Photophysics and photochemistry of the nitro derivatives of

complex of HBT was found to undergo nonadiabatic proton transfer from HBT to the solvent to form a “free" anion. The activation energy of the format...
0 downloads 0 Views 1MB Size
J. Phys. Chem. 1987, 91, 3517-3524 to also exist as a solute-solvent intermolecular complex through hydrogen bonding. Upon irradiation, the intermolecular hydrogen-bonded complex of S A was found to undergo anti-syn isomerization. The irradiated intermolecular hydrogen-bonded complex of H B T was found to undergo nonadiabatic proton transfer from HBT to the solvent to form a “free” anion. The activation energy of the formation of the anion was found to be dependent upon the hydrogen-bond strength of the solvent. No such anion was observed in trifluoroethanol compared to ethanol or methanol. Theoretical calculations indicate a substantial increase in the

3517

electron density on the nitrogen atom in the excited singlet state compared to the ground state. This is consistent with the fact that ESIPT occurs upon irradiation. Acknowledgment. The laser flash experiments were performed at the Center for Fast Kinetic Research at the University of Texas at Austin, which is supported by NIH Grant R-00886, the Biotechnology Branch of the Division of Research Resources, and the University of Texas. We also thank Dr. Michael Zerner at the University of Florida at Gainesville for kindly supplying us with one theoretical program.

Photophysics and Photochemistry of the Nitro Derivatlves of Salicylideneanillne and 2-( 2‘-Hydroxyphenyl)benzothiazole and Solvent Effects Ralph S. Becker,* Christian Lenoble, and Abudi Zein Department of Chemistry. University of Houston-University (Received: November 14, 1986)

Park, Houston, Texas 77004

The photophysics and photochemistry of the 3’- and 5’-nitro derivatives of salicylideneaniline and 2-(2’-hydroxypheny1)benzothiazole have been studied by microsecond laser flash photolysis and picosecond fluorescence spectroscopy. In all of these compounds there are at least three ground-state species which can be present in solution depending upon the solvents and the concentration. These are (1) the intramolecular hydrogen-bonded enol, (2) an aggregate proposed to be an enol-zwitterion mixed dimer, and (3) an anion. After laser excitation the nitro derivatives of salicylideneaniline preferentially undergo excited-state intramolecular proton transfer to form the zwitterion while the nitro derivatives of 2-(2’-hydroxypheny1)benzothiazole preferentially undergo excited-state intermolecular proton transfer to form an anion in any of the studied solvents. The excited-state proton transfer occurs on the picosecond time scale (kes> 10’’ s-’) while the reverse (ground state) proton transfer is slower by a factor of 106-107. The principal component of the overall activation energy requirement for back proton transfer probably is that required for conformationalchanges. The differencesin photochromism between these molecules are discussed, and it is found that the relative change in charge densities on the nitrogen and oxygen atoms between the ground and lowest excited single state is of importance as shown by quantum mechanical calculations. Competitivephotophysical and photochemical processes (to ESIPT) were also found to occur depending upon the solvent.

Introduction Salicylideneaniline and 2 4 2’-hydroxyphenyl)benzothiazole and their derivatives are well-known for their photochromic properties. It was shown that they undergo excited-state intramolecular proton transfer after irradiation in the crystalline in solution at low temperature,*v4” or in a solid matrix at low A recent picosecond fluorescence study showed that the excited-state proton transfer had no potential barrier and occurred in less than 5 ps even at 4 K.74 The proton transfer was examined in the crystalline state, in the glassy state, and in solution by proton-nitrogen I4N nuclear quadrupole double resonance spectroscopy,lOsiland direct proof of the proton transfer was obtained



(3) Williams, D. L.; Heller, A. J. Phys. Chem. 1970,74; 4473. (41 Cohen. M. D.: Hirschberg. Y.:Schmidt. G. M. J. J . Chem. SOC.1964.

-

2051,’ 2060. ( 5 ) Becker, R. S.; Richey, F. J . Am. Chem. Soc. 1967,89, 1298. (6) Richey, F.; Becker, R. S . J . Chem. Phys. 1968,49, 2092. (7) Barbara, P. F.; Rentzepis, P.M.; Brus, L. E. J . Am. Chem. SOC.1981, 103, 2156. (8) Ding, K.; Courtney, S . J.; Strandjor, A. J.; Flom, S . ; Friedrich, D.; Barbara, P. F. J. Phys. Chem. 1983,87, 1184. (9) Barbara, P. F..;Rentzepis, P. M.; Brus, L. E. J. Am. Ckem. SOC.1980, 102, 2786. (10) Hadjoudis, E.J . Photochem. 1981,17, 355.

0022-3654/87/2091-3517$01.50/0

very recently by picosecond infrared spectroscopy in solution at room temperature.’* Low-temperature (10 K) studies of salicylideneaniline showed that its photochromism is not restricted to only proton migration, corresponding to the keto-enol tautomerism, but in addition involves framework changes which disrupt the hydrogen bond between the oxygen and the nitrogen atom.13 Cohen and Schmidt proposed that the photochromism of the anils or the phenylbenzothiazole derivatives was a topochemically determined phenomenon; that is, there seemed to be no correlation of activity with substituents on either phenyl ring, but the packing arrangements in the crystal was of importance.’ Later, Richey and Becker did not observe any photochromism but did observe thermochromism of the 3’- and 5’-nitro derivatives of salicylicompounds.’,’ We recentlv studiedI4 in detail the DhotoDhvsics and Dhotochemistry o i salicylideneaniline ( S i ) ahd’ 2-(2’-hy&oxypheny1)benzothiazole (HBT) by means of laser flash techniques, (1 I ) Hadjoudis, E.; Milla, F.; Seliger, J.; Blinc, R.; Zagar, V. Ckem. Phys. Left. 1980,47, 105. (12) Elsaesser. T.: Kaiser. W. Chem. Phvs. Lett. 1986.128. 3. 231. (13) Higelin, D.; Sixl, H. Chem. Phys. i983.77, 391. (14) Becker, R. S.; Lenoble, C.; Zein, A. J . Phys. Chem., preceding paper in this issue.

0 1987 American Chemical Society

3518

Becker et al.

The Journal of Physical Chemistry, Vol. 91, No. 13, 1987

and thus it seemed natural to investigate the photophysics of the nitro derivatives of these compounds. In this work, we present the study of the 3’-nitro and 5’-nitro derivatives of salicylideneaniline (SA) and the 3’-nitro and 5’-nitro derivatives of 2 4 2’-hydroxyphenyl)benzothiazole (HBT).

I

18

-

:

12

‘E 0

Z

6

W

SA

HBT

All these compounds are studied by means of laser flash photolysis and picosecond fluorescence emission spectroscopy. Moreover, we will attempt to correlate the experimental results with PPP and INDO/S semiempirical calculations. We will show in the following sections that the nitro derivatives are indeed photochromic under laser flash conditions at room temperature but that the chemistry of the nitro derivatives is complicated by the existence of several species in the ground state and, in some cases, also in the excited state. The purpose of this investigation is (1) to identify the nature of all species thermochemically or photochemically produced, (2) to identify the possible transients in the process and their multiplicity nature, (3) to obtain kinetic data of the transients and/or of the photoproducts, and (4) to investigate what mechanisms determine the photochromism of the nitro derivatives of S A or 2-( 2’-hydroxyphenyl) benzothiazole.

0

2

280

310

340

370

400

430

460

490

520

550

-2 580

WAVELENGTH (nm)

Figure 1. (5’-Nitrosa1icylidene)aniline: (A) absorption spectrum in methylcyclohexane, 5.98 X IO4 M; (---) absorption spectra in pure ethanol, (1) 6.34 X IO4 M, (2) 6.34 X M, (3) 6.34 X IO4 M; (---) absorption spectrum in basic ethanol ([OH-] = 2.5 X IO-* M), 5.27 X lo4 M; transient absorption spectra ( 4 0 D ) in pure ethanol, 2.06 X IO4 M, ( 0 )2 ps and (0) 67 ps after 355-nm laser flash excitation (1 I-ns pulse); (-) emission spectrum in pure ethanol, 2.06 X IO4 M (arbitrary unit), with 355-nm excitation.

,I

3

Experimental Section

The nitro derivatives of salicylideneaniline were prepared as described earlier.5 The phenylbenzothiazole derivatives were prepared either by the reaction of salicylic aldehyde derivative with 2-aminothiophenol in the presence of sodium acetate or from salicylaldehydes and 2-aminothiophenol in the presence of phosphorus tri~hloride.~The compounds were recrystallized several times from acetic acid or toluene. All the solvents used were spectro-grade and were used without further purification. Absorption spectra were recorded on a Cary 15 UV/vis spectrophotometer. The microsecond laser experiments were carried out using the third harmonic (355 nm) of a Q-switched Nd:YAG laser with an 1 1-ns pulse width. The time resolution of the system is -200 ns. The kinetic absorption spectrometer used to detect OD changes (AOD) after excitation has been described previo~sly.’~The intensity of the laser radiation was controlled by the use of wire mesh screens. The common chosen energies were in the range of 5-10 mJ. In this range of energies, the change of optical densities was proportional to the laser intensity. Degassing of solutions was performed by bubbling nitrogen gas through each solutions; in addition, oxygen was bubbled in separate experiments to determine particularly the presence of triplet excited-state transients. The fluorescence lifetimes were measured by using the 355-nm excitation (30-ps fwhm) of a mode-locked Nd:YAG laser. The detection of the emission was performed by a Hamamatsu streak camera system with a time resolution of 10 ps. Molecular orbital calculations have been performed in the ?r-electron approximation introduced by Pariser, Parr, and Pople, with a singly excited CI calculation.I6 For the nitro derivative of salicylideneaniline the parameters were taken as proposed by Hinze and Jaffe,” while for the nitro derivatives of (hydroxypheny1)benzothiazolewe used a set of parameters developed earlier by Dorr et al.183’9 The two-center electron-repulsion integrals

B 280

310

340

370

-1

400

430

460

490

520

550

580

WAVELENGTH (nm)

Figure 2. (5’-Nitrosalicylidene)aniline: (- - -) absorption spectra in Me2S0, (1) 7 . 6 X lo4 M, (2) 7.6 X M, (3) 7.6 X 10” M; transient absorption spectra (4OD) in Me,SO, 1.9 X M, ( 0 ) 14 ps and (0) 230 ps after 355-nm laser flash excitation ( I 1-ns pulse; (-) emission spectrum in Me2S0, 1.9 X M (arbitrary unit) with 355-nm excita-

tion. were estimated by the use of the Mataga-Nishimoto approximatiom20 We also performed calculations of the intermediate neglect of differential overlap (INDO) type2’ using the INDO/S-CI mode1.22*23The configuration interaction (CI) consisted of 196 selected single excitations. All the molecules were assumed to be planar. For the nitro derivative of salicylideneanilinethe C-C, C 4 , C=N, C-N(NO,), and N-O(N02) bonds were taken to be equal to 1.40, 1.36, 1.40, 1.46, and 1.21 A, respectively, and all angles were taken to be 120O. The geometry of the benzothiazole moiety of the nitro derivatives of (hydroxypheny1)benzothiazole was taken in accord with the experimental values of the thiazole derivative^.^^ (20) Nishimoto, K.; Mataga, N. 2. Phys. Chem. (Munich) 1957, 12, 355; 1958, 13, 140.

(15) Becker, R. S.; Freedman, K. J . Am. Chem. SOC.1985, 107, 1477. (16) Pariser, R.; Parr, R. G. J. Chem. Phys. 1953, 21, 446. (17) Hinze, J.; Jaffe, M . H. J . Am. Chem. SOC.1962.84, 540. (1 8) Dorr, F.; Hohlneicher, G.; Schneider, S . Ber. Bunsenges. Phys. Chem. 1966, 70, 803. (19) Woolfe, C. J.; Melzig, M.; Schneider, S . ; Dorr, F. Chem. Phys. 1983, 77, 213.

(21) Pople, J. A,; Beveridge, D. L.; Dobosh, P. A. J . Chem. Phys. 1967, 47, 2026. (22) Ridley, J. E.; Zerner, M. C. Theor. Chim. Acra 1973, 32, 1 1 1. (23) Ridley, J. E.; Zerner, M. C. Theor. Chim. Acra 1976, 42, 223. (24) Nygaard, L.; Asmussen, E.; Hog, J. M.; MaheSwari, R. C.; Nielsen, C. H.; Petersen, 1. B.; Rastrup-Andersen, J.; Sorensen, G . 0 .J . Mol. Sfruct. 1971, 8, 225.

Nitro Derivatives of S A and HBT

Results and Discussion I. Solvent and ConcentrationEfiects on the Absorption Spectra of 3’- and 5‘-Nitro Derivatives of Salicylideneaniline and 2(2’-Hydroxyphenyl)benzothiazole.Figures 1 and 2 show the absorption spectra of 5’-nitroanil in methylcyclohexane, ethanol, and dimethyl sulfoxide as a function of concentration. In methylcyclohexane the absorption spectrum of 5’-nitroanil has its long-wavelength absorption band at 305 nm, with a shoulder at 333 nm and an onset at 380 nm. The absorption spectrum in methylcyclohexane was not seen to be sensitive to the concentration to 5 X 10” M. in the range of 5 X In ethanol, at a concentration of -6 X lo4 M the extinction coefficient at 305 nm was smaller than in methylcyclohexane and a new band appeared with a maximum at 430 nm. At lower concentration, keeping the product of concentration and path length constant, the intensity of the band at 305 nm decreased as the intensity of the long-wavelength absorption increased and underwent a blue shift. Finally, at low concentration (-6 X lo4 M) the absorption maxima were found to be at 406, 362, and 305 nm with an onset at 490 nm (Figure 1). N o isosbestic point was observed, indicating that more than two species were involved in the equilibrium. When sodium hydroxide (2.5 X M) was added to an ethanolic solution of 5’-nitroanil (6 X lo4 M), the anion produced had two absorption maxima, at 406 and 367 nm. Cohen and Flavian had also observed the same result in their study of 5’nitroanil in ethanoL2 They found that the low-concentration absorption spectrum of 5’-nitroanil in ethanol was very similar to the absorption spectrum of the chemically generated anion.2 Our results confirm their observation. Moreover, they found that when a concentrated solution (- lo4 M) of 5’-nitroanil was left in the dark for several hours, the final absorption spectrum obtained was very similar to that of the low-concentration solutions ( 10” M).2 We also studied the absorption spectra of 5’-nitroanil in acetonitrile, ether, and dimethyl sulfoxide. In acetonitrile and ether the long-wavelength absorption band (-430 nm) had a much weaker intensity than in ethanol; moreover, when the concentration was decreased, the intensity of the long-wavelength band showed almost no increase in intensity, in contrast to the ethanol result. In dimethyl sulfoxide, the long-wavelength absorption band (Amx 420 nm) increased in intensity as the concentration was decreased and underwent a red shift from 420 to 445 nm. Moreover, a new band was observed with a maximum at 379 nm. No isosbestic point was observed, see Figure 2. From the above results, we can draw the following conclusions. There are several species in the solvents studied. In a nonpolar non-hydrogen-bonding solvent such as methylcyclohexane, only one species is clearly observed, which is the intramolecular hydrogen-bonded enol with a maximum of absorption at 305 nm and a shoulder a t 333 nm. In polar solvents that can participate in hydrogen bonding such as ethanol or dimethyl sulfoxide, two more species can be characterized. At high concentration, an aggregate is observed with a maximum of absorption at 430 nm in ethanol and 422 nm in Me2S0. At low concentration, the aggregate dissociates and we observed the anion of the anil with two maxima of absorption at 367 and 406 nm in ethanol and 379 and 445 nm in Me2S0. These results are in harmony with the previous work of Cohen and Flavian.2 Note that previous absorption studies of nitro derivatives of anils by Becker et al. showed that, in 3-methylpentane at 77 K, a new absorption band appeared a t 430 nm compared to the room-temperature absorption spectrum.5,6 This absorption band (430 nm) was thought to belong to a cis-keto form of the nitro derivative^.^,^ We now, however, believe that the absorption band is due to the aggregate as shown by the similar location of the long-wavelength absorption band in hydrogen-bonding solvents at high concentration a t room temperature (or in nonpolar, non-hydrogen-bonding solvents at low temperature, 77 K). In polar but non-hydrogen-bonding solvents such as acetonitrile, we observed the aggregate with a maximum of absorption at 430 nm with low intensity. Upon dilution, the aggregates do not

The Journal of Physical Chemistry, Vol. 91, No. 13, 1987 3519

280

310

340

370

400

430

490

460

520

550

580

W A V E L E N G T H (nm)

Figure 3. (3’-Nitrosalicylidene)aniline: ( A ) absorption spectrum in M; ( - - - ) absorption spectra in pure methylcyclohexane, 5.78 X ethanol, (1) 5.7 X lo4 M, (2) 5.7 X IO“ M; (---) absorption in basic M), 5.24 X lo4 M; transient absorption ethanol, ([OH-] = 2.5 X spectra in pure ethanol, 2.8 X M, ( 0 ) 0.8 ps and (0)40 ps after 355-nm laser flash excitation (1 1-ns pulse); (-) emission spectrum in M (arbitrary unit) with 355-nm excitation. pure ethanol, 2.8 X 3

10

N

-

0

1 280

I 310

340

370

400

430

460

490

520

550

580

WAVELENGTH (nm)

Figure 4. (3’-Nitrosalicylidene)aniline: (- - -) absorption spectra in M, (3) 8.79 X IOd M; Me2S0, (1) 8.79 X lo4 M, (2) 8.79 X transient absorption spectra in Me2S0, 2.9 X lo4 M, ( 0 )8 ps and (0) 200 ps after 355-nm laser flash excitation (1 1-ns pulse); (-) emission spectrum in Me2S0, 2.9 X lo4 M, (arbitrary unit) with 355-nm excitation.

15

; i 0

10

0.5

0

250

280

310

340

370

400

430

460

490

520

550

WAVELENGTH (nm)

Figure 5. 2-(2’-Hydroxy-5’-nitrophenyl)benzothiazole: (- - -) absorption spectra in pure ethanol, (1) 6.83 X lo4 M, (2) 6.83 X M, (3) 6.83 X 10” M; (---) absorption spectrum in basic ethanol, ([OH-] = 2.5 X lo-* M), 5.73 X lo4 M; transient absorption spectra in pure ethanol, 2.67 lo4 M, ( 0 ) 1 ps and (0) 6.5 ps after 355-nm laser flash excitation (1 1-ns pulse); (-) emission spectrum in pure ethanol, 2.67 X M X

(arbitrary unit) with 355-nm excitation.

dissociate as shown by the absence of any significant changes in the absorption spectrum. Therefore, it seems that, in polar but non-hydrogen-bonding solvents, the anil cannot transfer its proton to the solvent and the aggregate is stabilized even at low concentration. The structure of the aggregate will be discussed later,

3520

The Journal of Physical Chemistry, Vol. 91, No. 13, 1987

r--

Becker et al. TABLE I: Calculated T,T* Electronic Transition Energies and Oscillator Strengths for the Enol and the Anion of the Nitro Derivatives of Salicylideneaniline and

2-(2’-Hydroxyphenyl)benzothiazole

,-.

-

5

-

0

E

-

0

5 1

0

w

transition energy, nm (osc str)

2

0

comDd 3‘-nitro

355 (0.95) 312 (0.16)

34

3’-nitro anion

1 7

5‘-nitro

0

5’-nitro anion

340 (1.02) 310 (0.03)

310

340

370

400

430

460

490

520

550

580

W A V E L E N G T H (nm)

Figure 6. 2-(2’-Hydroxy-5’-nitrophenyl)benzothiazole: (- - -) absorption spectra in Me2S0, (1) 6.2 X IOw4 M, (2) 6.2 X IOw5 M, (3) 6.2 X IOd M; transient absorption spectra in Me2S0, 1.16 X IO-” M, (0)3.5 ps and (0)62 ys after 355-nm laser flash excitation (1 I-ns pulse); (-) emission M (arbitrary unit) with 355-nm excispectrum in Me,SO, 1.16 X

tation. but we can already recognize the fact that the aggregate is not an enol dimer since it was not observed in hydrocarbon solvents. Figures 3-6 show the absorption spectra of (3’-nitrosalicylidene)aniline (3’-nitroanil) and 2-(2’-hydroxy-5’-nitrophenyl)benzothiazole in ethanol and dimethyl sulfoxide as a function of concentration. We can see that the latter compounds behave in a fashion very similar to that of the 5’-nitroanil. At least three different species can be characterized depending upon the solvent: (a) the intramolecular hydrogen-bonded enol, (b) the aggregate at high concentration, and (c) the anion at low concentration. The anion of 3’-nitroanil shows only one absorption maximum at 445 nm compared to the two absorption bands observed for the 5’-nitro derivatives. Moreover, for the 3’-nitro derivatives of ani1 or (hydroxyphenyl)benzothiazole, the shape of the absorption band showed a long tail on the long-wavelength side of the absorption spectrum in hydrocarbon solvents. We believe that in the case of the 3’-nitro derivatives more than the three species described above are involved and that the following complex equilibrium exists between species with an intramolecular hydrogen bond to the nitrogen atom or the oxygen atom of the nitro group:

370 333 492 366 361 300 448 392

380 (0.52) 320 (0.10)

3‘-nitro

318 476 355 5‘-nitro 355 (0.086) 335 334 (0.768) 318 289 (0.546) 288 5’-nitro anion 435 379

3’-nitro anion

(0.80) 351“

332d 303d 406,c 442’ 368,‘381f

19 475d 20 950d 17 250,’ 22 2 5 6 16200,‘ 1 3 9 5 6

(0.24) (0.48) (0.17) (0.65) (0.15) (0.51) (0.69) (0.3)

3059 623W 442,h 4601 9580,* 10210’ 33W

19475’

286‘ 18406‘ 406,k 445‘ 20200,k 29 190‘ 362,k 379’ 19800,k 18710’

TABLE 11: Calculated T,T* Electronic Transition Energies and Oscillator Strength for the Cis-Keto Form (PPP) or Cis Zwitterion ( I N W / S ) for the Nitro Derivatives of Salicylideneaniline and 2- (2’-Hydroxyphenyl)benzothiazole

transition energy, nm (osc str) PPP

compd

INDOjS

Salicylideneaniline 486 (0.67) 458 (0.52)

3‘-nitro

311 (0.41) 458 (0.59) 323 (0.20)

5‘-nitro

334 (0.20) 443 (0.48) 341 (0.16)

Xz,’nm 500b 500b

2-(2’-Hydroxyphenyl)benzothiazole

3’-nitro 3’-nitro anion

-

7960a 330“ 63 15“ 445,b 485‘ 13385? 10240‘

(0.14) (0.56) (0.20) (0.9) (0.49) (0.83) (0.25)

428 (0.57) 330 (0.20)

5’-nitro anion

-

mol-’ L cm-I

QMethylcyclohexane(5.78 X M). bEthanol (5.7 X IOd M). M). dAcetonitrile (5.54 X M). ‘Ethanol‘Me2S0 (8.79 X /NaOH (5.24 X M). fMe,SO (7.6 X 10” M). 9MethylcycloM). ‘Me,SO hexane (6.9 X IO-“) M). “thanol/NaOH (9.54 X (6.17 X ’Dichloromethane (5.54 X M). kEthanol (6.83 X M). ‘Me,SO (6.83 X M).

5‘-nitro

In order to aid in the assignments of the different species in equilibrium in solution, we performed semiempirical calculations for 3’-nitroanil, 5’-nitroanil, and 2-(2’-hydroxy-5’-nitrophenyl)benzothiazole. We calculated the transition energies of the enol, the anion, and the zwitterionic or keto form of the nitro derivatives using the INDO/S semiempirical scheme proposed by Ridley and Zerr~er.**.*~ The results are summarized in Tables I and 11. For the enol form of all the compounds studied, we can see that the predicted transition energies from both calculations are in good agreement with the experimental values. In the case of the nitroanil derivatives, the INDO/S calculations generally give more accurate results for the location of the a a* transitions as well as their relative intensities. Moreover, the INDO/S calculations predict that the transition of lowest energy originates from excitation to the ‘(n,a)* state (out of plane polarized). The existence of the n a* transitions is difficult to verify because of the spectral complications resulting from the multitude of species existing in solution. We also calculated the transition energies of the anion of the nitro derivatives (Table I). It is interesting to note that the

6.

2 4 2’-Hydroxyphenyl)benzothiazole 347 (0.55) 3519 38409

-1 7

280

hobs. nm

Salicylideneaniline

1

0

INDO/S

PPP

466 (0.65)

431 343 435 379 448 474

(0.46) (0.29) (0.69) (0.3) (0.51) (0.48)

440‘ 319‘ 490‘

Absorption maximum of the transient absorption spectrum. * I n methylcyclohexane. In dimethyl sulfoxide. calculated values are generally closer to those of the anion in dimethyl sulfoxide than to those found in ethanol (Table I). In order to obtain insight into the laser flash results of the nature of photophysics and photochemistry of the various nitro-substituted molecules, calculations on the ground- and excited-state charge densities of nitrogen and oxygen atom sites were done (PPP) (Table 111). For 3’-nitro- and 5’-nitroanil, the charge density on the nitrogen atom is found to increase significantly, from 0.08 to 0.16 e in the excited state, while the charge density on the oxygen atom is found to increase slightly by -0.02 e. This result is similar to the one found for the parent anil.I4 Therefore, it seems that the significant relative increase of the basicity of the nitrogen atom can account for the proton transfer in the excited state. For the 5’-nitro benzothiazole derivative, the PPP calculation and the INDO/S calculation differ significantly. The calculated lowest transition is found to be a R R* state at 355 nm for the PPP calculation with a nearby R R* transition at 334 and 335 nm for the INDO/S calculation. The value of the oscillator

--

The Journal of Physical Chemistry, Vol. 91, No. 13, 1987 3521

Nitro Derivatives of S A and HBT TABLE III: Calculated T Charge Densities in the Ground State (PPP, INDO/S) and in the First Excited State (PPP) of the Nitrogen and Oxygen Atoms of the Nitro Derivatives of Salicylideneanilineand 2-(2’-Hydroxyphenyl)benzothiazole compd/method QhN QsIN Qh0 Qs,’

3’-nitro/PPP 3’-nitro/INDO/S 5’-nitro/PPP 5’-nitro/INDO/S

Salicylideneaniline 1.232 1.327 1.290 1.237 1.392 1.295

1.872

1.903

1.894

1.886 1.901

2-(2‘-Hydroxyphenyl)benzothiazole 5’-nitro/PPP 1.664 1.668 1.898 5’-nitro/INDO/S 1.289 1.907 3’-nitro/PPP 1.660 1.689 1.885 3’-nitro/INDO/S 1.284 1.900

1.904

1.912 1.908

strength is found to be 0.086 for the 355-nm transition by the PPP calculation and 0.65 for the 3354x11 transition by the INDO/S calculation. However, the 334-nm PPP predicted transition has an oscillator strength of -0.77 while by INDO/S there is another transition at 318 nm with an oscillator strength of 0.18. Contrary to what was found for the nitroanil derivatives, the charge density, as calculated by the PPP method, on the nitrogen atom increases only very slightly (-0.004 e) in the first excited state while the charge density on the oxygen atom increases slightly by -0.01 e. Therefore, for this compound, the intramolecular proton transfer in the first excited state would not be favored. The first excited state as calculated by the PPP method shows a strong charge-transfer character, from the benzothiazole moiety to the nitrophenyl moiety. We will consider this result again relative to the laser flash experiments. For the 3’-nitro benzothiazole derivative the charge density in the first excited state, as calculated by the PPP method, on the nitrogen atom increases by -0.027 e in the first excited state while the charge density on the oxygen atom increases by -0.023 e. Again it seems that for the nitro benzothiazole derivatives the change in charge density in the first excited state would not favor an intramolecular proton transfer. We shall now consider the results obtained after laser flash photolysis of the nitro derivatives. II. Laser Flash Photolysis. The results of the laser flash photolysis are summarized in Table IV and in Figure 1-6. For the 3’-nitro- and 5’-nitroanil, all the solutions were nitrogen saturated prior to the laser excitation at 355 nm (11-ns pulse). When saturating the solutions with oxygen, we did not observe any influence on the formation, rise time, or decay time of the transients, and no triplet transients were observed in any solvents. The rise times of all the transients in any solvent were found to be detection limited, that is, C200 ns. The concentrations used M, the concentration at which aggregation occurred were in polar or hydrogen-bonding solvents. A . Laser Flash Photolysis of (5’-Nitrosalicylidene)aniline. We will first examine the results in methylcyclohexane where no aggregation was seen to occur (Table IV). After irradiation of a nitrogen-saturated solution of 5’-nitroanil in methylcyclohexane, we observed a transient with a maximum of absorption at 500 nm. This transient decayed to the base line and showed two first-order kinetic components. In the region of absorption of the enol (Le., < 380 nm), we observed a decay to a negative AOD showing two first-order kinetic components with identical lifetimes as observed at 500 nm. No further change of absorbance was observed up to 400 ps. The above result is very similar to the one We, therefore, observed for the parent sali~ylideneaniline.~~~~~~~~ believe that 5’-nitroanil undergoes intramolecular proton transfer in the singlet excited state to produce the zwitterion. The decay data indicate that the zwitterion exists in two different conformations. The main difference with the parent ani1 is that the zwitterion does not decay back to the original enol by back proton

-

~~

(25) Nakagaki, R.; Kobayashi, T.; Nakamura, J.; Nagakura, S. Bull. Chem. SOC.Jpn. 1977, 50, 1909. (26) Itoh, M ; Fujiwara, Y. J . Am. Chem. SOC.1985, 107, 1561.

TABLE I V Spectral Absorption Data and Photophysical Data of the Transients of the Nitro Derivatives of Salicylideneaniline and 2 4 2’-Hydroxyphenyl)benzothiazole Obtained after Laser Excitation at 355 nm (11-ns Pulse)

Salicylideneaniline 3‘-nitro methylcyclohexane 500 4 dichloromethane 520 16 ethanol 510 14 MezSO 520 33 5’-nitro methylcyclohexane 500 12 510 28 dichloromethane ethanol

510

MezSO

490

18 31

24

230

17 50

183

39 307 122

2 4 2’-Hydroxyphenyl)benzothiazole 490 3’-nitro MezSO 5‘-nitro dichloromethane 430 10 50 ethano 440 Me2S0 440

50

75 3.4 20

60

“Decay times of zwitterion. bDecay times of anion, or unidentified photoproduct (of the 3’-N02 anil). transfer since a long-lived species is observed after the back proton transfer. Since we observe a depletion in the region of absorption of the enol, we believe that the zwitterion decays by back proton transfer to a conformer of the original intramolecular hydrogen-bonded enol. In previous studies of the 5’4troanil in hydrocarbons, no net photochemistry at room temperature was ever o b s e r ~ e d therefore, ;~ we believe that this conformer ultimately thermally decays to the origirlal intramolecular hydrogen-bonded enol. We observed results very similar to the ones described above when examining a nitrogen-saturated dichloromethane solution of 5’-nitroanil. Recall that in dichloromethane very little aggregation is observed and no anion is observed. The transient, assigned as the zwitterionic form, had a maximum of absorption at 510 nm and decayed with two first-order kinetic components. The lifetimes were, however, 1 order of magnitude longer in dichloromethane than in methylcyclohexane (Table IV). When irradiating a nitrogen-saturated ethanol solution of 5’nitroanil, we also observed a zwitterion, with an absorption maximum at 480 nm, which decayed with one (vs. two for those above) first-order kinetic component. The lifetime of the zwitterion was shorter in ethanol than in dichloromethane or methylcyclohexane (Table IV). The zwitterion was also found not to return to the original enol. Moreover, in the region of absorption of the aggregate (-430 nm), we observed a net depletion still present 400 ps after the decay of the transient. In dimethyl sulfoxide, the results were slightly different from those in the previous solvents. The transient absorption spectrum showed three maxima at 380,440, and 490 nm. When monitoring the kinetic behavior of the transients, we found that the decay at 490 nm consisted of two first-order kinetic components with 31 and 122 ps, respectively, for lifetimes. The decays at 380 and 440 nm each consisted of one first-order kinetic component with a lifetime of 50 ps, indicating that these absorptions originated from a single transient species. We assign the two transients as follows. The transient with a maximum of absorption at 490 nm is the zwitterion, in parallel with the results in the other solvents. The transient with absorption maxima at 380 and 440 nm has very similar absorption maxima to the anion of 5’-nitroanil in Me,SO. We thus believe that in Me2S0 both the zwitterion and the anion are produced after laser excitation. As in the case of ethanol, after total decay of the transients, we observed a net depletion in the region of absorption of the aggregate (-430 nm). B. Laser Flash Photolysis of (3’-Nitrosalicylidene)aniline. After laser flash excitation at 355 nm (1 1-ns pulse) of a nitrogen-saturated methylcyclohexane solution of 3’-nitroanil, we observed a transient absorption spectra with two absorption maxima at 410 and 500 nm. At 500 nm, the transient decay contained two first-order kinetic components with lifetimes of 4

3522

The Journal of Physical Chemistry, VoL 91, No. 13, 1987

and 24 pus. Only one first-order decay was observed at 410 nm with a lifetime of 17 ps. The ground-state recovery at 360 nm contained one first-order component to a negative base line (depletion) and recovered with a lifetime of 17 pus. The results in dichloromethane were very similar to the ones described above with two distinct transients with absorption maximum at 420 and 520 nm, respectively. The lifetimes were, however, longer in dichloromethane than in methylcyclohexane by 1 order of magnitude. When irradiating a nitrogen-saturated ethanol solution of 3’nitroanil, we observed a transient with an absorption maximum at 510 nm which decayed with one first-order kinetic component with a lifetime of 14 ps. No other transient was observed. In the region of absorption of the aggregate (-440 nm) the transient decayed to a negative base line (depletion). When irradiating a nitrogen-saturated solution of 3’-nitroanil in Me2S0,we observed a transient with an absorption maximum at 520 nm which decayed with two first-order kinetic components to a negative base line (depletion) in the region of absorption of the aggregate. On the basis of the foregoing results, we believe that excitation of the 3‘-nitroanil results in the formation of the zwitterion with a maximum of absorption near 500 nm, the exact wavelength depending upon the solvent. The zwitterion exists in one conformation in ethanol while two conformers are observed in MefiO, methylcyclohexane, and dichloromethane. Moreover, in the latter two solvents, another transient is observed with an absorption maximum around 420 nm. Although we do not propose any specific structure for this transient, we note that the nitro group in the ortho position to the hydroxy group could also be an acceptor site for a proton transfer. Before examining the results obtained for the phenylbenzothiazole derivatives, we will discuss the geometry of the zwitterion and the structure of the aggregate using laser flash photolysis and semiempirical calculation results. C. Nature of the Zwitterion and the Aggregate for the 3‘-Nitro and 5‘-Nitro Derivatives of Salicylideneaniline. Using PPP and INDO/S semiempirical methods, we calculated the transition energies of the cis-keto and cis-zwitterionic form. For the cis-keto form we used the PPP calculation where all the bond lengths were as described in the Experimental Section except for the C=O bond with a bond length of 1.22 A. For the “cis”-zwitterion form we used the INDO/S calculation where all bond lengths were as described above except for the C-0- and N+=H bonds with lengths of 1.26 and 1.034 A, respectively. We found that for the INDO/S calculation the calculated K--R* transition energies were generally lower than the experimental values, from 1800 to 2500 cm-’. The PPP calculations gave values usually closer to the experimental values. If we used a geometry appropriate to a quinoid form, where the single and double carbon-carbon bonds alternated, the calculated values were lower by 5000 cm-‘ compared to the experimental values. The better agreement of the INDO/S calculation using a zwitterionic structure makes us believe that the colored forms are of more zwitterionic character than quinoid character. However, it is most likely that the colored form will have an electronic structure intermediate between a keto and a zwitterionic form. We therefore will refer to the colored form as zwitterion rather than keto forms. We believe that the discrepancy still found for the calculated values compared to the experimental values, especially for the INDO/S calculation, is due to the fact the zwitterion is not planar and also, of course, calculations are with reference to a gas-phase environment. A nonplanar character of the colored form was also proposed earlier by Rosenfeld et al. baaed on their PCILO semiempirical calculation on the “keto form” of ~alicylideneaniline.~’We will see later that the emission spectroscopic studies tend to substantiate the proposal that the zwitterion is not planar. Moreover, this proposal would explain why, after the thermal back proton transfer within the distorted zwitterion, a conformer of the enol form is obtained (which subsequently isomerizes back to the original ( 2 7 ) Rosenfeld, T.; Ottolenghi, M.: Meyer, A . Y . Mol. Photochem. 1973, 5 , 39.

Becker et al. intramolecular hydrogen-bonded enol). We also note that in any solvent the rate of thermal reverse proton transfer is observed to be much slower than the ESIPT (as observed for other classes of molecules showing ESIPT). The reverse proton transfer rate has an activation energy which is solvent dependent. However, it is important to keep in mind that conformational changes occur in the ground state and these certainly have an activation energy, probably solvent dependent. Thus, the activation energy for the overall process of obtaining the appropriate conformation (for proton transfer) and the proton transfer itself likely has the activation energy of conformational change as the principal contributor. Another interesting feature of the results of the laser flash photolysis experiments of the nitro derivatives of salicylideneaniline is the depletion in the region of absorption of the aggregate after total decay of the zwitterion. It is apparent that after excitation some of the aggregate dissociated. We propose the following mechanism to explain the experimental results. After irradiation, one or all of the components of the aggregates undergo photochemistry leading to the dissociation of the aggregate. We already known that the aggregate is not a dimer of the enol form. Let us suppose that since the nitro group enhances the acidity of the hydroxy group, some zwitterion is in thermal equilibrium with the enol. However, the laser flash results show that the zwitterion is not stable for more than a few microseconds; therefore, the zwitterion becomes relatively stabilized by aggregation with an enol molecule. Note that this mixed dimer ultimately is not stable since after several hours the aggregate dissociates, leading to the stable anion as found earlierS2This mixed dimer also dissociates upon irradiation since the enol molecule in the mixed dimer may still undergo photochemistry. The enol can either undergo anti-syn isomerization or excited-state intramolecular proton transfer. It is known that salicylideneaniline can isomerize around its C=N bond in hydrogen-bonding solvent or when it is protonated, that is, when the intramolecular hydrogen bond is p e r t ~ r b e d . ’ ~ ~ ~ ~ - ~ * We do not know the exact type of interaction responsible for the aggregation process, but in any event is is likely that the intramolecular hydrogen bond will be perturbed. We therefore believe that, after the irradiation of the mixed dimer, the enol molecule within the dimer undergoes antisyn isomerization leading to the dissociation of the mixed dimer. This mechanism would explain the observed result of the decrease in concentration of the dimer after laser excitation. Furthermore, the proposed structure for the aggregate explains why more aggregates are observed in hydrogen-bonding solvents or in nonpolar non-hydrogen-bonding solvents at low temperature where the keto-enol equilibrium is displaced toward the keto (zwitterionic) form to obtain the enol-keto/zwitterion aggregate. We shall now discuss the results obtained after laser flash photolysis of the nitro derivatives of 2-(2’-hydroxyphenyl)benzothiazole. D. Laser Flash Photolysis of 5’-Nitro and 3’-Nitro Derivatives of 2-(2’-Hydroxyphenyl)benzothiazole.After irradiation of a nitrogen-saturated solution of the 5’-nitro derivative of 242’hydroxypheny1)benzothiazole in dichloromethane, we observed immediately after the pulse (