2282
J. Phys. Chem. 1994, 98, 2282-2289
Photochemistry and Photophysics of Aromatic Sulfoxides. 1. Characterization of the Triplets at Cryogenic Temperatures William S. Jenks,. Woojae Lee, and David Shutters Department of Chemistry, Iowa State University, Ames, Iowa 5001 1-31 1 1 Received: December 27, 1993"
Aromatic sulfoxides are photochemically active molecules. Much of that reactivity has been attributed to the previously uncharacterized triplet states of these molecules. At 77 K in ether/isopentane/ethanol (EPA) glass, aromatic sulfoxides are shown to have weak phosphorescence. The triplet energies are a few kilocalories per mole higher than the corresponding ketones; the diaryl sulfoxides are about 3 kcal/mol lower than the corresponding ketones, and the diaryl sulfoxides are about 3 kcal/mol lower than the corresponding aryl methyl sulfoxides. For instance the triplet energy of diphenylsulfoxide is 78 kcal/mol, whereas the triplet energy of methyl phenyl sulfoxide is estimated to be 81 kcal/mol. The lifetimes of emission are generally under 100 ms. From the diffuse vibrational structure of the spectra, the lifetimes, and the effect of solvent on the triplet energy, it is concluded that the triplets are delocalized aromatic states that involve substantial charge transfer off the oxygen atom.
Introduction The sulfoxide group is extensively used as a synthetic intermediate and chiral auxiliary, and further elaboration of this chemistry is the subject of substantial continuing Sulfoxides are also known to undergo photochemical reactions of several types,*-ll but this chemistry has not been developed or generalizedin parallel fashion. As the foundation of a systematic study of the photochemistry and photophysics of aromatic sulfoxides, we have characterized the lowest triplets of a set of aryl alkyl and diaryl sulfoxides by studying their emission characteristicsin organicglassesat 77 K. In addition,wecompare the results for several sulfoxides to the corresponding sulfides and sulfones. We also discuss time resolved electron paramagnetic resonance (TREPR) spectra obtained at 15 K for several of these compounds. The class of photochemical reactions of sulfoxides that has been best documented involves cleavage of the S-C bond (acleavage) and yields products derived from the resulting radical pair or biradical. Strongly supporting the a-cleavage mechanism is the observation by steady-state EPR of the sulfinyl radicals themselves and their spin-trapping products.12-14 Products rationalized by a mechanism involving a-cleavage include racemates of starting material, sulfenic esters (R-S-0-R') and their photoproducts, products representing loss of SO, and other radical recombination and disproportionation products. Several reaction products have been rationalized on the basis of hydrogen abstraction by the oxygen atom, but there is no direct evidence supporting this mechanism. A third class of reaction, whose mechanism is not understood, is the photodeoxygenation of sulfoxides to sulfides. In most cases where an assignment has been made, the photoreactivities of aromatic sulfoxides have been attributed to their triplet ~tates.9J5-2~These assignments have generally been made using the ordinary criteria of triplet sensitization and quenching experiments. While characterization of the triplet states of ketones has played a critical role in the development of their photochemistry, the same cannot be said for sulfoxides. There are only three reports of phosphorescence from aromatic sulfo~ides,24~26~2~ no quantum yields of lifetimes, and no qualitative characterizations. Moreover, the little information that exists casts serious doubts that sensitization by relatively low energy molecules such as benzophenone and naphthalene is by the Abstract published in Advance ACS Absrracts, February 1, 1994.
standard energy-transfer mechanism.1s-19~24~2s~28 In this paper, we show that although the excitation of aromatic sulfoxides is not localized on the SO chromophore, the sulfinyl substitution is critically important to the nature of the lowest aromatic triplet. Phosphorescenceat 77 K is typically an inefficient process for the sulfoxides we examined,and the triplet lifetimes we could measure are generally in the range of 35-100 ms. The triplet energies of sulfoxides are between those of the corresponding ketones and the unsubstituted aromatic. Wealso show that thecharge density on the sulfoxidic oxygen is substantially reduced in the triplet state. Experimental Section General Instrumentation, All luminescence spectra and lifetimes were recorded with an Edinburgh Instruments FL900 spectrometer. A suprasil liquid nitrogen immersion dewar was used to hold samples at 77 K. The samples were contained in 5-mm suprasil cylindrical tubes within the dewar. Compounds were checked for purity using a Hewlett-Packard 5890 I1 gas chromatograph (GC) equipped with HP-1 or DB-17 capillary columns and a flame ionization detector. UV/visible absorption measurements were performed with a Shimadzu PC-3101 spectrophotometer. NMR spectra were obtained using either a Nicolet or Varian 300 MHz instrument. Compounds. Commercially available materials were purified by flash chromatography on silica, distillation, sublimation, or recrystallization until no impurities were detectable by GC (>99.9%). All of the rest of the compounds are known, and satisfactory spectroscopic data were obtained for each. Phenoxathiin (6-S) was prepared by the method of Suter.29 pFluoropheny1 methyl sulfoxide (2-SO),30,31 phenoxathiin Soxide, (6-SO),32 and dibenzotbiophene Soxide (7-S0)32 were obtained by oxidation of the corresponding sulfides29 with Bu4N+I04- and catalytic (5,10,15,2O-tetraphenylporphine)iron(111) chl0ride.3~ pMethoxypheny1 methyl sulfoxide (3-SOy3 was prepared by oxidation of 3-S by NaI04 in ~ a t e r . 3 ~ Dixylyl sulfoxide (9-S0),35dimesityl sulfoxide (lO-SO),35 dipfluorophenyl ~ulfoxide,3~ (11-SO), di-pbromophenyl sulfoxide ( 13-S0),37 and di-pmethoxyphenyl sulfoxide (14-S0)3* were prepared by condensation of the corresponding arene and thionyl chloride in the presence of AIC13. pMethoxypheny1 methyl sulfone (3-S02)39and phenoxathiin S,S-dioxide ( 6 - s O ~ ) ~were o prepared by oxidation of the corre-
0022-365419412098-2282$04.50/0 0 1994 American Chemical Society
Photoactivity of Aromatic Sulfoxides
The Journal of Physical Chemistry, Vol. 98, No. 9, 1994 2283
0
sso 0
*in /
79
740
11-80
12-80
0
0
0
0
/
Figure 1. Compounds used in the present study.
sponding sulfides with 30% aqueous hydrogen peroxide in refluxing acetic acid.", General Luminescence Methods. The solvent mixture for all luminescence experiments unless otherwise specified was a 5 5 2 mixture of ether, isopentane, and ethanol (EPA). Isopentane and methylcyclohexane were spectro grade and were used as received. Absolute ethanol was refluxed over CaHz for a t least 24 h to remove carbonyl-containingcompounds and then distilled freshly under Ar for each use. Ethyl ether was similarly handled. All samples were flushed with Ar for 10 min to remove oxygen. Luminescent lifetimes and spectra for triplet energies were obtained from samples that were ca. 1 mM in chromophore. Control experiments showed that the lifetime measurements did not change on dilution of the samples. The spectrometer uses a xenon lamp which emits a pulse of a few new microseconds in duration, and the decay is collected using multichannel scaling. Due to an instrumentally imposed limitation, the lamp cannot be pulsed at a rate lower than 1 Hz. This leads to some uncertainty in lifetimes longer than about 0.5 sand can lead to some artifactual long components in decays. Data from very few samples with lifetimes under 200 ms could be satisfactorily fit to singleexponential decays and were fit to distributions of exponentials using software provided by Edinburgh Instruments. Phosphorescence Quantum Yields. Quantum yield measurements were made using optically dilute samples. Solutions were prepared and the optical densities were measured using a standard 1-cm UV cell a t room temperature. The concentration of the chromophore was adjusted such that the optical density (in the 1-cm cell) was 10.200 a t the exciting wavelength, usually 265 nm. The samples were transferred to a suprasil5-mm cylindrical tube, and each was identically deoxygenated. A liquid nitrogen immersion dewar mounted in the spectrometer reproducibly positioned the samples. After the spectra were obtained and the spectral intensities were integrated, the data were corrected for variations in the measured optical density and compared to benzophenone (@p = 0.74 A 0.02).41 Each reported value is the
average of at least two measurements each on two independent samples, and the estimated error is *20%. Time-Resolved Electron Paramagnetic Resonance (TREPR). EPR data were obtained on a spectrometer a t Columbia University which has been described elsewhere.42 Briefly, a Bruker ESP380 EPR equipped with a dielectric ring cavity (ESP380-1052 DLQH ) was used as the spectrometer. The samples were about 50 mM in 2-methyltetrahydrofuran and were contained in quartz tubes whose temperature was controlled with an Oxford liquid helium cryostat. The samples were excited with 308.5-nm pulses from a Lambda-Physik LPX-100 laser running at 30 Hz. The signal was directly detected without field modulation by a boxcar integrator (PAR Model 4420). The microwave field was continuously on, but the signal was sampled only from 1 to 2 MS after each laser pulse. By sweeping the magnetic field an absolute (not derivative) spectrum was obtained. D and E values could be obtained by standard methods from the data or their derivatives.4345
Results Compounds. We chose to study compounds of two general classes: symmetrical diaryl sulfoxides and aryl methyl sulfoxides. We chose symmetrical diaryl sulfoxides in order to avoid localization of excitation on thelow-energy "half" of themolecule. In addition, we examined the luminescence of several of the corresponding sulfides and sulfones. The compounds used in the study are illustrated in Figure 1. Emission Spectra and SpectroscopicExcitation Energies. None of the sulfoxides examined exhibited any luminescence a t room temperature in a variety of solvents. There is one report of fluorescence from l - S O , 4 6 which we were not able to reproduce. It is likely that the reported fluorescence was from an impurity, most likely 14, vide infm. At 77 K, however, phosphorescence was observed from most of the compounds and fluorescence from a few. The luminescence
2284
Jenks et al.
The Journal of Physical Chemistry, Vol. 98, No. 9, I994
TABLE 2 Spectroscopic Triplet Energies for All of the Compounds Which Showed Measurable Phosphorescence at 77 K in EPA Glass triplet energy (kcal/mol) compd family sulfide sulfoxide sulfone 73' 75 73 75 73 66L 70
1 2 3
4 5
, 350
I
400 450 Wavelength (nm)
I
500
6 7' 8 9
I 550
Figure 2. Phosphorescence spectrum of di-p-tolylsulfoxide (8-SO) at 77 K in EPA.
TABLE 1: Spectroscopic Singlet Energies for Those Sulfoxides Which Showed Fluorescence at 77 K in EPA singlet excitation compound energy (kcal/mol) 3-SO
6-SO 7-SO
99 91 85
of all of the compounds were examined in 5:5:2 ethyl ether: isopentane:ethanol (EPA) glass. The sulfoxides were substantially more soluble in this mixture than in any of the standard hydrocarbons used for organic glasses. We were unable to completely strip 2-methyltetrahydrofuran of aromatic impurities that made its use impractical for many of the sulfoxides, which do not have substantial absorption coefficients beyond about 275 nm. Fluorescence. Among the sulfoxides, fluorescence at 77 K was observed only from threecompounds: 3-S0,6-S0, and 7-SO. The singlet energies are listed in Table 1. In each case, the fluorescence was a relatively minor part of the luminescence, perhaps no more than about 10% of the total. The fluorescence decay of 730,which was nonexponential but in the nanosecond regime, showed that the fluorescence was prompt. The millisecond lifetime (uide infra) of the phosphorescence and large energy gap between the fluorescence and phosphorescence made delayed fluorescence an unreasonable explanation. Triplet Energies. The phosphorescence spectra of most of the sulfoxides were very similar in appearance. The spectrum obtained from ditolyl sulfoxide, which is typical, is presented in Figure 2. There is a single shoulder on the high-energy side and then a relatively unstructured band. The triplet energies that appear in Table 2 are taken from the A, of the blue edge band. The principal variation among the phosphorescences of the sulfoxides is whether any additional (broad) structure is visible and the width of the band. Only the two tricyclic sulfoxides, 6-SO and 740,showed any complex vibrational structure. Weak S to TI excitation spectra were observed for a few of the sulfoxides; no large Stokes shifts were found. Again with the exception of the dibenzothiophene (7) system, all of the sulfides have triplet energies lower than the corresponding sulfoxides. The triplet energies of the sulfones are typically comparable to, but a little higher than, those of the sulfoxides. Figure 3 illustrates the spectra of diphenyl sulfide, sulfoxide, and sulfone (543, -SO, and -SO*). The aryl methyl sulfoxides are 78 and 80 kcal/mol, which is 3 kcal/mol higher than the corresponding diaryl sulfoxides. No experimental triplet energies are given for 1-SO or 4-SO. Only extremely weak spectra could be obtained from these compounds, and they were indistinguishable from thoseobtained from thecorresponding sulfide. We estimate the triplet energy of phenyl methyl sulfoxide, 1-SO, to be 81 kcal/mol, based on the three following criteria: (1) The known triplet energy of p-bromophenyl methyl sulfoxide, 15-S0, is 79 k c a l / m ~ l (2) ; ~ ~the triplet energy of diphenyl sulfoxide, 5-SO, is 2 kcal/mol higher than the corresponding bromo substituted
(81)bJ
82
78 80
81 75 d 80 79 64
b 78 79 61 77 76 76 75 76 76 77
10
11 12 13 14
0 74 kcal/mol in EPA at 77 K.47 Emission from these sulfoxides was very weak or non-existent. No spectrum, however weak, that differed from that of the corresponding sulfidecould be obtained. Triplet energy estimated to be 81 kcal/mol. Emission too weak to obtain a triplet energy. In agreement with literature reports in ethanol g l a ~ s . * ~ ~ ~
350
4bO 450 500 Wavelength (nm)
550
Figure 3. Phosphorescence spectra of 5-SO2,5-SO, and 5-S (diphenyl sulfone, sulfoxide, and sulfide) at 17 K in EPA. The spectra have been normalized for presentation.
TABLE 3 Triplet Energies in EPA and Methylcyclohexane (MCH), with Energy Shift' sulfides sulfoxides sulfoncs family EPA MCH AET EPA MCH AET EPA MCH AET 3 5 6 7 8
73 13 66 70
12
73 66' 10
80 78 19 61 77
15
-5
6
69 58
75 13
-10 -3 -5 -5 -1 -1 -3
14
-3
12
9
16
71
11 12 13 14
75 76 76 71
68
benzophenone 69.4 68.4 a
-1 0 0 0
75 80 79 64
74 79 74 64
-1 -1 -5 0
-1
All energies in kilocaloriesper mole. Too weak to obtain a spectrum.
65 kcal/mol in i~opentane.~~
compound, 13-SO; and (3) the 3 kcal/mol corrleation between the diaryl and aryl methyl sulfoxides places the triplet energy of 1-SOat 8 1kcal/mol. We are unable toestimate the triplet energy of 4-SO with any specificity. Solvent Effects on Triplet Energies. The sulfoxides with significant phosphorescence in EPA were also examined in methylcyclohexane (MCH),along with their corresponding sulfides and sulfones. The triplet energies of these compounds are shown in Table 3. Those of the sulfides were essentially insensitive of the solvent change. Sulfones showed only a small sensitivity, approximately -1 kcal/mol, whereas most of the sulfoxides showed solvent shifts, from -3 to -10 kcal/mol. The direction of the shift, from larger triplet energies in the more
Photoactivity of Aromatic Sulfoxides
TABLE 4 "pd family 1 2
3 4
5 6 7 8 9 10 11 12 13 14
The Journal of Physical Chemistry, Vol. 98, No. 9, 1994 2285
Quantum Yields of Phosphorescence in EPA phosphorescence quantum yield# sulfide sulfoxide sulfone 0.32
