum of tryptophan in EGW as a function of temperature. Spectra were taken in the time window 13.3-18.0 ns after excitation.

41 0

Figure 3. Time-resolved fluorescence spectrum of indole in EGW as a function of temperature. Spectra were taken in the time window 13.3-18.0 ns after excitation.

Figure 7. Time-resolved spectra of Trp in EGW as a function of temperature. Time window: (a) 0.1-1.2 ns, (b) 13.3-18.0 ns. TABLE I: Effect of Temperature on Indole Decav in EGWb fit

T/K

Figure 4. Raw data for time-resolved spectrum of indole in EGW at 251 K. Spectra were taken in the time window 13.3-18.0 ns after excitation.

I

93 93 105 151 181 215 229 231 241 251 273 295

rl/ns 5.3 5.5 5.4 5.0 5.5 2.0 4.7 2.7 6.2 5.5 5.1 4.6

h 0.1 k 0.1 0.1 f 0.1 f 0.1 k 0.1 0.1 0.1 f 0.1 0.1 0.1 0.1

* *

* * *

*

II 1.00 0.87 1.00 1.00 1.00 0.13 0.65 0.10 0.87 1.00 1.00 1.00

r2/ns

I2

3.1 f 0.1

0.13

6.0 f 0.1 12.3 f 0.1 6.8 f 0.1 11.4 0.1

0.87 0.35 0.90 0.13

acceptabld? X,,/nm ves

no Yes Yes Yes

*

no no no no

Yes Yes Yes

315 315' 320 3 20 >305 325 340 >305 >305 >305 >305 >305

'No emission polarizer. bExcitationwith laser at 285 nm. Observation through monochromator (IO-nm bandwidth) or 305-nm cutoff

filter, with emission polarizer set at magic angle.

hinm

Figure 5. Time-resolved fluorescence spectrum of tryptophan in EGW as a function of temperature. Spectra were taken in the time window 0.1-1.2 ns after excitation.

-

fluorescence stays almost constant at af 0.6 in this temperature range. The time-resolved fluorescence spectra of indole in EGW as a function of temperature are illustrated in Figures 2 and 3. A spectrum showing the raw data is illustrated in Figure 4. Figures 2 and 3 illustrate the change in the early time and late time spectra, respectively. At room temperature, there is no difference in the time-resolved spectra and there are small changes in the lowtemperature spectra. In the temperature region 190-250 K the early time spectrum is blue-shifted with respect to the late time spectrum. No change is observed in the steady-state spectrum below 180 K. Time-resolved spectra for Trp are presented in Figures 5-7. For Trp, the room-temperature time-resolved spect;a are not

identical, due to the difference in the emission between T~ and r2 (Figure 7). Since the r 1 spectrum is blue-shifted with respect to the r2 spectrum, the early time spectrum is also blue-shifted. The resolution of the Trp r 1 and T~ components at room tem. perature in EGW is particularly good. At low temperature, howevdr, there is little difference in the early and late time spectra; the spectra are red-shifted slightly compared to the indole spectra, as they are at room temperature. In the temperature range 190-250 K the time-resolved spectra are considerably shifted with respect to one another, as in the case of indole. Lifetime Results. Measurements of the lifetime of indole in EGW as a function of temperature are summarized in Table I. As the temperature is lowered, the lifetime lengthens at first. In the region 215-241 K, neither a single- nor a double-exponential fit is adequate. This is not an unexpected result, as this is in the region where the spectrum is shifting in time. Below 21 5 K the decay is single exponential, with a lifetime constant within error at about 5.3 f 0.2 ns. If, however, the emission polarizer set a t the magic angle is removed, the decay is no longer single exponential; this is due to rotational depolarization effects, which will distort the decay when the rotational reorientation time of the fluorophore is of the same order as the decay time. The use of the emission polarizer eliminates this distortion.

The Journal of Physical Chemistry, Vol. 88, No. 20, 1984 4629

Low-Temperature Fluorescence of Tryptophan

cx-15

TABLE 11: Effect of Temperature on Tryptophan Decay Parameters and Emission Maximum in EGW"

T/K

A,,

17.4

310

100.0 107.1 119.9 124.0 149.9 160.2 165.0 179.8 188.2 200.4 213.9 219.9 220.4 240.0 240.4 260.0 260.4 270.8 280.0 287.2 295.2

300.0

r,/ns 1.6 1.3

I,

r2/ns

I2

0.04 0.04

5.5 5.4

0.97 0.96

5.4 5.4

0.95

311

1.9

0.05

5.4

1.9 2.0

0.05 0.02

5.5 5.5

1.1

0.08 0.04

5.3 6.0 5.5

313

1.9 0.4

322

1.2

3.6 336

3 .O

340

1.8 2.6 2.2

341

2.6 2.3 1.4 1.5 1.o

342

1.1 0.9

0.07 0.01 0.18 0.18 0.02 0.10 0.03 0.19 0.07 0.11 0.09 0.09 0.11 0.06

8.0

8.6 7.7

8.3 7.8 7.8 6.9 6.5 5.7 5.0 4.2 3.4 3.5

1.oo 1.oo

0.95 0.98 0.92 0.96 0.93 0.99 0.82 0.82 0.98 0.90 0.97 0.8 1

0.93 0.89 0.91 0.91

0.89 0.94

ROTAMERS OF TRYPTOPHAN

fit acceptable? yesa no" Yes

no"

Yes no8 no no' no no' no no" no' no

I Il rn Figure 8. Newman projections of a-p rotamers of tryptophan derivatives.

yes" Y €3

yes" Yes yes"

Figure 9. The skeleton of the substituted 3-ethylindole molecule.

Yes

yes" yes" Yes

"Excitation with laser at 285 nm; emission observed at maximum with emission polarizer set at magic angle (10-nm observation bandwidth) except where noted. Errors on lifetimes f O . l ns. bFlashlamp excitation at 280 nm; observation at emission maximum with 25-nm observation bandwidth and no emission polarizer. Table I1 summarizes the measurements of the Trp decay in EGW as a function of temperature. Some measurements in this table were performed using flash lamp excitation without polarizers. The other measurements were performed using laser excitation and an emission polarizer set at the magic angle. The observation of the emission was at the emission maximum, as indicated in the Table 11. Down to about 280 K, there is no significant difference between measurements performed with and without polarizers. The values of r 1 and r z retrieved from the analysis lengthen as the temperature is lowered. In the region 150-270 K, the relative proportions and magnitudes of the two components differ with and without polarizers. The fits are generally unacceptable in this region. Measurements performed without polarizers below 150 K gave decays requiring a doubleexponential decay law; those with the emission polarizer gave a single-exponential decay of lifetime 5.4 f 0.1 ns, comparable with that of indole at low temperature. If one compares the Trp and indole results, as the temperature is lowered, the indole decay has become single exponential by about 215 K, while the Trp decay is not single exponential until below 150 K. The mechanism for the shift in the emission spectrum after excitation with time and temperature in the intermediate temperature region has been proposed to be solvent reorientation'-* (or complexation with solvent) after excitation, causing a change in the emitting state. The initially excited state is the 'Lb state, which is the emitting state at low temperature and in nonpolar solvents. In a fluid polar solvent, reorientation of and/or complexation with the solvent molecules occur after excitation, which stabilizes the 'La state below the 'Lb state. At room temperature this reorientation is much faster than emission. However, in the intermediate temperature range, solvent reorientation is slowed as the medium undergoes transition to a glassy state, emission occurs from both solvent-equilibrated and unequilibrated states, and the fluorescence decay cannot be described by simple exponential kinetics. Thus, for indole, solvent reorientation or complexation is completely frozen out by 215 K. However, the Trp decay only becomes single exponential below 150 K. Either it requires much lower temperatures for solvent reorientation to be frozen out for Trp compared to the case of indole or the decay would still be in-

herently double exponential above this temperature in the absence of solvent reorientation, as it is at room temperature. Discussion

The reason for the double-exponential decay kinetics of Trp and some of its derivatives in solution at room temperature has been a subject of much interest. Some of the models proposed to explain this phenomenon have been (i) simultaneous emission from uncoupled 'La and 'Lb states," (ii) diffusion-controlled quenching of the chromophore by the alanyl chain in the excited state,', (iii) emission from different ground-state CY-@ rotamers (the conformer m ~ d e l ) , ' ~and J ~ (iv) the modified conformer model (the MCM).I5 Explanations i and ii were later discarded by their authors. The MCM is an improvement on the original conformer model. The conformer rnodell3 was based on N M R results which showed the existence of three a-/3 rotamers, interconvertible on the N M R time scale, for the Trp as well as for other derivative~.'~J~ These a-/3 rotamers are illustrated in Figure 8. The two b-7 rotamers were not considered to be relevant to the fluorescence measurements. However, this model has shortcomings, and hence Petrich et al.15 have proposed the "modified conformer model" (the MCM). According to this model, the nonradiative process competing with fluorescence, intersystem crossing, and photoionization is intramolecular charge transfer (ICT) from the excited indole ring to electrophilic substituents on the 3-ethyl chain, where 3-ethylindole is considered to be the parent molecule. The electrophilic substituents are referred to here as R1 and Rz (see Figure 9); for example, for the Trp zwitterion R1 = NH3+ and R2 = COO-. The electrophilicity of the R group (expressed as the carbanion stability of the molecule H-CH,-R) was correlated with its indole quenching ability, either substituted (as indole-3-CHzCHR) or free in solution (as CH3R). The presence of single- or double(11) D. M. Rayner and A. G. Szabo, Can. J . Chem., 56, 743 (1978). (12) G. S. Beddard, G. R. Fleming, G. Porter, and R. J. Robbins, Philos. Trans. R. SOC.London, Ser. A , No. 298, 321 (1980). (13) A. G. Szabo and D. M. Rayner, J. Am. Chem. Soc., 102,554 (1980). (14) G. R. Fleming, J. M. Morris, R. J. Robbins, G. J. Woolfe, P. J. Thistlethwaite,and G. W. Robinson, Proc. Nutl. Acud. Sci. U.S.A.,75,4652 (1978). (15) J. W. Petrich, M. C. Chang, D. B. McDonald, and G. R. Fleming, J . Am. Chem. SOC.,105, 3824 (1983). (16) J. R. Cavanaugh, J . Am. Chem. SOC.,92, 1488 (1970). (17) P. Skrabal, V. Rizzo, A. Baici, F. Bangerter, and P. L. Luisi, Biopolymers, 18, 995 (1979). (18) J. Kobayashi, T. Higashijima, S . Sekido, and T. Miyazawa, Znt. J . Pept. Protein Res., 17, 486 (1981). (19) K. F. Turchin, M. N. Preobrazhenskaya, L. A. Savel'eva, E. S. Belen'kaya, N. P. Kostyuchenko, Yu.N. Sheinker, and N. N. Suvorov, Zh. Org. Khim., 7, 1290 (1971).

4630

The Journal of Physical Chemistry, Vol, 88, No. 20, 1984

Gudgin-Templeton and Ware

exponential decays was predicted on the basis of the relative electrophilicities of R1 and R2 and their proximity to the indole ring in each rotamer. Only the distance of the more electrophilic R group from the ring is considered important in determining the number of lifetimes. Thus, in Trp, the 0.5-ns lifetime would be due to the two rotamers in which the more electrophilic -NH3+ group is equidistant from the ring; the rotamer with only the less electrophilic -COO- group available for charge transfer has the longer 3 4 s lifetime. In molecules such as NATA and the Trp anion, Petrich et al. argue that R, and R2 are essentially equivalent in electrophilicity and hence only one lifetime is o b s e r ~ e d . ~ J ~ This scheme is attractive and correctly predicts the decay kinetics of most derivatives. One objection to the model, however, is that the model seems to predict that singly substituted derivatives which decay with a single-exponential decay law, such as the tryptamine cation, should exhibit double-exponential decays with a long component due to the conformer with no accepting group over the ring (the authors have attempted to rationalize this on the basis of 0-7 conformerslS). The major difficulty with the MCM is the same as with the original conformer model. The apparent relative populations giving rise to the two components do not seem to vary greatly among the derivatives. However, N M R results'* show that the ground-state population distributions for various derivatives can vary greatly, depending on the charge and nature of the substituents. This is inconsistent with the observed fluorescence results if excited-state populations are assumed to be determined by the ground-state populations of conformers. It would be more satisfactory to have a model which would explicitly predict a difference in the spectral distribution of the emission due to the two fluorescing species, as well as taking into consideration the presence of solvent molecules which may affect the nonradiative processes. The previously proposed models regard the molecule as an isolated entity without solvent involvement in the excited state. However, exciplex formation with polar solutes has been postulated to m r for various indole derivatives dissolved in nonpolar solvents.2"-2z The complexes formed with polar molecules in nonpolar solvents fluoresce with a broadened and red-shifted emission spectrum relative to the uncomplexed emission. Whether such complexes are true exciplexes or are already present in the ground state is open to q ~ e s t i o n ~since ~,'~ small changes in the absorption spectrum have been observed in some cases. However, there is no doubt that the interaction is stronger in the excited state. It has been proposed that there are two exciplexing sites in indole and 3-methylind01e,~'N1 (acidic) and C3 (basic), consistent with the ability of 1-methylindole to form exciplexes in spite of the absence of the N-H group. On the other hand, it has been calculatedz5that the 'La state contains contributions from a Rydberg state of the Cz-C3ethylenic double bond and that exciplex formation of indoles with polar solvents occurs through solvation of this state; charge transfer to solvent (CTTS) would be a likely solvent complexation mechanism. It should be noted that the possible existence of such complexes is closely linked to the observation of monophotonic photoionization (PI) occurring throughout the first absorption bands of indole derivative^.^^^^^ The presence of CTTS complexes between polar solvents and indole has been proposed to explain the low-energy threshold of PI.

It is thus apparent that the interplay between the R, and R2 substituents and the complexed solvent should be considered in any thorough explanation of Trp photophysics, and the low-temperature results reported here support this view. At low temperature, where the temperature-dependent nonradiative process which differs for the two fluorescing species no longer contributes, single-exponential kinetics are observed for Trp. In fact, Trp and indole fluoresce identically at low temperature, where the solvent is prevented from forming excited-state complexes before fluorescence occurs; there is no Stokes shift in the fluorescence, and the decays are identical. The effect of the 3-ethyl chain on the photophysics is thus negligible in the absence of excited-state solvent interactions. The solvent must be. important in determining the nonradiative rates in indole and 3-methylindole in fluid medium; hence, it seems likely that fluorescence for all of these derivatives is from a solvent-complexed state. From the work of Eisinger and Navon,' the quantum yield of fluorescence at 80 K in EGW for indole, 1-methylindole, Trp, 1-MeTrp, and 1,2-dimethylindole are the same within error (af= 0.6), implying that the methyl substituents do not effect the low-temperature nonradiative rates. The radiative lifetime in EGW at 77 K for these molecules calculated from T~ = T f / a f is approximately 9 ns, compared to 20-30 ns at room temperature. Meech et a1.* have suggested that this lengthening of the radiative lifetime is unlikely to be due to fluorescence from an unreconstructed 'La state but may arise from a state with internal C T character which would be stabilized in polar media, such as that suggested by the calculations mentioned p r e v i ~ u s l y .It ~ ~may be necessary for the excess electron density to be localized in the spatially extended C2-C3 Rydberg orbital before complexation giving rise to the double-exponential fluorescence decay can occur. The stoichiometry of complexation may differ; for example, measurements on complexing of 3-methylindole with 1-butanol in n-heptane" imply that the 1:l complex fluoresces at about 320 nm while the 1:2 complex fluoresces at 330 nm (the actual predications of stoichiometry may not be correct, since they depend on the validity of the assumptions used in the treatment of the data). Thus, excited-state complexes with solvent of differing stoichiometry may be responsible for differing emission spectra and multiexponential decay kinetics. It is also interesting to note that the calculated value for the exciplex dissociation (37 kJ mol-')21 is of the same order of magnitude as the activation energies for quenching of the lifetimes by the so-called CT process for tryptamine and NAT of 37 kJ The estimated rate constant of 10" s-l is rapid enough to be a possible quenching process;'' thus, excited-state complex dissociation could be the actual quenching mechanism in these molecules. Thus, it might be useful to consider a model in which the charge donor is the solvated indole chromophore (for example, a CTTS complex) and the acceptors are, as in the MCM, the electrophilic R1 and R2 groups (a solvated MCM model). The basic concepts of the MCM remain unchanged by this view, except that the actual quenching process in the altered model could instead be the exciplex dissociation resulting from the ICT. If C T interactions between the chromophore and the solvent are important in stabilizing excited-state CTTS complexes, the electrophilic nature of such substituents, their size, and their proximity to the center of C T could play a major role in complex formation and deactivation.

(20) R. Lumry and M. Hershberger, Photochem. Photobiol., 27, 819 (1978). (21) M. V. Hershberger, R. W. Lumry, and R. Verrall, Photochem. Photobiol, 33, 609 (1981). (22) N. Lasser, J. Feitelson, and R. Lumry, Isr. J. Chem., 16,330 (1977). (23) B. Skalski, D. M. Rayner, and A. G. Szabo, Chem. Phys. Lett., 70, 587 (1980). (24) V. N. Umetskaya, Zh. Fiz. Khim., 56, 2625 (1982). phys., 67, 3274 (1977); E.H.Strickland and c, (25) H. J. Billups, Biopolymers, 12, 1989 (1973). (26) M. Bazin, L. K. Patterson, and R. Santus, J . Phys. Chem., 87, 189 (1983). (27) R. F. Evans, c. A. Ghiron, R.R. Kuntz, and w. A. Volkert, Chem. Phys. Lett., 42, 415 (1976). (28) E. F. Gudgin Templeton and W. R. Ware, Chem. Phys. Lett., 101, 345 (1983).

Conclusions The aim of this work was to provide insight into the reason for double-exponential decay kinetics of Trp. It has been suggested that reorientation of or complexation with the polar solvent in the excited state, which lowers the 'La state below the 'Lb state, is necessary before double-exponential decay kinetics are observed, and hence Trp decays single exponentially at 77 K. The observation that the nonradiative rates at low temperature appear to be identical for a number of indole derivatives (including Trp) indicates that the interaction between the solvent and the chromophore in fluid medium depends on the substitution of the indole moiety as well as the nature of the solvent. Charge-transfer

J. Phys. Chem. 1984,88, 4631-4636 interactions between the chromophore and the solvent could be affected by substitution on the indole ring. The modified conformer model as proposed by Petrich et al.15 does not encompass all the observed results, as well as ignores excited-state solvent interaction. A more general view has been proposed which is consistent with the formation of two solvent complexes in the excited state. Charge transfer to solvent is proposed as a possible interaction mechanism, consistent with the ease of photoionization in highly polar solvents. It seems likely that there is a strong interplay among the chromophore, the substituents, and the solvent in determining the

4631

nonradiative rates and that intramolecular charge transfer as proposed by the MCM and CTTS are competing processes, depending on the charge-accepting abilities of the solvent and substituents.

Acknowledgment. We thank Dr. V. S. Sivasankar for his assistance in the low-temperature measurements and Dr. P. W. M. Jacobs for the use of the Oxford Instruments cryostat. The financial assistance of the Natural Sciences and Engineering Research Council of Canada is acknowledged. Registry No. Indole, 120-72-9;L-tryptophan,73-22-3.

lime-Resolved Electron Paramagnetic Resonance Investigation of Photochemical Reactions. Hydrogen Abstraction in Azaaromatics and Carbonyls Seigo Yamauchi* and Noboru Hirota Department of Chemistry, Faculty of Science, Kyoto University, Kyoto 606, Japan (Received: December 19, 1983)

Transient EPR spectra have been observed for the intermediate radicals in the photoinduced hydrogen abstraction reactions of azaaromatic molecules and carbonyls in isopropyl alcohol. Temperature dependence has been examined at -10 to -90 O C . The intermediate radicals are assigned from the spectra with partially resolved hyperfine structures. From the polarity of the signals the precursory excited states for the reactions are determined and their properties are discussed in terms of the possible CIDEP mechanisms. The reactions take place from the Tl(na*) states with negative D values or the Tl(ar*) states with positive D in the cases of azaaromatics. In aromatic carbonyls the reactions occur from the Tl(nr*) or T2(nr*) states with negative D values, while in aliphatic carbonyls from the Tl(nn*) states with positive D values.

Introduction Numerous studies have been made on transient EPR signals produced by chemically induced dynamic electron polarization (CIDEP) during photolysis and radiolysis. Most of these studies have been devoted to understanding the behavior of transient signals and to establishing the CIDEP mechanisms. Various factors affecting the CIDEP signals such as temperature and viscosities of solvents have been examined in great detail.Id5 Thanks to these investigations the CIDEP mechanisms as well as the behavior of transient CIDEP signals now seem to be well understood. One might then expect that transient EPR studies could be applied widely to obtain information about the details of many photochemical reactions. However, application of transient EPR studies to photochemical reactions seems to have been rather limited so far. Previous studies are mostly concerned with those of aromatic carbonyls, though there have been a few reports on azaaromatic molecules and ketones recently.68 We thought that the application of the laser photolysis EPR technique could be extended to many photochemical systems by using an excimer laser (1) R. Kaptein and V. L. Oosterhoff, Chem. Phys. Lett., 4, 195 (1969). (2) E. J. Adrian, J. Chem. Phys., 54, 3918 (1971). (3) S. K. King and J. K. Wan, J. Am. Chem. Soc., 94, 7197 (1969).

(4) P.K. Atkins, J. K. Dugger, and K. A. McLauchlan, Chem. Phys. Lett.,

.--

24., -34 (1975). -, ( 5 ) P. B. Ayscough, T. H. English, G. Lambelt, and A. J. Elliot, Chem. Phys. Lett., 34, 357 (1975). ( 6 ) S. Basu, K. A. McLauchlan, and G. R. Sealy, Chem. Phys. Lett., 88, ~

84 (1982). (7) S . Basu, K.A. McLauchlan, and A. J. D. Ritchie, Chem. Phys., 79, 95 (1983). (8) A. I. Grant and K. A. McLauchlan, Chem. Phys. Lett., 101, 120 (1983).

0022-3654/84/2088-4631$01.50/0

which provides UV light of shorter wavelength and higher intensity than a commonly used N2laser. We have examined the transient EPR signals of the radicals produced by photochemical reactions of the hydrogen abstraction in azaaromatic molecules and carbonyls. The main purpose of the present study is twofold. One goal is to observe the EPR spectra of the short-lived intermediate radicals and to ascertain the reaction mechanisms. The photochemical reactions of these systems have been studied with flash photolysis techniques using transient optical absorption. While this method is powerful in following the reaction kinetics, it is rather weak in identifying the intermediate species because of broad and structureless spectra. It seems worthwhile to observe the EPR spectra with resolved hyperfine structures to identify the radical species unambiguously. The CIDEP signals also tell whether the reactions take place from the excited singlet or triplet states, thus helping to establish the reaction mechanisms. The second purpose is to examine the nature of the reacting triplet states in relation to the signs of the zero-field splittings (zfs). When a CIDEP signal is produced by the triplet mechanism, the polarity of the signal depends on the signs of zfs. If we know the sublevels populated by the S, TI intersystem crossing (isc), we are able to determine the signs of the zfs from the polarity of the signal. This is useful because the determination of the sign of zfs is generally very difficult and has been made only in a few cases. Alternatively, we can determine the character (na* or a r * ) of the reacting triplet state from the polarity of the CIDEP signal. Investigation of azaaromatics and carbonyls was thought to be useful as an extension of our previous ODMR and EPR investigations on the nature of the T1 states of these molecule^.^-^^

-

(9) E. T.Harrigan and N. Hirota, Mol. Phys., 31, 663 (1976). (10) S. Yamauchi, T.Ueno, and N. Hirota, Mol. Phys., 47, 1333 (1982).

0 1984 American Chemical Society

~