2286
The Journal of Physical Chemistty, Vol. 82,No. 21, 1978
in agreement with published pulse radiolysis results. We find that the (Na+, esol;) species is formed in 110 ps.
References and Notes (1) G. Lepoutre and M. J. Sienko, Ed., "Metal-Ammonia Solutions", Colloque Weyl I,W. A. Benjamin, New York, N.Y., 1964. (2) J. J. Lagowski and M. J. Sienko, Ed., "Metal-Ammonia Solutions", Colloque Weyl I1 (IUPAC), Butterworths, London, 1970. (3) J. Jortner and R. N. Kestner, Ed., "Electrons in Fluids", Colloque Weyl 111, Springer-Verlag, Berlin, 1973. (4) S.Matalon, S.Golden, and M. Ottolenghi, J . Phys. Chem., 73, 3098 (1969). (5) M. T. Lok, F. J. Tehan, and J. L. Dye, J . Phys. Chem., 76, 2975 (1972). (6) S.H. Glarum and J. H. Marshall, J . Chem. Phys., 52, 5555 (1970). (7) 9. Bockrath and L. M. Dorfman, J . Phys. Chem., 77, 1002 (1973). (6) G. A. Salmon, W. A. Seddon, and J. W. Fletcher, Can. J . Chem., 52, 3259 (1974).
K. K. Smith and K. J. Kaufmann
(9) 9. Bockrath, J. F. Gavalas, and L. M. Dorfman, J. Phys. Chem., 79, 3064 (1975). (10) L. J. Gilling, J. G. Kloosterboer, R. P. H. Rettschnick, and J. D. W. van Voorst, Chem. Phys. Lett., 8,457 (1971). (11) J. W. Fletcher, W. A. Seddon, and F. C. Sopehyshyn, Chem. Phys. Lett., 18, 592 (1973). (12) N. R. Kestner and J. Jortner, J. Phys. Chem., 77, 1040 (1973). (13) N. R. Kestner and J. Logan, J . Phys. Chem., 79, 2815 (1975). (14) D. Huppett, W. S. Struve, P. M. Rentzepis, and J. Jortner, J . Chem. Phys., 63, 1205 (1975). (15) D. Huppert and P. M. Rentzepis, J . Chem. Phys., 64, 195 (1976). (16) I.Hurley, T. R. Tuttle, and J. Golden, ref 2, p 503. (17) S.L. Hager and J. E. Willarg, J. Chem. Phys., 61, 3244 (1974). (18)J. L. Dye, M. G. Debacker, J. A. Eyre, and L. M. Dorfman, J . Phys. Chem., 76, 839 (1972). (19) L. Gilles, J. E. Aldrich, and J. W. Hunt, Nature(London),Phys. Sci., 243, 70 (1973). (20) G. A. Kenney-Wallace, Can. J . Chem., 55, 2009 (1977). (21) T. R. Tuttle, Jr., J. Phys. Chem., 79, 3071 (1975).
Picosecond Studies of Intramolecular Proton Transfer Kevin K. Smith and Kenneth J. Kaufmann" Depatiment of Chemistry, University of Illinois, Urbana, Illinois 6 1801 (Received December 16, 1977)
The rate of intramolecular proton transfer in methyl salicylate was found to be greater than 10" s-l even at 4 K. The fluorescence lifetime of the zwitterion formed by the transfer of the proton was found to be strongly dependent on temperature. An energy of activation of 3.7 kcal/mol was found for the nonradiative decay rate of the excited zwitterion. The neutral excited state species had a fluorescence lifetime at room temperature which was a factor of 3 larger than the excited zwitterion. Therefore the two excited state forms of methyl salicylate are not in equilibrium as was previously believed.
Introduction Large Stokes shifts have been observed in the emission spectrum of a large number of aromatic ketones and alcohols.1,2 Forster postulated that the large red shifts were due to a change in the acid-base properties of these aromatic molecule^.^ The change in pK, arising from absorption of a photon can be estimated from the absorption and fluorescence maxima by use of the Forster ~ y c l e . ~It, ~can also be evaluated by a titration of the intensity of the anomalous fluorescence vs. hydrogen ion c~ncentration.~ Such studies of the excited state pK reveal a general trend in which aromatic alcohols become more acidic, while aromatic ketones become more basic in the excited statea5 Molecules which possess an intramolecular hydrogen bond in the ground state often demonstrate an even larger red shift than is seen in aromatic molecules with just one functional group. For example, in methyl salicylate (I) the 0
II
C-OC H3
I
fluorescence maximum is red shifted by about 10 000 cm-I from the absorption maximum, while in phenol it is red shifted by about 4000 ~ m - l . ~ The , ' large Stokes shift seen in the emission from methyl salicylate is due to intramolecular proton transfer in the excited state. Proton transfer is driven by a pK change of about -6 in the phenolic o ~ y g e nand ~ , ~a pK change which may be as large as +8 for the carbonyl oxygen of the carboxyl group.l0 This *Alfred P. Sloan Fellow. 0022-3654/78/2082-2286$01 .OO/O
results in the formation of a zwitterion which fluoresces in the blue a t about 450 nm. In addition to the blue fluorescence, there is a much weaker component a t about 340 nm. Since methyl o-methoxybenzoate (11) emits at 320
fi m C - O C H 3
w o w 3 II nm, the near UV fluorescence was believed to originate from excited molecules in which the proton remains bound to the phenolic oxygen. The 340- and 450-nm emissions were equally quenched by carbon disulfide.6 Thus it was speculated that the two forms of the excited molecule were in equilibrium. If they were in equilibrium, then the rate of proton translocation would have to be much faster than the deactivation of the singlet state. This set a lower limit of los on the intramolecular transfer rate. The relative contribution of the two components to the total fluorescence was strongly dependent on temperature. As the temperature was lowered, the ratio of the 450- to 340-nm emission increased. From the temperature dependence of the ratio of 450- to 340-nm fluorescence, a difference in enthalpy of about -1.0 kcal/mol between the neutral form and the zwitterion was calculated. Since the proton transfer appeared to proceed even a t 4 K, it was felt that it occurred via a tunneling mechanism.l' The relative intensity of the two emissions could also be altered by dissolving methyl salicylate in acetonitrile instead of cyclohexane." In this solvent, the ratio of 450to 340-nm emission decreases. Weller postulated that the neutral form had a dipole moment in the excited state 0 1978 American Chemical Society
Picosecond Studies of Intramolecular Proton Transfer
which was larger than that of the zwitterion and was therefore stabilized by the acetonitrile. Sandros found that the ratio of the UV to blue emission was very sensitive to the hydrogen-bonding ability of the solvent.12 The emission spectrum was also dependent on the excitation wavelength. This led him to speculate that ground state rather than excited state equilibrium was responsible for the distribution of the fluorescence intensity between the two wavelengths. In agreement with Weller, Sandros felt that intramolecular proton transfer led to a zwitterion which fluoresces at 450 nm. Naiboken has proposed a somewhat different mechanism for the dual fluorescence from methyl ~a1icylate.l~ Absorption of a photon results in an excess of vibrational energy. This excess energy ruptures the intramolecular hydrogen bond. After relaxation takes place in the excited state,the hydrogen bond can then be re-formed. Molecules which re-form the hydrogen bond in the excited state emit at 450 nm, while those that do not re-form the hydrogen bond will emit at 340 nm. The absorption maximum of methyl salicylate is found at 308 nm while the absorption maximum for methyl o-methoxybenzoate occurs at 293 nm. If dissociation of the hydrogen bond takes place during the absorption of a photon one might expect the absorption of methyl salicylate to be to the blue of the absorption of methyl o-methoxybenzoate which does not have an intramolecular hydrogen bond. Calculations of the changes in the absorption of carbonyl compounds placed in hydrogen bonding solvents have been carried 0 ~ t . lThey ~ predict that for n-a* transitions rupture of a hydrogen bond would lead to a blue shift with respect to n-a* transitions in which the ground state is not hydrogen bonded. Also, the difference in energy between the blue and ultraviolet emissions is much larger than that of a hydrogen bond. The separation between the maximum of the two fluorescence components corresponds to about 6600 cm-', while the hydrogen bond in methyl salicylate is estimated at about 2400 cm-l.I5 Even if one assumes vastly different Franck-Condon factors for the species with and without a hydrogen bond, one cannot account for the large difference in the energy of the two emissions. In contrast, it has been well documented that the protonation of carboxyl groups and the deprotonation of phenolic groups in the excited state lead to large changes in the emission spectrum.lS* In addition to its interesting photophysics methyl salicylate had been used to protect polymers against damage from ultraviolet radiation. It has since been replaced by a number of other molecules. However, many of these, such as o-hydroxybenzophenone, possess intramolecular hydrogen bonds in the ground state. In fact an empirical relationship exists between the strength of the intramolecular hydrogen bond in the ground state and the ability of the molecule to protect polymers from damage by solar radiation.16 Therefore, motion of the proton in the excited state appears to play a role in the quenching of singlet and triplet excitation. By studying the time-resolved fluorescence from methyl salicylate and salicylic acid we hoped to provide additional information on the mechanism for proton transfer in the excited state. Such measurements would also provide new data on the quenching of singlet excitation by molecules with intramolecular hydrogen bonds. Finally by specifically resolving the individual fluorescence components, we could ascertain the origin of the two emissions.
Experimental Section A mode-locked neodymium phosphate glass laser produced a train of pulses each about 8 ps in duration (Figure
The Journal of Physical Chemistry, Vol. 82, No. 27, 1978 2287
Computerl Figure 1. Block diagram of laser apparatus: F, filters;
,
2500
k
I
0
I
S,sample.
,
Averaged Data Single Run
A',
I
I
I
I %
' 0
200
400
6C0
800
1000
1200
TIME (picoseconds)
Figure 2. Signal-to-noise improvement by averaging 15 laser shots. Open circles represent average of 15 shots while solid circles represent one of those 15 runs.
1). A pulse selector extracted a single pulse which was amplified about 50-100 times. The laser light was frequency doubled to 528 nm with an angle-tuned KDP crystal (Easermetrics). It was then converted to 264 nm with a temperature-tuned ADP crystal. A small portion of the light was reflected into a photodiode which triggered a streak camera (Hamamatsu). The green and ultraviolet components of the pulse were separated with a dichroic beam splitter. Residual green light in the ultraviolet beam was filtered out with a polarizer, a second dichroic mirror, and a set of color filters. The green pulse passed along a path parallel to the ultraviolet pulse. The path length of the green pulse was adjusted so that it arrived at the sample prior to the ultraviolet pulse. The 528-nm light passed through the sample without being absorbed and was imaged onto the slit of the streak camera. The ultraviolet beam excited the sample and the fluorescence was collected and also imaged onto the streak camera slit. The phosphor of the streak camera was viewed with a SIT vidicon camera. The output of the camera's video preamplifier was amplified and digitized by a multichannel analyzer constructed by our electronics shop. The digital data was transferred to a minicomputer for storage and analysis. The green prepulse always arrived at the photocathode a fixed period before the fluorescence of the sample. Therefore it was an accurate time mark in each streak record. By using this time marker the computer was capable of averaging a number of streak records to improve the signal-to-noise ratio." The power of this averaging method is illustrated in Figure 2. In this figure the data from one laser shot are superimposed on the average of
2208
The Journal of Physical Chemistry, Voi. 82, No. 21, 1978 i
I
I
I
K. K. Smith and K. J. Kaufmann 2coi
,
I
Room Temperature
1500t
. 0
lO0Ot
0
0
0 . 0 . I
1
Figure 4. Plots of the fluorescence lifetimes observed for the 450-nm component of methyl salicylate at various temperatures.
15 such measurements. On can see from Figure 2 that data obtained from a single streak record were in good agreement with those made using the averaging procedure. Fluorescence lifetimes of less than 3 ns were recorded with the apparatus described above. A few subnanosecond measurements were also made with a cross correlation phase fluorometer. Lifetimes longer than 3 ns were measured by excitation with a 30-ps 264-nm pulse from a neodymium YAG laser system and recorded with a 1P28 photomultiplier attached to a transient digitizer (Tektronix). The photomultiplier had a response function with a full-width at half-height of about 3 ns. Consequently all data obtained with the photomultiplier transient digitizer combination were deconvoluted to obtain the lifetime. In the region of 2-4 ns, the deconvolution procedure was checked by comparing measurements made using the photomultiplier with those obtained with the streak camera. The streak camera was calibrated using a variably spaced etalon consisting of two 90% reflectors. Both the salicylic acid and the methyl salicylate used in this experiment were subjected to repeated vacuum sublimations until no changes were observed in their fluorescence spectra. Solutions of salicylic acid (0.01 M) were prepared in methanol, concentrated H2S04, diluted H2SO4(-0.05 N), 6 N KOH, methylcyclohexane, and slightly basic methanol (-0.05 N). For methyl salicylate, steady state fluorescence spectra as well as fluorescence lifetimes did not vary over the concentration range of 10-2-10-4 M. Lifetime measurements of methyl salicylate were performed in methylcyclohexane, heptylcyclohexane, and octadecane. Fluorescence was not observed from the neat solvents indicating the absence of spurious results due to solvent impurities or emission from the filters used to isolate the observed radiation. Results Methyl salicylate in methylcyclohexanewas excited with a 264-nm picopulse. Since the formation of the zwitterion is accompanied by fluorescence at 450 nm measurements of the rise of the fluorescence a t this wavelength should give a lower limit on the rate of intramolecular proton transfer. The results of such an experiment are shown in Figure 3. The recorded fluorescence signal is a convolution of the apparatus function with the fluorescence from the sample. In Figure 3 the solid dots represent the apparatus function, while the open circles represent the actual data.
The data are almost perfectly fit by the response of the streak camera. This response function is for light at 528 nm striking the photocathode. The wavelength dependent response of the streak camera is expected to broaden the apparatus function for emission at 450 nm.18 This might account for the discrepancy between the data and the apparatus function. Deconvolution of the data indicates that the transfer rate must be greater than 10" s-l. The data did not change even as the temperature was lowered to 4 K. Replacing the proton with a deuteron and cooling to 4 K also did not alter the data. The fluorescence lifetime of methyl salicylate in methylcyclohexane was found to be about 280 ps at room temperature. The lifetime measured was independent over the concentration range 10-2-10-4 M. This indicated that dimer formation did not affect the lifetime. Fleming et al. have found that stimulated emission can result in the measurement of erroneously short lifetimes.lg The excitation power was varied over about two orders of magnitude, yet the lifetime remained unchanged. Therefore the data were not affected by stimulated emission or some other nonlinear process. In addition, measurements with a cross correlation phase fluorometer set an upper limit of 380 ps on the fluorescence lifetime. Perrin showed that a rotating dipole excited with polarized light could appear to have a fluorescence lifetime which depended on the angle between the polarization of the exciting light and the polarization of the luminescence.20 This effect is due to rotational motion of the excited dipoles either into or out of the plane of observation. Tao showed that emission observed a t an angle of polarization of 54.7O with respect to the excitation polarization would be free of this effect.21 Thus observation of the emission using a polarizer at 54.7' with respect to the polarization of the exciting light would yield the correct fluorescence lifetime of the sample. Methyl salicylate was studied by observing the fluorescence at the magic angle of 54.7O as well as with the detected fluorescence polarized perpendicular or parallel to the polarization of the 264-nm picopulse. For all three angles of the polarization, the recorded lifetimes were nearly identical. The lifetime of the 450-nm fluorescence from methyl salicylate was measured over the range 40-353 K. The lifetime was nearly temperature independent between 40 and 160 K. Above 160 K the lifetime became shorter as the temperature was increased (see Table I and Figure 4). Over a temperature range 253-333 K the relative quantum
The Journal of Physical Chemistry, Vol. 82, No. 21, 1978 2289
Picosecond Studies of Intramolecular Proton Transfer
TABLE I
t
relative quantuma temp, K
r , ps
k,,(T),
160 184 197 212 213 233 253 273 296 303 313 323 333 343 353
8300 5020 4650 2250 2092 1189 803 470 280 239 218 176 150 124 113
7.87 X lo' 9 . 4 6 ~10' 3.24X 10' 3 . 5 8 ~l o 8 7.21 X l o 8 1 . 1 2 ~lo9 2.01 x lo9 3.45 x lo9 4.06 x lo9 4.47 X lo9 5.56 X lo9 6.55 X lo9 7.94 x 109 8 . 7 3 ~l o 9
s-I
at 450 nm at 340 nm
t
2.02 1.00 0.88 0.63 0.48 0.35
1.15 1.00 1.20 1.20 1.13 1.03
Normalized so that at this wavelength the quantum yield is 1 at 296 K.
X (nm) Figure 6. Excitation spectra of methyl salicylate in methylcyclohexane. Solid dots represent the 450-nm band, while open circles represent the 350-nm fluorescence.
TABLE I1 Amax, nm absorplifetime, tion fluorescence ns
17250 l
300
3 50
* 400
4 50
5 oc l
U T x lo3
Figure 5. Plot of In k, vs. 1/ T . Open circles represent data for three separate experiments with methyl salicylate. Solid dots represent one run with the hydroxy proton replaced with deuterium. Solid lines represent least-squares fit of the data for the normal and deuterated samples.
yield was found to decrease with increasing temperature. The change in the relative quantum yields of the 450-nm emission closely paralleled those observed in the lifetime data. This led us to believe that the radiative rate remained constant, while the nonradiative rate was a function of temperature. Using the 40-K fluorescence lifetime as an estimate for a temperature-independent radiative lifetime, we were able to calculate the temperature-dependent nonradiative rate using the formula 1/7f = h,
+ k,(T)
where Tf is the measured fluorescence lifetime, k, was chosen to be 1.2 X lo8 s-l from low temperature work, and k,(T) was the temperature-dependent nonradiative decay rate. For three separate runs a plot of the In h,(7') vs. 1 / T yields an activation energy of 3.7 kcal/mol (Figure 5). A single preliminary experiment for methyl salicylate in which the hydroxy proton has been replaced with a deuteron gives an identical energy of activation. One can also use the relative quantum yields and 1.2 X lo* s-l as an estimate of the radiative rate to calculate the temperature-dependent nonradiative decay rate from the equation h,, = 4F/kr - k,. One then obtains an activation energy of about 4.7 f 0.6 kcal/mol. Even though the temperature range used is much smaller and hence more prone to errors, this value is in good agreement with the activation energy calculated from the variation of
salicylic acid (MeOH M t H') salicylic acid (MeOH 10-3M t OH-) salicylic acid ( 6 N KOH l o + M ) salicylic acid (conc H,SO,) salicylic acid (MCH 10-3M ) methyl salicylate (MCH 10-4 M )
304
350 t 438
0.38
302
398
- 3.4
304
400
-3.5
305
410
0.33
31 3
415
0.62
308
350 t 450
0.28
fluorescence lifetime with temperature. Methyl salicylate was also dissolved in heptylcyclohexane and octadecane. At 293 K, the former solvent has a viscosity about a factor of 4 larger than that of methylcyclohexane, while the latter is a solid.22 The lifetime of the 450-nm fluorescence was measured to be about 350 and 450 ps, respectively, in these two solvents. Presumably these changes represent differences in the nonradiative rate. The steady state fluorescence of the 340-nm component did not change measurably in intensity from 253 to 333 K, while in contrast (see Table I) blue emission decreased by more than a factor of 6 as the temperature was increased. The excitation spectra for the 450- and 340-nm emission were slightly different as well (Figure 6). The amount of 340-nm emission from methyl salicylate in methylcyclohexane was too small to accurately measure with the streak camera. However in acetonitrile, we found that the 340-nm light had a decay time of about 1 ns, while the 450-nm luminescence lasted only 100 ps. Salicylic acid is more complicated than methyl salicylate since it can exist in several ionic forms in the ground and excited state. In addition, it has a much greater tendency to form dimers than the methyl ester. The fluorescence spectrum and the fluorescence lifetime of salicylic acid are strongly dependent on the solvent conditions. These data are summarized in Table 11. Unfortunately, the possible presence of several forms of the molecule makes the
2290
The Journal of Physical Chemistty, Vol. 82, No. 21, 1978
analysis of the salicylic acid data difficult.
Discussion The measurement of intramolecular proton translocation sets a lower limit of 10'' s-l on the transfer rate. Since we were unable to measure either a temperature effect or an isotope effect, we cannot confirm or deny the postulate that proton movement proceeds via a tunneling mechanism." This is in contrast to measurements made in substituted 6-(2-hydroxyl-5-methylphenyl)-s-triazine, where the rate of proton transfer could be m e a ~ u r e d It . ~was ~ ~ found ~ ~ to be relatively insensitive to temperature and independent of the nature of the isotope. This led the authors to state that in triazines proton transfer does not proceed via a tunneling mechanism. Similar conclusions cannot be reached for methyl salicylate until we can actually measure the rate of proton transfer. The excited state species which emit at 340- and 450-nm are obviously not in equilibrium, since their fluorescence lifetimes differ by about a factor of 10 in acetonitrile. In addition, while the fluorescence intensity of the 450-nm emission increases in parallel with an increased fluorescence lifetime as a consequence of lowering the temperature, the 340-nm emission remains relatively constant. The rapid formation of the 450-nm emission and the very different lifetimes for the two fluorescence components would also indicate that two species are formed immediately after absorption of a photon. Perhaps they originate from different ground state molecules. This is suggested by the observation of differences in the excitation spectrum of the two emissions (Figure 6). It does not seem possible that small differences in the ground state conformation or differences in the degree of vibrational excitation of the ground state would result in the irreversible breaking of the hydrogen bond within a few picosecond~.'~ Therefore the data suggest that in the ground state the location of the hydroxy proton with respect to the carbonyl oxygen must be different for molecules which will emit at 450 nm and those that will emit at 340 nm. The small change in the size of the 340-nm component with temperature suggests that the ground state equilibrium is not very sensitive to temperature. If this is true it would indicate that AH for the two ground state forms must also be small. This implies that both forms must be hydrogen bonded in the ground state or else AH would have to be a t least 3-4 kcal/mol. However, one form cannot be hydrogen bonded to the carbonyl oxygen, since we see 340-nm fluorescence. Perhaps it is hydrogen bonded to the ester oxygen. Further work on other systems and on the temperature dependence of the 340-nm emission lifetime and quantum yield are needed to substantiate this hypothesis. Since the two excited state species are not in equilibrium, movement of the proton between the two oxygen atoms could not be responsible for the rapid deactivation at room temperature. Also, the energy of deactivation is much less than the vibrational frequency of an OH bond. In addition, there appears to be no isotope effect on the activation energy. Therefore OH vibrations in the zwitterion are not responsible for the decay. It has been reported that intramolecular rotational motion in a variety of dye molecules results in a rapid rate for internal conv e r ~ i o n . For ~ ~ ,such ~ ~ a process, the fluorescence lifetime and the quantum yield depend strongly on the solvent viscosity. Changes in solvent viscosity with temperature are not responsible for the observed data, because only small changes in the recorded lifetimes were observed when the solid octadecane replaced methylcyclohexane as the
K. K. Smith and K. J. Kaufmann
solvent. It is possible however that intramolecular rotational motion could be responsible for the decay. The replacement of the methyl group with an isopropyl group does change the lifetime by nearly a factor of 2. Vibrations of the hydrogen and carbon atoms of the ring could also be responsible for temperature-dependent decay. Since the UV fluorescence from methyl salicylate and from o-methoxybenzoate are longer in lifetime a t 296 K than the blue fluorescence, the rapid deactivation appears to be associated with the zwitterion. Otterstedt has argued that the formation of the zwitterion decreases the energy gap between the excited singlet and the ground state.27 This decrease in the separation would thus speed up the rate of internal conversion. The triplet zwitterion will not be as stabilized with respect to the neutral triplet as is the singlet zwitterion with respect to the neutral singlet because pK changes observed in the triplet state of aromatic alcohols are usually much less than those in the singlet state.% Therefore proton transfer may result in decreasing the singlet-triplet gap for the zwitterions as well and could result in enhanced intersystem crossing. If intersystem crossing is the deactivation rate then conversion to the ground state from the triplet must also be fast to account for the inability to measure phosphorescence from methyl ~alicylate.~ Work is presently underway to identify the decay channel. These studies have provided us with new insight into the nature of the proton transfer reaction in the excited state. They have also uncovered an interesting excited state decay mechanism. Further work on the molecular details of this decay mechanism may have impact on the protection of synthetic and biological polymers against solar radiation.
Acknowledgment. The authors thank Dr. Gregorio Weber, Dr. Dave Jameson, and John Wehrly for their assistance in the steady state and cross-correlation phase fluorometer measurements. We also thank Larry Sorenson for help and encouragement. This work was supported by the Research Corporation, by the donors of the Petroleum Research Fund administered by the American Chemical Society, and by the National Science Foundation. References and Notes A. Weller, Prog. React. Kinet., 1, 188 (1961). E. Vander Donckt, Prog. React. Kinet., 5, 273 (1970). T. Forster, Z. Electrochem., 54, 43 (1950). A. Weller, 2. Phys. Chem. (Frankfurtam Main), 17, 224 (1958). W. Klopffen, Adv. Photochem., 10, 311 (1977). A. Weller, Z. Electrochem., 60, 1144 (1956). W. Bartok, P. J. Lucchesi, and N. U. Snider, J . Am. Chem. SOC., 84, 1842 (1962). G. M. Gantz and W. G. Summer, Textile Rev. J., 27, 244 (1957). P. J. Kovi, C. L. Miller, and S. G. Schulman, Anal. Chim. Acta, 81, 7 (1972). E. Vander Donckt and G. Porter, Trans. Faraday SOC.,64, 3215 (1968). H. Beens, K. H. aellmann, M. Gurr, and A. H. Weller, Discuss. faraday Soc., 39, 183 (1965). K. Sandros, Acta Chem. Scand., Ser. A , 30, 761 (1976). U. V. Nalboken, E. N. Pavlova, and B. A. Zaborozhnyi, Opt. Spectrosc., 6, 231 (1956). J. E. Del Bene, J . Am. Chem. SOC.,95, 6517 (1973). B. A. Zaborozhnyi and I. K. Ischenko, Opt. Spectrosc., 19, 306 (1965). J. H. Caudet, G. C. Newland, H. W. Peters, and J. W. Tamblyn, SOC. Plastics Eng. Trans., 1, 26 (1961). K. K. Smith, J. Y. Koo, G. B. Schuster, and K. J. Kaufmann, Chem. Phys. Lett., 48, 267 (1977). D. J. Bradley and G. H. C. New, Proc. I€€€, 62, 313 (1947). G. R. Fleming, J. M. Morris, and G. W. Robinson, Chem. Phys., 17, 91 (1976)
F.'Perrin, J . Phys. Radium, 5, 497 (1939). T. Tao, Biopolymers, 8, 604 (1969). American Peiroleum Institute Research Project 44 Table 23c-K, 1958. H. Shizuka, K. Matsui, Y. Hirata, and I. Tanaka, J. Phys. Chem., 80, 2070 (1976).
Photodissociation of Tetramethyldioxetane (24) H. Shizuka, K. Matsui, Y . Hirata, and I.Tanaka, J . Phys. Chem., 81, 2243 (1977). (25) G. Oster and J. Nishijima, J. Am. Chem. Soc., 78, 1581 (1956). (26) E. P. Ippen, A. Bergman, and C. V. Shank, Chem. Phys. Lett., 38,
The Journal of Physical Chemistry, Vol. 82, No. 21, 1978 2291
611 (1976). (27) J. E. Otterstedt, J . Phys. Chem., 58, 5716 (1973). (28) T. S.Godfrey, G. Porter, and P. Suppan, Discuss. Faraday Soc., 39, 194 (1965).
Photodissociation of Tetramethyldioxetane Kevin K. Smith, Ja-Young Koo, Gary 6. Schuster," and Kenneth J. Kaufmann' Department of Chemistry, University of Illinois, Urbana, Illinois 6 180 1 (Received December 16, 1977)
Tetramethyldioxetane was excited with a single 10-ps pulse at 264 nm. The resulting acetone fluorescence was observed and the rise time was found to be less than 10 ps. A mechanism accounting for this and previous experimental evidence is proposed,
Introduction Dioxetanes are an interesting class of molecules, for upon thermal decomposition an excited state product may be formed. For tetramethyldioxetane 1 the electronically 0-0
CH3 CH3
1
excited products of thermolysis have been identified as triplet and singlet acetone. When tetramethyldioxetane is excited with ultraviolet light it decomposes into either singlet or triplet excited acetone fragments.l The wavelength of the exciting light determines in part the ratio of singlet-to-triplet products. We were able to examine this reaction using picosecond spectroscopy and based upon our results we propose a mechanism which may explain both the thermal and photolytic decomposition.
Experimental Section The experimental apparatus (see Figure 1) uses a passively mode locked neodymium glass laser to generate a train of picosecond pulses which are separated by about 9 ns. A single pulse is selected by use of a nitrogen spark gap which triggers a Pockels' cell, allowing a single pulse to pass through crossed polarizers. The single pulse is amplified by two additional glass rods and converted to the green with an angle-tuned KDP crystal. After the remaining 1.06-pm light is filtered out, the pulse passes through an angle-tuned ADP crystal producing the fourth harmonic at a wavelength of 264 nm. The UV light is reflected by a dichroic mirror and is transmitted through a Glan-air polarizer to extinguish any residual green light in the excitation beam. The UV beam is again reflected by another dichroic mirror followed by a quartz lens which focuses the radiation onto the sample. The majority of the green light transmitted by the first dichroic mirror is reflected by an aluminum mirror on a path shorter, but parallel to the excitation beam. This green light is used to establish a consistent time base for the short time scale experiments and allows one to monitor the exciting pulse. The fluorescence was observed using a Hamamatsu streak camera. For short time scales the output of the streak camera was coupled into an RCA SIT camera. This camera was interfaced to a digitizer constructed by our electronics shop. The information from the digitizer was *Alfred P. Sloan Fellow, 0022-3654/78/2082-2291$01 .OO/O
then read into a Nova 3 computer. The dioxetane was synthesized by the procedure of Kopecky2and purified by sublimation and recrystallization from pentane. Acetonitrile was the solvent and showed negligible impurity fluorescence. No change in the absorption spectrum was seen following the picosecond experiments indicating that insignificant thermal or photochemical decomposition of the tetramethyldioxetane had occurred.
Results Photoexcitation of tetramethyldioxetane yields both triplet and singlet excited products. The singlet excited product observed in this experiment may be attributed to the formation of acetone. The fluorescence from excited dioxetane and excited singlet acetone have nearly identical lifetimes (Figure 2), -2 n ~ This . ~ value is in good agreement with the accepted values for a ~ e t o n e Although .~ acetone and dioxetane have quite different absorption spectra their fluorescence spectra are identical and chemical trapping experiments have shown that singlet acetone is formed upon photodecomposition of d i ~ x e t a n e . ~ These results demonstrate that the nanosecond luminescence from photoexcited dioxetane is due to acetone fluorescence. If any detectable intermediates were present in the photodecomposition of dioxetane to singlet acetone, they could best be observed as a change in the rise of the resulting acetone fluorescence. At the fastest streak camera speed, the signal-to-noise ratio was so low that no evaluation of the rise was possible. To improve the signalto-noise ratio, signal averaging was used. To average one must first establish a reliable time base to overcome the effects of jitter in the streak camera. Owing to the weakness of the signal, the beginning of the fluorescence could not be accurately determined. The prepulse was used to ensure proper addition of the signals. (See Experimental Section.) For the measurement of the fluorescence rise, the optical path of the prepulse was adjusted so that it arrived at the photocathode of the streak camera 50 ps before the onset of the tetramethyldioxetane emission. The conversion of the 528-nm picopulse to the 264-nm picopulse was carried out by a nonlinear process. Therefore the green prepulse could not be used to monitor the size of the excitation. The risetime experiments were carried out with a sweep speed of the streak camera which corresponded to 600 ps full scale. Over this period the fluorescence of acetone decays by only about 25%. 0 1978 American Chemical Society
